Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by ALBION REED HODGDON, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN ROBERT CRICHTON FOSTER Associate Editors ROLLA MILTON TRYON RADCLIFFE BARNES PIKE LORIN IVES NEVLING, JR. VOLUME 68 1966 The Nem England Botanical Club, Ine. Botanical Museum, Oxford St., Cambridge 38, Mass. FARLOW REFERENCE LIBRARY APR cc Arn å JOO Todora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by ALBION REED HODGDON, Editor-in-Chief ALBERT FREDERICK HILL y STUART KIMBALL HARRIS RALPH CARLETON BEAN ROBERT CRICHTON FOSTER \ Associate Editors ROLLA MILTON TRYON RADCLIFFE BARNES PIKE LORIN IVES NEVLING, JR. Vol: 68 January-March, 1966 No. 773 CONTENTS: A Multivariate Analysis of the Pinus chiapensis-monticola- strobus Phylad Jobn W. Andresen ............ een 1 Cytogeography of Orontium aquaticum (araceae) John W. Gream, JT. wcccccooccssscrsscsssersssssssssscscssccsssccscssesessacesesssenes 25 Authors of Plant Genera and the International Plant Index (Review) Reed C. Rollins ................ esee 35 Studies in the Capparidaceae VIII. Polanisia dodecandra CL.) DC.: Hugh H. Iltis ................. eren 41 Polanisia dodecandra in New Hampshire Edward J. Hehe ..........soososssoseocsoosescscecssussovossosossoscoensocscososcoseesos 48 | (Continued on Inside Cover) Cee tr P 1X0. The Nem England Botanical Club, iuc. Botanical Museum, Oxford St., Cambridge 38, Mass. ^. > " S ~ D CONTENTS: — continued Chromosome Numbers in Vernonia (Compositae) S. B. Jones, Jr. and W. H. Duncan ees 49 Natural Hybrids of Lady's Slippers (Cypripedium) in Manitoba H. H. Marshall, A. T. H. Gross and G, A. Stevenson ........ 58 A Study of Variation in the Cytotypes of Dryopteris spinulosa Rolla Tryon and Donald M. Britton ess 59 Varieties of Ballota nigra in the Eastern United States C. V. Morton esset 93 Paspalum minus (Gramineae) in Louisiana and Mississippi Donald J. Bankes ..sss e 94 Studies in the Flora of Bolivia, — IV. Gramineae Robert C. Foster sss 97 Errata lrer ett dl 120 — RHODORA.—A quarterly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid. in funds payable at par in United States currency in Boston; single copies (if available) $1.80. Back vol- umes 1-58, with a few incomplete, can be supplied at $5.00 per volume, Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes. relating directly or indirectly to the plants of North America, will be considered by the editorial com- mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout, Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Albion R. Hodgdon, Dept. of Botany, Nesmith Hall, University of New Hampshire, Durham, New Hampshire. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. Manufactured by THE LEXINGTON PRESS, INC. Lexington, Mass. QTRbooora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 68 January-March, 1966 No. 773 A MULTIVARIATE ANALYSIS OF THE PINUS CHIAPENSIS-MONTICOLA-STROBUS PHYLAD* JOHN W. ANDRESEN As considered here, Pinus chiapensis (= Pinus strobus var. chiapensis Martinez) is the southern member of the North American pine phylad Pinus chiapensis-monticola- strobus. Phytogeographically, this complex forms a broad discontinuous triangle with apices of species distribution in British Columbia, Newfoundland, and Guatemala (Fig. 1). Earlier, the taxonomic recognition of Pinus strobus var. chiapensis Martinez was questioned by several authors (Braun 1950: 483; Martin and Harrell 1957; Sharp 1946) who regarded it as a weakly-differentiated geographic form of P. strobus L. undeserving separate taxonomic rank, but others (Loock 1950: 117-119; Soto, Barrett, and Little 1962: 52-53; Standley and Steyermark 1958: 55-56) agreed with its varietal status. Gaussen (1960: 91, 198) proposed it as a distinct species but his new combination was not validly published because he omitted citation of the basionym (Art. 33, LC.B.N., Lanjouw, et al. 1961). These divergent opinions prompted the present biometric study to provide a statistical basis for a determination of the proper taxonomic disposition of this controversial taxon. In a companion paper (Andresen 1964), it is recommended that P. strobus var. chiapensis be elevated from varietal to This study was supported in part by funds from National Science Foundation Grant No. G-15879. bo Rhodora [Vol. 68 Pinus chiapensis è ipinus monticola ERE 4 — ' Pinus strobus NN m" SCALE Fig. 1.— Distribution of the phylad Pinus chiapensis-monticola- strobus in Central and North America. Adapted from unpublished map supplied by U. S. Forest Service. Plotted on Goode's Base Map No. 202 by permission of the University of Chicago Press, specific rank with the binomial Pinus chiapensis (Martinez) Andresen. The proposal was predicated on evidence ob- tained from the following study of morphologic and progeny data that were not available to previous authors. Following the analysis of this data, it became apparent that there was a much wider genetic and morphologic divergence than was previously suggested between P. strobus var. chiapensis and typical P. strobus. Recent conceptual advances in the field of biosystematies (Heslop-Harrison 1963, Heywood and Love, 1963), that presented a clearer perspective of the problem at hand, led 1966] Pinus — Andresen 3 Rogers (1963), and Sokal and Sneath (1963) to stress the value of reliable statistics that incorporate fundamental quantitative characters to determine taxa affinities or dif- ferentials. They coined the terms “taximetrics” and “nu- merical taxonomy” respectively, to focus attention on the statistical techniques designed to solve problems associated with experimental taxonomy. In this respect it is of pri- mary importance to the taxonomist that contemporary statistical analyses and those yet to be derived can be pro- grammed, especially with the aid of electronic computers, to assess the taxonomic value of any taxon trait or group of characters. As an example, Davidson (1963) developed an “Itemized frequency distribution" and set of association indices based on the F statistic, to discriminate Cirsium altissimum from C. discolor. In a study of variability of the Pacific Coast and Rocky Mountain forms of Pinus contorta, Jeffers and Black (1963) used three forms of multivariate analyses which incorporated the “Q-technique,” discriminant anal- ysis, and component analysis, all of which rely on various manipulations of the correlation coefficient and its components. In this study, to show the degree of discrimination within the Pinus chiapensis-P. monti- cola-P. strobus complex, I employed the coefficient of divergence (C.D.) technique (Clark 1952) which is based on the sum of mean differentials of multivariate characters. To evaluate and assign relative importance values to the 12 leaf and cone characters used in the C.D. analysis, the F Statistic, intraclass correlation (expressed as a reliability index, R,), and coefficient of variation were also used. Data for the multivariate character analysis of cones, leaves, seed, and seedlings were obtained from herbarium and living specimens. Sample sizes, characters examined, statistical techniques, and pertinent examples to illustrate a particular method used are included as footnotes to data tables or in the discussion of the characters. 4 Rhodora [Vol. 68 In the following text, Pinus chiapensis is coded as A; P. monticola Dougl. as B; and P. strobus as C, except where the Latin binomials are more useful for clarity or emphasis. COEFFICIENT OF DIVERGENCE VALUES Before the importance of individual characters and their component roles in species delimitation within this phylad are evaluated, an explanation of the simultaneous treatment of fundamental characters is in order. A useful measure that incorporates this technique is the coefficient of divergence which is used in association with F statistics and reliability indiees to summarize leaf and cone characters Q, to Q, species pairs AB, AC, and BC. The method (Clark 1952, Klauber 1940) of arranging or ranking taxa in relation to the ascending magnitude of their C.D.’s is especially ger- mane to studies utilizing objectively scored traits. Although Clark (1952) referred to herpetological data in his example, the extension of Klauber's (1940) one-dimensional coeffi- cient by Clark is adaptable to any numerical or coded array of biologic parameters computed as a series of multiple character means. In the present example (Table 1) four results are ap- parent: (1) About the same degree of information was provided by either 39 random samples or the total number of herbarium specimens examined (see Tables 1 and 2 to compare means of characters Q, to Q,. for the two sample sizes) ; (2) The C.D.’s derived from cone characters alone were greater in pairs AB and AC: (3) Leaf characters were more important than cone traits in computing the C.D. for BC; (4) With the incorporation of all 12 characters, BC is less than one-half as divergent as either AP or AC. This magnitude of divergence, or lack of it, is of special significance — for according to the C.D. ranking, two long- recognized species, P. monticola and P. strobus, are much more closely aligned to one another than the species P. strobus is to its supposed variety chiapensis. This evidence, primarily the strength of the following F and R, statistics 1966] Pinus — Andresen 5 for the leaf and cone characteristics, plus other pertinent data, motivated the author to elevate P. strobus var. chi- apensis to specific level (Andresen 1964). LEAVES Leaf length (Q,) was greatest for A, but leaf width (Q;) was least (Table 2). Thus, the resultant W/L value (Qs) for A was smallest. Conversely, the shorter but wider leaves of B had a high W/L ratio that was twice as great as the value for A. An intermediate value of Q; for C is a reflec- tion of the interjacent values of Q, and Q. The interpreta- tion of this value, however, should be tempered by the lack of significance in the F statistic and the low reliability index of .497 for Q, in species pair AC; thus the component of greater relative importance for Q, in C is leaf length. The role of leaf length in the ratio is emphasized because numerous authors (Loock 1950: 117; Martinez 1948: 133; Schwerdtfeger 1953; Sharp 1946; Soto, Vazquez, and Little 1962: 53) describe the leaves of P. chiapensis as “more delicate," “more slender," "finer," or “thinner” than P. strobus. These comparatives are misleading, however, when Q, and Q, are examined objectively. As shown earlier, there is a highly significant difference between A and C in regard to length but the width differential is non-significant. With these and other facts in mind, the following thoughts are offered to attempt an explanation of the reasoning of previous observers. In the field, the illusion of angustifoli- ation is probably created when observers familiar with leaf length in P. strobus subconsciously compare the much longer (and seemingly thinner) leaves of P. chiapensis and infer that the leaves of A are more delicate. But more important, because most descriptions are based on dry material, the gracile appearance of desiccated specimens is related to the herbarium artifact caused by severe de- hydration and subsequent collapse of the leaf mesophyll. This condition was suggested when dried leaf specimens of A were noted to be deeply concave on all three surfaces (turgid leaves bear slightly convex surfaces), were curled 6 Rhodora [Vol. 68 or twisted, and seemed lighter in weight than dried leaves of either B or C. To advance this postulation I (1) dried 125 leaves of A, B, C, and B X C grown under similar conditions near Placerville, California, and (2) calculated the proportionate weight loss of each set of samples by comparing fresh weight (weighed immediately after col- lection) to oven-dry weight. The differential percentages (Table 4) illustrate a much greater weight loss in A than the other three which were not significantly different from each other. Although tabulated at the 5% level of signifi- cance A was actually significant at the 2% level of con- fidence. The loss of 69% of the original weight in A is attributable mainly to the removal of fluids (mostly water) from the abundant intercellular space within the leaf meso- phyll. A comparison of intercellular spaces within the leaves of the four taxa (Table 4) illustrates the significant differential between A and the other three. Another feature which has previously been used to sep- arate A from C is the supposed higher number of foliar resin canals in the former (Gaussen 1960: 91; Loock 1950: 117, Martinez 1940, 1947: 134; Soto, Vasquez and Little 1962: 53). This group of authors contends that the more common number of resin canals in A is 3 and more often 2 in C. The analysis (Table 4) however, provides data to the contrary. Based on 500 observations, ¢ values at the 5% and 146 levels shows no significant differences between A and C. Mean values of 2.52 and 2.57 respectively, low standard deviations, and ranges of 2 to 3 resin canals for both species invalidate this character as a trait of differ- entiation. Although there is an overlap of range in number of resin canals between B and the former two, there is a significant difference between the means, so this feature could be used to separate B from A or C. Of interest is the broad range of numbers of resin canals in the hybrids of B and C and the large standard deviation, but this is an established pattern with numerous hybrid pines ( Keng and Little 1961). 1966] Pinus — Andresen zi Even though the original description (Martinez 1940) of A was predicated on the criteria of thinner leaves and a greater number of foliar resin canals than C, the data indicate that both characters are of little or no value in separating the two taxa. On the other hand, a character apparently overlooked in the past that has high diagnostic utility is the degree of leaf serration, Q, (Table 2). In fact, the F value for this character for species pairs AB or AC was the most sig- nifieant of any leaf parameter. Twice as many serrations per 5 mm interval were found on A when compared to C. The high reliability index of .995 for Q, of pair AC also demonstrates the dependability and high level of confi- dence to be expected when using this trait for species sep- aration. Also, a low coefficient of variability of 14.3 for A in contrast to 25.4 for C and 57.5 for B is added measure of reliability for this character in A. In a progeny study of juvenile forms of P. strobus, Mergen (1963) found an in- crease in serration number in relation to progression from southern to northern sources of origin. We found just the opposite trend in mature field specimens, and observed a random pattern in juvenile specimens. The weakest leaf character of the five examined (Table 2) was the number of stomatal rows borne on the ventral (adaxial) surface, Q.. Although there was a significant difference between A and C for Q., the reliability index was lowest in comparison to the other significant features of leaf characters. Stomata on the dorsal (abaxial) surface, although not included in the character analysis, were absent on all 440 leaves examined of A and all 3408 leaves exam- ined of C. Dorsal stomata were present, however, on B leaves with a mean of 0.28 and a range of 0 to 4 based on a 1279 leaf-count. When present, this is a reliable character to separate B from C (Harlow 1947, Sargent 1897: 23, Shaw 1914: 34). In two studies of pine leaves however, (Doi and Morikawa 1929, Sutherland 1934) reported an absence of dorsal stomata in P. 8 Rhodora [Vol. 68 Fig. 2. — Comparison of cone characteristics of Pinus chiapensis (A), P. monticola (B), and P. strobus (C). PLATE 1318 CONES Since the value of the mean for ovuliferous scale number (Q;) was highly dissimilar between the three species and C.V.'s were relatively small, large F statistics and strong reliability indices were derived. The highest number of scales were found on B (Fig. 2, Table 2) and even though there were about 25% more scales per cone in A than in C, cone lengths were not significantly different. Precise cone- phyllotaxis of the three species was very difficult to deter- mine in mature open cones, so this trait was ignored. 1966] Pinus — Andresen 9 Shaw (1914: 12) observed that this pattern of cone scale arrangement presented an indefinite phyllotaxy for the sub- genera Diploxylon and Haploxylon and the only importance of phyllotactic differential is that it separates the two sub-genera on the basis of a higher order-fraction in the former group. Nevertheless he commented that in cones of equal size, B had an obviously higher phyllotaxy than C. a phyllotaxy of A was also higher than C but lower than Apophysis width (Q;) in both A and C was similar, but length (Q.) was more divergent, (Table 2, Fig. 2). Even though Q; and Qs were dimensionally greater in B than in the other two taxa, the resultant ratio (Q,) of the three species was largest for A because of its higher propor- tionate width. A high F statistic and reliability index of .959 of Q, for AC indicate the potential diagnostic value of this trait. In pair BC, the ratio differential was non- significant with a low R,, and so was of little value in separating the cones of B and C. Length of peduncle (Qı) was greatest in A, but shorter and about the same for B or C. Although F and R, values of Q, were high for pairs AB and AC, the low values in association with BC indicate the similarity of this trait in B and C. Character Q.» is, at times, difficult to measure for the peduncle is brittle and tends to fracture easily. Caution must be exercised to carefully collect and store the cones to avoid negation of this character. Another reliable cone trait to isolate A from B or C is the degree of scale margin undulation (Q;, ). In A, the apo- physis margin with inflexed umbo-tip forms several involu- tions (Fig. 2) rarely found in B or C. As with peduncle length, scale margin bore differentially high statistics for pairs AB and AC, but was useless to segregate B from C in which the umbos were simply concave. Martinez (1940) and Soto, Barrett, and Little (1962: 53) described the umbo as having ". - - undulating edges turned inward.” but in their work this conspicuous feature was not com- 10 Rhodora [Vol. 68 Fig. 3. above — Basal scale configuration of Pinus chiapensis (A), P. monticola (B), and P. strobus (C). Fig. 4. below — Ovuliferous scales (adaxial surface) of Pinus chiapensis (A), P. monticola (B), and P. strobus (C). PLATE 1319 pared to other species. Also, the scale apex of A is truncate versus the rounded apices of P and C (Fig. 4). The most outstanding difference in cone morphology and the character that was related to the highest F and R, val- ues of all leaf and cone characters for pairs AB and AC was the relative number of basal reflexed scales (Qi.). This curving of scales contiguous to the peduncle of taxa A and B is obvious in numerous illustrations (Gaus- sen 1960: 199; Harlow and Harrar 1958: 57; Sargent 1897: plates 539, 541; Sudworth 1908: Fig. 2). In Fig. 3, 1966] Pinus — Andresen 11 a comparison of the number and presence of these scales in B and C is made with the lack of reflexing in A. In our samples of A, none of the cones bore basal scales that were fully reflexed, and only a few were weakly curled. Since F and R, values were high for all three pairs, an exam- ination of Q,. alone would suffice to determine individual members of the phylad. In addition to the foregoing values obtained from leaf and cone measurements, the data following the discussion of sample size also add confirming evidence that P. chia- pensis is strongly divergent from P. strobus. SIZE OF SAMPLE In biometric analyses, achievement of the highest degree of precision commensurate with judicious use of time is a goal of primary importance. In Table 1, it is demonstrated that approximately the same coefficient of diver- gence can be calculated by measuring relatively large num- bers of specimens or by a less time-consuming random sample of 39. The question then is how many samples are required to provide a desired level of accuracy in separat- ing these taxa? Numerous models and examples are offered by Cochran and Cox (1950), Dixon and Masey (1957), and Snedecor (1956), but I prefer the technique derived by Wright and Freeland (1960). Their method, in which it is advantageous to use small sample numbers, employs stand- ard deviations and means of the components of any species pair. When substituted in the formula in Table 3, these values multiplied by an appropriate “t” provide a predict- able sample size for any required level of accuracy. Since the minimum number of complete samples within A, B, or C was 39 and a recommended minimum sample (Freese 1959) should approximate 30, we used a set of 39 observations for the 12 characters to obtain the s, X, F, and R, data in Tables 2 and 3. Also, from past experience with the group Strobi, it seems that 30-40 random samples are required for an initial character survey. Estimates of sampling (Table 3) for further studies of characters Q, to Qi: for taxa pairs 12 Rhodora [Vol. 68 AB, AC and BC indicate that a relatively small number of specimens is required to exhibit statistical significance for most traits. A maximum of ten samples of each species within a pair would be required to detect a difference at the 5% level of confidence in 28 of the 36 possible character- taxa pair combinations and 20 samples to detect differences at the 1% level for 28 out of 36. These are reasonable sample sizes with which to work, and the number of char- acters, if definitive, could be expanded to give a more pre- cise evaluation of a total difference. Note the direct rela- tionship (Table 3) between the reliability index and the number of samples required. Low or negative R, values correspond to higher sample numbers. Large sample sizes, e.g. 1552 and 227 for BC-Q,, and Q,, would require a prohibitive amount of observation time, so these characters are best not measured.. The time saved could be better spent on other observations such as those which follow. SEED AND SEEDLING CHARACTERISTICS Differences in seed weight of the three species used in the progeny tests are given in Table 5. To determine the accurate weight of sound seed each lot was first floated in 95% ethyl alcohol to separate the denser, sound seed from the buoyant, partially filled seed. After winnowing, the seeds were floated and it was discovered that about 10% of the seed were blank. Based on this unit of malformed seed and experiences with the other group Strobi seed, we con- verted the reported weights (Anon 1948: 269) of P. monticola and P. strobus to arrive at 53,460 per kg which compared very closely to our data for the two species. The large number of 60,000 seed per kg for A is related to the smaller size of the individual seed. The smaller seed may be linked to the ability of propagules with reduced endo- sperm to survive in the favorable growing and germinating conditions within the natural range of A. The seed of all three species, in either the stratified or unstratified condition, began to germinate within 10 to 16 days after sowing (Table 5). In contrast to B and C, the 1966] Pinus — Andresen 13 germination period for A had ceased 37 days after initiation of the test— also, the germination values for either treatment were identical. On the other hand, un- stratified seed of B and C had considerably lower values than those stratified and some seed of both treatments had not germinated at the end of the test. The marked con- trast in germination values between A and B or C is related to a major physiological difference within the phylad ger- mination regime. Under uniform growing conditions in a greenhouse the seed sources of A produced seedlings with longer hypo- cotyl and cotyledon length. A was significantly larger than B and C, but B was smallest. Larger numbers of samples are required to determine the variability within B and C, but the 30 sources of A seemed adequate to assess some in- fraspecific variation. Cotyledon numbers were not signifi- cantly different from one another although C had a slightly higher range and mean. Englemann (1880) reported 6-9 for B and 7-11 for C. Of all the seedling characteristics examined in this study, cotyledon number was most uni- form and of least diagnostic value. Shaw (1914: 1) suggested that cotyledon morphology and number are unreliable for species identification unless there are distinct numerical differences between species. In a recent study of the Pinus flexilis - P. strobiformis com- plex (Andresen and Steinhoff 1965), a significant differ- ence was found in numbers of cotyledons with a mean of 8.8 for P. flexilis and a mean of 11.5 for P. strobiformis. In addition to the above seedling characteristics, there was a difference in growth of secondary leaves and bud set- ting. As might be expected, because of its southern origin (Mirov 1962), A produced secondary leaves much earlier than either B or C. No secondary leaves were formed on either B or C during the first 200 days of observation. Also, A did not produce any dormant buds, for as soon as a bud was formed it produced continuous growth of leaves. Definite dormant buds did form on B and C. In the course of my 14 Rhodora [Vol. 68 field work in December 1962, I observed numerous P. chia- pensis south of Sola de Vega, Oaxaca, Mexico, which had produced several increments of cones the past growing season and with a few trees undergoing anthesis. Mr. Boone Hallberg (personal correspondence) has noted this phenomenon for several years and reports that an average stand in Oaxaca has the following periodicity of cone matu- ration: Percent of Trees Date of Maturation 20 20 July + second crop about 20 November 10 1-20 August 60 25 August - 15 September 10 15 September - 10 October Mirov (1962) also observed that P. oocarpa in Nicaragua and other tropical pines are characterized by accelerated and uninterrupted growth rhythms. Only one year is re- quired for seed set after pollination. CONCLUSIONS Coefficient of divergence values indicate a much closer morphological affinity between Pinus strobus and P. monti- cola than between P. strobus and the former variety chiapen- sis. The wide divergence between P. chiapensis and P. strobus provides one form of biometric evidence to argue against the opinions of earlier writers advocating that the two taxa should be conspecific. With the available C.D. values and the sustantiating data that follow, little doubt re- mains that P. chiapensis is a distinct species. An analysis of the more statistically significant characters that discrimi- nate P. chiapensis from P. strobus reveals that P. chiapensis differs by having: (1) 25% more leaf serrations per unit length, (2) leaves that are 31% longer, (3) cone-scale apo- physes that are truncate and 20% shorter, (4) extremely thin and wavy apophysis margins, (5) no reflexed scales contiguous to the peduncle, (6) 13% more seed per kg, (7) seed that germinates rapidly and uniformly without after- 1966] Pinus — Andresen 15 ripening, and (8) continuous growth in the primary seed- ling stage without formation of dormant buds. Ecologically, P. monticola and P. strobus grow within similar sub-boreal to cool temperate habitats. Although there undoubtedly are clinal or ecotypic variants (Hanover 1962; Mergen 1963; Wright, Lemmien and Bright 1963) within the geographic distribution of the two species, three omnipotent growth factors are present: (1) both species are influenced by annual photoperiodic cycles of short and long days that occur between 34° and 52° N. Lat., (2) pre- cipitation ranges from 500 to 2000 mm (with up to 3450 in Coastal Washington) with pronounced dry periods during the summer months, and (3) temperatures are low enough to induce marked winter dormancy of three to six months. On the other hand, P. chiapensis is subjected to sub-equa- torial insolation from 15° to 20° N. Lat. and is under the influence of a humid, warm-temperate climate associated with abundant precipitation concentrated in the summer months. In addition, this effective precipitation is further enhanced by frequent mountain fogs. Leopold (1950) de- scribed the P. ckiapensis phyto-association as a tropical cloud forest while Beard (1944) termed it a temperate rain forest. Even though most P. chiapensis forests are limited to subtropical or warm-temperate climates there are excep- tional isolated stands that occasionally are exposed to freez- ing temperatures. In addition to the foregoing morphological and ecological divergences there also are ancient phytogeographical divari- cations that are extremely important when considering the origin of Pinus chiapensis. I believe that the original dis- persal of the phylogenitors of P. chiapensis, P. monticola, and P. strobus was accomplished at least by early Eocene and that later but similar forms, morphologically and phy- siologically allied to the three contemporary taxa, were isolated from one another by late Pliocene. My reasoning follows: Substantial evidence (Gausen 1960) of an ancient Hap- 16 Rhodora [Vol. 68 loxylon pine flora richer and more diverse than the sub- genus now extant has been revealed in the fossil record of the late Mesozoic and Arcto-Tertiary forests. With the advent of accelerated climatic change in the pre-Pliocene (Good 1953: 263), xeric climatic bands evolved (Schwartz- bach 1963) that were unfavorable to the mesophytic plants of the southern portion of the Arcto-Tertiary biome. These latitudinally oriented zones contributed to the formidable hiatus which first isolated the early Quaternary progenitors of such contemporay pine relicts as Pinus ayacahuite, P. chi- apensis, P. griffithü, P. parviflora, and P. peuce. In the early Pliocene of North America pronounced orogenesis con- tributed to a secondary discontinuity that further divided the forests above 259 North Latitude into eastern and west- ern disjuncts. It was then, in the forested and mountainous areas of the Appalachians, the Pacific North West, and the southern Mexican-Guatemalan highlands, that the immedi- ate ancestors of P. chiapensis, P. monticola, and P. strobus persisted in competition with their associates. Thus, prior to the continental and montane glaciation of the Pleistocene, isolation of the phylad segregates was com- plete and to this day final. The waves of subsequent Pleis- tocene ice with their periglacial climatic regimes undoubtedly caused areal concentration of both P. monticola and P. strobus. However, the cooler temperatures and neo- pluvial conditions that were contemporary with glaciation did not, as Deevey (1949), Dressler (1954), and Sharp (1953) suggest, create an ameliorated environment across the semi-arid Texan hiatus that would have permitted the mesophytic phyto-associations of the southeastern United States to reach and penetrate the montane regions of east- ern and southern Mexico. Martin and Harrell (1957) observed that further de- tailed ecologic and taxonomic revisions should be made of the vicariads of these two biotas. For only with new evi- dence can the theory of mid-Cenozoic disjunction be ex- panded, 1966] Pinus — Andresen 17 Finally, the divergence patterns and biometric differen- tials within the Holocene Pinus chiapensis-monticola-strobus phylad provide one such example of floristic confirmation of pre-Pleistccene isolation of the montane forests of south- eastern Mexico, as suggested by McVaugh (1952) ; Martin and Harrell 1957; and Steyermark 1950. DEPARTMENT OF FORESTRY, SOUTHERN ILLINOIS UNIVERSITY CARBONDALE, ILLINOIS ACKNOWLEDGEMENTS I am grateful to the personnel of the Western Institute of Forest Genetics and the Moscow (Idaho) Field Service Unit of the U. S. Forest Service, to Drs. John C. Genys, University of Maryland, How- ard C. Kriebel, Ohio State University, and Mr. Boone Hallberg, Oaxaca, Mexico for fresh foliage and seed. The curators of the following herbaria kindly provided specimens for the study: A, F, MEX “F” (Instituto Nacional de Investigaciones Forestales), MEXU, MSC, NA, MO, US. In addition I am grateful to Drs. John H. Beaman, J. L. Ruby, and J. W. Wright, and Mr. W. Rubel for their helpful comments. Special recognition is given to Dr. Nicholas T. Mirov who originally suggested this investigation. Mr. Nathan Shier, laboratory technician, Michigan State University, performed the majority of the leaf and cone measurements, LITERATURE CITED ANDRESEN, J. W. 1964. The taxonomic status of Pi Phytologia, 10: 417-421. . and R. J. STEINHOFF. 1965. The taxonomy of the Pinus flexilis-P. strobiformis complex. [Manuscript] ANONYMOUs 1948. Woody-plant seed manual. U. S. Dept. Agr. Misc. Pub., No. 654, vi 4- 416 pp. illus. BEARD, J. S. 1944. Climax vegetation in tropical America. Ecology 25: 127-158. BRAUN, E. L. 1950. Deciduous forests of eastern North America. Philadelphia, Blakiston Co. xiv + 596 pp. CLARK, P. J. 1952. An extension of the coefficient of divergence for use with multiple characters. Copeia 2: 61-64. CocuRAN, W. G. and G. M. Cox. 1957. Experimental designs. ed. 2. New York, John Wiley and Sons, Inc., 611 pp. Davipson, R. A. 1963. Initial biometric survey of morphological varia- tion in the Cirsium altissimum-C. discolor complex. Brittonia 15: 222-241. nus chiapensis. 18 Rhodora [Vol. 68 DEEVEY, E. S. 1949. Biogeography of the Pleistocene. Bull. Geol. Sco. Amer. 60: 1315-1416. Dor, T. and K. MoRIKAWA. 1929. An anatomical study of the leaves of the genus Pinus. Kyushu Imp. Univ. Jour. Dept. Agr. 2: 149- 198. Dixon, W. J. and F. J. Massey, JR. 1957. Introduction to statistical analysis. ed. 2. New York, McGraw-Hill Book Co., Inc. xiii + 488 pp. DRESSLER, R. L. 1954. Some floristic relationships between Mexico and the United States. Rhodora 56: 81-96. FREESE, F. 1959. A guide book for statistical transients. U. S. Dept. Agr., Southern For. Exp. Sta. 77 pp. ENGELMANN, G. 1880. Revision of the genus Pinus and a description of Pinus elliottii. Trans. St. Louis Acad. Sci. 4: 161-189. GAUSSEN, H. 1960. Les Gymnospermes actuelles et fossiles. Chapitre 11. Généralités, Genre Pinus. Trav. Lab. Forestier Toulouse, tome II, sect. 1, vol. 1. pt. 2, 272 pp. illus. Goop, R. D. 1953. The geography of flowering plants. ed. 2. New York, Longmans, Green. 452 pp. HaRLow, W. M. 1947. The identification of the pines of the United States, native and introduced, by needle structure, New York State Coll. For. Tech. Publ. No. 32, 19 pp. illus. . and E. S. Harrar. 1958. Textbook of den- drology. ed. 4. New York, McGraw-Hill Book Co, Inc. xi + 561 pp. illus. HESLOP-HARRISON J. 1963. Species concepts: theoretical and practical aspects. pp. 17-40, In: Swain, T. (Ed). Chemical plant Tax- onomy. New York. Academic Press ix + 543 pp. HEYwoop, V. H. and A. LOVE, (Eds). 1963. Symposium on biosyste- maties. Regnum Vegetabile 27: 1-72. HANOVER, J. W. 1962. Clonal variation in western white pine. U. S. Dept. Agr. Intermountain Forest and Range Exp. Sta. Res. Note 101, 4 pp. JEFFERS, J. N. R. and T. M. BLACK. 1963 An analysis of variability in Pinus contorta. Forestry 36: 199-218. KLAUBER, L. M. 1940. Two new subspecies of Phyllorhynchus, the leaf-nosed snake, with notes on the genus. Trans. San Diego Soc. Nat. Hist. 9: 195-214. KENG, H. and E. L. LiTTLE, JR. 1961. Needle characteristics of hybrid pines. Silvae Genetica, 10: 131-146. LANJOUW, J.et al. 1961. International code of botanical nomenclature. Utrecht, Netherlands, International Bureau for Plant Taxonomy and Nomenclature, 372 pp. LEOPOLD, A. S. 1950. Vegetation zones of Mexico. Ecology 31: 507-518. 1966] Pinus — Andresen 19 Loock, E. E. 1950. The pines of Mexico and British Honduras. Union South Africa, Dept. Forestry Bull. No. 35, x + 244 pp. Martin, P. S. and B. E. HARRELL. 1957. The Pleistocene history of temperate biotas in Mexico and eastern United States. Ecology 38: 468-480. MaRTINEZ, M. 1940. Pinaceas Mexicanas, An. Inst. Biol. Mex. 11: 57-84. McLemore, B. F. and F. J. CZBATOR. 1961. Length of stratification and germination of loblolly pine seed, Jour. For. 59: 267-269. McVaucu, R. 1952. Suggested phylogeny of Prunus serotina and other wide-ranging plylads in North America. Brittonia 7: 317- 346. MERGEN, F. 1963. Ecotypic variation in Pinus strobus L. Ecology 44: 716-727. Mirov, N. T. 1962. Phenology of tropical pines. Jour. Arnold Arb. 43: 218-219. Rocers, D. J. 1963. Taximetrics — new name, old concept. Brittonia 15: 285-290. ScuwARTZBACH, M. 1963. Climates of the past; an introduction to paleoclimatology. London, D. Van Nostrand Co., Ltd. xii + 328 # pp. SCHWERDTFEGER, F. 1953 Informe al gobierno de Guatemala sobre la entomolgia forestal de Guatemala. Vol. I. Los pinos de Guate- mala. United Nations-Rome, informe FOA/ETAP No. 202, 59 pp. SARGENT, C. S. 1897. The silva of North America. Vol. XI Coniferae. New York, Houghton, Mifflin and Co. 163 pp. illus. Suaw, G. R. 1914. The genus Pinus. Harvard University, Arnold Arb. Publ. No. 5, 96 pp. illus. SHARP, A. J. 1946. Pinus strobus south of the United States. Elisha Mitchell Sci. Soc. Jour. 62: 229-230. . 1953. Notes on the flora of Mexico: world distribution of the woody dicotyledonous families and the origin of the modern vegetation. Jour. Ecology 41: 374-380. SNEDECOR, G. W. 1956. Statistical methods. ed. 5. Ames, The Iowa State College Press, xii + 534 pp. SoKAL, R. R. and P. H. A. SNEATH. 1963. Principles of numerical taxonomy. San Francisco. W. H. Freeman and Co. xvi + 359 pp. Soto, J. VASQUEZ, W. BARRETT, and E. L. LITTLE, JR. 1962. Botany pp. 1-64. In: Anon. Seminar and study tour of Latin-American conifers. FAO of UN, English Edition No. 1, x + 209 pp. STANDLEY, P. C. and J. A. STEYERMARK. 1958. Flora of Guatemala. Fieldiana: Botany, Part 1. ix + 478 pp. STEYERMARK, J. A. 1950. Flora of Guatemala. Ecology 31: 368-372 SupworTH, G. B. 1908. Forest trees of the Pacific Slope. U. S. Dept. Agr., For. Serv. 441 pp. illus. 20 Rhodora [Vol. 68 SUTHERLAND, M. 1934, Microscopical study of the structure of the leaves of the genus Pinus. Trans. Proc. New Zealand Inst. 63: 517-568. WRIGHT, J. W. and F. D. FREELAND. 1960. Plot size and experimental effieiency in forest genetic research. Michigan Agr. Exp. Sta. Tech. Bull. 280, 28 pp. Table1. Computation and Comparison of Coefficients of Divergence within the Pinus chiapensis — P. monticola — P. strobus Complex. Based on Means! of Characters Q, to Qe for Taxa A to C. Taxa Parameter Means Qi Q. Q: Q: Q: Q: Q: Qs Qo Qi Qu Q: A 104.6 0.65 0.0062 12.9 4.3 87.7 17.7 9.9 1.79 22.4 2.5 0.04 B 68.1 0.92 0.0135 3.1 4.0 110.8 20.0 14.2 1.41 16.1 1.0 13.10 C 72.7 0.67 0.0092 6.1 3.4 67.6 16.4 12.4 1.32 16.4 1.1 7.00 Coefficients of Divergence, C. D., derived from the above data and means of 39 randomly selected samples from same populations values: Taxa with all samples with 39 samples species pair agegre-| aggre- Leaf cone} gate ||eaf | cone gate BC .87|.155| .169 |.189 | .145 |.165 P. monticola vs, P. strobus AC .206 .414 | .343 | -207 | 418 .346 P. chiapensis vs. P. strobus | | AB 343/424|.392 ||.348 | .423 |.393 P. chiapensis vs. P. monticola ' In the above summary of means let: A denote 60 foliage samples and 40 cone samples of Pinus chiapensis collected from individual trees found in southern Mexico, and northern Guatemala. B denote 120 foliage samples and 45 cone samples of Pinus monticola collected from individual trees found in southwestern Canada and the northwestern United States. C denote 300 foliage samples and 47 cone samples of Pinus strobus collected from individual trees found in the Lake States, and the northeastern and southeastern United States, and let following Q, to Qi» represent: Qı total leaf length in mm Q: leaf width in mm Q: ratio of leaf width to leaf length Q./Q, Q. number of serrations per 5 mm interval of leaf edge at leaf center 1966] Pinus — Andresen 21 Qo Qio Qu Qu To number of stomatal lines on vertral leaf surface total number of ovuliferous scales of mature cone width of scale apophysis in mm Length of scale apophysis in mm ratio of apophysis width to length Q:/Qs length of cone peduncle in mm involutions of edges of central scales, range from 0 to 4 number of reflexed basal-scales contiguous to peduncle; range from 0 to 18 calculate any C. D., the following general formula was used: IC BIET tas — bot Laie = bt CD. -4/ Babee 2) +... + (am ELEM k Ay ; ur Bi where: a, = AB) b= A. 4- B, for Q, etc. k — no. of characters, or 12 in this study Table 2. Means and standard deviations of characters Qi. . . Qu: for Pinus chiapensis (A), P. monticola (B), and P. strobus (C), derived from 39 random samples. F statistics for species pairs AB, AC, and BC incorporate 1 and 38 d.F. Char. chi- acter Pinus apensis Pinus monticola Pinus strobus F Statistics’ X g X s X s AB AC BC Qı 106.51 14.44 69.21 16.42 77.97 12.29 113.51 88.30 7.138 Q: 0.66 0.06 0.95 0.12 0.68 0.08 203.47 2.97 142.09 Q: 0.0062 0.0008 0.0148 0.0042 0.0089 0.0013 154.83 121.90 69.46 Q: 12.92 1.84 3.15 1.84 5.69 1.45 548.54 370.60 45.61 Qs 4.33 0.83 4.09 0.89 3.46 0.77 1.46 23.20 11.41 Qe 87.72 9.08 114.95 29.68 68.36 12.90 30.03 58.76 80.85 Q: 17.74 1.82 20.10 3.14 16.49 1.50 16.53 11.08 42.17 Qs 9.95 1.15 14.33 2.14 12.54 1.74 126.87 60.02 16.44 Qs 1.80 0.22 1.41 0 23 1.33 0.17 55.03 106.02 2.75 Qi 25.10 5.84 15.77 4.66 16.10 3.10 60.82 72.26 0.14 Qu 2.43 0.72 1.10 0.31 1.05 0.22 113.66 132.28 0.71 Qu 0.03 — 12.54 2.39 7.41 1.68 1064.13 794.19 122.13 ‘Values under 2.98 are non-significant. Value of 7.13 is signifi- cant at 5% confidence level. Values greater than 11.07 are significant at 196 confidence level. "The coefficient of variation (C. V.) discussed in the text is cal- culated from the formula C. V. = sX (100). For example in A — Q C. V. = (14.44)?/106.51 (100) = 13.6 22 Rhodora [Vol. 68 Table 3. Character reliability indices (R:)* and number of samples? of each taxon required to detect significant differences at the 5% and 1% level within species pairs AB, AC, and BC. Char- Reliability index Samples needed to detect differences acter AB AC BC AB AC BC 5% 1% 5% 1% 5% 1% Qı 983 978 754 3 5 4) 6 22 39 Q, .990 497 986 2 3 50 82 3 4 Q. 987 984 972 2 4 3 4 5 6 Q, 996 995 957 2 3 2 3 7 8 Q: 189 .917 .839 99 157 8 14 15 26 Qs 936 965 976 7 1 6 7 5 6 Q: 883 835 954 10 17 16 27 7 8 Q. 984 967 885 3 4 5 T 10 17 Qs 964 959 288 6 7 4 5 68 108 Qio 968 973 —.758 6 7 5 6 1852 2160 Qi: 983 985 —169 4 5 3 4 227 854 1 1 1 1 3 4 12 .998 | .997 .984 'Although Ri, as the intraclass correlation, is more often used to compare characters between genetically related individuals, it also serves here as a statistic to evaluate character weight, It is com- puted from the model R: = o o + g where oc? is the be- b B Ww B tween taxa component of the total variance and c? is the within wW component of the total variance for any one character. In AB—Q;, .983 means that 98.3% of the proportionate variance is related to differences between taxa rather than within taxa. High R+! values indicate “strong” characters of differentiation. ?Sample sizes were determined by solving for n in the formula n = 2 Vt/(X, — X.) which was derived from the model of “stu- dent's" t. In our sense, n — degrees of freedom or sample size, V — pooled variance of a species pair, t —estimated value from "student's" t-distribution, X: + X. = means of character X in species 1 and 2 respectively. "To calculate the number 4 proceed as follows: The variance (o?) of Q: for A is 207.36 and for C is 151.29 (pooled Vi + V; should be divided by 2), the mean leaf length for A is 106.5 and 78.0 for C; these values substituted in the n formula give: n — 2/2 (358.65) t?/ (28.5)? = .442t. Since .978 is a high R; and indicative of a low n, substitute n — 2 to find an empirical t; (t — 2n — 2): 1966] Pinus — Andresen 23 with 2 df so try 6 df so try 3 df 2 = .442U 6 — .442U 32> 4470 t for 4—2 df t for 12—2 df for 6—2 df at 5% — 2.78 at 5% = 4.30 at 5% — 2.23t 3 = .442 (1.13) 2 — .442(18.50) 6—.442(4.97) 3— 83.42 rounded = 4 2-—S8 S 6 — 2.20 so the conservative n — 4 but 218.18 but 612.20 with 4 df the “t” side was only 2.63, so the appropriate n is reached the first time the “t” side of the equation exceeds the “n” side. Table4. Internal leaf characters: number of resin canals and suc- culence represented by percents of moisture and intercellular space of P. chiapensis, P. monticola, P. strobus, and hybrid P. monticola X strobus (D). Resin Canals Succulence Taxa Mean Std. %o % No. dev. Range | Taxa Moisture Intercellular sp. D 1.89" 581 0-4 A 67.75** 407E B 2.25 .266 2-3 B 58.89 10 A 2.52 .207 2-3 C 58.76 10 C 2.57 176 2-3 D 58.26 10 ‘Taxa D and B or A and C not significantly different from each other in number of resin canals but former pair is significantly dif- ferent from latter at 1% level using the value of t. **Taxa A values significant at 1% level, B to D non-significant. Table 5. Seed and Seedling Characteristics of Pinus chiapensis, P. monticola, and P. strobus. Taxa Seed per kg Seed Germination Germination rs rate in days values' X Range Unstrat.* Strat. Unstrat. Strat. A 61,100 46,900 to 104,200 | 10 to 37 10to 37 10.35 10.35 B 52,300 39,400 to 67,800|10t01864- 13 to186+ 0.04 0.48 C 53,800 47,000 to 66,700|16t0186-- 13 to186+ 0.48 1.28 1Unstratified seed stored dry at 4°C. for 60 days. ? Seed stratified in moist peatmoss at 4°C. for 60 days. °Germination value = greatest cumulative number of seed that germinated in fewest days multiplied by total percent germinated, divided by total number of days in test period (McLemore and Cza- bator 1961). Rhodora [Vol. 68 Seedlings Hypocotyl Cotyledon Length mm Length mm Number Xx Range X Range X Range 39 20 to 70 28 20 to 40 8 6 to 10 20 15 to 35 18 14 to 25 8 7 to 10 31 18 to 44 25 15 to 30 10 8 to 12 CYTOGEOGRAPHY OF ORONTIUM AQUATICUM (ARACEAE)! JOHN W. GREAR, JR.* The Golden Club, Orontium aquaticum L., constitutes a monotypic genus of subfamily Calloidae Engl., an alliance also including the genera Symplocarpus, Lysichitum and Calla. These genera are interpreted as temperate or sub- arctic extensions of the predominantly tropical family Ara- ceae. All but Orontium are represented in eastern Asia and in North America, and thereby lend strength to the affinities existing between the floras of these regions. Morphologically, Orontium is regarded as a derived, highly modified genus without any near relatives. The geo- graphical distribution of the genus and its allies lends con- siderable support to this view. Orontium aquaticum is a perennial herb possessing 4 deep-seated, firmly-anchored, vertical rhizome. Its distri- bution (Fig. 1) is restricted to the eastern United States, where it occurs on the Coastal Plain in aquatic or semi- aquatic habitats such as bogs, marshes, streams and pools, and in similar habitats of the inland physiographic prov- inces, sometimes up to 2800 ft. in the mountains. However, lack of suitable habitats has localized its distribu- tion in the Appalachian Highlands. To date there has been no comprehensive morphological study of Orontium. Harshberger (1916), in his account of the vegetation of the New Jersey pine barrens, described and illustrated the anatomy of the leaf in consider- able detail. In a meticulous study of the floral structure, *Present Address: New York Botanical Garden 1Condensed from a thesis submitted to the Department of Biology in partial fulfillment of the requirements for the degree of Master of Science in the Department of Biology, Vanderbilt University, January 1963. 25 26 Rhodora [Vol. 68 Fig. 1. Map showing distribution of Orontium based upon records documented by herbarium specimens. The southernmost extent of glaciation is shown by the dotted line. The heavy broken lines sepa- rate respectively the Coastal Plain, Piedmont and Appalachian high- lands, including the Blue Ridge, Ridge and Valley and Appalachian Plateaus provinces. Schaffner (1937) pointed out inaccuracies of previous descriptions and reported a wide range of variation among the flowers on a single spadix. In a taxonomic study of the temperate North American Araceae, Huttleston, (1953, unpublished) found the chromosome number of Orontium to be 2n—28, a number different from that previously found by Malvesin-Fabre (2n—24, 1945, un- 1966] Orontium — Grear 27 published) see Darlington and Wylie, 1956. In the treat- ment of the Arales for the generic flora of the Southeastern United States, Wilson (1960) summarized the existing in- formation concerning Orontium, indicating that the mor- phological and biological details of this plant warranted further investigation. The present investigation was prompted by conflicting statements about the chromosome number as well as the relationship between this data and the distribution of the plant in all major physiographic provinces of the Eastern United States. It was further felt that by correlating cyto- logical with geographical data, some insight into migra- tional patterns might be gained. Through the correlation of past geological history with cytological, morphological, taxonomical and ecological evidence, it is sometimes possible to postulate an earlier distribution and possible migration route, which would ex- plain the present distribution of Orontium. The Araceae are a predominantly tropical family and during the Ter- tiary period, when tropical conditions prevailed farther northward, a wide range of plants such as temperate coni- fers, deciduous hardwoods and warm-temperate plants, including araceous plants, ranged to high northern lati- tudes, far beyond the limits of the present tropics. In Symplocarpus and Lysichitum, the two genera most closely related to Orontium, the migration route to present ranges appears to have been from the region of the Arcto- Tertiary flora of the then tropical boreal regions of west- ern North America and eastern Asia (Rosendahl, 1911). Subsequent climatic changes forced them to migrate south- ward. Symplocarpus apparently migrated to eastern North America while Lysichitum remained in the West. As with Orontium, plants in both of these genera require a very moist or wet habitat. Symplocarpus foetidus (L.) Nutt., constituting a monotypic genus, occurs in two widely dis- junct areas — eastern Asia and eastern United States. Lysichitum consists of two closely related species: L. cam- tschatcense (L.) Schott, the Asiatic one, occurs in Japan 28 Rhodora [Vol. 68 and Kamtschatka, and its American counterpart, L. ameri- canum Hultén and St. John, is distributed in western North America between California and Alaska (Hultén & St. John, 1931). Fossil evidence substantiating a more extensive dis- tribution for araceous plants has been found in Spitsbergen (presumably as imprints) referable to extant species. An aquatic aroid from the Tertiary flora, Acorus brachysta- chys, has been described recently by Schloemer-Jager from the Brogger Peninsula of the same island (Andrews, 1961). Cockerell (1926) gives evidence supporting a once more westerly distribution of Orontium in North America. The relevant specimen consisted of a spadix, incomplete at the distal, fertile end, and its sterile stalk. Although the identity of the specimen is uncertain, it was sufficiently characteristic, according to the author, to be referred to this genus. The only character found to separate the fossil plant from the modern species was a lack of thickening of the scape below the spadix. Cockerell states that “this fossil may take the name Orontium fossile n. sp.; it adds one more to the numerous examples of genera now existing in the Eastern and Southern states, but found in the Rocky Moun- tain Region only in the fossil state.” If Orontium and related genera migrated from the re- gion of the Arcto-Tertiary flora, then they were probably distributed over a wide range during the late Cretaceous and early Tertiary, from the east coast of the continent to the region of Colorado or perhaps even further west, under the influence of the equable climate of the era. At that period, the Appalachian Uplands were essentially a pene- plain, affording ideal conditions for coastal-plain types such as Orontium in the streams and swampy areas. Uplift of this old Cretaceous (Schooley) peneplain during the Mio- cene epoch, and the accompanying climatic changes and erosion cycles resulted in greatly altered conditions. The climate to the west became much more arid, eliminating possible populations of Orontium west of the Mississippi River. A more restricted distribution was placed upon the 1966] Orontium — Grear 29 species in the Appalachian Region and only relict popula- tions survived in the favorable mesophytic and hydro- phytic habitats of the newly elevated highlands. With the later recession of the sea, the newly-emergent Coastal Plain became available to the migrating species from the up- lands, which found optimum conditions in the bog-marsh habitats. An account of the Coastal Plain element in the flora of the Appalachian uplands, correlated with physio- graphic history of the provinces concerned, has been de- scribed thoroughly by Kearney (1900), Harshberger (1903), Fernald (1931, 1937), Braun (1937), Carr (1938), Core (1938), et al. Observation of the map (Fig. 1) shows that the present range of Orontiwm has extended into the formerly glaciated region in some of the Northeast. Peattie (1922) reports the occurrence of Orontium from as far inside the glacial boun- dary as Lake Oneida, New York. During the glacial and interglacial stages of the Pleistocene, the climatic fluctua- tions probably caused a retreat of the species to refugia farther south of the advancing ice. From these refugia, when the ice receded for the final time, probably it migrat- ed to its present northernmost limits. As stated earlier, varying chromosome numbers for Or- ontium have been reported. Huttleston (1953) has counted 28 somatic chromosomes from root tip squashes of four plants, two from Duval County, Florida, and two from Middlesex County, New Jersey. Both of these locations are in the Atlantie Coastal Plain. A different count, of 24 chromosomes, has been reported by Malvesin-Fabre (1945) in a thesis submitted to the University of Bordeaux, France. As this work was unobtainable, the source and the state of the studied plant material are not known to the present writer. However, the possibility of different numbers sug- gested that some correlation might exist between chromo- some number and geographical distribution. It was planned originally to work with meiotic material, but flowers from plants collected in early April of 1961 30 Rhodora [Vol 68 Fig. 2. Chromosomes of Orontium. First meiotic metaphase, Ma- terial from Fentress County, Tennessee. X approx. 500. had already undergone meiosis. In collections made in early September of the same year, meiosis had already occurred in young flowers that would not make their appearance until the following spring. It appears that the greater part of the meiotic process probably occurs sometime during the summer season, more than eight months in advance of of anthesis the following spring. Only one determination of chromosome number was made from meiotic material. This was obtained quite by chance from the tip of an inflorescence collected in Fent- ress County, Tennessee, April 1961, in which meiosis was complete. A few pollen mother cells of the young sporo- genous tissue showed 26 chromosomes at the first meiotic metaphase, the number not yet having been reduced. The inflorescence was collected in Newcomer’s solution and the squash was made using acetocarmine. Some of the chromo- somes exhibited a tendency to stick together, as evidenced by strand-like connections between them (Fig. 2). Because of the uncertainty relating to the exact time of meiosis, all other work was done with root tip chromosomes. Living samples of Orontium, variously consisting of fruits, seedlings and mature plants, were obtained from various sources for root tip material. The samples were maintained in the green house in concrete water tanks dur- ing the course of investigation. Fresh, vigorously-growing 1966] Orontium — Grear 31 root tips about 1 cm. in diameter were excised immediately before treatment. The live tips were pretreated with a-bromonapthalene to inhibit spindle formation, shrink the chromosomes and render the centromeres more evident. The method of Tijo and Levan (1950), utilizing 8-hydroxyquinoline as a pre- treatment agent, proved less effective in fixing the meta- phase than a-bromonapthalene even though other conditions were constant. The root tips were then killed and fixed with aleohol-acetic acid (3:1) and stained with aceto-orcein after the method of LaCour (1941). The meristematic tip of the root was then excised and squashed in a drop of 45% acetic acid. Mitotic chromosome counts were obtained from 24 dif- ferent populations and in every instance the diploid number was 26. Neither of the previosly reported numbers (2n — 24, 28) was ever found, although attempts to verify Huttleston's Duval County report were made using material collected from the same site. The chromosomes are very large, intensely-staining and easily observed where the squash preparation is good. The position of the majority of the centromeres is median (iso-brachial), a feature re- garded by some workers (e.g. Levitzky, 1931) as represent- ing the generalized or primitive condition. The chromosomes of one pair have clearly discernable satellites (Fig. 3). A tendency toward fragmentation at the centromere may ac- count for Huttleston's report of 2n = 28. However, this was a random occurrence in the present study. Because the chromosome number appeared the same in all plants examined, regardless of source, consideration was given karyotypic variation among populations. Graphic portrayal of chromosome arm length was attempted, but, because of differential shrinkage and swelling of the chromo- some arms it was not possible to match homologous chromo- somes. Therefore, no correlation was found between quantitative chromosomal data and the geographical source of the samples studied. The only two chromosomes that could be paired with any degree of certainty were the two 32 Rhodora [Vol. 68 = Ó Ea Dm Z e c 3 Fig. 3. Chromosomes of Orontium. Mitotic chromosomes. Material from Cameron Parish, Louisiana. X approx. 700. with satellites. Until a more reliable procedure is developed, permitting assessment of differences in linear dimensions, the quantitative information must be considered insignifi- cant. It appears that Orontiwm chromosomes would be par- tieularly suitable for differential staining (see Fukuda and Kozuka, 1958, for work on Trillium), and that the homolo- gous pairs of chromosomes might thereby be identified. There was some indication of differential stainability of the chromosomes of Orontium during the present investigation and this suggests the possibility of relating the populations to distribution. SUMMARY Information obtained from this study contributes to knowledge of the cytology of Orontium, a monotypic genus of the Araceae. The diploid chromosome number of all populations studied is 26, irrespective of geographical 1966] Orontium — Grear 33 source. Two satellites or secondary constrictions are evident in every karyotype and the chromosomes are extremely large (ca. 204), exhibiting essentially median centromeres. DEPARTMENT OF BIOLOGY VANDERBILT UNIVERSITY, NASHVILLE, TENNESSEE ACKNOWLEDGEMENTS Appreciation is due the following persons, who supplied living ma- terial as well as valuable information regarding distribution: Dr. Wade T. Batson (University of South Carolina), Dr. Harold D. Ben- nett (West Virginia University), Dr. Edward Davis (University of Massachusetts), Mr. J.B. Ebert (Pembroke State College), Dr. Robert Kral (Virginia Polytechnic Institute), Dr. Carl Monk (University of Florida), Dr. Herman O'Dell (University of East Tennessee), Dr. James D. Ray Jr. (University of South Florida), Dr. William Reese (University of Southwestern Louisiana), Drs. Elsie Quarterman, R. B. Channell, Howard F. L. Rock, Messrs Yoshimichi Kozuka, H. Oliver Yates and William Brode (Vanderbilt University). I am particularly indebted to the late Dr. Howard F. L. Rock, supervisor of this work, and to Dr. R. B. Channell, who suggested the problem, for their advice and encouragement, and to Dr. Howard S. Irwin for critical evaluation of the paper. The writer wishes to acknowledge the curators of the following herbaria for the loan of material: ALU; BUS; CINC; DUKE; FSU; GA; GEO; GH; IND; KY; MASS; MISSA; NCU; NO; NY; SMU; TENN; TEX; UARK; US; USF; VDB; WVA; UNIVERSITY OF SOUTHWESTERN LOUISIANA. LITERATURE CITED ANDREWS, H. N. 1961. Studies in Paleobotany. New York: John Wiley & Sons. p. 381. BRAUN, E.L. 1937. Some relationships of the flora of the Cumberland Plateau and Cumberland Mountains in Kentucky. Rhodora 39: 193-208. Carr, L.G. 1940. On the “residual” interpretation of coastal floras in the Appalachian uplands of eastern United States. New Phytol. 39: 129-132. COCKERELL, T.D.A. 1926. A miocene Orontium (Araceae). Torreya 26: 69. Cork, E.L. 1938. Plant migrations and vegetational history of the southern Appalachian region. Liloa 3: 5-29. 34 Rhodora [Vol. 68 DARLINGTON, C.D. & A.P. WYLIE. 1956. Chromosome Atlas of Flower- ing Plants. New York: The Macmillan Co. p. 374. FERNALD, M.L. 1981. Specific segregations and identities in some floras of eastern North America and the old world. Rhodora 33: 25-63. —————————— 19387. Local plants of the inner Coastal Plain of southeastern Virginia. Rhodora 39: 266-321, 379-415, 438-459, 465-491. FUKUDA, I. & Y. Kozuka. 1958, Evolution and variation in Trillium, V. A list of chromosome composition in natural populations of Trillium kamtschaticum Pall. Jour. Fac. Sci. Univ. Hokkaido Ser, V. 6: 2779-319. HARSHBERGER. J.W. 1903. An ecological study of the flora of moun- tainous North Carolina. Bot. Gaz. 36: 241-258, 368-383. 1916. The vegetation of the New J ersey Pine Bar- rens. Philadelphia: Christopher Sower Company. pp. 265-260. HULTEN, E. & H. St. JoHN. 1931. The American species of Lysichitum. Sv Bot. Tidskr. 4: 453-464. HUTTLESTON, D.G. 1953. A Taxonomic Study of the Temperate North American Araceae. Cornell. (Thesis). KEARNEY, T. H. 1900. The lower austral element in the flora of the southern Appalachian region, Science 12: 830-842. LEVITZKY, G. A. 1931. The karyotype in systematics. Bull. Appl. Bot. Genet. Plant Breed. 27: 220-240. MALVESIN-FABRE. 1945. Contribution a la caryologie des Araceae. Bordeaux. (Thesis). PEATTIE, D.C. 1922, The Atlantic Coastal Plain element in the flora of the Great Lakes. Rhodora 24: 57-70, 80-88. ROSENDAHL, C.O. 1911. Observations on the morphology of the under- ground stems of Symplocarpus and Lysichiton, together with some notes on geographical distribution and relationship. Minn. Bot. Stud. 4: 137-152. SCHAFFNER, J. H. 1937. The flowers of the golden-club. Am. Bot. 43: 99-103. WILson, K.A. 1960. The genera of the Arales in the southeastern United States. Jour. Arnold Arb. 41: 47-72. AUTHORS OF PLANT GENERA AND THE INTERNATIONAL PLANT INDEX REED C. ROLLINS An impressive volume printed with the aid of computers and assembling in one place a lot of information heretofore somewhat scattered has recently appeared.’ However, most botanists will not find the information to be of great in- trinsic value and an occasional consultation of the work will probably suffice for their purposes. The bulk of this paperbound book is devoted to an alpha- betical listing of the “authors” accompanied by four columns under the headings, AUTHOR CODE, BORN, COUNTRY, and DIED. This is followed in double-column format by 62 pages of “author” listings, giving the author code in each instance. The final section gives a three-column listing of name ab- breviation, author name and author code in a double-column format, followed by an addenda to each of the three lists. If botanists did indeed take up and use the author code- order index for citation of author names of plants, as urged by Mr. Gould, it would be a radical departure from past practice. It would also involve giving up to the machine the essences of familiarity which link plant names with the men who gave them to the plants. We now have no diffi- culty in remembering that Schlecht. stands for Schlechtendal or Schott for Schott but how many botanists will remember that SCH794 is Schlechtendal and SCO794 is Schott? How can one be expected to remember that SHO904 is A. J. Sharp and SHA904 is W. M. Sharp? These unnecessary abbrevi- ations reflect the choice by the authors of an inadequate system to handle the information they are trying to present. :Authors of Plant Genera, by Sydney W. Gould and Dorothy C. Noyce. International Plant Index 2: 1-336. 1965. The New York Botanical Garden, New York; The Connecticut Agricultural Ex- periment Station, Box 1106 (IPIx) New Haven, Connecticut 06504; $6.00 in U.S.A. and Canada, $6.50 in other countries. 35 36 Rhodora [Vol. 68 It can be argued that having the names of the men of historical botany closely associated with the names of the plants they helped reveal to the public is of no great moment. However, it is not as simple as that. The name of the author, if it is known and recalled, helps put the name of the plant into its historical setting, which in turn helps with the geography and other matters concerning the plant itself. The richness of the presently used name content, with ci- tations that are well understood and appreciated, will be greatly denuded by the device offered in the book under review. In addition, one will be forced to look up, on every occasion, the correct author code. This will be a nuisance for taxonomists; for non-taxonomists, it will make author citations both incomprehensible and ridiculous. Are citations such as Rhododendron canadense (LIN707)BSP859 or Buchloé dactyloides (N UT786) ENG809 to the advantage of either taxonomy or plant sciences generally? many persons who have not named any plant taxon what- ever. Many are not botanists at all. One finds librarians, zoologists, friends of the authors and the authors themselves, none of whom named genera, Is this the kind of capricious- ness that should characterize an accurate scholarly work? Do the authors think they are playing games with the ma- 1966] Plant Index — Rollins 37 terials of science? What then are the criteria for the in- clusion of a name in the “Authors of Plant Genera”? ACCURACY AND COMPLETENESS: One reliable and readily available source that should have been a major check-point for “Authors of Plant Genera" is Index Nominum Generi- corum, published by the International Association for Plant Taxonomy. Apparently this recent authentic work, now only partially published but having over 21,000 cards in it, was hardly utilized at all. This circumstance permits us to use it as a check for the accuracy of the book under review. The alphabet from Aa to Af, amounting to approximately 400 cards of Index Nominum Genericorum, was checked against “Authors of Plant Genera” by a colleague. This number turned up six errors Or omissions, of which five were complete omissions of botanists who had described genera of plants. If this rate of omission holds when checked against the rest of the Index Nominum Genericorum presently available, it means that over 250 omissions of authors who named plant genera were carelessly left out of the book being billed as “the most complete work of its kind.” If Index Nominum Genericorum were complete, one wonders how many thousands of names that legitimately should have been included in “Authors of Plant Genera" would be shown up by a check against it. Checking for accuracy is a time consuming and unin- spiring enterprise and one tends to turn to the quickest way of seeing whether the data given in à book stands up or not. We noticed that Leslie Andrew Garay and Clarence Garay are the same in the author code, page 86, and that he is alleged to be from Canada. Leslie Garay, who works nearby, tells me that he has never published under the name of Clarence Garay and has not been known under that name. Furthermore, he has been in the United States since 1957. and was born in Hungary. Although he did spend some years in Canada, the information on Garay is both inac- curate and misleading. We know that Thomas Coulter was born, lived and died in Ireland, but on page 69 he is listed as England. On the 38 Rhodora [Vol. 68 other hand, N. L. Bor is given as Ireland, although he Specialized on the grasses of southeast Asia, spent much time there, and for many years he has been at Kew in England, where he has been professionally identified right down to the present. Oddly enough, although Edward Perceval Wright was born, worked and died in Ireland, ap- parently on the basis of a visit to the Seychelles Islands for six months in 1867, "Seychell. has become his "country of principal work or home." In short, every entry under this column must be checked by the user to determine its meaning and accuracy, if these are important to him. Under the column "Country" the strangest designations appear. For example, S. America is given as a country many times, but so also are Argentina, Brazil, Colombia, etc., the true countries of South America. It must seem strange to the well known Argentinian botanists, Professors Arturo Burkart and Lorenzo R. Parodi, to read that their country is "S. America." I am Surprised to see listed as countries: Patagonia, Sahara, Transvaal, South West Africa, etc. Certainly these subversions cannot be read as a sign of erudition on the part of the authors. One is led to ask, how can such sloppiness be supported as a crusade ostensibly to help librarians, taxonomists and plant scien- tists ? THE INTERNATIONAL PLANT INDEX: The two volumes so far published by IPIx, as poorly prepared as they are, can scarcely be said to have done any real harm. This is so because the content of each of them is readily found else- where in reliable sources and the real scholar will soon discover that the volumes are not worth bothering with, in spite of the exorbitant claims to the contrary. However, I am alarmed by the forecast that the third volume will be "Genera of the Plant Kingdom." If such a work is published with the same lack of understanding of botanical nomen- clature and taxonomy, and the same disregard for the In- ternational Code of Botanical Nomenclature that is evident in the published work so far done by the people at Inter- national Plant Index, it will be a disaster much worse than 1966] Plant Index — Rollins 39 the appearance of the infamous volumes of Otto Kuntze? in the 1890's. We have no quarrel with the use of computers or other data-processing machines to help organize and produce an index of plant names. Properly handled, such a single com- plete index would not only be extremely useful in itself but it could be the basis for encoding a wide range of botanical information that could then be made readily available. We part company with the organizers of IPIx when they insist on going beyond a simple index, yet use only books (and secondary sources, at that) as the basis for value judge- ments concerning the taxonomy of the plants whose names they are handling. In Family Names of the Plant Kingdom,’ it is a simple fact that a value judgement was made every time a name was selected for use as as operative name. A taxnomic judgement was made every time a name was listed as a synonym. These are facts even though the authors specifically state that the contrary is the case. The hundreds of value judgements made by the author of *Family Names" were completely unsupported and evi- dently were made without any real knowledge of the plants involved or of any reasons why one course of action should be taken in preference to another. The use of the word "operative" or some other designation such as “correct” or “in use” does not lessen the responsibility for making an acceptable choice. To be acceptable the name chosen has to stand up to the requirements of the science. It is not the business of an indexer to make choices where taxonomic judgements are involved, unless he can produce the scientific evidence to back up the choice. Quite obviously, Family Names of the Plant Kingdom is not an unbiased index and it cannot, therefore, be reliably used as an index. As à HE orcum ?Revisio Generum Plantarum. Pars I-III plus supplement. 1891- 1898. *Family Names of the Plant Kingdom, by Sydney W. Gould. Inter- national Plant Index 1: 1-111. 1962. For a review see Botanical Nomenclature, Punched Cards, and Machines, by C. E. Wood, Jr., R. S. Cowan and G. Buchheim. Taxon 12: 2-12. 1963. 40 Rhodora [Vol. 68 reliable source of the proper name of a family, it is com- pletely hopeless. In his writing, Mr. Gould has lectured to taxonomists, quoting his favorite authors and giving naive analogies to telephone communication and space science. Unfortunately he has kept himself immune from even an elementary under- standing of the basic ideas and tenets of taxonomy and nomenclature, the subjects in which he is apparently trying to attain an authoritative position. The real danger that lies ahead arises from the fact that those most deeply committed to IPIx are untrained for the task they are trying to perform and will not listen to any- one knowledgeable enough to help them. Their supporters and advisers are equally lacking in an understanding of the problems involved. If the International Plant Index continues in the path so far laid down, taxonomists and botanists generally will be saddled with a nomenclatural mess, a heritage largely from those who have willfully or unwittingly given their support to IPIx. GRAY HERBARIUM HARVARD UNIVERSITY STUDIES IN THE CAPPARIDACEAE' VIII. POLANISIA DODECANDRA (L.) DC.: HucH H. ILTIS? FURTHER NOTES ON ITS TYPIFICATION. In two studies in this series (Iltis 1954, 1958) it was maintained that the common American Polanisia graveolens Rafinesque should properly be called P. dodecandra (L.) DC., the latter based on Cleome dodecandra of Linnaeus's Species Plantarum (1753). However, in a recent generic review of the Caper family for the Southeastern United States, Ernst (1963: 90-91, footnote 6) casts serious doubts on the validity of this disposition, suggesting instead that the C. dodecandra L., based as it is on a description in the Flora Zeylanica (No. 242, p. 109), a work based in turn on Paul Hermann's Ceylon plants and drawings, is actually a hitherto unrecognized or misunderstood Asiatic species which should not be equated with the well-known American Polanisia graveolens Raf., the two being presumably totally different taxa. The present paper attempts to show that on the basis of the critical collections housed in London (and 1[Ed.: Published as Capparidaceae rather than Capparaceae at the request of the author.] ?I am much obliged to the Botanical Society of America for a grant- in-aid to attend the Xth International Botanical Congress in Edin- burgh, and thus be privileged to visit London and its herbaria, and to Dr. Wallace Ernst, U. S. National Herbarium, Smithsonian Insti- tution, Washington, for the careful presentation of his arguments in correspondence. The staffs of the Linnaean and British Museum herbaria, especially J. E. Dandy, Keeper of Botany, N. K. B. Robson, and D. Hillcoat have been most generous with their help, as has been my colleague John W. Thomson and particularly Mr. Dandy for con- structive and instructive criticisms of the manuscript. To them, my very sincerest thanks. This research was supported in part by Re- search Committee of the University of Wisconsin on Funds from the Wisconsin Alumni Research Foundation. 41 42 Rhodora [Vol. 68 thus on somewhat different grounds than previously), and on a careful reexamination of the texts involved in this argument, my original conclusions must still be considered valid: namely that Cleome dodecandra L. and Polanisia graveolens Raf. are synonymous. Therefore, P. dodecandra (L.) DC. is still the correct name for the American taxon. The original Linnaean citation (Sp. Pl. ed. 1, 2: 672) of Cleome dodecandra reads as follows: "dodecandra. 5. Cleome floribus dodecandris, foliis ternatis. Fl. Zeyl. 242*". As was fully discussed in earlier studies (Iltis 1954, Ernst 1963), the two other following citations refer to other species. The star after Flora Zeylanica indicates the pres- ence of a good description in that work. As to the origin of the species, he reported it as “Habitat in Indiiis" meaning by that the warm regions of both Asia and America, as shown by the inclusion in synonomy of a Sloane plant from Jamaica. As was pointed out previously (Iltis 1954), the short de- scription in the Flora Zeylanica is nevertheless a vivid. and. excellent one (apparently having been made from living plants), and, with its reference to 8 stamens, 3-foliolate leaves, emarginate white petals, red pistil, unilateral gland and thick, hispid capsules, fits exactly only one taxon in the world, the American Polanisia graveolens. Thus, while Lin- naeus' total conception of his species was confused if his diverse synonomy is considered, his own personal descrip- tion of the plant in question is brilliantly clear. "While the third and last of these Linnaean polynomials clearly refers to the New World C. serrata Jacq., the second polynomial (and reference to “Burm. Zeyl. 216, t. 100 f. 1") was thought to refer to perhaps C. burmannii W. and A., or to C. rutidosperma DC., though, truly, the drawing is unidentifiable. In the Geneva Herbarium are several "Herbier Burmann" collections, including one that matches this drawing — a depauperate miserable specimen, about 15 cm tall, with one hidden flower, labelled “Cleome dodecandra" [by Burmann?], a plant that conceivably could be a young C. viscosa or C. aspera, or [doubtfully], as Briquet had annotated it, C. burmannii. What- ever taxon this specimen belongs to, the plant is much too poor for Linnaeus to have used in drawing up the excellent Flora Zeylanica description, and can thus not be considered type material. 1966] Polanisia — Iltis 43 However, the Flora Zeylanica is not based on American collections, but on collections of plants and drawings made in Ceylon in ca. 1670-1677. But how can we reconcile a tem- perate American plant in tropical Ceylon? Fortunately, these drawings and plants of Hermann’s are still preserved in the British Museum of Natural History, where they may be studied. They are thus of immense scientific value. In the Flora Zeylanica, to quote Stearn (1957 "1183"... only phrase-names are used but, as pointed out by Trimen (1887), when dealing with Ceylon plants in the Species Plantarum, ‘Linnaeus was careful to quote under them the number of the Fl. Zeylan., and thus the specimens of Her- mann’s herbarium become types for many of Linnaeus’s species . . . especially as the large majority of the species in Hermann’s herbarium are unrepresented in Linnaeus’s own collection.’ " Thus, when typifying a species described in the Flora Zeylanica, one must consult the Hermann her- barium! Recently an opportunity presented itself to make a search for the type of Cleome dodecandra, and, with the generous help of Miss D. Hillcoat, the four volumes of the Hermann herbarium were examined page by page, specimen by specimen. Of the Cleome species cited in the Flora Zey- lanica, the following were found to be represented by speci- mens: C. gynandra, Fl. Zeyl. No. 239 (in vol. I, p. 1, and vol. Hb: 754 C. viscosa, Fl. Zeyl. No. 241 (in vol. I, p. 75, and vol. III, p. 2); C. monophylla, Fl. Zeyl. No. 243 (in vol. I, p. 52 and 57, and vol. III, p. 2). There were no specimens for C. icosandra, Fl. Zeyl. No. 240, and, more importantly, there were none for C. dodecan- dra Fl. Zeyl. No. 242! Next, an examination of Hermann's Icones volume showed but two Cleome illustrations, one on page 421 (numbered 242 — was this number entered there 44 Rhodora [Vol. 68 by Linnaeus.'), the other on page 24, neither of which could be C. dodecandra. Both appear to represent normal plants of Cleome aspera, the latter figure (p.24) but a small- scale copy of the first, and matching very well such Cleome aspera collections as Beddome 203 from Madras, India (BM!) or the C. aspera type (K!). Thus, neither Her- mann's drawings nor his herbarium contain any material referable to Linnaeus's Cleome dodecandra. Futhermore, we can follow neither Trimen (1887:146), who equated No. 242 of the Flora Zeylanica (i.e. C. dodecandra, with the first of these drawings (No. 242), as “C. viscosa L. var.", nor Ernst (1963), who, following Trimen, suggests that this drawing should be taken as a type of C. dodecandra. For these draw- ings, both evidently of the same plant, represent C. aspera, a 6-staminate, glandless, slender-fruited species, and in no way agree with Linnaeus's excellent descriptions of C. dode- candra. As a matter of fact, how could Linnaeus have obtained all this detail, as well as flower and pistil color, from this simple drawing? At this point Mr. Dandy very kindly (and literally) came to the rescue, for in examining the Flora Zeylanica text (p. 109) he immediately noted a very significant omission in the Cleome dodecandra description, one that neither I nor Ernst was aware of: namely that Linnaeus does mot cite (as he usually does in the Flora, Zeylanica) any collec- tion or plate of Hermann’s Musaeum Zeylanicum following the description of C. dodecandra! This can only mean that Linneaus did not have a Hermann specimen in front of him when he drew up this description, but rather some plant of different origin, a plant which he for some reason must have believed should belong in a book on Ceylon plants. Very possibly, he equated one of his garden plants with Burmann's Ceylon drawing which, being based on a Ceylon plant, was to be included in Flora Zeylanica whether or not ‘According to Trimen (1887:130), “In the [Hermann] herbarium itself he [Linnaeus] has added to Hermann's labels a reference to the number of the species in his own ‘Flora Zeylanica. " 1966] Polanisia — Iltis 45 Hermann had collected it. In any case, he must have had his reasons, for the inclusion in the Flora Zeylanica of plants other than those collected by Hermann in Ceylon is a rare (though not unknown) occurrence. That an error on the part of Linnaeus is involved here was deduced earlier on somewhat different grounds (Iltis 1954: 67, footnote). Since Linnaeus evidently had a plant available when writing the description of C. dodecandra, we are now again forced to look for a specimen, and the logical place to seek is in the Linnaean Herbarium. There, the one plant that clearly fulfills all requirements is sheet 850.12 (Savage 1945). In my earlier studies I could only guess at the iden- tity of this plant from the poor photograph then available. Now, however, after an examination of the actual speci- men, there is no doubt that this plant is indeed the same as Polanisia graveolens Raf., with the emarginate petals, and at least 8 stamens (inflorescence very young and not vigor- ous) clearly visible. Since Linnaeus mentions “capsula crassa, hispida”, he must have had in hand (or in memory) the characteris- tically thick and hispid-fruited plants of this species, prob- ably from his Uppsala garden, where the species was evidently in cultivation, the specimen in question (850.12) having a small tag glued to its stem that reads “H. U. 12 andr” [Hortus Upsaliensis dodecandra]. This cultivated plant, therefore, we must take as the type of Cleome dode- candra L.: it is clearly the only specimen that agrees (and then exceedingly well) with the original description, and is furthermore labelled by that name with a tag that we have reason to believe is contemporaneous with the publication of the Flora Zeylanica (1747) or earlier. As to the country of origin, we can only guess the Northeastern United States or adjoining Canada. Linnaeus evidently did not know the source, but believed it to be identical to that of the Burman and Boerhaave plants from tropical Asia cited in Flora Zeylanica. Complicating this story are the whole series of errors that occurred in the labelling (contemporaneously or subse- 46 Rhodora [Vol. 68 quent to 1753?) of the Cleome collections in Linnaeus's herbarium as well as in his texts. Thus: 1) he was mistaken in later writing on sheet 850.12 the name “viscosa” (though the older tag label says *H. U. 12 andr"!). Whoever scrib- bled this reidentification on the bottom of this sheet (and presumably it was Linnaeus himself ), he must have suffered a lapsus calami, for this specimen is so unlike any of the other sheets labelled “viscosa” that the identification ap- pears absurd; 2) he was mistaken when he wrote ''dode- candra" on sheet 850.14 which is clearly C. viscosa, and which in every respect matches the other C. viscosa sheets in the Linnaean Herbarium (e.g. 850.11, 850.13) ; 3) he was mistaken when he wrote “dodecandra” on sheet 850.7, which carries a good specimen of C. serrata Jacq. (C. polygama L.), only later to scratch this name out, and to substitute the correct name “polygama” after it; 4) lastly he made the curious error of misquoting the word octandris as dodecan- dris when copying the original definition (polynomial) of No. 242 from the Flora Zeylanica to the Species Plantarum, and then taking the specific epithet dodecandris from this miswritten adjective. It is an error easy to understand, however, for the young specimen (850.12) that he studied, as young specimens of that species are apt to, had 8 stamens, while no doubt on the somewhat older plants in his garden he observed the commoner higher stamen number, hence the later and very descriptive epithet “dodecandra.” SUMMARY Cleome dodecandra L., though published in Flora Zey- lanica, is based on cultivated American materia] grown in Hortus Upsaliensis by Linnaeus and now preserved in the Linnaean Herbarium. This material is conspecific with Polanisia graveolens Raf. The correct name of this taxon is, therefore, as was maintained earlier (Iltis 1954) on less certain evidence, Polanisia dodecandra (L.) DC., or, if re- tained in Cleome, C. dodecandra L. 1966] Polanisia — Iltis 4T POSTSCRIPT THE SUBSPECIES OF POLANISIA DODECANDRA. A reinterpretation of evidence presented earlier (Iltis 1958, esp. Figs. 6 and 9-21), and observations of some of these taxa in the field, demands that the Mexican and South- western P. uniglandulosa (Cav.) DC. should be considered a perfectly distinct species on morphological as well as geo- graphic grounds, and not as a subspecies of C. dodecandra,” though the two are evidently closely related. On the other hand, P. dodecandra and the western P. trachysperma be- have like geographical subspecies, the former evidently a taxon of more recent evolutionary origins, restricted to the glaciated northeastern United States and southern Quebec, but intergrading with the latter in a broad belt from Mis- souri to Minnesota. The species is therefore composed of P. dodecandra (L.) DC. ssp. dodecandra and P.d. ssp. trachysperma (T. & G.) Iltis stat. nov. (— P. trachysperma T. & G. Fl. N. Am. 1: 669. 1840). HERBARIUM, DEPARTMENT OF BOTANY UNIVERSITY OF WISCONSIN, MADISON, WISCONSIN LITERATURE CITED BURMAN, J. 1737. Thesaurus Zeylanicus. Amsterdam. EnNsT, W. R. 1963. The genera of Capparaceae and Moringaceae in the Southeastern United States. Jour. Arnold Arb. 44: 81-95. Iuris, H. H. 1954. Studies in the Capparidaceae I. Polanisia dodecan- dra (L.) DC., the correct name for Polanisia graveolens Rafin- esque. Rhodora 56: 65-70. Pl. 1201. . 1958. Studies in the Capparidaceae IV. Polanisia Raf. Brittonia 10: 33-58. LINNAEUS, C. 1747. Flora Zeylanica. Amsterdam. . 1753. Species Plantarum. Holmiae. SAVAGE, S. 1945. A Catalogue of the Linnaean Herbarium. London. STEARN, W. T. 1957. Introduction, in Linnaeus's Species Plantarum. Facsimile of first edition. Ray Society 140: 1-176 (esp. p. 119). TRIMEN, H. 1887. Hermann's Ceylon Herbarium and Linnaeus's Flora Zeylanica. Jour. Linn, Soc. London, Bot. 24: 129-155. 5T am obligated to Dr. W. L. Brown, Jr., Museum of Comparative Zoology, Harvard University, for calling some of the pertinent rea- sons for this interpretation to my attention. POLANISIA DODECANDRA IN NEW HAMPSHIRE" Polanisia dodecandra (L.) D. C. of the Capparaceae, not previously recorded in New Hampshire, has been found growing in Durham, New Hampshire. In New England, this species was reported only from the Lake Champlain region of Vermont by Bean, Hill, and Eaton (Rhodora, 63:53, 1961). The records, however, are based essentially on specimens in the Harvard Herbaria. In Connecticut, Weatherby reported it from the banks of the Hockanum River in East Hartford, and Driggs, from the vicinity of the Connecticut River at Hartford. These reports are cited in The Catalogue of Flowering Plants and. Ferns of Con- necticut, State of Connecticut Geological and Natural His- tory Survey, Bull. 14:211, 1910. I have been unable to locate any specimens collected in Connecticut in the her- baria at the University of Vermont, University of Con- necticut, Yale University, and the Connecticut Botanical Society. I found the plants growing in crushed granitic rock of the B&M railroad bed about two hundred yards north of the old Durham depot, on September 4, 1965. The sandy soil beneath the rock was fairly moist despite a dry super- ficial appearance. On September 23, 1965, Dr. A. R. Hodgdon, Mr. Wendell Berry, Jr., and I examined the station more closely and estimated that betwen two and three thousand plants were growing in an area about three hundred feet long and thirty feet wide. The plants ranged in height from less than an inch to two feet. The collections from this station have been de- posited in the University of New Hampshire Herbarium and in the New England Botanical Club Herbarium at Harvard University, Cambridge, Massachusetts EDWARD J. HEHRE DEPT. OF BOTANY UNIVERSITY OF NEW HAMPSHIRE 'I wish to acknowledge the assistance of the following: Mr. Frank Seymour, Mr. Leopold Charette, Dr. George Torrey and Dr. James Neale. Their help in my search for herbarium Specimens is much appreciated. 48 CHROMOSOME NUMBERS IN VERNONIA (COMPOSITAE) S. B. JoNES, JR. AND W. H. DUNCAN The genus Vernonia is à member of the tribe Vernonieae of the Compositae, which is not as well known cytologically as the other tribes in the family. It is the largest genus in the tribe with about 1000 species of temperate and tropical America, Asia and Africa. The members of the genus are herbs or shrubs, and nearly all are perennial. The North American species of Vernonia were monographed by Gleason (1906) who also treated them in “North American Flora" in 1922. Chromosome numbers reported in Vernonia are summarized in Table 1. This list includes both Old and New World species. Gametic numbers of 9, 10, 16, 17, 18, 20, and 27 are indicated. The first known determination was in 1933 by G. W. Bohn, who counted n=18 in V. bald- winii (personal communication to the senior author, April, 1962). Grant (1953) suggested that the basic number for the genus may be 9 based on his record for V. cinerea, a spe- cies considered by Gleason to be one of the more primitive Vernonia. Additional work is needed, however, to establish Grant's hypothesis firmly. Hunter (1964), ina recent paper, has presented chromosome counts of n—1'7 for seven spe- cies and one hybrid. In connection with taxonomic studies in the genus, the chromosome numbers were obtained for 13 taxa and these are reported in this paper. Meiotic chromo- some counts were obtained from pollen mother cell squashes. Buds were collected in the field or from transplants, and killed and fixed in modified Carnoy's (4 parts chloroform : 3 parts: absolute alcohol: 1 part glacial acetic acid). All material was squashed and stained in acetocarmine. Voucher specimens were made and deposited in the University of Georgia Herbarium, Camera lucida drawings and photo- micrographs were used to record the observations which were made from fresh mounts. 49 50 Rhodora [Vol. 68 All plants studied had a gametic number of n=18. Differ- ences in chromosome size and morphology were not discern- ible. Meiotic figures appeared normal. This was true even in those individuals suspected of being of hybrid origin. Hunter (1964) reports a gametic number of n=17 or ca. 17 for three of the taxa reported in this paper. The reason for, or the significance of this variation is not known at this time. It is possible that both n—17 and n=18 exist within the populations and taxa of this genus. Hybridization frequently occurs in Vernonia (Gleason, 1906) and those that we have studied are no exception. Hy- bridization studies by Jones (1964a) demonstrated that all combinations of taxa attempted were cross-fertile. Certainly chromosome number would not be a limiting factor pre- venting hybridization. The taxa are usually ecologically or geographically isolated; however, these external barriers are not completely effective, producing much of the varia- tion we find in populations of Vernonia. Field cllected in- termediates, as well as artificial F, hybrids, appear to undergo normal meiosis during pollen formation. Since chromosome numbers are known for only about 30 of the possible 1000 species, it is obvious that additional cytotaxonomic work is needed in Vernonia before a realistic concept of cytological relationships in the genus may be established. Since the genus is widespread in temperate and tropical America, Asia, and Africa, geographic and taxo- nomic correlations of chromosome numbers might prove of interest. DEPARTMENT OF BIOLOGY, UNIVERSITY OF SOUTHERN MISSISSIPPI, HATTIESBURG DEPARTMENT OF BOTANY, UNIVERSITY OF GEORGIA, ATHENS, Table 1. Chromosome Numbers in Vernonia (Compositae). Taxon Chromosome Author(s) or Number Vouchers* V. acaulis (Walt.) Gleason n—17 Hunter 1964 7,—18 22394, Duncan, Elbert Co., Ga. —— ———— — . . *Vouchers collected by senior author unless otherwise noted. 1966] SS S NEXSAEN . angustifolia Michx. . angustifolia Michx. var. angustifolia . angustifolia Michx. var. mohrii S. B. Jones . angustifolia Michx. var. scaberrima (Nutt.) A. Gray . altissima Nutt. . anthelmintica (L.) Willd. . baldwinii Torr. . blodgettii Small . cinerea (L.) Less. . colorata Drake X concinna Gleason (pro sp.) . conferta Benth. . X dissimilis Gleason (pro sp.) . glabra Vatke greggii Gray . guadalupensis Heller . interior Small . karvinskiana DC. . kotschiana Sch. Bip. V V V. V V. guineensis Benth. V V V . lindheimeri Gray Vernonia — Jones & Duncan 51 n—ca.17, 17 Hunter 1964 n—18 n—=18 n—18 n—18 n—27 n—18 n—18 2n—18 2n—18 9, —18 2n—18 n—18 2u—20 n—=17 n—17 2n—20 1055, McDuffie Co., Ga.; 1089, Camden Co. Ga.; 1119, Tel- fair Co., Ga.; 1112, Lowndes Co, Ga.; 1116, Coffee Co., Ga.; 1130, Wilkinson Co., Ga. 1091, 1092, Duval Co., Fla.; 1059, Bibb Co., Ala.; 1060, 1061, Autauga Co., Ala.; 1054, Talbot Co., Ga.; 1096, Putnam Co., Fla.; 1102, Lake Co., Fla.; 1105, Alachua Co., Fla.; 1111, Lowndes Co., Ga. 1075, Emanuel Co., Ga.; 1122, Hancock Co., Ga; 1420, D. J. Banks, Claredon Co., S. C. 2595, Perry Co., Ala. Prameshwar 1960” Bohn 1933 1099, 1100, Charlotte Co., Fla. Mangenot et al. 1957”; Miege 1960"; Grant 1953 Mangenot et al. 1957” 1093, Putnam Co., Fla.; 1104, Lake Co., Fla. Mangenot 1957” 1063, Lee Co., Ala.; 1066, Harris Co., Ga. Riley and Hoff 1961” Hunter 1964 Hunter 1964 Mangenot 1958" n— ca.171, 17 Hunter 1964 n—ca.17 2n—18 Turner et al. 1962" Miege 1960” n—ca.17,17 Hunter 1964 V. marginata (Torr.) Raf. n=17 Hunter 1964 "Cited in “Index to Plant Chromosome Numbers," M. S. Cave, Editor. 52 Rhodora [Vol. 68 V. noveboracensis (L.) Michx. n=18 2604, Gwinett Co., Ga. V. nudiflora Less. n—=16 Covas and Hunziker 1954” V. ovalifolia T. & G. »—18 1095, Putnam Co., Fla.; 1104, Lake Co., Fla. V. perottetii Sch. Bip. 2n—18 Miege 1960” V. platensis Les. n—=20 Cabrera 1944 V. primulina O. Hoffm. n—20 Mangenot et al. 1958" V. pulchella Small n—18 1084, Tattnal Co., Ga.; 1085, Long Co., Ga. V. X recurva Gleason (pro sp.) n—18 1077, Emanuel Co., Ga.; 1083, Tattnal Co., Ga. V. texana (A. Gray) Small »—17 Hunter 1964 n=18 2292, Franklin Co. Miss.; 2293, Amite Co., Miss. LITERATURE CITED CABRERA, A. L. 1944. Vernoniae Argentinas (Compositae). Darwiniana 6: 265-379. CAVE, M. S. (Editor). Vol. I, 1956-1959. Index to plant chromosome numbers. Univ. of N. C. Press, Chapel Hill, N. C. Vol. II, 1960-62. Index to plant chromosome numbers. Univ. of N. C. Press. Chapel Hill, N. C. GLEASON, H. A. 1906. A revision of the North American Vernonieae. Bull. N. Y. Bot. Gard. 4: 144-243. . 1922. Vernonieae. North American Flora. N. Y. Bot. Gard. 33: 52-95. GRANT, W. F. 1953. A cytotaxonomie study in the genus Eupatorium. Am. J. Bot. 40: 729-749. HUNTER, G. E. 1964. Chromosome Numbers in Vernonia: Section Lepidoploa: Subsection Paniculatae Verae. Southwestern Natu- ralist, 9: (4): 289-244, JONES, S. B. 1964a. Taxonomy of the narrow-leaved Vernonia of the Southeastern United States. Ph. D. dissertation, Univ. of Georgia. 1964b. Taxonomy of the narrow-leaved Vernonia of the Southeastern United States. Rhodora, 66: 382-401. . In Press, Experimental hybridizations in Vernonia (Compositae). Brittonia. NATURAL HYBRIDS OF LADY'S SLIPPERS (CYPRIPEDIUM) IN MANITOBA H. H. MARSHALL, A. T. H. GROSS AND G. A. STEVENSON A small yellow lady's slipper, Cypripedium Calceolus L. var. parviflorum (Sal.) Fern., occurs frequently in low meadows in southern Manitoba, Canada, while C. candidum Muhl., a tiny white lady’s slipper, is more restricted and has been reported from only three Manitoba locations, Wood- lands, Brandon and Aweme (Scoggan, 1947). In 1954 an in- vestigation of cream colored lady's slippers found in associ- ation with C. Calceolus var. parviflorum and C. candidum, was conducted by the authors. From 1954 to 1964 annual surveys of three areas of 5 to 20 acres between Brandon and Aweme revealed similar lady’s slippers but profuse flower- ing was recorded only in 1954. These hybrids were similar to those described by Fuller (1932) and Curtis (1932) as C. X Andrewsii Fuller. Seedlings of C. Calceolus var. parviflorum are influenced greatly by local environmental conditions (Curtis, 1943). He reported that plants grown from seed may bloom in 9 years under favorable conditions but may take 15 years to bloom and C. candidum probably requires 12 or more years to bloom. Plants of both species are long-lived. Lady’s slippers in nature are pollinated by bees of the genus Andrena (Darwin 1890). The sticky pollen does not permit wind pollination and the location of anthers and stigma precludes self pollination. Hybrid lady’s slippers can therefore occur naturally if suitable pollinators are present and cross-compatible species occur in the same area and have overlapping periods of bloom. Fuller (1932) states that it is very likely wherever C. candidum and C. Calceolus var. parviflorum grow in the same vicinity, hy- brids will be found. Stebbins (1957) describes a method of indexing plant characteristics to identify hybrid populations. Each trait 53 54 Rhodora [Vol. 68 is indexed on a scale of from 1 to 5 although the range may be either reduced or extended. Plants with the lowest total for indexed characteristics are considered of the one pa- rental type and those with the highest total for indexed characteristics are of the other parental type. This paper reports characteristics of natural hybrids of C. candidum X C. Calceolus var. parviflorum in Manitoba and presents an hypothesis to explain the origin of the hybrids. In June 1954, 55 specimens of Cypripedium were collected from a low lying meadow in section 11 township 10 range 18 west of the principal meridian southeast’ of Brandon, Manitoba. Measurements or descriptions of 10 character- istics,, i.e. height, leaf number, length and width of the largest leaf, length and width of lateral petals, color and length of lip, and length and width of the upper sepal, were recorded. Plant characteristics were indexed according to the method described by Stebbins (1957) to identify hybrid populations. The smallest measurement, fewest number and least intensity of color Was assigned the lowest index number. The specimen plants with the lowest total for indexed characteristics are most similar to the C. candidum parent and those with the highest totals are most similar to the C. Calceolus var. parviflorum. parent. In 1963, controlled crosses were made between C. candi- dum and C. Caleeolus var. parviflorum to establish whether the purported parents were cross-compatible and could give rise to hybrids. Viable hybrid seed was readily produced by hand polli- nation of C. candidum and C. Calceolus var. parviflorum thus indicating that they are cross-compatible and could have given rise to hybrids. Native C. candidum commences bloom two weeks earlier than C. Calceolus var. parviflorum at Brandon, Manitoba, hence, their bloom periods seldom coincide and rarely overlap for more than a few days. This is at variance with Fuller (1932) who stated that they bloom at the same time. The bloom period of C. X Andrewsii commences 10 days after C. candidum, thus it overlaps the bloom period of C. Calceolus var. parviflorum to a greater 55 Table 1. Plant measurements (cm) of the two parental species, C. candidum and C. Calceolus var. parviflorum. 1966] Lady’s Slippers — Marshall, Gross & Stevenson Plant C. candidum C. Calceolus var. parviflorum Characteristic E 2. 1. 2. Plant height 15 - 40 14 - 22 20 - 70 17.2 - 32.5 Leaf-length - 9.5 - 12.5 11 - 20 12 - 15.5 -width 1-4 2.0- 3.5 4-12 3.8- 6.7 Sepal-length 2-3 21- 2.8 4-7 3.8- 4.8 -color Petal-length -color Lip-length -color greenish spotted with purple 2.5 -3.5 greenish spotted with purple 1.8 - 2.5 waxy white, purple veined green with purple streaks 2.2- 3.2 green with purple streaks 1.6- 2.0 white, purple veins greenish yellow streaked purple 5-9 3-5 golden yellow greenish yellow streaked purple 4.8- 6.1 2.9- 3.5 yellow 1. Data taken from Seoggan (1957) and Fernald (1950. 2. Brandon data. 56 Rhodora [Vol. 68 extent than that of C. candidum. Bloom appears to be influ- enced greatly by local environmental conditions as indicated by Curtis (1943), Only in 1954 was profuse flowering observed. In 1964 late spring frosts on May 30 (25°F), May 21 (27°F) and June 1 (26°F) destroyed flower buds and bloom was observed in only the most sheltered places. Spring frosts of 27°F or lower were recorded after May 20 in 1958, 1959, 1961, and 1963 which could account for the restricted amount of flowering observed in those years. Measurements of C. candidum and C. Calceolus var. parviflorum (author’s collection No.’s 884 and 883) are similar to those published by Scoggan (1957) and Fernald (1950). The C. candidum plants were smaller than the C. Calceolus var. parviflorum specimens in all 6 of the measure- ments recorded (Table 1). Indices for the naturally occurring Cypripedia indicated 8 plants resembled C. candidum, 8 resembled C. Calceolus var. parviflorum and 39 were intermediate between these two species and resembled C. X Andrewsii (Table 2). This classification was corroborated by the identifications re- ported by Boivin (1960). The index totals for the 8 C. candidum plants ranged from 10 to 14 and did not overlap with the 39 C. X Andrewsii specimens which ranged from 19 to 38. The 8 C. Calceolus var. parviflorum specimens index totals ranging from 30-47 formed a continuous series with C. X Andrewsii. The parental species were not readily distinguishable on the basis of leaf number and width of lateral petals, hence, these two characteristics were of no value in indexing the specimens. Lip color of the hybrids ranged from white of the C. candidum parent to yellow of the C. Calceolus var. parviflorum parent with many creamy intermediate types. The seven other plant characteristics revealed a complete overlap of the parental species by the hybrids. The hybrids were readily separated from the C. candidum. plants using either the indexing method of Stebbins (1957) or the stand- ard taxonomic bases of specimen identification, but they were less readily distinguished from the C. Calceolus var. 1966] Lady's Slippers — Marshall, Gross & Stevenson 57 parviflorum parent. This is in agreement with Stebbins (1957) that hybrids tend to resemble one parent more than another. It is postulated that the hybrids are of relatively recent origin. The largest plants are the parental types and the hybrids are either single stemmed or small clumps. The slow growth and infrequency of bloom of Cypripedia in- dicates that the hybrid swarms could have developed since 1900. This would explain why J. H. Macoun and H. Criddle (Scoggan 1947), who observed the parental species in the Brandon — Aweme district did not report hybrid swarms of creamy colored Cypripedium. These hybrid swarms probably developed as a result of a change in the environment in which the parent species grew. Hybrid swarms usually appear in a disturbed habitat but this is not possible with Cypripedia as they are never found in cultivated land, rarely in pastured areas, and are confined to areas that have been undisturbed. It is suggested Table 2. Frequency of plant measurements of C. candidum, C. Calceolus var. parviflorum and C. X Andrewsti indexed according to Stebbins (1957). l C. candidum C. X Andrewsii C. Calceolus var. parviflorum ^. Index No. > 7 mcd ess d n Plant characteristic frequency frequency frequency Plant height CERE RU b 812' 9 5b 0.0.1 93 4 Largest leaf-length bun2 un go 0. NT 050 2 2 4 -width (ed Ub t & 8 7 15 b 00 0 8 5 No. of leaves’. Si. i. 727 5 — — 0 6 2 — — Upper sepal-length 7 1 0 9 11-10-81 0290770::35* 0 -width 8 0 0 6 6 14 10 3 000 2 6 Lateral petals-length 7 1 0 2 13-14 8 4 QOO 1-1 -width* 3 5 0 5 17 16 1 — oe 0 2 4 — Lip-length 8 0 0 312 7 8 9 0008 5 -color 8 0 0 3 6 9 3 18 0.0008 Range of totals 10 to 14 19 to 38 39 to 47 Plant average 12.6 29.1 42.4 Indexed on a 1 to 3 scale. ?Indexed on a 1 to 4 scale. 58 Rhodora [Vol. 68 that although native Andrena bees were the probable natural pollinators, the introduction of the honeybee made the hy- brid swarm possible. The introduction of honeybees in the early 1900's coincides with the estimated time when the hybrids originated and they could have been the insect vector which brought about the development of hybrids. CANADA DEPARTMENT OF AGRICULTURE EXPERIMENTAL FARM, BRANDON, MANITOBA REFERENCES BorviN, B. 1960. Centurie de plantes Canadiennes III. Naturaliste Canadien, 87: 25-49. Curtis, J. T. 1932. A new Cypripedium hybrid. Rhodora. 34: 239- 242. Curtis, J. T. 1943. Germination and seedling development in five species of Cypripedium L. Am. J. Botany, 30: 199-206. DARWIN, C. 1890. 2nd ed. The various contrivances by which Orchids are fertilized by insects. John Murray, Albemarle Street, London, England. FERNALD, M. L. 1950. 8th ed. Gray’s Manual of Botany. American Book Co., New York. FULLER, A. M. 1932. A natural Cypripedium hybrid from Wisconsin. Rhodora 34: 97-101. GARAY, L. A. 1953. A natural Cypripedium hybrid from Ontario. Can. J. Bot. 31: 660-662, SCOGGAN, H. J. 1957. Flora of Manitoba. Canada Department of Northern Affairs and National Resources. National Museum of Canada. STEBBINS, G. L. JR. 1957. Variation and evolution in plants. Colum- bia University Press. New York. A STUDY OF VARIATION IN THE CYTOTYPES OF DRYOPTERIS SPINULOSA. ROLLA TRYON! AND DONALD M. BRITTON?” Among the groups of ferns in which cytological and evo- lutionary studies are being pursued most vigorously is the genus Dryopteris (sens. strict. of North America and Europe. In eastern North America (see Bibliography) thirteen fertile cytotypes have been recognized, and among these six are diploid (n—41), six are tetraploid and one is hexaploid. Sixteen sterile hybrids of these fertile cytotypes have been examined cytologically and four of them are di- ploid, nine are triploid, two are tetraploid and one is penta- ploid. These previous studies have been concentrated on chromosome counts of the taxa and hybrids to determine the level of ploidy and on chromosome pairing analyses of natural and synthetic hybrids as evidence of their genome relations. There have been no extensive cytological surveys of any of the cytotypes nor analyses of their variation. It was against this background, and our belief that cytological data can be fully exploited in evolutionary studies only when accompanied by data on morphological variation, that the present work was initiated. The data that we have obtained on variation of the cytotypes of Dryopteris spinulosa is pertinent to the problem of the origin of the tetraploid var. spinulosa and to the problems of the evolutionary potential of triploids and the variation of derived hexaploids. Dryopteris spinulosa var. intermedia (2 X) and var. spin- ulosa (4X) were chosen as especially suitable taxa for investigation because of their abundance and because of the abundance of their sterile triploid hybrid. These taxa pre- sented an opportunity to study and compare variation, in a related diploid and tetraploid and their hybrid, on a broad "Gray Herbarium, Harvard University. "Department of Botany, University of Guelph. This study was sup- ported, in part, by grants from Sigma Xi-RESA to D. M. Britton. 59 60 Rhodora [Vol. 68 basis. Dryopteris spinulosa is a rather common Species in northeastern United States and adjacent Canada where it usually grows on wooded hillsides, or in wooded swamps and less often in shrubby swamps, or along damp roadsides. Variety spinulosa (figs. 9, 10) grows more commonly in wet, often water-logged sites, while var. intermedia (figs. 5, A 3 PLATE 1320 Figs. 1-4. Illustrations of characters c to f. Fig. 1 — character c, basal pinnae of intermedia, each inferior basal pinnule, A and B, is shorter than the adjacent one, C and D. Fig. 2 — character d, basal pinnae of the hybrid, five basal tertiary segments, A to E, on the acroscopic side of the inferior pinnules, are parallel to the pinna-rachis. Fig. 3 — character e, basal pinnae of the hybrid, three of the inferior pinnules, A to C, are longer than the basal superior pinnule, D. Fig. 4—character f, central pinnae of intermedia, on each of the four pinnae, the basal inferior pinnule, A to D, overlies the rachis, E. 1966] Dryopteris spinulosa — Tryon & Britton 61 PLATE 1321 Figs. 5-6. Lamina of Dryopteris spinulosa var. intermedia, diploid, x 14. Fig. 5— from Plant 110. Fig. 6 — from Plant 357. o bo Rhodora [ Vol. AAAS M > AYA PLATE 1322 Figs. 7-8. Lamina of Dryopteris spinulosa var. intermedia, X var. spinulosa, triploid, X 4. Fig. 7 — from Plant 487. Fig. 8 — from Plant 268. 6) is most often in moist but well drained sites, but both exhibit considerable latitude in their habitat. The two varieties frequently grow intermixed or in close proximity to each other. In our experience, their hybrid (figs. 7,8) is found whenever the two parents are in some abundance (and sometimes when one is rare) and it is usually frequent or even more abundant than either parent. 1966] Dryopteris spinulosa — Tryon & Britton 63 PLATE 1323 Figs. 9-10. Lamina of Dryopteris spinulosa var. spinulosa, tetraploid, x 14. Fig. 9— from Plant 272. Fig. 10— from Plant 294. The classifieation and nomenclature of this group Is sub- ject to considerable uncertainty and disagreement. While this paper does not intend to deal with either of these matters in a definitive manner, we do present some reasons later for the classification adopted. It is well established that var. spinulosa and var. intermedia are taxa (major evo- lutionary units) and disagreements about their taxonomic rank is of secondary importance. Some of the difficulties involved in the correct nomenclature are so involved and technical that we wish to do no more than mention their 64 Rhodora [Vol. 68 existence, The following synonymy will serve to orient the reader with other names currently used for these plants. Dryopteris spinulosa (O. F. Muell.) Watt, var. spinulosa (Dryopteris spinulosa; Dryopteris spinulosa, “typical”; Dryopteris austriaca (Jaeq.) Woynar, var. spinulosa (Muell.) Fiori). In the remainder of the paper this will be referred to as spinulosa. Dryopteris spinulosa var. intermedia (Willd.) Underw. (Dryopteris intermedia (Willd.) A. Gray; Dryopteris austriaca var. intermedia (Willd.) Morton). In the remainder of the paper this will be referred to as intermedia. Dryopteris spinulosa var. intermedia X var. spinulosa (Dryopteris intermedia X spinulosa; Dryopteris spinulosa var. fructuosa (Gilbert) Trudell; Dryopteris austriaca var. fructuosa (Gilbert) Morton; Dry- opteris X triploidea Wherry). In the remainder of the paper this will be referred to as the hybrid. This study was initiated in 1959 and has continued through June of 1965. Rolla Tryon has made the analyses of the characters, Donald M. Britton the cytological determi- nations and the collections of specimens; both have con- tributed to the discussions and conclusions. MATERIALS AND METHODS OF STUDY. Most of the plants used in this study (155) were obtained from southern Ontario (west of 799 W. Long. and south of 46? N. Lat.). Twenty-five others were from the Algoma District of Ontario (east of Lake Superior) and eight were from Pennsylvania and Virginia. The material for cyto- logical study was usually collected in late May to the middle of June. At that time a voucher leaf was taken from the plant and prepared as an herbarium Specimen. In many cases, the plant also was marked with the collection number and additional leaves were collected from some of these plants in September of the same year and also in the follow- ing year. Individual plants, therefore, have been repre- sented by one to sixteen leaves and often by sets of leaves collected in successive years. Eighty-nine plants of inter- media were determined as diploid ( n—41); forty-seven plants of the hybrid. were determined as triploid (n—123) ; 2 and fifty-two plants of spinulosa were determined as tetra- ploid (n=82). Voucher specimens of these cytologically 1966] Dryopteris spinulosa — Tryon & Britton 65 determined plants have been deposited in the Herbarium, Department of Botany, University of Guelph and a selection of them in the Gray Herbarium, Harvard University. Not all of the 188 plants and 746 leaves available have been used in this study. Some leaves were defective and others were too small (less than 25 em. long) to be included in case they would bias the results. About 175 plants and 550 leaves were available for most characters; the number studied for each is given as part of Charts 1 and 2. Thirty-four characters were considered during the course of our work. Eleven of these were studied in some detail and the results of these analyses are presented. Preliminary studies were made of eighteen other characters but these were not continued because the characters were either too variable to be of particular value or because no significant differences between spinulosa and intermedia were observed. Among these, a few of the variable characters were the following: complexity of the lamina, shape of the pinnae, tapering of the apex of the pinnae and size of the indusium. The variation of four other characters was not analysed principally because of technical difficulties in their study. An initial survey of these indicated that they probably be- haved in the same manner as some of the characters ana- lyzed in detail. These characters were the orientation of the rhizome, the color of the petiole scales, the evergreen quality of the leaves and the kind of habitat. Finally, the apex of the spines of the spores was not studied simply because of the small size of the character, the spines them- selves being only about 1 micron long. The illustrations have been prepared from a selection of specimens of the cytologically determined plants. These specimens were photographed and the negative printed as a silhouette. The specimens were relaxed in warm water, prior to photographing, so that they were substantially restored to their fresh condition. This is of some importance because minor shrinking, wrinkling and curling of the teeth and ultimate lobes can alter the total aspect of a pinna or lamina to a considerable degree. 66 Rhodora [Vol. 68 ANALYSIS OF CHARACTERS, The following comments on the individual characters ex- plain the character, when necessary, and the manner in which it was studied. The original measurements or counts for six of the characters are presented in Charts 1 and 2 and in Graph 1. In order to compare the range of variation of the different characters and their means, this data was ad- justed to a scale of 0 to 100 for each character (the minimum count or measurement—0, the maximum- 100). For ex- ample, character e varies from a minimum count of 1 to a maximum of 8. Then 1—0 and 8—100 and counts of 2 to 7 —14, 28, 43, etc., respectively. For characters b and ec the score was then reversed (0—100, 100—0) so that the scores for all of the characters would be high for intermedia and low for spinulosa. These adjusted scores are presented in Chart 3 and in Graph 2. Selections of the data are presented in Charts 4 and 5 and in Graph 3 to illustrate special features of the variation of the cytotypes. a. Glandularity of the base of the pinna. The lamina of intermedia is glandular while that of spin- ulosa is usually glabrous. The glands are small, stalked and have a more or less Spherical cap cell. They are usually most abundant at the base of the pinnae. There is some tendency for them to be deciduous as the leaf ages, although they will persist on leaves that have lived over the winter. In order to obtain a measurement of this character, the cytological voucher leaves (collected in early summer) were used and a single central pinna was chosen for observation. The glands that could be seen on the lower surface on one linear millimeter of the pinna stalk were counted. The data obtained from sixty plants are presented in Chart 1. b. Length of the guard cells of the stomata. This character was studied only in a preliminary manner although the results indicate that extended data would be of much interest. A portion of a leaf collected in late summer was relaxed in warm water and then placed in a lactic acid mount. Measurements were made, in micra, of the maxi- 1966] Dryopteris spinulosa — Tryon & Britton 67 mum length of the guard cells from stomata between the veins on the lower surface and the mean from ten measure- ments was obtained for each leaf. The data obtained from 27 plants are presented in Chart 1. c. Length of the inferior basal pinnules on the basal pin- nae in relation to the adjacent ones. Each basal pinna of intermedia usually has the inferior basal pinnule shorter than the adjacent one (fig. 1), while spinulosa usually has the inferior basal pinnule longer than the adjacent one (fig. 16). CHART 1. Analysis of variation in Dryopteris spinulosa, original counts and measurements of characters a and b. (See Graph 1.) Range of variation of Mean of all leaves. all leaves. (Total number of leaves.) a. Glandularity of the base of the pinna: number of glands on one linear mm. of pinna-stalk, each leaf from a different plant. intermedia 10-75 45.6 (20) hybrid 9-55 29.2 (20) spinulosa 0-9 1.6 (20) b. Length of the guard cells of the stomata: mean of ten measure- ments for each leaf, each leaf from a different plant. intermedia 34-53 micra 42.5 micra (9) hybrid 40-58 micra 49.5 micra (9) spinulosa 44-68 micra 55.7 micra (9) The character was measured by taking the average of the length of both basal pinnules and both adjacent ones because the relative length of these two pinnules may differ on each basal pinna. The measurements are in millimeters and the average is expressed as the basal pinnules being longer, or shorter, than the adjacent ones. The data ob- tained from 546 leaves are presented in Chart 2; illus- trations of the variation are given in figs. 11-16. 68 Rhodora [Vol. 68 16 PLATE 1324 Figs. 11-16. Basal pinnae, X 1/3. Figs, 11-12. Dryopteris spinulosa var. intermedia, diploid. Fig. 11 — from Plant 28, the inferior basal pinnules are 15 mm. shorter than the adjacent ones. Fig. 12 — from Plant 110, the inferior basal pinnules are 7 mm. longer than the adjacent ones. Figs. 13-14. Dryopteris spinulosa var. intermedia x var. spinulosa, triploid. Fig. 13 — from Plant 38, the inferior basal pinnules are 7 mm. shorter than the adjacent ones. Fig. 14 — from Plant 502, the inferior basal pinnules are 10 mm. longer than the adjacent ones. Figs. 15-16. Dryopteris spinulosa var. spinulosa, tetraploid. Fig. 15 — from Plant 68, the inferior basal pinnules are equal in length to the adjacent one. Fig. 16 — from Plant 36, the inferior basal pinnules are 15 mm. longer than the adjacent ones. d. Number of parallel tertiary segments. In intermedia there is a Strong tendency to have many of the basal tertiary segments, on the acroscopic side of 1966] Dryopteris spinulosa — Tryon & Britton 69 the pinnules, disposed more or less parallel to the pinna- rachis. This character is illustrated (by the hybrid) in fig. 2. In spinulosa (fig. 15) there is a strong tendency to have rather few such segments parallel to the pinna rachis. This character has been counted on one arbitrarily chosen basal pinna and on the pinnules on its inferior side, as illustrated in fig. 2. The data obtained from 565 leaves are presented in Chart 2. This character may be correlated with the size of the pinna and for this reason small leaves were not used in its study. In leaves of moderate or large size there is no abso- lute correlation with the size of the pinna as the following counts wil show. In a plant of spinulosa, a basal pinna of one leaf was 9.5 cm. long and had 1 parallel tertiary segment on its inferior side. A basal pinna, from another leaf of the same plant, was g cm. long and had 2 parallel tertiary segments on its inferior side. e. Number of longer inferior pinnules. The pinnae in spinulosa tend to be relatively broad at the base and to be abruptly reduced in width. In intermedia there is a tendency for the pinnae to be hardly, or not at all, broadest at the base and to be gradually reduced in width. These tendencies have been reduced to an easily observed character by using one basal pinna and comparing the length of the inferior pinnules with the basal superior one. The character was taken from the basal pinna with the shortest basal superior pinnule and, in reference to this pinnule, the number of inferior pinnules that were longer was de- termined. The data obtained from 546 leaves are presented in Chart 2; and the character is illustrated (by the hybrid) in fig. 3. As in the study of the previous character, small leaves were not used because of the possibility that small pinnae would bias the data. In leaves of moderate to large size, there is no absolute correlation with the size of the pinna as the following counts will show. A basal pinna of one leaf of a plant of intermedia was 11.5 cm. long and had 5 70 Rhodora [Vol. 68 longer inferior pinnules. A basal pinna from another leaf of the same plant was 10 cm. long and had 6 longer inferior pinnules. CHART 2. Analysis of variation in Dryopteris spinulosa, original counts and measurements of characters ¢ to f. (See Graph 1.) Range of variation Mean of all leaves. All leaves. Plant means from (Total number of leaves.) plants with 4 or more leaves. (Total number of plants: total number of leaves.) c. Length of the inferior basal pinnules of the basal pinnae in relation to the adjacent ones. intermedia 16 mm. shorter to 10 mm. shorter to 4.5 mm, " mm. longer. 2 mm. longer. Shorter (255) (31: 172) hybrid 7 mm. shorter to 1.5 mm. Shorter to 1.7 mm. 11 mm. longer. 5 mm. longer. longer (119) (16: 87) spinulosa equal to 16 mm. 3.7 mm. longer to 6.4 mm. longer. 9 mm. longer longer (108) (11:60) d. Number of parallel tertiary segments. intermedia, 3-10 4.0-6.8 5.7 (296) (19: 107) hybrid 1-8 2.5-4.7 3.4 (137) (18: 95) spinulosa 1-5 1.6-3.5 2.5 (132) (14: 82) e. Number of longer inferior pinnules. intermedia 2-8 3.4-6.2 4.8 (278) (34: 212) hybrid 2-5 2.0-3.6 2.9 (135) (17: 96) spinulosa 1-4 2.0-3.1 2.1 (133) (14: 84) f. Number of pinnules that overlie the rachis. intermedia, 0-10 1*-2.9-10 7.8 (299) (30: 162) hybrid 0-7 0-3.0 1.0 (155) (20: 107) spinulosa 0-3 0-0.75 0.07 (138) (17: 97) *Plant 357 with two leaves. 1966] Dryopteris spinulosa — Tryon & Britton 71 f. Number of pinnules that overlie the rachis. In intermedia, the pinna, the basal inferior pinnule and its basal segment nearest the rachis are disposed so that the basal segment is relatively close to the rachis and most often so that it overlies it (figs. 4 and 5.). In spinulosa these same structures are disposed so that the basal segment is relative- ly far from the rachis (figs. 9 and 10). This character was analyzed by the number of basal inferior pinnules that overlie the rachis on the lowest ten pinnae. Fig. 5 shows a lamina with 10 overlying pinnules and figs. 9 and 10 show laminae with no overlying pinnules. Only 10 pin- nae were used in order to avoid bias due to high counts in unusually large leaves. Bias due to low counts in small leaves was eliminated (or reduced) by taking only those leaves with a lamina at least 30 cm. long. In order to confirm that this character was not related to changes in orientation due to pressing of the leaf, counts were made on the same leaves when they were fresh and after they were pressed. No difference was observed. The data obtained from 592 leaves are presented in Chart 2. g. Position of the pinnae in relation to the rachis. The pinnae may be patent (more or less at right angles to the rachis) or they may vary to ascending (at an acute angle to the rachis). This character was originally studied because it has been one that has been prominently used for the identity of the taxa, and especially of the hybrid. How- ever, so much variation was soon observed in both inter- media and spinulosa that the character was neither studied further nor refined. Figures 17 and 18 of intermedia and figs. 31 and 32 of spinulosa illustrate some of the variation observed in the pinna-angle. Ninety-seven leaves of inter- media had patent pinnae, 55 leaves had them slightly ascend- ing and 23 had them definitely ascending. The same range of variation, but in somewhat different proportions, was seen in spinulosa in which 13 leaves had patent pinnae, 47 leaves had slightly ascending pinnae and 23 leaves had definitely ascending pinnae. The character was also studied on plants from which several leaves were available but these 72 Rhodora [Vol. 68 PLATE 1325 Figs. 17-23. Central pinnae of Dryopteris spinulosa var. intermedia, diploid, X 1/3. Fig. 17 —from Plant 105. Fig. 18 — from Plant 357. Fig. 19 — from Plant 335. Fig. 20 —from Plant 108. Fig. 21 — from Plant 29. Fig. 22 — from Plant 104. Fig. 23 — from Plant 111. 1966] Dryopteris spinulosa — Tryon & Britton 73 results confirmed the previous data, i.e., although there was a slight statistical difference between intermedia and spin- ulosa it was not a sufficient one to invite further analysis. E LO pis 1 9 C " `. ^ Nd B S 24 È 27 PLATE 1326 Figs. 24-29. Central pinnae of Dryopteris spinulosa var. intermedia X var. spinulosa, triploid, X 1/3. Fig. 24 — from Plant 86. Fig. 25 — from Plant 84. Fig. 26 — from Plant 562. Fig. 27 — from Plant 87. Fig. 28 — from Plant 358. Fig. 29 — from Plant 32. 74 Rhodora [Vol. 68 30 33 35 PLATE 1327 Figs. 30-35. Central pinnae of Dryopteris spinulosa var. spinulosa, tetraploid, X 1/3. Fig. 30 — from Plant 33. Fig. 31 — from Plant 370. Fig. 32 — from Plant 79. Fig 33 — from Plant 305. Fig. 34 — from Plant 488. Fig. 35 — from Plant 289, h. Aspect of the lamina. The aspect of the lamina is usually different in spinulosa and intermedia. This may be readily observed in fig. 5 of intermedia and fig. 9 of spinulosa which are characteristic of each taxon. One of the initial objectives of this study was to analyze the general character of aspect into its individual components. This objective was partially but not wholly 1966] Dryopteris spinulosa — Tryon & Britton 15 realized. Characters c to g are individual characters of as- pect and our analyses of these, therefore, also represent an analysis of aspect. Other individual characters have not been analyzed and among these may be mentioned shape, symmetry, spacing and complexity of the pinnae, of the pinnules and of the tertiary segments, and the length, shape and orientation of the marginal teeth. Figures 17 to 35 illustrate the central pinnae of leaves that were selected to present the variation in aspect of the three cytotypes. Common aspects of intermedia, are illus- trated by figs. 21 and 23, of the hybrid by figs. 25 and 27, and of spinulosa by figs. 31 and 35. The other figures of 17-35 represent less common variations. The greatest resemblance of intermedia. to spinulosa. is seen in. figs. 20 and 34. The aspect of the hybrid. seems to intergrade with that of both of its parents. Figure 17 of intermedia is close to fig. 29 of the hybrid and fig. 19 of intermedia is close (except in the orientation of the pinnae) to fig. 24 of the hybrid. The aspect of the hybrid and spinulosa is similar in fig. 27 of the hybrid and fig. 33 of spinulosa, and in fig. 26 of the hybrid and fig. 34 of spinulosa. i. Glandularity of the indusium. Glands of the same type as those on the lamina also occur on the indusium of intermedia and the hybrid. The indusium of spinulosa is glabrous. Eighty-one per cent of the 89 plants of intermedia had the indusia definitely glandular and nineteen per cent had them slightly glanduiar; eighty-five per cent of the 47 plants of the hybrid. had the indusia defi- nitely glandular and fifteen per cent had them slightly glandular. These observations were made only on the cyto- logical voucher leaves, collected in early summer, since the glands tend to be deciduous with age. However, in both intermedia and the hybrid, glands were also observed on the indusia of leaves that had lived over the winter. j. Spores. Our observations of the spores of numerous plants ot intermedia and spinulosa showed them to be normal in ap- 16 Rhodora [Vol. 68 pearance and presumably viable. Only rarely was an abor- tive spore seen. The spores of the hybrid, on the other hand, were uniformly shriveled (abortive). No normal appearing spores were seen among those of the hybrid al- though some effort was made to find them. k. Chromosome number. The chromosome counts, already mentioned, may be sum- marized here for the sake of a complete list of the characters studied, intermedia: diploid, n = 41 (89 plants). hybrid: triploid, n = 123 (47 plants). 2 spinulosa: tetraploid, n = 82 (52 plants). DISCUSSION. Variation in the cytotypes. — In relation to the numerical as well as the geographical limitations of our sample, it is desirable that some assess- ment be made of the parameters of the variation that have been established in our study. An effort was made to obtain a cytological analysis of as many variants as possible and we believe that most of the variation was obtained in the areas sampled. Nevertheless, it is patent from a study of materials in the Gray Herbarium and the Herbarium of the New England Botanical Club that there is greater vari- ability in the Dryopteris spinulosa group than in our cyto- logically determined samples of it. For this reason we are reluctant to place reliance on our parameters as those of the taxa and their hybrid. We feel that most of the param- eters can be extended by further cytological sampling, especially in other regions. The means of the characters, however, and the mean of the means, can probably be ex- trapolated with some confidence as representative of the taxa and the hybrid because of the inertia of the mean. The range of variation of the original counts and measure- ments is presented in Graph 1. It is portrayed more clearly in Graph 2 where the original data has been adjusted for comparative purposes. Considering the data from all leaves, 1966] Dryopteris spinulosa — Tryon & Britton 77 CHART 3. Analysis of variation in Dryopteris spinulosa, adjusted scores (see text) of characters a to f. (See Graph 2.) Character. a to f *Plant 357 with two leaves. intermedia, hybrid spinulosa intermedia hybrid spinulosa intermedia hybrid spinulosa intermedia hybrid spinulosa intermedia hybrid spinulosa intermedia hybrid spinulosa intermedia hybrid spinulosa Range of variation. All leaves. 13-100 12-62 0-12 47-100 29-82 0-71 28-100 16-72 0-50 22-100 0-77 0-44 14-100 14-57 0-43 0-100 0-70 0-30 Mean range of variation. 21-100 12-70 0-42 Mean of plants with 4 or more leaves. 44-81 34-55 22-38 33-64 17-41 7-28 34-74 14-37 14-30 10*-29-100 0-30 0-7.5 Mean of all leaves. Mean of means. 64 94 Lf it will be seen in Graph 2 that the range of variation (mini- mum to maximum score) of each cytotype overlaps that of the others in characters b to f. In character a, the range of variation of intermedia and the hybrid overlap while that of spinulosa is contiguous, or nearly so, with. the range of those two. The range of variation of the mean of plants with four or more leaves is presented for characters c to f. In 78 Rhodora [Vol. 68 o co 30 20 10 -10 -20 eG — O Oo Qo 4 GRAPH 1. Analysis of variation in Dryopteris spinulosa, original counts and measurements of characters a to f, from Charts 1 and 2. White dot: mean of all leaves. Bars indicate range and extent of variation of a character. Solid bar: all leaves. Outline bar: plant means from plants with four or more leaves (extension by dash of character f in intermedia includes Plant 357 with only two leaves). I: intermedia, H: hybrid, S: spinulosa. 1966] Dryopteris spinulosa — Tryon & Britton 79 X [^4] Uu -— M [ 2] C) = - C) (0) le o T € = = era BS oO me} OQ "TO o GRAPH 2. Analysis of variation in Dryopteris spinulosa, adjusted scores of characters a to f, from Chart 3. White dot: mean of all leaves. Bars indicate range and extent of variation of a character. Solid bar: all leaves. Outline bar: plant means from plants with four or more leaves (extension by dashes of character f in intermedia includes Plant 357 with only two leaves). I: intermedia, H: hybrid, S: spixulosa. 80 Rhodora [Vol. 68 the means of plants the range of variation of spinulosa and intermedia is discrete in each character, while that of the hybrid. overlaps the variation of both parents in each of them. The extent of variation (maximum score minus the minimum) of the means of plants is about half that of the extent of the variation of single leaves. The mean of each character is strikingly different in spinulosa and intermedia. The mean of the hybrid falls between those of the parents. In characters a to c it is more or less midway between the parental means, while in charac- ters d to f it is closer to the mean of spinulosa. The mean of the means of the six characters demonstrates that, in these, the hybrid falls somewhat closer to spinulosa than to inter- media. We do not wish to imply, however, that this may represent the actual genetic position of the hybrid for in other characters it is closer to intermedia. It will be noted also in Graph 2 that in most cases the mean of a character is more or less centrally located between the extremes, while in a few cases it is notably close to one extreme. This is especially true of character a in spinulosa and of character f in all of the cytotypes. The genetic factors which might control these characters would be worthwhile to investigate. At the moment nothing is known of the inheritance of any character in Dryopteris. Information concerning numbers of pairs of genes, dominance relation- ships and factor interaction is lacking. It can be seen in Graph 2 that the extent of variation in a character is usually greater in intermedia than in spin- ulosa. In characters a, e and f the extent of variation in intermedia is twice as great, or more, than in spinulosa. In characters ¢ and d it is about half again as great in inter- media than in spinulosa. Only in character b is the extent of variation greater in spinulosa than in intermedia. In each of these characters the extent of variation in the hybrid is equal to or greater than that in the less variable parent and smaller than that in the more variable parent. This com- parison of variation is summarized in Graph 2 by the bars that indicate the mean range (and extent) of variation of 1966] Dryopteris spinulosa — Tryon & Britton 81 CHART 4. Analysis of variation in Dryopteris spinulosa. Range and extent of variation between extreme leaves and plants in the means of adjusted scores (see text) of characters a and c to f. (See Graph 3.) Character and score. Mean of Range and scores. (extent) a c d e T of variation of mean. Leaf from inter- Plant..viv 13 48 33 57 40 40 media Leaf from 40-79 (39) Plant 104 66 75 78 86 90 19 Leaf from Plant 126 19 . 99 338 144 0 16 hybrid Leaf from 16-47 (31) Plant 95 44 471 18 B: Ge Leaf from spinu- | Plant 91 0 8 22 14 0 9 losa Leaf from 9-24 (15) Plant 370 6 84 33 48." I0 724 Character and mean of Mean of plant means. a c d e f Plant 101 inter- 4 leaves 13 52 33 64 57 44 media Plant 113 44-71 (27) 4 leaves 55 80 56 74 92 "1 Plant 284 6 leaves 12 38 30 29 3 22 hybrid Plant 116 22-40 (18) 4 leaves 13 50 33 36 10 40 Plant 36 spinu- 13 leaves 3 27 13 14 (-- 11 losa Plant 43 11-19 (8) 8 leaves 5 38 21 30 09 all six characters (mean of the minimum scores to mean of the maximum). The extent of variation is almost twice as great in intermedia (79) as in spinulosa (42), while in the hybrid it is quite intermediate (58) between the two. 82 Rhodora [Vol. 68 | | l | l | | | 0 10 20 30 40 50 60 70 80 90 GRAPH 3. Analysis of variation in Dryopteris spinulosa, range and extent of variation in means of adjusted scores of characters a and € to f, from Chart 4. Solid bar: variation between extreme leaves. Outline bar: variation between extreme plants. I: intermedia, H: hybrid, S: spinulosa. The greater extent of variation of the characters of inter- media is consistent with the opportunities for recombination and mutation to be expressed phenotypically at the diploid level and the smaller extent of variation of spinulosa is con- sistent with the restrictions on their expression at the tetra- ploid level. The variation as represented by extreme leaves and ex- treme plants and the extent of variation between them is given in Chart 4 and in Graph 3. It will be seen that the extent of variation among extreme leaves and extreme plants is more than twice as great in intermedia as in spin- ulosa while in the hybrid it is about midway between that of the parents. The fact that the extent of variation in individual characters, in extreme leaves and in extreme plants of the Aybrid, is so nearly intermediate between that of the parents implies that the hybrid plants we have ana- lyzed have come from the whole spectrum of variation of the parent plants as we have seen it. One of the most interesting aspects of the variation of the cytotypes is their intraplant variability. Data on this variation, from selected plants, are presented in Chart 5. 1966] Dryopteris spinulosa — Tryon & Britton 83 The last column of the chart gives the extent of the intra- plant variation as a per cent of the total extent of variation in the cytotype. Most commonly the intraplant variation is rather small and this is illustrated by plants 27 and 301 which show the variation in the leaves of a plant in one year (one crown) and plants 36 and 38(1959, 1960) which show the variation in the mean of crowns in successive years. The frequent extent of intraplant variation of a character is about 15%, or less, of that of the cytotype. Rarely the intraplant variation may be great and this is illustrated in plants 38(1959) and 504 which show the variation in the leaves of a plant in one year and plants 32 and 117 whick show the variation in the mean of crowns in successive years. These examples illustrate that rarely the extent of variation of a character in one plant may be 70%, or more, of that of the cytotype or, as in plant 117, may even exceed that of the cytotype. (This is due to the fact that in Chart 3 the mean of plants is derived from all availa- ble leaves of a plant rather than only those of a single crown.) The genetic implications of the great intraplant vari- ability suggest that the genes controlling these characters are not markedly different in intermedia and spinulosa. They also suggest that the expressivity of the genes is extremely variable, providing that no unusual mitotic divisions are involved and that the cells of a plant all have a characteristic genotype. Similar variability in gene expression is well known in other plants, for example, the leaves of Morus alba. The taxonomic implications are that it vitiates, to à degree, the differences between the taxa brought out by the difference in their means of these characters. The only character that we have studied in which no vari- ation was observed within a cytotype was the chromosome number. Although both the spores and the indusia have been grouped into two discrete classes: normal-abortive and glabrous-glandular respectively, there is in fact variation in both of these characters. We have not studied the proportion of normal to abortive spores in spinulosa and intermedia, 84 Rhodora [Vol. 68 CHART 5. Intraplant variation in Dryopteris spinulosa. Range and extent of variation of single characters in selected plants. Data on individual leaves are in roman type, data of means of plants are in italic. Plant, cytotype Range of Variation. Extent of Variation and material, In example as 96 of Character. cytotype. S [^ Original — Adjusted, . E data. scores. " 8 HS A > ov o Plant 301, 0.5 mm. longer hybrid, 5 c to 41-48 7 56 12% leaves, 1960 3 mm. longer Plant 27, intermedia d 5 to 6 44-56 12 78 15% 4 leaves, 1960 Plant 504, 4 mm. longer spinulosa, c to 3-38 35 50 70% 4 leaves, 1961 15 mm. longer Plant 38, 6.5 mm. shorter hybrid 5 c to 20-70 50 56 90% leaves, 1959 9.5 mm. longer Plant 36, 7.2 mm. longer spinulosa, 4 c in '59 to 7.4 27-28 1 16 6% leaves, 1959, mm, longer in 5 in 1960 60 Plant 38, 1 mm. longer hybrid 5 c in '59 to 1.8 44-47 3 21 14% leaves, 1959, mm. longer in 3 in 1960. '60 Plant 32, 2 in 1959 hybrid, 4 f to 20-43 23 30 77% leaves, 1959, 4.3 in 1960 3 in 1960 Plant 117, 3.66 in 1959 hybrid, 3 e to 9-38 29 23 126% leaves, 1959, 1.66 in 1960 3 in 1960 1966] Dryopteris spinulosa — Tryon & Britton 85 being content with the observation that the abortive spores are rare. The glandularity of the indusium of intermedia and of the hybrid is clearly variable and although this has not been studied in detail it seems to show a range of vari- ation similar to that of the pinna base in intermedia and the hybrid. Identification. — Identification is most readily made during the rather brief time when mature spores are present in abundance. There is a possibility of contamination when there are only a few spores. During this two or three week period, the hybrid is readily identified by its abortive spores, intermedia by its normal spores and glandular indusium and spinulosa by its normal spores and glabrous indusium. Before, or after, this period fertile leaves of spinulosa may be identified by their glabrous indusium but intermedia. and the hybrid cannot be distinguished from each other by the indusium. Sterile ma- terial of spinulosa or intermedia cannot always be dis- tinguished from the hybrid. The difficulties of identification, then, center on the many herbarium specimens and plants in nature that are collected or observed at a time when uncontaminated mature spores are not present. Other characters must then be employed, if possible, for identifi- cation and it is with these that this study has largely been concerned. We have mentioned our conclusion that the parameters of the characters we have studied are undoubtedly greater in the taxa and their hybrid than in our sample of them. For this reason we do not believe that any of the data on single charaeters (Charts 1, 2 and 5, Graphs 1 and 2) afford a basis for the identification of all plants of the cytotypes. In all of the characters, except a, the range of variation of one cytotype overlaps that of another (or both) and in character a, the range of spinulosa is very close to that of the other two. The data on extreme plants (Chart 4, Graph 3), derived from the mean of 5 characters, shows that the range of variation of each cytotype is distinct from that of the other 86 Rhodora [Vol. 68 two. However, the degree of distinction is small: a differ- ence (in the mean of the adjusted scores) of 3 between plant 43 of spinulosa (19) and plant 284 of the hybrid. (22) and a difference of 4 between plant 116 of the hybrid (40) and plant 101 of intermedia (44). We do not believe that these small differences afford a basis for the identification of all plants of the cytotypes. It seems worthwhile to emphasize the variability of character c, the relative length of the inferior basal pinnules on the basal pinnae and the adjacent ones, because this character has been much relied upon in the literature for purposes of identification. Chart 2 and figs, 11-12, 15-16 demonstrate the variability in this character. Intermedia usually has the inferior basal pinnules shorter than the ad- jacent ones (fig. 11), but they may be longer (fig. 12) and spinulosa, usually has the basal inferior pinnules longer than the adjacent ones (fig. 16), but they may be equal in length (fig. 15). The great variation among the leaves of one crown can be seen in Chart 5, Plant 38 (1959) in which one of the five leaves had the inferior basal pinnules 6.5 mm. shorter than the adjacent ones and another leaf had them 9.5 mm. longer. The evergreen quality of the leaves has also been empha- sized in the literature and it is generally true that leaves of spinulosa are deciduous and those of intermedia are ever- green. However, our limited observations have indicated that this character is a variable one especially influenced by the environment of the individual plant. In most of the plants of the hybrid that were studied the leaves were de- ciduous or only partially evergreen but in some they were evergreen as is typical intermedia. Other characters employed in the identification of the taxa and their hybrid that we have also found to be too vari- able to be utilized are the position of the pinnae in relation to the rachis (see Analysis of Characters, g.) and the as- pect of the lamina (see Analysis of Characters, h). Con- sidering the variability discovered in all of the characters mentioned, we feel that others such as the orientation of 1966] Dryopteris spinulosa — Tryon & Britton 87 the rhizome, the color of the petiole scales and the type of spines on the spores cannot be accepted as affording a dis- tinction of the taxa until their constancy has been confirmed by an extensive study of plants of known identity. There remain then as wholly applicable characters for identification of the cytotypes only the glandular or glabrous indusium, the abortive or normal spores and the chromo- some number. Certain extremes of other characters do seem sufficiently distinctive to allow some confidence in their use and these have been added to the key. It will be obvious that the key will not allow the identification of all material. Key. Indusium glandular. Spores normal; n—41. (Any leaf with 9-10 basal inferior pinnules on the lowest 10 pinnae overlying the rachis, or any plant with four or more leaves averaging 6-10 overlying pinnules probably belongs here.) .......— eene intermedia Spores abortive; n—123. (Any plant with a glandular in- 2 dusium and four or more leaves averaging 0 cverlying pin- nules probably belongs here.) ...- Hmmm hybrid Indusium glabrous; spores normal; n—82. (Any leaf with the pinna-bases glabrous probably belongs here.) .....- spinulosa Classification. — The authors are in amiable, but firm, disagreement about the proper taxonomic status of intermedia and spinulosa. Donald M. Britton is of the opinion that, in these taxa, the cytogenetic evidence must be decisive. Intermedia is a basic diploid taxon genetically isolated from spinulosa which is a derived polyploid taxon. Accordingly, the status of species is most appropriate for each of them. Rolla Tryon is of the opinion that the taxonomic status of intermedia and spin- ulosa must be determined by a comparision of the two to each other and to other taxa in their alliance on the basis of all evidences of similarity and difference. He is of the opinion that the similarities and differences brought out in the pres- ent study make the category variety most appropriate, at the same time admitting that this status is tentative pending adequate comparative studies of other allied taxa. 88 Rhodora [Vol. 68 It is difficult to decide whether or not the hybrid. should be treated as a taxon and then be designated by a binomial. Rather than explore this complex question here we prefer not to recognize it as a taxon but to formally refer to it by the appropriate formula. This is advantageous because it directly conveys the parentage which is the most im- portant information about a hybrid. The triploid hybrid. — There are several aspects of the natural history of the sterile hybrid. that deserve some additional discussion. Al- though its distribution is not well known, its presence has been cytologically confirmed in Ontario, Massachusetts, New Jersey, Virginia and Michigan. It is probably to be found throughout northeastern United States and adjacent Canada within the range of the two parents and in most localities where they both grow. We have found it at all (seven) localities where cytological sampling revealed the presence of both intermedia and spinulosa. In Massachusetts, it has been found at each of several localities where a search was made for it. Most of these plants were identified by their abortive spores; a few were identified by their chromosome number. The data in Chart 6 suggest the abundance of the hybrid in a few localities and also demonstrate how it may be more abundant than either parent. We have observed in the same locality that the triploid plants often have larger leaves and more leaves than those of the parents. While this might be due to environmental conditions that are more favorable for it than for either parent, it seems more likely that it is due to the phenomenon of hybrid vigor. The hybrid plants may arise in two ways. The plant may come from the fertilization of a gametophyte of one parent by one of the other parent, or it may come from an existing plant by vegetative reproduction. Although we have seen some evidence of clonal development in the hybrid we have not been able to assess the relative importance of this in the formation of populations, The principal reason being that although a close similarity of plants that are near each other may be taken as evidence of a clone, it does not follow 1966] Dryopteris spinulosa — Tryon & Britton 89 CHART 6. Relative abundance of Dryopteris spinulosa at four locali- ties, from random samples of the populations. % of Locality. Number of plants. population. ‘High and Dry’, Wellington Co., Ontario. intermedia 16 19 hybrid 56 65 spinulosa 14 16 Cockburn Swamp, Wellington Co., Ontario. intermedia 3 11 hybrid 14 52 spinulosa 10 37 Sodom, Halton Co., Ontario. intermedia 16 31 hybrid 20 38 spinulosa 16 31 Littleton, Massachusetts. intermedia 26 2T hybrid 62 64 spinulosa 9 9 that dissimilar plants always represent independent hybrids. The data in Chart 5 on variation in a crown from one year to the next (plants 32, 117) are sufficiently suggestive of the variability that may be present in a clone to make its identification in the field most uncertain. Some estimate of the extent of clonal development may be made by measuring the growth of the rhizome apex. The growth of the adult rhizome extends the apex some two or three centimeters a year and when the apex occasion- ally branches two separate plants are soon produced. Later- al stolons are also commonly produced from between the persistent leaf bases, on the older portions of the rhizome. These are slender and grow three or four centimeters in length before the internodes shorten and a crown of small leaves is produced. The base of the stolon soon dies and the rhizome of the small separate plant then may spread away from the parent. The spread of a clone by these two methods is rather slow and it would not occupy more than fifty to a hundred square meters of area in a few hundred years. In many localities where the hybrid. grows, it occupies an area of several thousand square meters. Even allowing 90 Rhodora [Vol. 68 for maximal clonal development, each of these large popu- lations must represent a few dozen, or even a few hundred, original hybrid plants. Such populations must occur fre- quently over an extensive area, at least from Virginia to New England, Ontario and Michigan, and they must repre- sent several hundred thousand F, plants. This frequency of crossing of the parental gametophytes can be evidence of the breeding behavior of the parents themselves. Since the gametophytes of the two taxa can cross with some frequency, those from different plants of the same taxon should do so with equal or greater frequency. Although we do not have sufficient data on the population Structure of intermedia and spinulosa to bring to bear as evidence of predominant outcrossing, our data on variation are consistent with the idea that outcrossing occurs. It seems appropriate to bring out this indirect evidence for relatively frequent outcrossing since so little is known about the natural breeding behavior of ferns. The relative greater abundance of the hybrid than either parent (Chart 6) is unusual, perhaps unique, in sterile hy- brid ferns. Most of these in Dryopteris are rare and occur only as an occasional plant or small colony in a very few of the places where the parents grow together, for example, D. marginalis X spinulosa var. intermedia and D. Goldiana X marginalis. A few are more common and occur in many localities where both parents grow, for example, D. cristata X spinulosa and D. cristata X spinulosa var. intermedia. The latter, perhaps more familiar as D. X Boottii, is the most common hybrid, except for intermedia X spinulosa, but it is never as abundant as its parents. We believe that the great abundance of intermedia X spinulosa is due to the large number of hybrid sporelings that are formed and to their intermediate physiological traits. Their physiology may adapt them to a large number of niches favorable for immediate survival. It may also provide them, and older plants, with the capacity to survive drought and pluvial periods, that might diminish the numbers of spinulosa, which prefers wet conditions, and also of intermedia, which prefers drier ones. 1966] Dryopteris spinulosa — Tryon & Britton 91 When we first became aware of the frequency of the hy- brid, we fully expected that further sampling would reveal a hexaploid. Although this expectation was not realized there are some aspects of the triploid hybrid that have a pertinence to studies of naturally occurring hexaploids. One of these aspects is the proximity of extreme plants of the hybrid. (Chart 4 and Graph 3) to certain plants of the parents. A hexaploid produced from one of these plants would be much closer to one parent than to the other and if the triploid became extinct (or was unknown) the rela- tions of the hexaploid might be correspondingly obscured. Further diffieulties in the analysis of the hexaploid would result if it were produced more than once and from plants at the opposite ends of the spectrum of variation of the triploid. The origin of the tetraploid. — The hybrid of spinulosa and intermedia has approximate- ly 41 pairs of chromosomes and 41 single chromosomes and on the basis of this evidence, Walker (1961) has concluded that the parents share a common genome. Therefore, he considers spinulosa to be an allopolyploid with intermedia as one of its ancestral parents. Among the known diploids only intermedia is a credible parent for spinulosa and, if spinulosa is an allopolyploid, it is evident that the other parent is either extinct or is as yet undetected. It must also be considered, however, that spinulosa may be an autopoly- ploid between two members of an original diploid spinulosa taxon which was closely related to intermedia. There are two characters of spinulosa that are not con- sistent with Walker’s hypothesis. These are the glabrous indusium and the only slightly and infrequently glandular lamina. A hybrid with typical intermedia would be ex- pected to have a glandular indusium and a more gland- ular lamina as, for example, in D. cristata X spinulosa var. intermedia and in D. cristata var. Clintoniana X spinulosa var. intermedia. However, it must be admitted that we can not be certain that the gladularity of intermedia, is always strongly dominant in its hybrids and so the slight 92 Rhodora [Vol. 68 glandularity of spinulosa can only be a qualified objection to Walker's hypothesis. Our data on the breadth of variation in intermedia, and in the hybrid, suggest that Walker's interpretation may be correct under certain cireumstances. The present data would be sufficiently consistent with his hypothesis if one of the parents of the original hybrid spinulosa were a plant of intermedia that was rather distant from the central varia- tion of the taxon in the direction of spinulosa. It would also be consistent if one of the parents of the original hybrid spinulosa was a member of an ancestral intermedia that was closer to the present spinulosa and which has since diverged from it. The variation that we have observed in spinulosa and in intermedia suggests that the origin of spinulosa may be a complex problem and that further studies are desirable to substantiate the claim of intermedia as one of its parents. BIBLIOGRAPHY. BRITTON, D. M. 1953. Chromosome studies in ferns. Amer. Jour. Bot. 40: 575-583. 1961. The problems of variation in North American Dryopteris. Amer. Fern Jour. 51: 23-30. 1962. Dryopteris dilatata (Hoffm.) A, Gray in North America. Rhodora 64: 207-212. and J. H Soper. 1966. The cytology and distribution of Dryopteris species in Ontario. Can. Jour. Bot. 44: 63-78. MANTON, I. and S. WALKER. 1953. Cytology of the Dryopteris spin- ulosa complex in eastern North America. Nature 117: 1116-1118. WALKER, S. 1959. Cytotaxonomic studies of some American Species of Dryopteris. Amer. Fern Jour. 49: 104-112, . 1961. Cytogenetic studies in the Dryopteris spinulosa complex, II. Amer. Jour. Bot. 48: 607-614. — ————. 1962. The problem of Dryopteris Leedsii. Amer. Jour. Bot. 49: 971-974. WAGNER, W. H., JR. 1963. Pteridophytes of the Mountain Lake area, Giles County, Virginia, including notes from Whitetop mountain. Castanea 28: 113-150. and D. J. HAGENAH. 1962. Dryopteris in the Huron Mountain Club area of Michigan. Brittonia 14: 90-100. VARIETIES OF BALLOTA NIGRA IN THE EAST- ERN UNITED STATES — The Black Horehound, Ballota nigra L., common in western Europe both native and as a weed, is an uncommon weed in the eastern and south- eastern United States, probably originally introduced mostly on ballast. Although no varieties or forms are mentioned in the eighth edition of Gray's Manual, it does occur in two rather easily distinguishable forms, which are called in the New Britton and Brown Illustrated Flora var. foetida Vis. and, var. ruderalis Koch, but these names are not correct. The former has the calyx lobes short and abruptly cuspidate at the apex, whereas the latter has them long-triangular and gradually acuminate. Dr. A. Patzak, of the Naturhistorisches Museum, Wien, published some time ago a revision of Ballota in which seven subspecies of B. nigra L. are recognized. Recently, a number of plants introduced into the United States were submitted to Dr. Patzak for study, and came back annotated partly as subsp. foetida Hayek and partly subsp. nigra. There can be no question as to the correctness of the latter name; as a variety it would be called var. nigra, of which var. ruderalis (Swartz) Koch is a synonym. The former name is not a new entity with Hayek (Repert. Sp. Nov. Fedde Beih. 30 (2): 278. 1929), but was based ultimately on Ballota foetida Lam. (Fl. Franc. 2: 381. 1778), which was merely a change of name for B. nigra L., and was thus superfluous when published and illegitimate. However, an originally illegitimate epithet like foetida can be validated in another category by a later author, who in this instance appears to be Koch, who published the variety B. nigra L. var. foetida Koch (Syn. Fl. Germ. Helv. 572. 1837), which must be considered as a new variety and not a transfer of B. foetida Lam. The authority for the combination B. nigra subsp. foetida should therefore be (Koch) Hayek, rather than merely Hayek. “Revision der Gattung Ballota Section Ballota," Annal. Naturhist. Mus. Wien 62: 57-86. 1958. 94 Rhodora [Vol. 68 When this plant is considered to be a variety, as such plants commonly are in floras of the eastern United States, the epithet will not be the same. Hayek was the first (ap- parently, according to Patzak’s synonymy) to propose a sub- specific epithet, and foetida thus has priority in this rank, but there were earlier varietal names proposed for this plant. Linnaeus himself realized, after the publication of B. nigra in the Species Plantarum of 1753, that there were two entities involved, and for the second he proposed Ballota aiba L. (Fl. Suec. ed. 2, 206. 1775), which is the plant with short, cuspidate calyx lobes that is called subsp. foetida by Patzak. As a variety its correct name will be B. nigra L. var. alba, (L.) J. E. Smith (Brit. Fl. 2: 635. 1804). C. V. MORTON SMITHSONIAN INSTITUTION, WASHINGTON, D. C. PASPALUM MINUS (GRAMINEAE) IN LOUISIANA AND MISSISSIPPI: Recently I reported Mobile, Alabama as a new location for Paspalum minus Fourn. (Banks, 1964). In the United States the species had been known previously in eastern Texas, near Fannett, Beaumont, and Anahuac. According to Jones et al. (1961), P. minus was collected by Silveus in Aransas county in southeastern Texas. My Mobile collec- tion extended the range of the species considerably but ap- peared to be isolated from the other locations. Last summer, while I was collecting plants for my Pas- palum studies, I found P. minus growing in Louisiana and Mississippi. The Mississippi collection was made two days after the Louisiana one during a field trip with Dr. Thomas Pullen, Department of Biology, University of Mississippi. — Contribution No. 64 from the Stephen F. Austin State College Department of Biology. 1966] Paspalum — Banks 95 Dr. Pullen arranged for lodging facilities at the University Forest Lands and provided local transportation to the collec- tion site for which I am grateful. These new locations help to bridge the gap between the Texas and Alabama stations and indicate that the species is probably more widespread in the Gulf Coastal Plain than we have realized. I contacted the following curators in Texas, Louisiana, and Mississippi to determine if other locations for the spe- cies might be known: Joseph Ewan (No), Frank Gould (TAES), G. W. Johnston (MISSA), Thomas Pullen (UNIV. OF MISS.) , Claude McLeod (SHST), Lloyd Shinners (SMU), John L. Strother (TEX), Clair A. Brown (LSU) and John Thieret (LAF). I wish to thank the above persons for checking the locations of P. minus in their herbaria and for allowing me to include the information in this paper. This correspondence did not reveal additional locations in Louisiana or Mississippi. Additional locations in Texas are, according to Strother’, Smith Point, and according to MeLeod?, Weldon, Lovelady, Diboll, Huntsville, and Grove- ton to Trinity. I do not know how P. minus was introduced into the United States. It is possible that hurricanes carried seeds from the West Indies. Another possibilty is that it may have been introduced in foreign imports of seeds of Pas- palum notatum Fligge (Bahiagrass). P. notatum is be- coming widespread in its distribution in the southeast, especially along highways and in pastures where it is often planted for erosion control and for forage. P. minus closely resembles P. notatum, the former being only slightly smaller in most characteristies, and probably accounts for its hav- ing been unnoticed. Closer observation may reveal addi- tional specimens of P. minus in colonies of P. notatum. Cytological examination of the Louisiana and Mississippi material using the aceto-carmine squash method revealed 10 chromosomes in the microspores and meiosis appeared normal. See Figure 1. Gould (1958) reported n=20 in the Huntsville material which was collected by McLeod?. My ?Personal communication. 96 Rhodora [Vol. 68 Fig. 1. Meiotic chromosomes of Paspalum minus (Banks 3500). n = 10. report therefore represents a new chromosome count for P. minus and indicates that some races are diploid. My collection data for P. minus in Louisiana and Missis- sippi and the disposition of specimens are as follows: LOUISIANA. ST. TAMMY PARISH: Edge of Pine woods along a wet ditch about 1.6 miles east of Abita Springs along Louisiana State Highway 36. Scattered clumps. D. J. Banks 3395. July 16, 1964. (ASTC, GH, LSU, MO, NO, NY, SMU, US). MISSISSIPPI. JACKSON COUNTY: In low pinelands in a wet ditch about 8 miles south of Vancleave. Plants erect. D. J. Banks 3500. July 18, 1964. (ASTC, GH, MISSA, MO, NY, SMU, University of Missis- sippi, University of Southern Mississippi, Us). DONALD J. BANKS STEPHEN F. AUSTIN STATE COLLEGE NACOGDOCHES, TEXAS LITERATURE CITED BANKS, D. J. 1964. Notes on Paspalum minus and Andropogon per- tusus in the United States. Rhodora. 66: 273-274. GOULD, F. W. 1958. Chromosome numbers in southwestern grasses. Am. Jour. Bot. 45: 757-767. JONES, F. B., C. M. ROWELL, JR. and M. C. JOHNSTON. 1961. Flower- ing plants and ferns of the Texas Coastal Bend counties. Welder Series B-1. Sinton, Texas. STUDIES IN THE FLORA OF BOLIVIA — IV. GRAMINEAE RoBERT C. FOSTER To anyone who has made use of Hitchcock’s treatment of the central and northern Andean grasses (Contrib. U. S. Nat. Herb. 24, pt. 8 [1927]), my indebtedness will be obvi- ous. I have followed it in the arrangement of genera and species, with such interpolations as time and the accretion of further material have made necessary. Since Hitchcock’s work was published almost forty years ago, it is not sur- prising that genera and species not known to him from Bolivia have now been found there. In the matter of species- concepts and synonymy, too, Hitchcock has been followed, almost without exception. This is conservative, perhaps ultraconservative, but that seems preferable to over-recogni- tion of categories of possibly uncertain value. It also means that if Hitchcock made errors they are perpetuated here. A non-agrostologist is perhaps too greatly impressed by the number of exceptions and qualifications made necessary in writing keys if the genera are first divided into tribes. Consequently, a continuous key is provided for the 109 genera treated in this paper. It is based, like the generic descriptions, on the species found in Bolivia, not on extra- territorial species. Certain points as to its terminology should be noted. A fertile floret may be either perfect or pistillate; when necessary, specific use is made of one term or the other. A sterile floret may be staminate, neuter, or a rudiment. Most often, the term is applied to an empty lemma. Again, a primary division is made on the basis of the disarticulation of spikelets — above the glumes and not above the glumes. This phraseology has been deliberate- ly used rather than above the glumes and below the glumes. It is made necessary by the fact that in some genera both glumes are lacking. Disarticulation may also occur in the pedicels or in the rachis itself, which is somewhat too far 97 98 Rhodora [Vol. 68 below the glumes to permit even a broad use of that ex- pression. Once again, this paper is based on the collections of the Gray Herbarium (GH). These have been supplemented by material borrowed from the incomparable collection of Gramineae in the United States National Herbarium (US). I am much indebted to Dr. W. L. Stern and Dr. T. R. Soder- Strom for their generous coóperation. In some cases, where Bolivian material has not been seen, collections seen by previous workers have been cited, with that worker's name in parentheses. For corrections and additions to this paper, I shall be most grateful. KEY TO GENERA 1. Inflorescences not surrounded by a hard bony involucre, 2. Disarticulation above the glumes. 3. Spikelets 1-flowered. 4. Inflorescence a solitary, terminal, pedunculate spikelet. MM 44. Aciachne. 4. Inflorescence not a solitary, terminal, pedunculate spike- let. 5. Rachilla prolonged as a bristle behind the palea. 6. Inflorescence not digitate. ........ 37. Calamagrostis. 6. Inflorescence digitate. occ... 57. Cynodon. ct Rachilla not prolonged behind the palea, except in the central spikelets only in Hordeum. 7. Inflorescence a terminal spike. 8. Spikelets 3 at a node, not unilateral on the rachis. sessen. 26. Hordeum. 8. Spikelets solitary at nodes, unilateral on the rachis, sees 58. Microchloa. 7. Inflorescence a panicle. 9. Spikelets paired, 1 pistillate and subsessile, 1 staminate and pedicellate, ................ 67. Pharus. 9. Spikelets not paired. 10. Awn of lemma excentric (in side view). 11. Lemma enclosing the membranous palea. sse 46. Nassella. 11. Lemma not entirely enclosing the coriaceous bicarinate palea. ................ RR 47. Piptochaetium. 10. Awn of lemma not excentric, or lemma awnless or cuspidate. 1966] Flora of Bolivia — Foster 99 19. Awn of lemma trifid. ........ 49. Aristida. 12. Awn of lemma, if present, undivided. 13. Lemma shorter than glumes. 14. Lemma neither fusiform nor indurate. ..............-. 38. Agrostis. 14. Lemma fusiform, indurate. ....-... si a Uo 48. Stipa 13. Lemma equal to glumes or longer than at least 1 glume. 15. Lemma awnless. s.es. EU RPM E ir e a 45. Sporobolus. 15. Lemma awned or cuspidate. 16. Awn dorsal „e.e. mE e MARO on 43. Triniochloa. 16. Awn terminal, if present, otherwise the lemma cuspi- date. ........ 42. Muhlenbergia. 8. Spikelets more than 1-flowered. 17. Spikelets with fertile florets only. 18. Rachilla villous. 19. Lemma with the true awn 1.1-2.2 cm. long. .....- NNNM 16. Cortaderia. 18. Rachilla glabrous. 20. Florets distant, glumes minute. ...Áná8 NNNM 7. Aphanelytrum. 20. Florets not distant, glumes not minute. 21. Spikelets and glumes over 1 cm. long. .......—.« NECS te 14. Arundo. 21. Spikelets and glumes less than 1 cm. long. 29. Glumes less than 2 mm. long. erer MM 8. Puccinellia. 29. Glumes more than 2 mm. long. 23. Glumes shorter than lemmas. ......-- vecececeausesecsanecececeseesseeesseeeess 19. Anthochloa. 23. Glumes longer than lemmas, seee NNNM 22. Triodia. 17. Spikelets with fertile and sterile florets, or the florets all pistillate or all staminate. 24. Plants dioecious. 25. Rachilla villous. ....MMÁ 16. Cortaderia. 25. Rachilla glabrous. 26. Leaf-blades about 1 mm. wide; inflorescence a rather condensed panicle. .... 13. Distichlis. 26. Leaf-blades to 4-6 cm. wide; inflorescence a plumose panicle. „e.e 17. Gynerium. 100 Rhodora [Vol. 68 24. Plants not dioecious. 27. Fertile florets above the sterile florets in the Spikelet. 28. Lateral nerves of first glume with a row of tubercles bearing long golden-brown hairs. RM 70. Tristachya. 28. Lateral nerves of first glume not tuberculate, not with long golden-brown hairs. 29. Spikelets strongly flattened and imbricate. np MM 64. Phalaris. 29. Spikelets not strongly flattened and imbri- cate. 30. Spikelets with more than 1 fertile floret. 31. Inflorescence a plumose panicle. RR 15. Phragmites. 31. Inflorescence not a plumose panicle. 32. Leaf-blades linear, to 6 mm. wide. ........... 1. Arthrostylidium. 32. Leaf-blades lanceolate, to 3 em. wide. ................ 2. Arundinaria. 30. Spikelets with only 1 fertile floret. 33. Fertile branches fascicled at nodes of main plant-axis; terminal floret perfect. sees 3. Chusquea. 33. Fertile branches not fascicled at nodes of main plant-axis; terminal floret pistillate, 34. Lemma of pistillate floret glab- rous or scabridulous. eee. TTT etn 68. Arundinella. 34. Lemma of pistillate floret densely long-pubescent. sess. HM 69. Trichopteryz. 27. Fertile florets below the sterile florets in the spikelet. 35. Inflorescence condensed into small glomerules more or less hidden by leaf-sheaths. ................ 5LIMMMEA 63. Munroa. 35. Inflorescence not condensed into glomerules. 36. Inflorescence digitate or paniculate (in one instance a panicle with whorls of digitately arranged branches). 37. Inflorescence digitate. 38. Rachis prolonged beyond the upper- most spikelet as a short subulate point, ................ 56. Dactyloctenium. 1966] Flora of Bolivia — Foster 101 38. Rachis not prolonged beyond the uppermost spikelet. 39. Spikelets with more than 1 per- fect floret. .........- 55. Eleusine. 39. Spikelets with only 1 perfect floret. .. RH 60. Chloris. 37. Inflorescence a panicle. 40. Lemma wing-margined. ... 40. Lemma not wing-margined. 41. Rachilla glabrous. 42. Glumes or lemmas awned. 43. Spikelets unilateral in 2 rows on the rachis. .......-- Et 61. Trichloris. the rachis. 44. Lemmas longer than glumes. 45. Inflorescence a spi- cate, dense, inter- rupted panicle. ...... Rude 31. Trisetum. 45. Inflorescence an open racemose pani- cle. .... 53. Gouinia. 44. Lemmas shorter than glumes. eA "— 59. Gymnopogon. 42. Glumes and lemmas awnless. 46. Spikelets unilateral on the the rachis in 2 rows. — 52. Leptochloa. 46. Spikelets not unilateral on rachis in 2 rows. 47. Glumes longer than lemmas. 48. Spikelets paired, 1 sessile and 1 pedi- cellate. .. 34. Holcus. 48. Spikelets not pair- ed. 49. Inflorescence spicate, dense. .. .... 29. Dissanthe- lium. 102 Rhodora [Vol. 68 49. Inflorescence not Spicate, the branches dis- tant, spreading. ee 18. Melica. 47. Glumes equal to or shorter than lemmas. 50. Leaf-blades long- petiolate, lance- ovate. eese EM 21. Zeugites. 50. Leaf-blades not long-petiolate ^ nor lance-ovate. 51. Paleas persist- ent on rachilla after fall of lemmas. ........... . 12. Eragrostis. 51. Paleas not per- sistent. .. 10. Poa. 4l. Rachilla not glabrous. 52. Lemma not bidentate. 53. Lemma apically dissected into numerous awns. ........ MM 23. Pappophorum. 53. Lemma not apically dis- sected, 54. Paleas persistent on rachilla after fall of lemmas. nsere ee 12. Eragrostis. 54. Paleas not persistent. 55. Spikelets 1.5 cm. long or longer, ... — 4. Bromus. 55. Spikelets rarely longer than 1 cm., usually shorter. 56. Glumes longer than lemmas. .. serere 18. Melica. 56. Glumes equal to or shorter than lemmas. 1966] Flora of Bolivia — Foster 103 57. Lemmas awnless. ......- S 10. Poa. 57. Lemmas AWE, ........- .. 6. Festuca. 52. Lemma bidentate or 4- toothed. 58. Lemma-awn dorsal. 59. Spikelets 1.5-2.5 cm. long. s 33. Avena. 59. Spikelets 4-5 mm. long. 2d 32. Deschampsia. 58. Lemma-awn from base of sinus between teeth. 60. Spikelets about 2.5 cm. long. ienn 4. Bromus. 60. Spikelets 1.5 cm. or less. 61. Glumes shorter than lemmas. .......- . 81. Trisetum. 61. Glumes longer than lemmas, 62. Spikelets 1.5 cm. long. 63. Leaves glab- TOUS, nenne 9. Amphibro- mus. 63. Leaves papil- late - hirsute. . 85. Dan- thonia. 62. Spikelets about 7 mm. long. ...... . 96. Lampro- thyrsus. 36. Inflorescence neither digitate nor pani- culate, but a solitary terminal spike or raceme. 64. Inflorescence a raceme with few, distant, ultimately spreading spikelets. MEM 5. Brachypodium. 104 Rhodora [Vol. 68 65 Spikelets placed edgewise to the rachis, sesse. 27. Lolium. 65. Spikelets not placed edgewise to the rachis. 66. Glumes awned. ........ 25. Elymus. 66. Glumes awnless. 67. Glumes shorter than lemmas. — HR 54. Tripogon. 67. Glumes not shorter than lemmas. 68. Plants 1-2 dm. high; spikes 1-3 cm. long. ... -~ 62. Bouteloua simplex. 68. Plants to 1-2 m. high; spikes 7-20 em. long. ...... MEM 24. Agropyron. 70. Spikelets with unisexual or neuter florets. 71. Inflorescence a spicate panicle or raceme, 72. Inflorescence a raceme. 73. Spikelets in whorls around the rachis. .......... mE 28. Pariana. 73. Spikelets not in whorls around the rachis. .... IM 51. Aegopogon. 72. Inflorescence a Spicate panicle. ........ 41. Lycurus. 73a. Pistillate and staminate spikelets in separate inflorescences on the same plant. ...... 66. Luziola. 73a. Pistillate and staminate spikelets not in separate 74. Spikelets not paired. co.cc, 94. Olyra 74. Spikelets paired. ................ 100. Hyparrhenia. 70. Spikelets with perfect florets. 75. Inflorescence a raceme or spike. 76. Inflorescence a raceme, ............... 51. Aegopogon. 76. Inflorescence a Spike, eee 50. Tragus. 75. Inflorescence a panicle. Ti. Glumes lacking. Jess 65. Leersia. 77. Glumes present. 78. Spikelets paired. 79. One spikelet perfect, the other staminate or neuter. arcen. 41. Lycurus. 79. One spikelet perfect and sessile, the other pistillate and pedicellate. 97, Eriochrysis. 78. Spikelets not paired. 80. Spikelets awnless. ............ 20. Orthoclada. 1966] Flora of Bolivia — Foster 105 80. Spikelets awned. 81. Inflorescence not interrupted. .............- ERE A DER OREL 39. Alopecurus. 81. Inflorescence interrupted. ............- MUI EEUU 40. Polypogon. 69. Spikelets with 2 or more florets. 82. Fertile floret below the sterile floret. 83. Inflorescence a panicle; spikelets usually solitary. Codb LER ements 20. Orthoclada. 83. Inflorescence of pectinately arranged spikes; spike- lets not solitary. ..... MM 62. Bouteloua. 82. Fertile floret above the sterile floret. 84. Spikelets subtended by bristles, these separate, or at least partially united and forming a bur. 85. Spikelets and bristles forming a BUTS e terre devesccssvscassscsccscccscesenauacnscessonscacossssscsonasees 93. Cenchrus. 85. Spikelets and bristles not forming a bur. 86. Bristles persistent, ....... 91. Setaria. 86. Bristles deciduous. ...........--- 92. Pennisetum. 84. Spikelets not subtended by bristles. 87. First glume absent. 88. Spikelets without unisexual florets. 89. Spikelets not unilateral on the rachis. .... E E 71. Leptocoryphium. 89. Spikelets unilateral on the rachis. 90. Second glume and fertile lemma turned away from the rachis. .... 80. Axonopus. 90. Second glume and fertile lemma turned toward the rachis (first glume rarely present but minute). 91. Fruit flexible, cartilaginous, usually dark. HM 13. Digitaria. 91. Fruit rigid, chartaceous, usually light. ........m HH 81. Paspalum. 88. Spikelets with unisexual florets; pistillate and staminate florets in separate spikelets. ........ vcunueeuusensecencnaneceesseceeeussssessseeeeeeeeeeeseeeeeeee 94. Olyra. 87. First glume present. 92. Second glume ventricose and uncinately spiny at maturity. ee 75, Pseudechinolaena. 92. Second glume not ventricose nor uncinately spiny at maturity. 93. Spikelets paired, with 1 spikelet sessile, the other pedicellate. 94. Both spikelets, or only 1, with a per- fect floret. 95. Both spikelets with a perfect floret. 106 Rhodora [Vol. 68 96. Spikelets awnless. sesse. — ÉE— 96. Saccharum. 96. Spikelets awned. .. 98. Erianthus. 95. One spikelet with a perfect floret. 97. Perfect spikelet pedicellate. aravassonaiesereseressenne 104. Trachypogon. 97. Perfect spikelet sessile. 98. Disarticulation immediately below glumes. ............. een —— 99. Andropogon. 98. Disarticulation not immedi- ately below glumes, but in rachis or pedicel. 99. First glume concealing the rest of the spikelet. ........ TE 107. Hackelochloa. 99. First glume not conceal- ing spikelet. 100. Inflorescence a soli- tary raceme (occa- sionally more than 1). 101. Spikelets awned. .. . 103. Hetero- pogon. 101. Spikelets awnless. 102. Glumes pubes- cent, hirsute or nurus. 102. Glumes glab- YOUS. secssssivecessss suris, 100. Inflorescence panicu- late. 103. Pedicellate spike- let staminate. ...... .. 101. Sorghum. 103. Pedicellate spike- let reduced to a pedicel, .................. eene 102. Sorghas- trum. 94. Neither spikelet with a perfect floret. —ÓÉORRE 108. Tripsacum. 93. Spikelets not ‘paired, or if in pairs, not 1966] Flora of Bolivia — Foster 107 with 1 spikelet sessile and the other pedi- cellate. 104. Inflorescence of solitary, or mostly so, racemes. 105. Spikelets paired. ..... 74. Thrasya. 105. Spikelets unpaired, but sometimes in 2 rows. 106. Spikelets (pistillate only) sunken in rachis. e.i iinsert. NMERRUNUR-: Su 108. Tripsacum. 106. Spikelets not sunken in rachis. 107. Sterile lemma and second glume with tufts of hairs. .. NEUE BIET 77. Mesosetum. 107. Sterile lemma and second glume glabrous. ............ MEE 79. Brachiaria. 104. Inflorescence a panicle. 108. Spikelets paired. 109. Glumes awnless. 110. Sterile lemma and second glume alike. ...... 72. Trich- achne. 110. Sterile and fertile lemmas alike, unlike second glume. ——Ó 95. Imperata. 109. Glumes awned or mucronate. 111. Ring-callas on rachilla at base of spikelet. ................- NEMPE 78. Eriochloa. 111. Ring-calus absent from rachilla. ...... 89. Oplismenus. 108. Spikelets unpaired or clustered. 112. Spikelets clustered. 113. Spikelets awnless, not plano- convex. .... 86. Hymenachne. 113. Spikelets awned, plano- COnVeX. ...... 90. Echinochloa. 112. Spikelets unpaired. 114. Spikelets placed obliquely on pedicels, ............ 84. Lasiacis. 114. Spikelets not placed obliquely on pedicels. 115. Spikelets awnless. 116. Second glume in- flated, saccate. .......... M eibi 85. Sacciolepis. 108 Rhodora [Vol. 68 116. Second glume not in- flated nor saccate. 117. Lemma stipitate. .. . 83. Ichnanthus. 117. Lemma not stipi- tate. 118. Spikelets with 2 perfect flo- rets, subglo- DOSE, ................ .. 88. Isachne. 118. Spikelets with 1 perfect floret, not subglobose. 119. First glume as long as Spikelet, ...... ..87. Homo- lepis. 119. First glume not as long as spikelet. ... 82, Pani- cum. 115. Spikelet awned or mucro- nate. 120. Ring-callus on rachil- la at base of spikelet. NEN 78. Eriochloa. 120. Ring-callus absent. 121. Glumes very un- equal. .............. ees sen 76, Rhynchely- trum. 121. Glumes equal or subequal. ................ ... 89. Oplismenus. 1. Inflorescences, at least the pistillate, surrounded by a hard bony involuere. essre 109. Coix. 1. Arthrostylidium Rupr. Arthrostylidium racemiflorum Steud. Syn. Pl. Glum. 1: 336 (1854). Perennial, climbing or sprawling, to 5 m. long; main axis to 5 mm. thick, the fertile branches clustered at the nodes. Leaf-blades linear, to 6 mm. wide, often short-pubescent above and beneath, with scattered longer hairs. Inflorescence racemose, the racemes to 8 cm. long, the 1966] Flora of Bolivia — Foster 109 rachis pubescent. Spikelets few-flowered, the lowest lemmas empty; disarticulation above the glumes. Lemma to 6 mm. long, 7-nerved, awned, the awn 1-2 mm. long. Palea longer than the lemma, the keels ciliolate. SANTA CRUZ: ICHILO: Río Yapacani, 350 m., Steinbach 7587 (GH). This collection is a sterile sheet identified by Mrs. Agnes Chase as “probably” this species. 2. Arundinaria Michx. Arundinaria Herzogiana Henrard in Meded. Rijks Herb. Leiden, no. 40: 75 (1921). Shrubby perennial. Leaf-sheaths long-fimbriate at the apex, the fimbriae to 2 cm. long; blades lanceolate, acute, to 25 cm. long and 3 cm. wide, glaucous, glabrous above, scabrous and more or less hirtellous beneath. Inflorescence a terminal, lax, racemose panicle to 3 dm. long, hardly, if at all, exserted from the uppermost leaves, the pubescent branches distant on the axis. Spikelets 2-3 cm. long, pedicellate, about 5-flowered; disarticulation above the glumes. First glume absent; second glume about 1 mm. long, 3-nerved, the margin ciliate, with a 3-4 mm. awn. Sterile lemmas 2, awned, the first pubescent on the upper portion. Fertile lemmas to 1 cm. long, 7-nerved, shortly puberulent, with a 2-3 mm. awn. LA PAZ: INQUISIVI [?]: Río Saujana, 3200 m., Herzog 2396 (US; type-number). There is no information available as to the height of the plant. 3. Chusquea Kunth Shrubby perennials, sometimes low-growing, or the culms 2-7 m. long, slender to stout, woody. Leaf-blades lanceolate to linear- lanceolate, 2-18 cm. long, acute, disarticulating from the persistent sheaths. Fertile branches loosely to densely fascicled at the nodes, or the inflorescences terminal on the culms; inflorescence an open to rather crowded panicle, or spicate in appearance, spikelets few to numerous in an inflorescence. Spikelets pedicellate, small; glumes present or absent; sterile lemmas 2, below 1 perfect floret; disarticu- lation above the glumes. a. Glumes absent or less than 0.5 mm. long. b. Glumes absent. ....... eem 2. C. delicatula. b. Glumes present, less than 0.5 mm. long. c. Culms 2-7 m. long, 1 cm. thick; fertile stems densely clustered at nodes, with numerous spikelets... 4. C. peruviana. c. Culms much shorter, 4 mm. thick; fertile stems few at nodes, spikelets few. eene emen 1. C. uniflora. a. Glumes present, 1 mm. long, or more. d. Rachis, branches and pedicels glabrous. „eee 5. C. serrulata. d. Rachis, branches and pedicels puberulent. 110 Rhodora [Vol. 68 e. Spikelets mostly distinctly pedicellate, the panicle somewhat open; lemmas distinctly awned. ......... 6. C. scandens. e. Spikelets rarely pedicellate, the inflorescence densely spicate in appearance, lemmas shortly mucronate. ............ 3. C. spicata. 1. Chusquea uniflora Steud. Syn. Pl. Glum. 1: 827 (1854). C. longipendula O. Ktze. Rev. Gen. 3(2): 348 (1898). Leaf-blades pu- bescent or glabrous beneath. Spikelets to 7 mm. long, sterile and fertile lemmas strongly carinate, the keel prolonged as a distinct scabrous awn. LA Paz: NOR YUNGAS: Bella Vista, Hitchcock 22753 (GH). 2. C. delicatula Hitchcock in Contrib. U. S. Nat. Herb. 23: 309 (1927). Leaf-blades glabrous beneath. Spikelets 5-6 mm. long; sterile lemmas ending in a short subulate point. LA PAZ: NOR YUNGAS: Bella Vista, Hitchcock 22748 (GH; type-number). 3. C. spicata Munro in Trans. Linn. Soc. 26: 60 (1868). Leaf- blades glabrous beneath, the margins densely ciliolate. Spikelets about 6 mm. long. LA PAZ: LARECAJA: Mapiri, Rusby 196 (Hitchcock). 4. C. peruviana E. G. Camus, Bamb. Monogr. 88, t. 53B (1913). Leaf-blades sparsely pilose beneath, the apex long-setaceous. Spikelets glabrous, to 10 mm. long, the lemmas shortly mucronate. LA PAZ: NOR YUNGAS: Bella Vista, Hitchcock 22746 (GH), 22744 (GH). 5. C. serrulata Pilger in Engler, Bot. Jahrb. 25: 719 (1898). Leaf- blades glabrous beneath. Spikelets about 7 mm. long, the lemmas mucronate. LA PAZ: MURILLO: La Paz, Rusby 10 (Hitchcock); NOR YUNGAS: Bella Vista, Hitchcock 22750 (GH) ; Coroico, Bang 2348 (GH) ; SUR YUNGAS: Colaya, 1810 m., Mexia 4298 (GH). 6. C. scandens Kunth, Syn. Pl. Aequin. 1: 254 (1822). C. quitensis var. patentissima Hack. in Fedde, Rep. Sp. Nov. 6: 161 (1908). Leaf- blades somewhat pilose beneath, the apex setaceous. Spikelets about 8 mm. long. LA Paz: NOR YUNGAS: Unduavi, 3300 m., Buchtien 4185 (GH); Bella Vista, Hitchcock 22747 (GH). 4. Bromus L. Annual or perennial, the culms erect or spreading, sometimes prostrate. Inflorescence a terminal racemose panicle, the spikelets several- to many-flowered, the puberulent rachilla disarticulating above the unequal glumes, the lemmas awned or awnless, glabrous or pubescent, the uppermost floret a sterile lemma or a rudiment. a. Glumes scabridulous to scabrous, but not puberulent nor pubescent. b. Lemmas pubescent, awn twisted. sss. 1. B. Trini b. Lemmas not pubescent or only very slightly so, awn not twisted. c. Leaf-blades to 1 cm. wide; lemma tuberculate-scabridulous. .... uM 2. B. segetum. c. Leaf-blades to 5 mm. wide; lemma not tuberculate-scabridu- lous. Meme eene S. B. catharticus. a. Glumes puberulent to pubescent. 1966] Flora of Bolivia — Foster 111 d. Inflorescence-rachis and pedicels short-pubescent. .... 4. B. lanatus. d. Inflorescence-rachis and pedicels scabridulous but not pubescent. rue d evotn ceklestesevi cobobuis TTT 5. B. pitensis. 1. Bromus Trinii Desv. in Gay, Fl. Chile, 6: 441 (1858). Annual, to 6 dm. high, the culms, inflorescence-rachis and pedicels glabrous. Leaves sparsely pilose on the sheaths, the blades pilose, at least when young. Spikelets to a little over 2 cm. long, about 7-flowered, the uppermont floret reduced to a narrow sterile lemma. First glume about 1.1 em. long, 1-nerved; second glume about 1.3 cm., 3-nerved, the nerves green. Lowest lemma about 1 cm. long, 2-toothed, the teeth very evident, the awn about 1.4 em. long. Palea about 9 mm. long, bicarinate, the keels ciliate near the apex. LA PAZ: LARECAJA: Sorata, Mandon 1370 (US). 2. B. segetum HBK. Nov. Gen. & Sp. 1: 151 (1816). Perennial, the panicle to 2 dm. long, culms, rachis and pedicels glabrous. Leaves glabrous, the blades scabridulous beneath. Spikelets to 1.5 em. long, 2.4-flowered, the uppermost floret reduced to a rudiment. First glume 5 mm. long, 1-nerved, the second glume to 8 mm. long, 3-nerved. Lowest lemma about 1.2 cm. long, shortly bidentate, awn about 2 mm. long. Palea as long as lemma, bidentate, bicarinate, the keels ciliolate. COCHABAMBA: CHAPARE: La Aduana, 3000 m., Steinbach 9533 (GH). 3. B. catharticus Vahl, Symbol. Bot. 2: 22 (1791). B. unioloides HBK. Nov. Gen. & Sp. 1: 151 (1816). B. angustatus Pilger in Engler, Bot. Jahrb. 25: 719 (1898). Annual or perennial, the culms glabrous, the rachis and pedicels scabridulous, the open panicles to 3 dm. long. Leaf-sheaths glabrous, blades scabridulous beneath. Spikelets about 2.5 cm. long, with up to 9 florets, the uppermost a rudiment. First glume to 9 mm. long, 3-nerved, the second glume to 1 cm., 5-nerved. Lowest lemmas to 1.4 cm. long, not toothed, awn about 5 mm. long, upper lemmas awnless. Palea to 1.3 em. long, bidentate, bicarinate, the keels long-ciliate. LA PAZ: MURILLO: La Paz, 3700 m., Buchtien 8533 (GH). COCHABAMBA: CERCADO: Valle de Cochabamba, Steinbach 8798 (GH). TARIJA: CERCADO: Calderillo, 3200 m., Fiebrig 2904 (GH). Very variable in size, spikelet-length and awn-length. Many lemmas awnless. 4. B. lanatus HBK. Nov. Gen. & Sp. 1: 150 (1816). B. Pflanzü Pilger in Engler, Bot. Jahrb. 49: 189 (1912). Perennial, to 6 dm. high, the leaf-blades somewhat villous. Panicle very lax, 10-20 cm. long. Spikelets to 2 cm. long, with about 6 florets, the uppermost a rudiment. First glume to 1 cm. long, 3-nerved, the second glume to 1.3 cm. long, 5-nerved. Lowest lemmas about 1.2 cm. long, densely villous along the margins, shortly bidentate, the awn 4-5 mm. long. Palea about 1 cm. long, bicarinate, the keels ciliate. LA PAZ: INGAVI: Comanche, Asplund 6459 (Hitchcock). TARIJA: CERCADO: Calderillo, 3200 m., Fiebrig 2904 in part (GH). 5. B. pitensis HBK. Nov. Gen. & Sp. 1: 152 (1816). B. Buchtieniüi ssssssesssesses 112 Rhodora [Vol. 68 Hack. in Fedde, Rep. Sp. Nov. 11: 30 (1912). Perennial, to 1.5 m. high. Leaf-blades glabrous or somewhat pilose, Panicles lax, up to 3 dm. long. Spikelets to 2.5 cm. long, up to 9-flowered. First glume about 6 mm. long, l-nerved, the second glume about 9 mm. long, 3- nerved. Lowest lemmas to 1.3 cm. long, bidentate, sparsely pubescent or villous on the lower half, the awn 4 mm. long. Palea about 1.1 em. long, bidentate, bicarinate, the keels ciliate. Anthers villous at base. LA PAZ: LARECAJA: Sorata, 2600-2800 m., Mandon 1365 (GH); MURILLO: La Paz, 3800 m., Buchtien 518 (GH). 5. Brachypodium Beauv. Brachypodium mexicanum Link, Hort. Berol. 1: 41 (1827). Peren- nial, straggling, to 6 dm. high. Leaves glabrous, the blades usually much shorter than the inflorescence, scabridulous beneath, acute, to 1.2 cm. wide, usually narrower. Inflorescence a solitary, terminal raceme of a few distant spikelets, these at first appressed, ultimately spread- ing or deflexed, the raceme to 10 cm. long. Spikelets to 2.5 cm. long (excluding the awns), subterete, to 10-flowered, the glabrous or upper portion and downward along the lateral nerves, awned, the minutely puberulent rachilla disarticulating above the glumes; glumes unequal, the first about 6 mm. long, the second to 7 mm. long, both awnless; lemmas about 9 mm. long, 5-7-nerved, puberulent on the awn about 4-6 mm. long, the 2 uppermost florets reduced and sterile; palea as long as the lemma, bicarinate, the keels densely ciliate. LA PAZ: LARECAJA: Sorata, Holway 544 (US). COCHABAMBA: CERCADO: Cochabamba, Tequifia, Hitchcock 22855 (US), Holway 409 (vs). 6. Festuca L. Annual or perennial, of varying heights, Leaves flat or involute, shorter than or exceeding the inflorescence. Inflorescence a terminal racemose panicle, the spikelets few. to several-flowered, the puberu- lent rachilla disarticulating above the unequal or subequal glumes. Lemmas firm to somewhat indurate in texture, awned, awn-tipped or awnless, the uppermost reduced and sterile, a. First glume 1 mm. long or less; lemmas ciliate at apex, tuber- culate-scabridulous. ...........ssssseee 1. F. megalura. a. First glume at least 1.5 mm. long; lemmas not ciliate at apex, not tuberculate-scabridulous. b. Lemmas with awns at least 5 mm. long. ............ 2. F. ulochaeta. b. Lemmas awn-tipped (about 1 mm.) or awnless. c. Rachis and pedicels scabrid to pubescent; lemmas not pubes- cent along margin. d. Lemmas (at least some) awn-tipped. e. Leaf-blades flat. sss 3. F. procera. e. Leaf-blades involute and pungent. f. Lemmas glabrous or puberulous at apex. .. 4. F. sublimis. 1966] Flora of Bolivia — Foster 113 f. Lemmas scabridulous at apex. seese 5. F. rigescens. d. Lemmas awnless ........../"7 DEA ES 1. F. dolichophylla. c. Rachis and pedicels glabrous; lemmas pubescent along margin. vosueceussssecensssscsnsccnccsensceascscsesssconscsassnecsonssonsesnasseeeees 6. F. orthophylla. 188 (1848). Annual. Panicles narrow, usually over 10 cm. long, often exceeded by the uppermost flat glabrous leaf-blades. Spikelets (excluding awns) about 1 cm. long, with 4-5 florets. Glumes unequal, the first about 1 mm. long or less, the second, 2.5-3 mm. long. Lowest lemmas to 5 mm. long with a 1.5 cm, scabrid awn, upper lemmas long-ciliate near the apex, most lemmas tuberculate-scabridulous (at least in dried material). Palea as long as lemma, bicarinate, the ciliate keels nearly marginal. LA PAZ: LARECAJA: vic. Sorata; Ullon- tiji, 2650-2900 m., Mandon 1363 (GH). COCHABAMBA: CHAPARE: Inca- chaca 2250 m., Steinbach 9498 (GH); CERCADO: south side of Mt. Tunari, 3300 m., Steinbach 9850 (GH). Potosi: CERCADO: Potosi, Cár- denas 203 (GH). TARIJA: CERCADO: Calderillo, 3200 m., Fiebrig 3159 (GH). 2. F. ulochaeta Steud. Syn. Pl. Glum. 1: 305 (1854). Perennial, up to 1 m. high, the culms, rachis and pedicels scabrid. leaf-blades sometimes exceeding the inflorescence, to 5-8 mm. wide, scabrid be- neath. Panicle fairly lax, to 2 dm. long. Spikelets (excluding awns) about 8 mm. long, 4-5 flowered, Glumes 1.5 and 2.5 mm. long. Lowest lemmas 5-6 mm. long, convex to rounded, 3-nerved, the central nerve scabridulous at apex, awns very filiform, flexible, 5-8 mm. long. Palea as long as lemma, bicarinate, the keels ciliolate. LA PAZ: MURILLO: Unduavi, Buchtien 6415 (Hitchcock). COCHABAMBA: CHA- PARE: Incachaca, 2300 m., Steinbach 8976 (GH). 3. F. procera HBK. Nov. Gen. & Sp. 1: 154 (1816). F. Fiebrigü Pilger in Engler, Bot. Jahrb. 37: 510 (1906). Perennial, to 1-2 m. high, the panicle to 3 dm. long; rachis and pedicels short-pubescent, leaf-blades narrow, flat or somewhat folded, scabrid beneath. Spike- lets about 8 mm. long, 3-5-flowered. Glumes 4 mm, and 5 mm, long, the first 1-nerved, the keel scabrid near the apex, the second 3-nerved, at least the keel scabrid above. Lowest lemmas to 6-7 mm. long, rounded, several-nerved, scabrid to scabrid-puberulent, especially on the upper portion, awn-tipped. Palea as long as lemma, bicarinate, the keels ciliolate, bidentate, the apex puberulous. TARIJA: AVILES: Pinos, 2800 m., Fiedrig 3117 (GH; type number of F. Fiebrigii), 3118 (GH). 4. F. sublimis Pilger in Engler, Bot. Jahrb. 25: 718 (1898). Perennial, caespitose, to 1 m. high, the involute, glabrous to scabrid, pungent leaf-blades sometimes exceeding the inflorescence. Panicle fairly lax, to 2.5 dm. long, rachis and pedicels scabrid to short- pubescent, Spikelets about 1 cm. long, 4-flowered (in most). Glumes 4 mm. and 5 mm. long, the second minutely ciliolate and sometimes 114 Rhodora [Vol. 68 scabridulous. Lowest lemmas about 7 mm. long, awnless, the upper ones awn-pointed (1 mm.), glabrous to puberulous. Palea as long as lemma, bidentate, puberulent, bicarinate, the keels ciliate. LA Paz: LARECAJA: vic. Sorata; Chileani, Mandon 1362 bis (GH). 5. F. rigescens (Presl) Kunth, Rév. Gram. 1: suppl. XXXI (1830). Caespitose perennial, to 4 dm. high, leaf-blades involute, pungent, glabrous, sometimes exceeding the narrow 5-20 cm. panicle, the rachis and pedicels puberulent. Spikelets to 1.1 cm. long, about 4-flowered. Glumes 6 mm. and 7 mm. long, the second scabridulous near the apex. Lowest lemmas to 8 mm. long, rounded, the apex scabridulous, some- times slightly 2-toothed, awnless to awn-pointed. Palea shorter than lemma, puberulous on both surfaces, bicarinate, the keels ciliolate. LA Paz: YUNGAS, Bang 173 (GH); MURILLO: Palca, 3700 m., Hitchcock 22560 (GH). Potosi: CERCADO: Potosi, 4000 m., Cardenas 208 (GH). 6. F. orthophylla Pilger in Engler, Bot. Jahrb. 25: 717 (1898). F. orthophylla var. boliviana Pilger in Engler, Bot. Jahrb. 37: 508 (1906). Caespitose perennial, to 5 dm. high, leaves involute, glabrous, firm, pungent, often equaling or exceeding the inflorescence. Panicle narrow, the branches ascending, rachis and pedicels glabrous. Spike- lets about 1 cm. long, 4-6-flowered. Glumes about equal, 6 mm. long, the first l-nerved, the nerve scabrid, margins slightly ciliate, with an apical tuft of hairs, the second 3-nerved, broader than the first but otherwise similar. Lowest lemmas about 7 mm. long, awnless, pubescent along the margin. Palea as long as lemma, pubescent at apex, bicarinate, the keels ciliate, at least on the uppper half. TARIJA: AVILES: Puna Patanca, 3700 m., Fiebrig 3192 (GH; type-number of F. orthophylla var. boliviana). 7. F. dolichophylla Presl, Rel. Haenk. 1: 258 (1830). F. laeteviridis Pilger in Engler, Bot. Jahrb. 37: 510 (1906). F. Buchtienii Hack. in Fedde, Rep. Sp. Nov. 6: 160 (1908). F. Pflanzii Pilger in Engler, Bot. Jahrb. 49: 188 (1912). Caespitose perennial, 4-10 dm. high, the involute pungent leaves glabrous or scabrid, occasionally equaling the inflorescence. Panicle narrow, the branches appressed or spreading, to 15 em. long, the rachis and pedicels scabridulous. Spikelets to 1.1 cm. long, 3-4-flowered. Glumes unequal, the first to 4.5 mm. long, apically scabrid, the second to 7 mm. long, ciliolate at base, apically scabrid. Lowest lemmas to 7 mm. long, awnless, scabrid toward apex, but otherwise glabrous. Palea as long as lemma or longer, bidentate, bicarinate, the keels minutely ciliolate. LA Paz: LARECAJA: vic. Sorata; Millipaya, 3700-4200 m., Mandon 1361 (GH) ; MURILLO: Palca, 3700 m., Hitchcock 22568 (GH); La Paz, 3300 m., Bang 33 (GH), 4100 m., Buchtien 428 (Gu). COCHABAMBA: CERCADO: Cerro Molle-Molle, Stein- bach 4093 (GH), 3800 m., Steinbach 4062 (GH); AYOPAYA: Tabacruz, 4200 m., Steinbach 9760 (GH); south side of Mt. Tunari, 3600 m., Steinbach 9787 (GH), 9784 (GH). POTOSÍ: SUR CHICHAS: Tupiza, 2700 m., Fiebrig 2955 (GH; type-number of F. laeteviridis). 1966] Flora of Bolivia — Foster 115 7. Aphanelytrum Hack. Aphanelytrum procumbens (Hack.) Hack. in Oesterr. Bot. Zeitschr. 52: 13 (1902). A straggling perennial, culms decumbent, sometimes stoloniferous, to 8 dm. high. Leaves glabrous, not scabridulous, much shorter than the inflorescence, the short acute blades about 3 mm. wide. Inflorescence a lax racemose panicle up to 2 dm. long, the very slender branches distant and bearing 2-3 spikelets at their ends. Spikelets about 1.2 cm. long, pedicellate, 2-flowered, the naked rachila disarticulating above the glumes, the second floret distant from the first; glumes greatly reduced, minute; lemmas to 7 mm. long, long-acute but not awned, more or less hyaline, several-nerved, the nerves green; palea similar to the lemma but shorter, 1-nerved, the nerve green. LA PAZ: MURILLO: Unduavi, 3300 m., Buchtien 4268 (US); NOR YUNGAS: Bella Vista, Hitchcock 22756 (us). 8. Puccinellia Parl. Perennials, low-growing, the more or less decumbent culms up to 20 em. high. Leaves glabrous, short, very narrow, the uppermost blade on a culm equaling or sometimes exceeding the inflorescence and often enveloping its base. Inflorescence a short, dense, race- mose panicle up to 12 cm. long, the branches erect, appressed, the pedicels often (or usually) short, thick, sulcate. Spikelets 2-3- flowered, all florets perfect, glumes present, the naked rachilla dis- articulating above the glumes; lemma rounded, the nerves obscure, glabrous, but sometimes puberulent at the base, awnless. Spikelets about 2 mm. long; lemma about 1 mm. long. .... 1. P. parvula. Spikelets 4-5 mm. long; lemma about 2.5 mm. long. .... 2. P. oresigena. 1. Puccinellia parvula Hitche. in Contrib. U. S. Nat. Herb. 24: 325 (1927). Spikelets 2.3.flowered. Glumes unequal, the first about 0.5 mm. long, the second, 1 mm. long. Lemmas elliptic-oblong, the apex obtuse or lacerate; palea about as long as the lemma. PoTosi: SUR CHICHAS: Atocha, 3300 m., Hitchcock 22878 (US; type). 2. P. oresigena (Phil). Hitche. in Contrib. U. S. Nat. Herb. 24: 326 (1927). Spikelets 3-flowered. Glumes unequal, the first about 1 mm. long, the second, 1.5 mm. long. Lemmas acute. PoTosi: SUR CHICHAS: Atocha, Hitchcock 22879 (GH). 9, Amphibromus Nees Amphibromus scabrivalvis (Trin.) Swallen in Amer. Journ. Bot. 18: 413 (1931). A rhizomatous perennial, the culms erect (the base sometimes decumbent), to 1 m. or more in height. Leaves glabrous, the lowest blade to 40 cm. long, the uppermost reduced, 2-6 mm. wide. Inflorescence a rather lax racemose panicle to 20 cm. long, the slender branches often spreading or drooping. Spikelets pedicellate, up to 7-flowered, the florets distant, the uppermost floret sterile, to 1.5 cm. 116 Rhodora [Vol. 68 long (excluding the awns), the rachilla unilaterally short-villous, with an apical ring of hairs below each floret, disarticulating above the glumes. Glumes unequal, the first about 4 mm. long, the second, 5 mm. long, the apex erose. Lemmas about 7 mm. long, bifid to the middle or below, the 1.2 em. twisted geniculate awn arising from the base of the sinus, up to 9-nerved, scabrid, the apices erose; palea shorter than the lemma, bicarinate. LA PAZ: MURILLO: La Paz, Buchtien 6422 (us). 10. Poa L. Annual or perennial, one species dioecious. Culms usually but not always exceeding the leaves. Leaf-blades flat or involute, usually scabridulous beneath. Inflorescence a terminal racemose panicle, narrow and condensed or more open and spreading, the rachis and pedicels glabrous, puberulent or scabridulous, sometimes sparsely so. Spikelets 2-6-flowered, the glumes usually unequal, the glabrous or puberulent rachilla disarticulating above the glumes; fertile lemmas below the sterile lemma or rudiment. a. Plants dioecious. l, 1. P. Buchtienii. a. Plants not dioecious. b. Dwarf plants, about 4 cm. high; glumes and lemmas subquadrate. Tete tte e eee e tere eH MILLIA d. Spikelets to 6 mm, long, 4-5-flowered; annual. .... 9. P. annua. d. Spikelets to 3 mm. long, 3-flowered; perennial. nes. 10. P. denticulata. c. Rachis and pedicels puberulent or scabridulous. e. Spikelets subquadrate, as broad as long. ...... 5. P. Hieronymi. e. Spikelets longer than broad (but note no. 11). f. Spikelets 2-flowered. sss 12. P. candamoana. f. Spikelets almost always more than 2-flowered. g. Lemmas not villous on any part. h. First glume about 2.5 mm. long. ............ 6. P. Lilloi. h. First glume about 4.5 mm. long. ........ 4. P. gymnantha. g. Lemmas with at least basal villosity. i. Rachilla puberulent. ........ s 8. P. asperiflora. i. Rachilla glabrous. j. Spikelets to 7 mm. long, often almost as broad as long. sese 11. P. horridula. j. Spikelets 4-5 mm. long, definitely longer than broad. k. Glumes equal or subequal. ........ 7. P. pratensis. k. Glumes unequal. ss 2. P. scaberula. 1. Poa Buchtienii Hack. in Fedde, Rep. Sp. Nov. 11: 29 (1912). P. Buchtienii var. subacuminata Hack. in Fedde, Rep. Sp. Nov. 11: 30 (1912). Perennial, to 8 dm. high, the leaf-blades flat, to 3 mm. wide. 1966] Flora of Bolivia — Foster 117 Rachis and pedicels scabrid. Staminate spikelets: about 5 mm. long, 3-5-flowered, the rachilla glabrous. First glume 2.5 mm. long, 1- nerved, the nerves scabrid, at least on the upper portion, the second glume about 3.5 mm. long, broader than the first, obscurely 3-nerved, the keel scabridulous. Lowest lemmas about 3.5 mm. long, obscurely nerved, the mid-nerve scabridulous on the upper part, the apex erose. Palea 3 mm. long, the 2 keels ciliolate. Pistillate spikelets: about 6-7 mm. long, 5-flowered, the rachilla glabrous. First glume about 25 mm. long, 1-nerved, the keel scabridulous, the second glume 3 mm. long, 3-nerved, the keel scabrid. Lowest lemmas about 4.5 mm. long, densely long-pilose at base and on the lower half of the keel, the apex erose to acute. Palea shorter than the lemma, the keels long-pilose below and ciliate above. LA PAZ: MURILLO: La Paz, 3700 m., Buchtien 2468 (GH) ; Palea, 3700 m., Hitchcock 22559 (GH), 22570 (GH). One culm on the Gray Herbarium sheet of Buchtien 2468 bears staminate spikelets with the basal villosity of the pistillate spikelets. 2. P. scaberula Hook. f. Fl. Antarct. 378 (1847). Annual, to 5 dm. high, the leaf-blades to 2 mm. wide and scabrid beneath, the rachis and pedicels densely scabrous. Spikelets to 4 mm. long, about 4- flowered, the rachilla glabrous. Glumes 2 and 2.5 mm. long, the keels scabrid. Lowest lemmas 2.5 mm. long, villous at base and halfway up the keel, the lateral nerves often somewhat villous basally. Palea much shorter than the lemma, the keels ciliate. LA PAZ: LARECAJA: vic. Sorata; Lancha de Cochipata, 2000-3700 m., Mandon 1336 (GH); MURILLO: La Paz, 3670 m., Buchtien (GH). 3. P. humillima Pilger in Engler, Bot. Jahrb. 37: 378 (1906). Very dwarf, caespitose perennial, the leaf-blades hardly 2 cm. long, involute, glabrous, the panicle not over 1 em. long, rachis and pedicels puberu- lent. Spikelets to 5 mm. long, 3.4-flowered, the rachilla glabrous. Glumes equal, 3.5 mm. long, subquadrate, apex coarsely and erosely toothed, obscurely 3-nerved. Lemmas about 4.5 mm. long, subquad- rate, 5-nerved, the apex erosely toothed. Palea shorter than lemma, bicarinate. LA PAZ: LARECAJA: vic. Sorata; near Anilaya, 4500 m., Mandon 1353 (GH). 4 P. gymnantha Pilger in Engler, Bot. Jahrb. 56, Beibl. 123: 28 (1920). P. perligulata Pilger in Notizbl. 11: 779 (1933). P. gym mantha var. aperta Pilger in Notizbl. 11: 780 (1933). Caespitose perennial to 20 cm. high, leaf-blades flat or involute, glabrous, 5-10 em. long, the panicles 3-7 cm. long, rachis and pedicels scabridulous- puberulent. Spikelets 5-7 mm. long, 2-3 (-4)-flowered. Glumes un- equal, the first to 4.5 mm. long, 1-nerved, scabridulous at apex, the second to 5 mm. long, 3-nerved, scabridulous at apex. Lemmas acute to subacute, 5.5 mm, long, 5-nerved, the keel scabrid-ciliate, the rest 118 Rhodora [Vol. 68 scabridulous. Palea as long as lemma, scabridulous, the 2 keels scabrid- ciliolate. LA PAZ: LARECAJA: vic. Sorata; Llachisani, 4200 m., Mandon Mandon 1351 (GH); MURILLO: vie. La Paz, 5000 m., Mandon 1352 ma COCHABAMBA: AYOPAYA: Tabacruz, 4200 m., Steinbach 9761 GH). 5. P. Hieronymi Hack. in Oesterr. Bot. Zeitschr, 52: 380 (1902). P. myriantha Hack. in Stuckert in Anal. Mus. Nac. Buenos Aires, 13: 517 (1906). Perennial, to 2 m. high, leaf-blades Scabrous, the flat blades 8-10 mm. wide, scabrous beneath, the very lax open panicle to 4 dm. long, the ultimate branches extremely filiform, the rachis and pedicels sparsely scabridulous. Spikelets, subquadrate, to 5 mm. long and broad, 5-6-flowered, the rachilla glabrous. Glumes equal, about 5 mm. long, the first l-nerved, the nerve scabrid, the margin scabrid- ciliolate, the second 3-nerved, the nerves scabrid. Lemmas acute, to 3.5 mm. long, the nerves Sscabrid, the margins ciliolate, a small tuft of tangled hairs at base of central nerve. Palea decidedly shorter than the lemma, the 2 keels ciliolate. COCHABAMBA: CHAPARE: Inca- chaca, 2700 m., Steinbach 8956 (GH), 2500-3000 m., Steinbach 9517 (GH). 6. P. Lilloi Hack, in Stuckert in Anal. Mus. Nac. Buenos Aires, 21: 153 (1911). Perennial, 10-20 (-30) em. high, leaves glabrous, the very narrow blades involute, panicle 2-4 cm. long, rachis and pedicels Scabrid, sometimes sparsely so. Spikelets about 5 mm. long, about 4- flowered, rachilla glabrous to slightly puberulent. First glume 2.5 mm. long, 1-nerved, the keel scabrid, the second glume to 3 mm., 3- nerved, scabridulous. Lemmas to 3.0 mm. long, 3-5-nerved, scabri- dulous, especially on the nerves. Palea nearly as long as the lemma, the 2 keels ciliolate, PoTosí: sUR CHICHAS: Atocha, Hitchcock 22931 (Hitchcock). The description given is based on an isotype, Lillo 5619 (GH). 7. P. pratensis L., Sp. Pl. 67 (1753). P. boliviensis Hack. in Fedde, Rep. Sp. Nov. 11: 25 (1912). Rhizomatous perennial, to 5 dm. high. Leaf-blades to 3 mm. wide, glabrous or scabridulous beneath. Panicles to 10 em. long, the rachis and pedicels scabridulous. Spikelets about 5 mm. long, 3-5-flowered. The rachilla glabrous or sparsely and minute- ly puberulent. Glumes subequal, about 3 mm. long, the keels scabri- dulous, the first l-nerved, the second 3-nerved. Lowest lemmas about 3 mm. long, 5-nerved, the keel basally long-villous, the upper half scabrous, marginal nerves basally villous. Palea nearly as long as the lemma, the 2 keels ciliolate. LA PAZ: MURILLO: San Jorge, 3550 m., Buchtien 236 (GH). 8. P. asperiflora Hack. in Fedde, Rep. Sp. Nov. 11: 28 (1912). P. Pflanzii Pilger in Engler, Bot. Jahrb. 49: 187 (1912). Caespitose perennial to 5 dm. high, the involute leaf-blades sometimes exceeding 1966] Flora of Bolivia — Foster 119 the inflorescence, scabrid. Panicles fairly open, generally 10 cm. or less in length, the rachis and pedicels scabridulous. Spikelets about 7 mm. long, 4-flowered, the rachilla puberulent. First glume about 3.5 mm. long, 1-nerved, the second glume about 5 mm. long, 3-nerved, the keel slightly scabridulous. Lowest lemmas about 5 mm. long, 5-nerved, the keel basally short-villous, with scattered pubescence on the lateral nerves. Palea about 4 mm. long, the keels ciliolate. LA Paz: LARECAJA: vic. Sorata; Lacatia, 3360-4200 m., Mandon 1341 (GH); MURILLO: Pilaya, 4000 m., Hitchcock 22587 (GH). 9. P. annua L. Sp. Pl. 68 (1753). Annual, the leaf-blades to 3 mm. wide, sometimes scabridulous beneath, the panicles open, to 6 cm. long, the rachis and pedicels glabrous. Spikelets to 6 mm. long, 4-5- flowered, the rachilla glabrous to sparsely puberulent. First glume about 2.5 mm. long, 1-nerved, faintly scabridulous near the apex, the second glume 3 mm. long, 3.nerved. Lowest lemmas 3 mm. long, 5-nerved, the keel with few to many hairs at base and about halfway up, but not forming a basal cobweb. Palea nearly as long as lemma, bicarinate. LA PAZ: LARECAJA: Vic. Sorata; Achacache, 2650-3950 m., Mandon 1335 (GH). COCHABAMBA: CERCADO: Cochabamba, 2600 m., Steinbach 8787 (GH); CHAPARE: Incachaca, 2250 m., Steinbach 9496 (GH). SANTA CRUZ: VALLEGRANDE: Cerro de San Mateo, Comarapa, 3400 m., Steinbach 8486 (GH). 10. P. denticulata Hack. in Fedde, Rep. Sp. Nov. 11: 27 (1912). Perennial, to 25 dm. high, leaf-blades 2-3 mm. wide, scabrid beneath. Panicles lax, to 8 cm. long, primary branches whorled, rachis and pedicels glabrous. Spikelets 4 mm. long, 3.flowered. Glumes 2 and 2.5 mm. long, the first 1-nerved, the second, 3-nerved. Lowest lemmas to 3 mm. long, ovate, obtuse, the apex erosely toothed, 5-nerved, the basal hairs reaching about one-third the length of the lemma. Palea as long as the lemma, bidentate, the keels scabridulous. LA PAZ: MURILLO: Unduavi, 3200 m., Buchtien 2584 (US; type-number). 11. P. horridula Pilger in Engler, Bot. Jahrb. 37: 506 (1906). P. androgyna Hack. in Fedde, Rep. Sp. Nov. 6: 159 (1908). P. dume- torum Hack. in Fedde, Rep. Sp. Nov. 11: 27 (1912). P. dumetorum var. unduavensis Hack. in Fedde, Rep. Sp. Nov. 11: 27 (1912). Perennial, to 1-1.5 m. high, leaf-blades to 8 mm. wide, scabridulous beneath. Panicle to 2 dm. long, the rachis and pedicels scabridulous. Spikelets to 7 mm. long, often about as broad as long, 4-5-flowered, the rachilla glabrous. First glume to 2.5 mm. long, 1-nerved, the nerve scabridulous, the second to 3 mm. long, 3-nerved, the nerves scabridulous. Lowest lemmas to 3.5 mm. long, 5-nerved, pilose on lower half of keel and along lower marginal nerves, the keel scabridulous on upper half. Palea as long as the lemma, the keels ciliate. LA PAZ: LARECAJA: vic. Sorata; Cochipata, 2700-3900 m., Mandon 1339 (GH) ; MURILLO: La Paz, Buchtien 846 (GH; type-number of P. androgyna), 3800 m., Buchtien 168 (GH) ; Unduavi, 3200 m., Buchtien 2583 (GH; 120 Rhodora [Vol. 68 type-number of P. dumetorum var. unduavensis) ; NOR YUNGAS: Bella Vista, Hitchcock 22754 (GH). 12. P. candamoana Pilger in Engler, Bot. Jahrb. 37: 381 (1906). Perennial, leaf-blades to 3 mm. wide, scabridulous beneath. Panicle to 15 em. long, but usually much Shorter, rachis and pedicels scabrid. Spikelets to 5 mm. long, 2-flowered, the rachilla sparsely villous. Glumes 3 and 4 mm. long, acute, the second much broader than the first. Lowest lemmas about 4 mm. long, villous on the lower half, the first floret sometimes not perfect but staminate. Palea shorter than the lemma, the keels ciliate. LA PAZ: MURILLO: La Paz, 3800 m., Buchtien 8536 (GH); NOR YUNGAS: Pongo, 3640 m., Hitchcock 22769 (GH). Errata: Volume 67 - p. 351 line 22 Daura should read Datura - p. 356 line 12 Gleichena should read Gleichenia / p. 868 Fig. 4 Caption should read Mean temperature for January in ^F, (after Hare 1952) - p. 372 line 27 sessiliflora should read sessilifolia p. 373 line 13 macrocarpa should read macrocarpon p. 409, 411 running head thyoites should read thyoides Volume 68, No. 773 including pages 1-120, was issued April 6, 1966. “wo OW REFERENCE LIBRARY ^. JUL 1 X 1966 LA TES — Dodova JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by ALBION REED HODGDON, Editor-in-Chief ALBERT FREDERICK HILL > STUART KIMBALL HARRIS RALPH CARLETON BEAN ROBERT CRICHTON FOSTER | Associate Editors ROLLA MILTON TRYON RADCLIFFE BARNES PIKE LORIN IVES NEVLING, JR. Vol 68 >`- April-June, 1966 No. 774 CONTENTS: New Data on North American Oak Ferns, Gymnocarpium Warren H. Wagner, Jm. ssssesrssessssesssesronsseereesesoessovereeeesereeseeeveese 121 Lactuca muralis in New England James P. Poole ................ 138 Cuchumatanea A New Genus of Compositae (Heliantheae) C. E. Seidenschnur and J. H. Beaman ............ eere 139 Seed Germination of Shortia galacifolia T. & G. Under Controlled Conditions Mary H. Rhoades ..................... 147 The Application of the Linnaean Names of Some New World Species of Euphorbia Subgenus Chamaesyce Derek Burch 1... eerte eee eee ee eese ta on ase tetra eee eP Pase eene eee eaae sepan ea en 155 (Continued on Inside Cover) The Nem England Botanical Club, Inc. Botanical Museum, Oxford St., Cambridge 38, Mass. iini BEREN 1 CONTENTS: — continued Identity of Witchweed in the Southeastern United States C. Earle Smith, Jv. ceccccscccsssssssssssscsssssescsssescsesesescscsescscseseseseseees 167 Bromus mollis and Allies in New England Frank C. Seymour |... eerte eher eren 168 Contribution to the Fungus Flora of Northeastern North America. IV Howard E. Bigelow and Margaret E. Barr |... 175 Notes on Rafinesque's Species of Lechea (Cistaceae) Robert L. Wilbur |... esee eren enne etat 192 Gentiana puberulenta sp. nov., A Known but Unnamed Species of the North American Prairies James S. Pringle ............ eee ettet ettet ntn tntntnta 209 New Chromosome Numbers in Zinnia and Sanvitalia Andrew M. Torres vicccccccssssssssssssssscsssssssssssessecsscecssssessccssseesssassce 215 Two New Naiads from Illinois and Distributional Records of the Naiadaceae Paul L. Fore and Robert H. Mohlen- brock csssiceranensncetscsantensseocivusiasisossscsabovssnnevosensesspeazsacaesmenmerserseecsvienes 216 Aquatic Vascular Plants New for Illinois Glen S. Winterringer crsccccsscscsssssessscscsssscsscsssssscsssscssasescscsssscesece 221 Studies in the Flora of Bolivia-IV (Continued) Robert C. Foster... eterne then 223 QTRbooora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 68 April-June, 1966 No. 774 NEW DATA ON NORTH AMERICAN OAK FERNS, GYMNOCARPIUM WARREN H. WAGNER, JR.' The oak ferns, Gymnocarpium Newman, include few taxa. Except for one or more separate and endemic elements in eastern Asia (cf. Pichi-Sermolli, 1965. p. 148), there are only two basic species over most of the circumboreal range. Nevertheless, there are controversial points regard- ing both the generic treatment and the species taxonomy. The common oak fern, Gymnocarpium dryopteris (L.) New- man has broadly triangular leaf blades with the axes so sparsely glandular that they are usually designated as “glabrous.” The “limestone” oak fern, G. robertianum (Hoffm.) Newman, is far rarer and more sporadic in most of its range. It has narrower leaf blades, and the blade axes (and the lamina) are densely glandular. Both of these ferns are variable, both in habitat and morphology. Dwarfed plants of G. robertianum, especially those growing on dryish rock cliffs, may tend to mimic the leaf outline of G. dry- opteris. Fully developed, however, the characteristics of "Research supported by National Science Foundation Grant GB-2025. I wish to acknowledge the aid of Miss Sandra Smith and Mrs. Katherine Lim Chen, and the suggestions of Doctors E. T. Wherry, Eric Hultén, and Conrad V. Morton. The following herbaria kindly lent materials: Cranbrook Institute of Science, Gray Herbarium, Missouri Botanical Garden, U. S. National Museum, and the Natur- historiska Riksmuseet, Stockholm. Dr. J. H. Soper made it possible for me to study the oak ferns in the herbarium at the University of Tcronto. 121 122 Rhodora [Vol. 68 OWesernvess op Horass, Tu Uswis ot Michasaw GF WASHINGTON Gymnocarpium dryopteris New, Kl > Deer conirereua woods: at Creek turnorf, 9.5 ig, Sneg * OnRt'l, Forest, Abundant in we June 25, 1957, M. i . Vagner 53057 å, By Eruickeberg sna PLATE 1328 Fig. 1. Gymnocarpium dryopteris var. disjunctum. Voucher speci- men, n — 40, Wagner 63087 and Kruckeberg (MICH). 1966] Oak Ferns — Wagner 123 Eoo eL £ * 4 atej 1 ‘ Con -— a hit Bg Coinimetey, He c d. WARREN H WRONER, fa te 28 iiem afa fl We, (PH! PLATE 1329 Fig. 2. Gymnocarpium heterosporum n. sp. Type specimen. Wagner 283 (MICH). 124 Rhodora [Vol. 68 these plants are distinctive and most authors have upheld them as separate species. Good keys and brief descriptions are found in familiar manuals and guides (e.g., Tutin et al., 1964; Wherry, 1961); and the taxonomic characters are fairly well established. At least two recent writers (Lawalrée, 1950; and Boivin, 1962) have questioned whether the two widespread taxa are specifically different. This question will be discussed further below. Also, there has been some disagreement about the generic placement of the oak ferns, but this matter will not be dealt with. Suffice it to say, the combi- nation of characters does seem very well-marked, and in the past few years there has developed a widespread tendency to recognize Gymnocarpium as a distinct genus. ne question is what its relationships are. The base number of x — 40 is quite different from such thelypteroid ferns as Phegopteris (x = 30), Cyclosorus (x = 36), and such species of Thelypteris as T. palustris (x = 35) and 7. "l 3 1 | Fig. 3. Distribution map. Stars = very small spored G. dryopteris (less than 23.9 X 1.18 microns). Dots = G. heterosporum. 1966] Oak Ferns — Wagner 125 noveboracensis (x = 27). The chromosome number of Gymnocarpium resembles that of typical aspidioid ferns (e.g., Dryopteris, Cystopteris, Polystichum, Tectaria, Cteni- tis, etc.) with x = 40, 41, or 42. Elizabeth Root (1961) studied intermediates between G. dryopteris and G. robertianum and pointed out, for the first time, that “the majority of the spores of all intermediate specimens were found to be abortive." Since her study we have learned that the “hybrid” Gymnocarpium is much more frequent and widespread than had been heretofore realized, as is illustrated in the map (Fig. 3, dots). We have also discovered that there are three ploidal levels present in North American Gymnocarpium — the diploid, triploid, and tetraploid. These facts will be examined below in connection with the following questions: Are G. dryopteris and G. robertianum distinct species? Should the western representative of G. dryopteris be recognized as a separate variety or subspecies? What is the nature of the plants which are morphologically intermediate between G. dry- opteris and G. robertianum? THE DISTINCTNESS OF G. DRYOPTERIS AND G. ROBERTIANUM Boivin (op. cit.) is the most recent worker who en- deavored to interpret taxon robertianum as merely a variety of G. dryopteris. He did not find it possible to uphold taxon robertianum as a species because “Tous les caracteres donnés chevauchent largement, sauf celui de la glandulosite qui est assez net si l'on excepte 6 ou 8 spécimens (sur en- viron 200 examiné). Cette glandulosité est si fine et ces deux espéces sont si semblables que, sur une cinquantaine de spécimens glanduleux examinés, un bon quart étaient mésidentifiés." I should point out that herbarium collections of ferns are often misidentified. The fact that certain species (e.g., in Gymnocarpium, Dryopteris, and Botrychium) are commonly misidentified should not be used as a basis for taxonomic judgments. The reasons for upholding G. robertianum as a distinct species, rather than a variety of G. dryopteris, fall into two 126 Rhodora [Vol. 68 classes: (1) The morphology of well-developed specimens is unquestionably distinct, including characters of the over- all blade outline, the details of segment form, and the glandularity. Small, ecologically depauperated or juvenile, specimens may of course cause difficulties if only gross characters are used. But the glandularity of G. robertianum is evidently always a dependable character. (Some of the problems of detecting the extent of glandularity in these plants could be alleviated by eliminating the unfortunate herbarium practice of pasting down fronds, thus coating the delicate glands with a thin “varnish” which makes them practically invisible.) The only really intermediate plants in respect to glandularity are those to be discussed below, but these plants are intermediate in other respects also between G. dryopteris and G. robertianum and they are probably of hybrid origin. (2) The second reason for upholding G. robertianum is simply the behavior of the natural populations. The two species grow together in numerous localities in the Great Lakes area, in exactly the same habitats and growth conditions, and their distinctness is immediately evident to the field worker. The fact that these two taxa can exist together sympatrically without merging, and the fact that there is no genomic difference (both have n — 80 in this region) in ploidal level, shows that using the varietal rank to express their relationships would be stretching that category too far (cf. Wagner, 1960). The differences between these ferns are best seen in the rich, shady swamps of northern Michigan where they develop to the maximum size in the luxuriant, moist con- ditions. Mixed populations are well known in several locali- ties near the University of Michigan Biological Station (Cheboygan, and nearby Presque Isle Cos., Michigan). GYMNOCARPIUM DRYOPTERIS VAR. DISJUNCTUM The western North American representatives of G. dry- opteris, especially in Idaho, Oregon, Washington, U.S.A. and British Columbia, Canada, tend to develop unusually large fronds. The area is roughly that shown by the stars 1966] Oak Ferns — Wagner 127 Fig. 4. Tracings of spores. Inner line — exospore; outer line — perispore. R. Gymnocarpium robertianum, Mich., Voss 4709 (MICH). X. G. heterosporum, Pennsylvania, Darling in 1956 (MICH). D. G. dryopteris, Mich., Voss 4708 (MICH). of especially small-spored plants in Figure 3. In the most extreme form these plants become thrice-pinnate and the larger segments are more toothed than in typical G. dry- opteris. Under the guidance of Dr. Arthur R. Kruckeberg of the University of Washington, I had an opportunity to study this form in the field. The best plants (Fig. 1) were seen in the deep coniferous woods at Denny Creek turnoff along U.S. 10, in the Snoqualmie National Forest, King Co., Washington. Here the "giant" G. dryopteris flourishes in damp, mossy forest with such other ferns as Dryopteris dilatata. and Athyrium filix-femina, the latter two species also becoming remarkably large. Chromosome materials were obtained and squashes of meiosis showed clearly n = 40 pairs (Figs. 6 and 7, top illustrations). This is the first discovery of the diploid condition in the genus Gymno- carpium, all previous counts for which indicated the tetra- 128 Rhodora [Vol. 68 x A ; x G. ROBERTIANUM x x x x x x x x x x x x x x x x x x x x x x x X E x x x x | & e x x x x * t t t t t ———— G. DRYOPTERIS V. DRYOPTERIS +k RK KAR RAR KK RARER KK EK RA TA AX OO X OO XXX TERK AK AK XOOD X X OX ROC X FARK KX RA KKK PRM Ke 3. DRYOPTERIS V. DISJUNCTUM KE X X X x x x x + D G. DRYOPTERIS -- ALASKA AND YUKON 20 25 3o Conversion: n X 1. 18 microns Fig. 5. Frequency histograms of spore lengths (exospore only). Each cross equals a 10-spore sample from a separate collection (Dots = spore samples of G. continentale). A. G. robertianum, all col- lections. B. G. dryopteris, all collections from Colorado east to Europe. C. G. dryopteris, all collections from British Columbia and Oregon east to Alberta and Montana. D. G. dryopteris, all collections from Alaska and Yukon. (Conversion: n X 1.18 microns). 1966] Oak Ferns — Wagner 129 ploid condition of n = 80. (According to spore size, as shown in Fig. 5, A, the chromosome number of the Siberian G. continentale may also turn out to be diploid.) For several reasons — the diploid chromosome number, the smaller spores (Fig. 5), the larger, more divided leaf (Fig. 1) and the distinctive western range — I choose (with some reluctance, however) to uphold the plants with this combination of characters as a separate variety. Morton (1941, p. 217) pointed out that the name disjunctum was based upon specimens from Sitka, Alaska, “and refers to certain large, lax forms that are essentially tripinnate, with the lower tertiary pinnules somewhat spaced out." At the time of his writing, Morton was of the opinion that “they do not seem to be worth nomenclatural recognition . . . " Boivin (op. cit.), on the contrary, upheld the large form as a distinct variety. Although I am inclined to agree now that var. disjunctum (Led.) Ching should probably merit recognition, I must point out that the situation becomes very complicated in Alaska and the Yukon, where both small- and large-spored types occur in the same general area (note the frequency histograms in Fig. 5, D). In this region also, we find numer- ous dwarfed individuals of both types, so that the gross appearance typical of *'var. disjunctum” becomes modi- fied and indistinguishable in many specimens from *'var. dryopteris," using the respective spore sizes as the criterion. Thus many specimens from the far northwest of North America cannot be readily identified. Dwarfing obscures the gross characters of the fronds; and the overlap of spore sizes between the diploid and tetraploid conditions is such that many specimens could not be identified without chromosome counts. Our spore surveys have uncovered still other compli- cations, the full meaning of which must await further study. It turned out that a number of collections of what I had interpreted to be ordinary "var. dryopteris" have abortive spores which in appearance match those of the presumed hybrid fern to be discussed below (like those in Fig. 4, X). 130 Rhodora [Vol. 68 PLATE 1330 Fig. 6. New chromosome conditions in Gymnocarpium. Top: G. dryopteris var. disjunctum diploid, n — 40 pairs, Wagner 63087 and Kruckeberg (interpretation top left in Fig. 7). Bottom: G. hetero- sporum triploid, 3 x = c. 120, Wherry s.n., from Blair Co., Penn- sylvania, in 1956 (bottom left in Fig. 7). 1966] Oak Ferns — Wagner 131 f T ^ y% wis 2 * M w Pr `a ^ x é Ps » x 4». e ? .* $ à % > 4, oc em % ane Ny uw V. fe a> » "m na E" ^ 4 è wet f $ b a ef ut ox D Iw ro ow ap o Ly Use Pixs 4 48. y ; JJ ww ^ ett Y 4 n me Ap ata + + 9F $3 s, | 4 a * °F D "A . rl » d. » * 5 19 S ¢ o a ` o » 4^2? % 9796 o2 * g A uem ek OS $m e^ $$ k ` 45 E E h ene > 090 of r A = Y | 3. Rr i 1.3 ^. € eu We o A PES M Y 9, EN ò s oe P 4 oo T v vw & $ , RAE M 4o. ake. Ere d 2 ¥ de rom $90 o, ql s Fig. 7. Camera lucida interpretations of chromosome squashes. Top: Diploids, same locality as in Fig. 6. Bottom: Triploids, same locality as bottom photograph in Fig. 6. Miss Virginia M. Morzenti kindly examined these collections and confirmed that the morphology of the abortion is very similar in the “sterile” G. dryopteris and the putative cross, G. heterosporum. It seems possible that at least some of these collections may be apomictic derivatives of G. dry- opteris var. disjunctum (the diploid) X var. dryopteris (the tetraploid), assuming here, as elsewhere in this dis- cussion, that the type of var. disjunctum corresponds to the diploid, large-leaved western plant. We failed to find a single example of G. robertianum in all those that we studied which showed a similar abortion of spores. Thus, if one argued that these "sterile" G. dryopteris plants might re- present mutations that are propagated locally by rhizomes, we wonder why similar mutations were not found in the related G. robertianum. Those collections we found of pre- sumably "sterile" G. dryopteris will not be enumerated in detail here, but their geographical distribution is extremely wide as shown by the following records: Alaska (2 col- lections), Alberta (1), British Columbia (1), Wisconsin (1), Michigan (5), New York (4), Vermont (2), Maine 132 Rhodora [Vol. 68 (1), New Hampshire (1), Quebec (1), Newfoundland (2), and Nova Scotia (1). Further research on these curious “sterile” oak ferns with the morphology of G. dryopteris is greatly to be desired, especially to determine their cytologi- cal features and the mode of their reproduction. THE APPARENT CROSS OF G. DRYOPTERIS AND G. ROBERTIANUM This is the plant discussed by Root (op. cit.). As stated above, the apparent cross is much more frequent and wide- spread than we had previously realized. Its occurrence, especially in the western Great Lakes area, indicates that it is more than a rare hybrid, and I have concluded therefore that it should be given a binomial designation as an apomic- tic species, as follows: Gymnocarpium heterosporum n. sp. W. H. Wagner. Planta frondis segmentique forma et glandularitate intermedia inter G. dryopteridem et G. robertianum; spori atro-brunnei, magnitudine et forma valde irregulares. Type: PENNSYLVANIA: BLAIR CO., Canaan Station, limestone slope, W. H. Wagner 283 (transferred from University of Pennsylvania to MICH — 2 sheets, one fertile, the holotype (Fig. 2), and one sterile). Representative Collections: U.S.A.: PENNSYLVANIA: BLAIR CO, 2 mi. n. w. of Hollidaysburg, among limestone rocks, T. Darling, Jr. in 1956 — same locality as the type (MICH). MICHIGAN: MARQUETTE CO., ca. 6 mi. n. w. of Ishpeming, near Ropes Gold Mine, n. w. % sect. 29, T 46 N, R 27 W, in woods on igneous rock cliffs, E. G. Voss 4707, W. H. Wagner and D. J. Hagenah 9415 (MICH). WISCONSIN: BAYFIELD co., Orienta Falls, G. H. Conklin and M. F. Somerville 1127 (WIS); BARRON CO., Barron Hills, just n. w. of Lehigh at the base of quart- zite talus, N. C. Fasset 15817 (wis), R. M. Tryon, Jr. 4154 (WIS, GH, MO). MINNESOTA: CARLTON CO., Carlton, Carlton-Thompson Gorge, St. Louis R., M. F. Somerville in 1928 (MICH); ST. LOUIS CO., obser- vation tower at Ash River, Wagner 9034.5a (MICH); LAKE CO., Gooseberry Falls State Park, 15 mi. n. e. of Two Harbors, in crevice of diabase cliff, damp and shady, R. M. and A. F. Tryon, A. C. Faber 4889 (MO); 45 mi. n.e. of Two Harbors, Manitou Falls near Lake Superior, under ledge of diabase, damp and shady, R. M. and A. F. Tryon, A. C. Faber 4895 (M0). CANADA: ONTARIO. ALGOMA DISTRICT, rocky cliffs at edge of Soulier Lake, vicinity of Michipicoten Harbor, R. C. Rosie, H. M. Harrison, E. O. Hughes 1092 (TRT, GH). ONTARIO, THUNDER BAY DISTRICT, 1 mi. n. of Little Pigeon Bay, Crooks Twp., talus slope, mostly shale, common, C. E. Garten 1933 (TRT); ca. 19 mi. n, of Nipigon, spruce-fir thickets on damp moss-covered talus 1966] Oak Ferns — Wagner 133 at foot of cliff, E. G. Voss 10335 (MICH, TRT) ; 15 mi. e. of Port Arthur, 2 mi. w. of Silver Islet, shady sandstone talus, R. M. and A. F. Tryon, A. C. Faber 4963 (MO). U.S.A.: ALASKA: WISEMAN, on Middle Fork of the Koyukuk River, ca. 67° 30' N, 150° W, Edith Scamman 913 (Gn). This fern, being intermediate between the two familiar species of Gymnocarpium, and having obviously peculiar spores might more appropriately be designated as merely a sporadic hybrid (e.g, as G. dryopteris X robertianum). There are several reasons why this was not done. (1) The plants develop large colonies locally, and behave as an apomictic species. (2) As shown on the range map (Fig. 3) plants conforming to this description run from Alaska southeast through the western Great Lakes area (the region, evidently, of its greatest abundance) down to central Penn- sylvania. (3) Cytogenetically (see below) at least one population, the type, is triploid, and not tetraploid as would be expected from our present knowledge of the parents as they occur in the eastern United States. The nearest dip- loids (judging from spore sizes and assuming correlation with the single diploid count) seem to be limited to the western part of the continent. (4) Finally, the intermediate is evidently to be found in at least some localities where one of the putative parents (G. robertianum) is unknown. The holotype, but not the cotypes, of Dryopteris linnaeana C. Chr. f. glandulosa R. M. Tryon, Jr. (Amer. Fern Jour. 29: 5, 1939) is evidently the same as the plant described above. The author wrote of his new “form” that the “rachis varies in glandularity, some specimens having only a few glands, while others have relatively many and rarely the blade is slightly glandular.” Judging from the specimens cited, the author included two kinds of populations — the form of G. dryopteris in which the glands are more frequent and conspicuous than usual, and the plant described above. As to the glandular form of G. dryopteris itself, I am not sure that it is worthy of recognition. In her survey (1961, p. 18) Mrs. Root concluded that “It is generally stated that ‘Additional collection localities not cited here but plotted in Fig. 3 were supplied by Dr. Erie Hultén. 134 Rhodora [Vol. 68 the fronds and rachises of G. dryopteris range from glabrous to rarely slightly glandular. However, upon close exami- nation of specimens from various localities it was found that every individual of this species examined has glands.” In synonymizing “forma glandulosa” with G. robertianum (as “G. dryopteris var. pumilum (DC) Boivin”), Boivin (1962) was in error. Neither the type nor the cotypes of Tryon’s form conform to the characters of G. robertianum. Specimens of G. heterosporum in herbaria have previously been determined as one or the other of the putative parents, but most commonly as “G. robertianum" or “G. dryopteris f. glandulosa.” The recognition of G. heterosporum is usual- ly fairly easy in herbarium material, so long as the speci- mens have ripe spores. These tend to be discharged and to lie around the axes of the frond, sticking to the sheet. If the spores appear blackish and strongly irregular in size and shape, this is an excellent diagnostic feature, in combination with the presence of the intermediate con- dition of glandularity. The spores can be seen even with a dissecting microscope or a high power hand lens; and once their appearance is learned it becomes easy to pick out other collections with the same spore conditions, so a compound microscope is unnecessary. The spores of typical specimens of G. robertianum and G. dryopteris seem always to have a “glassy,” tan or brownish color, and are strikingly regular. Both the soriferous specimens, as well as those which lack sori, of the intermediate oak fern will show the intermediate glandularity of the axes, especially on the upper parts of the rachis (if they have not been imbedded in herbarium paste). The glands of G. heterosporum are fewer and more widely spaced than they are in the densely glandular G. robertianum. In the field it is helpful to hold up the living fronds in the sun's rays and examine them with a good hand lens. The glands will glisten and thus be conspicuous, so that even small, sterile fronds of all three taxa, if growing together, may be separated. Wherry (1942) described the habitat of the type popu- lation as follows (italics mine): “... the base of a steep 1966] Oak Ferns — Wagner 135 northwest-facing wooded slope, above a brook, the under- lying rock being limestone which breaks into small angular blocks, from the crevices of which moisture oozes out, and by evaporation keeps the rocks cool." The Michigan station near Ishpeming, Marquette Co., is on steep shaded rock cliffs. In the latter area, the rocks contain serpentine. Garten took his specimens on a talus slope comprising most- ly shale. The Lake Co., Minnesota, material was recorded from crevices of a “diabase cliff," while that from Barron Co., Wisconsin, came from “quartzite talus." The speci- mens from east of Port Arthur, Ontario, were found on *shady sandstone talus." Thus, ecologically, the substratum itself does not seem to matter much. The constant features of the habitats of G. heterosporum seem to be (a) shadiness, (b) moisture and humidity, (c) presence of rock substratum of some kind, the plants growing either directly from crevices of cliffs or upon the fallen talus. One or both of the presumed parental species may be present at the same locality. CHROMOSOMES AND SPORES The chromosomes and spores seem to have significance in this group of plants because of their variability and their correlation with each other and with other systematic data such as distribution and morphology of the fronds. There are three chromosome conditions. The tetraploid with n = 80 or near that number is characteristic of G. robertianum as found by Manton and myself (Chiarugi, 1960; Fabbri, 1963). Judging from spore sizes (see Fig. 5) all populations of G. robertianum are probably tetraploid and the curve is normal. In Siberia, however, the rather similar G. con- tinentale (Petrov) Pojark. has small spores (indicated by dots in the graph) and may bea diploid. For G. dryopteris we first plotted all spore size averages and produced a broad, apparently undifferentiated curve. However, when we selected specimens from the area of presumably typical var. disjunctum (where we know there are diploids) an entirely different curve separated out. This area extends from British Columbia and Oregon east to 136 Rhodora [Vol. 68 Alberta and Montana, and the spore sizes average between 23-25 X 1.18 microns’ (Fig. 5, C) in maximum exospore diameter, in contrast to between 27-29 X 1.18 microns for var. dryopteris east of Colorado and including Europe (Fig. 5, B). Thus in G. dryopteris there is a correlation between spore size and polyploidy, as has been found in a number of other fern groups. The “sterile-spored” G. dryopteris commented upon above may represent triploid populations, derived by interploidal hybridization and spreading by some means of spore dis- seminule not yet fully understood, The question is where the diploid and tetraploid may meet geographically. The graph of spore samples from Alaska (Fig. 5, D) shows values all the way from 21 to 29 micrometer units, strongly indicating that in this area, G. dryopteris is a mixture of ploidal levels. It was extremely interesting to find that the type popu- lation of G. heterosporum (lower illustrations in Figs. 6 and 7) is triploid. Even though the pairing is very irregular, by totaling the estimated bivalents and univalents re- spectively of four sporocytes the following numbers re- sulted: (29) + 63, (33) + 50, (33) + 55, and (39) + 42. Their average was (33.5) + 52.5, or 119.5 chromo- somes in all. Assuming that G. heterosporum is truly of hybrid origin and that both G. dryopteris and G. roberti- anum in the eastern U. S. are tetraploid, then it is possible that the type population originated as 2x G. dryopteris var. disjunctum X 4x G. robertianum. Further cytological sur- veys of G. heterosporum in other areas are highly desirable. It is possible that 4x forms of it exist. I must alert collectors to the need of obtaining fertile materials of these plants of which the sori contain ripe spores. Out of 500 specimens that we examined, 150 proved to be sterile or too young for spore observations. Also, I should point out that it is not correct that the spores of ‘Conversion factor: one micrometer unit — 1.18 microns. Fig. 5 is plotted in micrometer units only, from spores taken from herbarium specimens and mounted directly in diaphane. 1966] Oak Ferns — Wagner 197 Gymnocarpium lack perispores. Ching’s description of “spores bilateral, warty, and without perispore" (quoted in Morton, 1941) surely does not conform to our observations. As shown in the spore tracings in Figure 4, there are two clearly defined layers — the outer, perisporial covering which is irregular, and the inner, exosporial boundary which is smooth. This is not different from other members of this group of ferns associated with Thelypteris and Dry- opteris. The perispores of Gymnocarpium show up especial- ly well in the abortive spores of “sterile G. dryopteris" and of G. heterosporum, where they actually are more or less exaggerated in development. This excessive perisporia] thickness is probably the factor which produces the darker, nearly black, appearance of the spores as seen en masse. UNIVERSITY OF MICHIGAN, ANN ARBOR REFERENCES CITED Borvin, BERNARD 1962. Etudes ptéridologiques. Il. Gymnocarpium Newman. Bull de la Soc. bot. France 109 (nos. 5-6): 127-128. CHIARUGI, ALBERTO 1960. Tavole cromosomiche delle Pteridophyta. Caryologia 13 (no. 1): 27-150. FABBRI, FERNANDO 1963. Primo supplemento alle Tavole cromoso- miche delle Pteridophyta di Alberto Chiarugi. Carologia 16 (no. 2): 287-335, LAWALRÉE, ANDRE 1950. Ptéridophytes. In Flore Générale de Bel- gique. Jardin Bot. de l'Etat, Brussels. Morton, C. V. 1941. On the name of the oak fern. Rhodora 43: 216-219. PICHI-SERMOLLI, RODOLFO E. G. 1965. Index filicum, suppl. IV. Intern. Bur. Pl Tax. and Nomencl., Utrecht. Root, ELIZABETH EICHSTEDT. 1961. Hybrids in North American Gymnocarpiums. Amer. Fern Jour. 51 (no. 1): 15-22. Tryon, ROLLA M., JR. 1939. Notes on the ferns of Wisconsin. Amer. Fern Jour. 29 (no. 1): 1-9. TUTIN, T. G. et al. (eds.) 1964. Flora europaea. Vol. I. Lycopodi- aceae to Platanaceae. Cambridge Univ. Press. WAGNER, W. H. JR. 1960. Evergreen grapeferns and the meanings of infraspecific categories as used in American pteridophytes. Amer. Fern Jour. 50 (no. 1): 32-45. 138 Rhodora [Vol. 68 WHERRY, EDGAR T. 1942. The ferns and lycosphens of Pennsylvania. Bartonia, no. 21: 11-63. 1961. The Fern Guide. Northeastern and Midland United States and Adjacent Canada. Doubleday and Co., Garden City, N.Y. LACTUCA MURALIS IN NEW ENGLAND Recently a number of duplicates from a collection of New Hampshire plants by Andrew P. N elson, Assistant Professor of Biology at Dartmouth College, were sent to Professor Albion R. Hodgdon at the University of New Hampshire. Among these was a specimen of Lactuca muralis (L.) Gaertn. (A. P. Nelson 1059) that was collected by the side of a gravel road between Cornish Mills and Plainfield in the town of Cornish, Sullivan County, N. H. on August 6, 1964. Professor Hodgdon states that this is the first report for this species in New England. Gray's Manual, 8th edition, describes the species as growing on roadsides and in waste places, local, w. Que., and e. N. Y. to Mich. (Adv. from Eu.). In Gleason and Cronquist Manual of Vascular Plants (1963) it is listed as a native of n, Eu. now known from N. Y. and Que. Voucher specimens have been deposited in the her- barium of the New England Botanical Club, Cambridge, Mass. and in the Pringle Herbarium at the University of Vermont in addition to the specimens in the University of New Hampshire Herbarium and the Jesup Herbarium. JAMES P. POOLE, CURATOR. JESUP HERBARIUM, DARTMOUTH COLLEGE CUCHUMATANEA A NEW GENUS OF THE COMPOSITAE (HELIANTHEAE)' C. E. SEIDENSCHNUR AND J. H. BEAMAN Cuchumatanea Seidenschnur & Beaman, gen. nov. Herba annua, minuta, depressa. Folia opposita, spathulata, obscure trinervia, basi leviter connata. Rami oppositi ad nodos. Capitula homogama, sessilia vel subsessilia, terminalia, solitaria, cum pare foliorum clavato-spathulatorum cireumclusa. Phyllaria 2, late oblongo- obovata, basi subite attenuata, membranacea valde concava, laxe flosculos involventia. Receptaculum conicum, paleaceum, paleis mem- branaceis flosculos non amplectentibus, exterioribus oblongo-ovatis, interioribus gradatim angustioribus usque ad centrales lineares. Flosculi 5-10 in capitulo, hermaphroditi actinomorphi. Corolla anguste tubuloso-campanulata, 3-4 lobata, infra flava, sursum purpurea. Sta- mina 3-4, antheris connatis basi sagittatis, apice cum appendice late ovata vel suborbiculata ornata. Stylus infra ramos contractus, ramis oblongo-ellipticis, acutis, lato margine dorsoque dense papilloso. Achaenia ellipsoideo-oblonga deorsum paulo attenuata levissime com- pressa, striolata, brunneo-nigra. Pappus nullus. Minute, depressed annual. Leaves opposite, spatulate, inconspicu- ously 3-veined, the bases slightly connate. Each node with two axillary branches. Heads homogamous, sessile or subsessile, solitary, terminal, surrounded by a pair of spatulate-clavate leaves. Phyllaries 2, broadly oblong-obovate, subpetiolate, membranaceous, strongly concave, loosely enclosing the florets, the outer oblong-ovate, the inner becoming pro- gressively narrower, the innermost linear. Florets 5-10, hermaphrodi- tic, actinomorphic. Corolla tubular, narrowly campanulate, 3-4 lobed, yellowish below, purplish above. Stamens 3-4, anthers connate, sagit- tate, with broadly ovate to suborbicular apical appendages. Style constricted just below the branches, the branches oblong-elliptical, acute, papillose to the base. Achenes ellipsoid-oblong, weakly com- pressed, brownish-black, striate. Pappus absent. Cuchumatanea steyermarkii Seidenschnur & Beaman, sp. nov. Herba annua, minuta, depressa usque ad 1 cm. alta, radice palari, saepe caespitosa. Caulis plus minusve sparse hirtellus vel glabratus, "The genus was discovered in the course of field work supported by National Science Foundation Grant G-9045. The research has been conducted in connection with NSF Grant GE-4051 for Undergraduate Research Participation. We are indebted to Dr. José Cuatrecasas for helpful suggestions and for preparing the Latin diagnoses and to Dr. D. C. D. De Jong for making the cytological preparations. 139 [Legend for figures 1-11] Fig. 1-4, 6-11. Cuchumatanea steyermarkii, based on Beaman 3962. Fig. 5. Jaegeria hirta, based on Beaman 4532. Fig. 1-3. Pales (X 15). Fig. 4-5. Stamens ( 40). Fig. 6. Head and subtending leaves (X 7, L, leaf; Ph, phyllary). Fig. 7. Phyllary (X 15). Fig. 8. Floret with achene (X 20). Fig. 9. Style branches (X 75). Fig. 10. Diagram of a terminal portion of the branching system (B, bud; H, head; L, leaf). Fig. 11. Diagram of a head with subtending leaves (B, bud; F, floret; L, leaf; Pa, pale; Ph, phyllary). 1966] Cuchumatanea — Seidenschnur & Beaman 141 viridis vel purpurascens composite dichasiale ramosus. Folia opposita decussata, basi paulo connata, saepe in pare inaequilonga, 3-7 mm longa 0.8-3 mm lata, spathulata, integra, margine parce revoluta sparseque ciliata, 3-nervata sed nervis plerumque inconspicuis, supra glabra, subtus praecipue in nervo medio sparse pubescens. Capitula homogama, sessilia vel subsessilia, terminalia, solitaria, campanulata, circa 2 mm alta 1 mm lata, cum duobus foliis brevibus clavato- spathulatis amplectentibus, 2-3 mm longis 1.2 mm latis, adpresse subtendentia partimque tecta. Phyllaria 2 circa 2.5 mm longa 1.5 mm lata plerumque inaequilonga, oblongo-ovata basim angustata, mem- branacea sed apice recurva herbacea leviter lacerata, 8-venulosa, dorso ad venulam mediam parcis pilis, valde concava, laxe flosculos in- volventia. Receptaculum 0.2-0.5 mm altum, conicum, paleaceum. Paleae membranaceae persistentes flosculos non involventes, 1.5-2 mm longae, venula media conspicua aliquando 2-3 venulis lateralibus fere inconspicuis, glabrae, sursum margine minute laceratae deorsum integrae, exteriores oblongo-ovatae, ceterae gradatim angustiores usque ad centrum lineares. Flores omnes hermaphroditi, 5-10, plerum- que 7 in capitulo. Corolla tubulosa 1 mm longa, limbo subcampanulato, . 3-4 lobato, lobis ovatis acutis purpureis glabrisque, tubulo 0.5 mm longo flavo, sparse pilosulo. Stamina 3-4 non semper pari quam corollae lobi, antheris basi sagittatis, apice appendice ovata vel subor- biculata, thecis 0.4 mm longis parietibus crassis. Stylus infra lobos constrictus, ramis oblongo-ellipticis, acutis, marginibus dorsoque papil- loso, uno saepe quam altero longiore. Achaenia 1.1 mm longa, ellip- soideo-oblonga, basim versus attenuata, levissime compressa, basi obtusa, brunneo-nigra, striata, glabra. Pappus nullus. Chromosoma m = 8. Minute, depressed, taprooted annual herb, 1 em or less high, oc- curring in patches and tending to form open mats. Stems glabrate to sparsely hirtellous, greenish or purplish, with two axillary branches at each node, one branch system arising from each leaf axil. Leaves opposite, decussate, 3-7 mm long, 0.8-3 mm wide, one member of each pair usually larger than the other, spatulate, entire, the bases slightly connate, margins sparsely ciliate, somewhat revolute, with 3 inconspicuous veins, glabrous adaxially, sparsely pubescent mostly on the midvein below. Heads homogamous, sessile or subsessile, solitary, terminal, campanulate, ca. 2 mm high and 1 mm wide, closely subtended and partly covered by a pair of small spatulate-clavate leaves 2-3 mm long and 1.2 mm wide. Phyllaries 2, ca. 2.5 mm long, 1.5 mm wide, one usually larger than the other, oblong-obovate, narrowing basally, mostly membranaceous but with herbaceous reflexed tip, lacerate near the apex, 3-veined, with a few hairs abaxially on the median vein, strongly concave, loosely enclosing the florets. Receptacle 0.2-0.5 mm. high, conical, paleaceous. Pales membranaceous, not sur- rounding the florets, 1.5-2 mm long, 1-veined, sometimes 2-3 incon- 142 Rhodora [Vol. 68 Spicuous lateral veins present, glabrous, minutely lacerate above, entire below, the outer oblong-ovate, the inner becoming progressively narrower, linear at the center of the head, persistent. Ray florets absent, Disk florets 5-10, commonly 7, hermaphroditie, Corollas tubular, narrowly campanulate, ca. 1 mm long, with 3-4 ovate acute lobes, tube 0.5 mm long, yellow, sparsely pilose, limb and lobes purplish, glabrous. Stamens 3-4, usually of the same number as the corolla lobes, anthers connate, sagittate, with broadly ovate to suborbicular apical appendages, thecae 0.4 mm long, with thick walls. Style constricted just below the branches, style branches oblong-elliptical, acute, papillose to the base, one branch often longer than the other. Achenes 1.1 mm long, ellipsoid-oblong, weakly compressed, with an obtuse base, brownish-black, striate, glabrous. Pappus absent. Chromosome number n = 8, based on Beaman 3962. Fig. 1-11. GUATEMALA. HUEHUETENANGO: Sierra de los Cuchumatanes, ca. 3 kms south of road between Llano de San Miguel and Todos Santos, from a point 2.5 miles west of Llano de San Miguel, near the highest point in the Cuchumatanes, collected from about the 3,680 to 3,740 meter level, in open Pinus rudis forest, occurring in patches in black soil near limestone outcrops, frequent in a local area, 2 August 1960, J. H. Beaman 3962 (Msc 171943, holotype; F, GH, K, TEX, UC, US isotypes) ; vicinity of Chémal, summit of Sierra de los Cuchumatanes, 3,700-3,750 m, Steyermark 50260 (F). The existence of this remarkable tiny plant was first pointed out by Dr. Julian A. Steyermark to Beaman in con- nection with the planning of a collecting expedition to Guatemala in 1959. Dr. Steyermark found it on August 8, 1942, in the vicinity of Chémal, Sierra de los Cuchumatanes, Department of Huehuetenango, Guatemala. In his field book he noted that it was annual, with the corolla yellow, the leaves green and shining, fleshy, fairly membranaceous, growing in bunches and patches. His collection was subse- quently examined by Dr. S. F. Blake who suggested to Steyermark that it probably represented a new genus near Jaegeria or Schkuhria. Blake pencilled on the sheet: “Jae- geria? Material too scanty & depauperate." In early August, 1959, and late July, 1960, Beaman made unsuccess- ful searches for the species in the vicinity of Chémal. But on August 2, 1960, it was discovered a few kilometers south- west of Chémal in one of the highest and most remote regions of the Cuchumatanes. In publishing the species it is a pleasure to associate the name of Dr. Steyermark with 1966] Cuchumatanea — Seidenschnur & Beaman 143 that of the Sierra de los Cuchumatanes. He has been the leading contributor to the understanding of the exceptionally diverse and unusual flora of this mountain range. Cuchwumatanea steyermarkii is an extremely reduced species, especially characterized by reduction of the plant body and of the florets. There is also slight asymmetry of the pairs of axillary branches, leaf and phyllary pairs, and style branches, with one member of each pair generally somewhat smaller than the other. In spite of reduction in size, however, the species has retained a complex branching system (see fig. 10). Two branches generally arise from each of the 3-4 nodes (except the first which is often un- branched). Additional axillary branches or buds are also present at the nodes of the primary branches. A solitary head terminates each branch. The heads are enclosed and somewhat concealed by sub- tending pairs of leaves. The two phyllaries follow the decus- sate pattern of the leaves (see fig. 6 and 11) and resemble immature leaves. In immature heads the subtending leaves are differentiated from the phyllaries mainly by the ciliate margins of the former and by their position. The two phyllaries partially surround the head and each subtends a pale and a marginal floret which is more advanced in de- velopment than other florets of the head. The number of corolla lobes varies from 3 to 4, which is also true for the stamens. While the number of corolla lobes and stamens is often the same, there are sometimes florets with 4-lobed corollas and 3 stamens. Florets with 3.]obed corollas and 4 stamens have not been found. In one floret with two normal stamens a third stamen was found with the filament of twice the normal width and two fused anthers. Such a condition may support the contention that the 3- and 4-merous florets are the result of a reduction process which is still in progress. Our findings confirm Blake's suggestion that the genus is related to Jaegeria. Among species of this genus it ap- pears to be closest to J. hirta (Lag.) Less. The habit of Cuchumatanea is very similar to that of a collection of V. [Vol. 68 Rhodora 144 SUME snonproop c-e snoiqu[d ‘Buorqo -prosdio ‘poj3ue -G ‘ƏY ‘ystumorg sesepuodde ojeAo Á] -peoiq YIM *'o39e3318€s Aysys sioqjue ‘e sdi} poyjnj *ie[n3uve -L3 yy” 192urT] po3orijsuoo JON gU ÁA[MO.L1€U ‘pajaay JON S9LIoS p jnoq* ur payeo -Hqui f'peoiq Alaa ayeIpey quesqy snoiqe[s ‘suo[qo-plosdyja — *oje -H3S ‘yorlq-ystumorg sogepuoedde aqyvao YUM ‘UIII 'ojejjrdes Apu3is sioqgjue ‘e əseq oui 0} oso[ -ded ‘aynow ‘avout U | pojorijsuoo 40N OF "8D pə JON iouur Jo S9LI9$ g pue i2jno c ojerpeqr quesqy snoiqes *Buo[qo-prosdi[[o *oje -H8 ‘youlq-ystuamorg sodup -uedde .e[norq.i0 q1r« '93e331d€s SloyjuR ‘e aseq a} 03 oeso[[rd -ed 'ojnow *oje[ooouv' po3orrjsuoo JON p9[99xM g ƏFeIpey juosqy snoiqeys "Suo[qo-prosdi[|p ‘ale “13S *xoe[q-usiua oq Sooup -uedde .1epnorq.toqns 0} 97840 A[peoiq YB 'ojej1pLdvs SIIYJUL *p-g aseq əy} 0} oso[[rded '[enboun uo3jo *o3noe 'eordi[[o-Suo[qo) sououe.tq əy} MOTE pojorijsuo;) p-e OT-S pooox 40N Z proostq snddeq sausyoy Suaure1g Saqout.q ƏS 9[ As Soqo[ e[[0.100 ysIp Jo ioquinN S]2.10g XSIp jo ioquinN sored serte[Auq SpeaH (0fzf ununag ‘naurbpjunjd 'g) D0112 Ş (6698 unwunoq ‘suadat 'g) D1322082D218 (GEST unuinag ‘pga p) Dia han Sf (96& umwv»og ‘yumu '7)) DauDJDUNYIND ""w1ouo3 poje[o1 YPM p2wpjpumion;) jo uosrieduro? T AQEL 1966] Cuchumatanea — Seidenschnur & Beaman 145 hirta (Beaman 4532) from La Cima, Federal District, Mexico. This specimen also has sagittate anthers with orbieular appendages which correspond closely to those of Cuchumatanea, although the latter has shorter thecae with thicker-walled cells along the margins. In another specimen of J. hirta (Bourgeau 1553 from the Valley of Cordoba, Mexico) the anthers seem to differ from those of Cuchuma- tanea mainly in not having thick-walled marginal cells. The genera are also similar in achenial structure, both having ellipsoid-oblong, slightly compressed, striate, glabrous achenes without pappus” Although the features noted above indicate a close re- lationship of Cuchumatanea with Jaegeria, they differ by characters of a magnitude equal to those by which other Verbesinoid and Galinsoginoid Compositae are distin- guished. The most important characters which separate the genera are as follows: (1) The heads of Cuchumatanea are discoid. Only J. gracilis Hook. f. from the Galapagos Islands is without ray florets (the type of J. discoidea Klatt, as noted by Robinson (1900), is radiate). (2) Cuchumatanea has only two phyllaries which do not surround the marginal florets but partially enclose the head. Jaegeria has 5 or more phyllaries each of which encloses an ovary or achene of a ray floret with slightly overlapping marginal membranes. (3) The tubular corollas of Cuchumatanea have 3-4 lobes rather than 5 as in Jaegeria, and the number of stamens in the former is also reduced. (4) The pales are unkeeled in Cu- chumatanea and keeled in Jaegeria. (5) The chromosome number in Cuchumatanea is n = 8, while counts reported for four species of Jaegeria have all been on the base of x = 9 (Turner et al., 1962; Beaman and Turner, 1962; Turner 2After submitting the manuscript for publication we sent a copy of it and some of the material of Cuchumatanea to Dr. A. M. Torres who is currently monographing the genus Jaegeria. He notes that the specimen of Jaegeria hirta (Beaman 4532) which we have used for comparison is at the extreme end of variation (a reduced form) in this species and that more typical specimens appear less similar to Cuchumatanea. He considers J. axillaris Blake, known only from the type collection from Colombia, to be closer to Cuchumatanea. 146 Rhodora [Vol. 68 and King, 1964). A further comparison of similarities and differences of Cuchumatanea and Jaegeria, in addition to the characters of two other related genera, is given in Table 1. We have not found any features to support Blake’s tenta- tive suggestion that Cuchumatanea might be related to Schkuhria (Helenieae). Because of its alliance with Jaegeria and other genera placed by Hoffmann (1894) in the subtribe Verbesininae, we tentatively assign Cuchumatanea to this subtribe next to Jaegeria. A more satisfactory classification may ultimately bring certain genera of the Galinsoginae into closer as- sociation with this group of the Verbesininae. DEPARTMENT OF BOTANY AND PLANT PATHOLOGY, MICHIGAN STATE UNIVERSITY, EAST LANSING, LITERATURE CITED BEAMAN, J. H., and B. L. TURNER 1962. Chromosome numbers in Mexican and Guatemalan Compositae. Rhodora 64: 271-276. HOFFMANN, O. 1894. Compositae in Engler & Prantl, Natürl. Pflan- zenfam. 4°: 87-391. ROBINSON, B. L. 1900. Synopses of the genera Jaegeria and Russelia. Proc. Amer. Acad. 35: 315-321. TURNER, B. L., A. M. POWELL, and R. M. KING. 1962. Chromosome numbers in the Compositae. VI. Additional Mexican and Guate- malan species, Rhodora 64: 251-271. TURNER, B. L., and R. M. KING. 1964. Chromosome numbers in the Compositae. VIII. Mexican and Central American species. South- west. Nat. 9: 27-39. SEED GERMINATION OF SHORTIA GALACIFOLIA T. & G. UNDER CONTROLLED CONDITIONS' Mary H. RHOADES” Until 1936 it was belived that Shortia galacifolia could not reproduce by seed because the scapes decayed and fell over before the seeds were mature (Kelsey, 1902). Observa- tions by Ross (1936) disclosed that Shortia seeds do reach maturity and germinate, but in an unusual manner. Germi- nation occurs in the capsule, the seeds being held against the placenta by the firm, partially opened capsule walls. In this position the seeds do not readily fall out when ripe, and are kept fairly moist because the capsule and cup-like calyx retain water. Seedlings reach the ground when the scape decays and falls. There they have little chance of survival because of predation by insects and because of the dense mass which Shortia plants form by vegetative reproduction. Attempts by Ross to germinate seeds on soil in the labora- tory were unsuccessful, and he concluded that there was little possibility for the spread of Shortia plants by seed. Ross's work suggests that seedlings of Shortia would be rare in nature, and that most of them would be near mature Shortia plants. On the contrary, several people have re- ported finding abundant seedlings in the field, both under- neath (Crandall, 1956) and several feet from patches of mature plants (Crandall, 1956; Vivian, 1965). However, it was not determined whether these seedlings were from seeds which germinated in the capsule or on the ground. Since it is more likely that seeds are transported rather than seedlings, it appeared worthwhile to investigate the possibility of seed germination outside the capsule. Capsules of Shortia galacifolia were collected on May 15, 1964, at an elevation of 1440 feet on the northwest side of "Part of a thesis submitted in partial fulfillment of the requirements for the M.A. degree at Duke University, Durham, North Carolina “Present address: Department of Biology, Virginia Polytechnic Institute, Blacksburg, Virginia 147 148 Rhodora [Vol. 68 Crossroads Mountain in Transylvania County, N.C. The capsules were open and most of them contained many seeds, but no seedlings were found in them. Immediately when picked they were wrapped in paper towels and put in plastic bags to prevent drying. A previous experiment showed that freezing the seeds did not impair germination, so the capsules were kept frozen until the seeds could be removed and counted. When the seeds were removed from the capsules, those with any visible defect, such as mold, discoloration, or extreme flatness, were discarded. The sound seeds were then put into envelopes of 50 seeds each and refrozen until all seeds were sorted. To prevent the seeds from drying out, only about six capsules were taken from the refrigerator at a time for removal of the seeds. Eight environmental factors were tested for their effects on germination. Temperature, moisture, and light were combined into one experiment, so there were six experiments in all. For all but the first experiment, germination was tested in an air-conditioned room with a temperature range of 20° to 24°C and a light intensity of 30 ft-c. Germination was tested in all treatments, except those involving soil, by sprinkling the seeds on moist Whatman No. 1 filter paper in a clean 9 em petri dish. The paper was moistened once a day with distilled water from a medicine dropper. Each treatment included six replications of 50 seeds each. The seeds were assigned to a particular treatment by numbering the envelopes of seeds and then using a table of random numbers to select the envelopes for a prearranged order of treatments. The number of seeds germinated was recorded every 4 days up to 32 days, and then once again after 50 days. The first appearance of the radical was the criterion for germination. To determine the effect of the interaction of temperature, moisture and light, germination was tested at 10°, 20°, and 30°C, in the light and the dark, and at three moisture levels. The moisture levels were “abundant,” in which the filter paper was saturated with water; “sufficient,” in which it 1966] Shortia — Rhoades 149 was thoroughly moistened but not saturated; and ‘‘de- ficient,” in which it was never thoroughly moistened. Both incandescent and fluorescent lamps provided illumination. The combinations of three temperatures, two light con- ditions, and three moisture levels made 18 treatments. Two other treatments tested the effect of a changing temperature regime on germination. In one treatment seeds were placed alternately between 10°C and 20°C every 48 hours. In the other treatment seeds were moved pro- gressively from 10°C, to 20°C, to 30°C. After 5 days at each of these temperatures, the seeds were put at 22°C for the remaining time. For both of the changing temperature treatments the seeds were kept in the light and given “suf- ficient” moisture. Since Shortia seeds germinate within the capsule, it was decided to test the effect of capsule parts on germination. For example, the calyx from about 10 capsules was cut up and the pieces were sprinkled on the filter paper along with the seeds.. The placenta, calyx, and ovary wall were tested for their effects in this way. To obviate the effect of moisture retention by capsule parts, a control treatment was set up in which pieces of filter paper were sprinkled in the dishes. The cumulative effect of the capsule parts was tested by two treatments: putting seeds into empty, intact capsules and putting pieces of all capsule parts with the seeds. Three other experiments were designed to test the effect on germination of substrate type, acid, and soaking. Steam sterilized sand, sphagnum, and a mixture of two parts sandy loam, one part sand, and one part sphagnum were the sub- strates tested. Petri dishes were half filled with the sub- strate and the seeds were sown on the surface, which was kept moist. The acid solutions tried were glacial acetic acid of pH 3.5 and two leachates of pH 4.5, one of moss and the other of leaf litter collected at the site where the seeds were obtained. Soaking periods used were 1, 3, 6, and 9 days in distilled water. Seeds kept at 22°C in the light and with “sufficient” moisture were used as the control for all tests. 150 Rhodora [Vol. 68 In the final experiment six lots of 200 seeds each were put in petri dishes and allowed to air dry for 1 to 6 days before they were transferred to moist filter paper. Dupli- cate sets of 100 seeds each were also air dried, then weighed and oven dried at 105°C to a constant weight. In this way moisture content of the seeds, as percentage oven-dry weight, was determined for each of the air drying periods. Although eight tests were designed for statistical treat- ment of the results, for some tests the great variation in number of seeds germinated among replications or the low total number of seeds germinated made a meaningful sta- tistical analysis impossible, Even the mean and the standard deviation cannot be given for some treatments because they indicate the possibility of a negative number of germina- tions, which is meaningless. Table 1 gives the results for the effect of temperature, moisture, and light on germination. Many more seeds germi- nated at 20°C than at 30° or 10°C, and more germinated in the light than in the dark. No germination occurred at the “deficient” moisture level in any temperature regime. At 20°C more seeds germinated in the “abundant” moisture condition than in the "sufficient." The difference is signifi- cant at the 2.5% level using a Student's t-test (Abundant, mean number of germinations/50 seeds = 4.33 + 1.51; Suf- ficient, mean = 2.17 + 1.33). The effect of the changing temperature treatments on germination is also shown in Table 1. Almost identical re- sults were obtained for the two treatments. A constant temperature of 20°C was as good for germination and per- haps better than, the two changing temperature regimes. The presence of the placenta, the calyx, the ovary wall, or all parts improved germination about fivefold over that of the control. An analysis of variance using the 2.5% significance level showed that there was no difference among the treatments involving capsule parts. Adding filter paper instead of capsule parts did not improve germination. Sand and sphagnum were superior to the sandy loam mixture or filter paper as substrates for seed germination. 1966] Shortia — Rhoades 151 Table 1. Effect of Temperature, Moisture, and Light On Seed Germination of Shortia galacifolia Temperature Number of Seeds Germinated After and Moisture 50 Days (300 seeds/treatment Light Dark 10° C Abundant 0 0 Sufficient 0 1 Deficient 0 0 20° C Abundant 26 2 Sufficient 13 0 Deficient 0 0 30° C Abundant 1 0 Sufficient 0 0 Deficient 0 0 Changing temperature Alternating 10^, 20° 4 Increasing 10° to 20° to 30° to 22° 3 Constant 20° 13 Using water with leaf litter in it instead of plain distilled water may have improved germination slightly, but using water with moss or glacial acetic acid added did not. Better germination was obtained for all soaking treatments than for the control, and the period of soaking had no effect. The average moisture content of Shortia seeds before they were air dried was 12% oven-dry weight. No loss of germi- native capacity occurred for seeds air-dried for up to 6 days. However, no loss of moisture occurred, because the vapor pressure deficit of the air remained low over the whole period. A separate set of 100 seeds was made up to be left until a significant amount of moisture was lost. The seeds were allowed to air-dry for about 3 days. Since no moisture loss had occurred after this time, the seeds were put in an oven at 30°C for 12 hours. Their weight was reduced 4% during this treatment, which was three times the weight lost by any of the air-dried lots of seeds. After removal from the oven, the seeds were left in the air for 20 days. 152 Rhodora [Vol. 68 They had regained their original weight when they were put on moist filter paper to test. germination. None of the seeds germinated. For 21 treatments the number of seeds germinated was recorded every 4 days. For most of the treatments the first germination occurred between 13 and 16 days from the start. Soaking caused germination to begin 4 to 8 days earlier. The distribution of germination over the 32 day period followed no particular pattern. A final count for all treat- ments was made after 50 days, and a few germinations were recorded at this time. Most germinations, however, had oc- curred before 32 days. Seeds of Shortia galacifolia germinated in the laboratory under a wide variety of conditions, indicating that the seeds do not have to be in the capsule to germinate. The factors which appear to be most significant in influencing germi- nation of Shortia seeds are temperature, moisture, light, and presence of capsule parts. The optimum temperature for germination is probably around 20°C, because substantially more germination oc- curred at this temperature than at either 10°C or 30°C. An abundant moisture supply is necessary for good germi- nation. No seeds germinate when moisture is always de- ficient, but some may germinate when moisture is alternately available and unavailable. Although the evidence is not conclusive, it also appears that Shortia seeds lose their germinative capacity if their moisture content falls below a certain critical level, even if they later regain moisture. Soaking seems to improve germination slightly, and even 9 days of soaking does not impair germination. Light appears to be beneficial for germination, since many more seeds germinated with 900 ft-c illumination than in the dark or with 30 ft-c. Although any capsule part improved germination when present with the seeds, combining parts did not produce a cumulative effect. Using leachate from leaf litter to water the seeds did not improve germination as much as the 1966] Shortia — Rhoades 153 presence of capsule parts, indicating that there is some or- ganic substance peculiar to Shortia which aids germination. From the results of the various tests it might be predicted that the best germination of Shortia seeds would be obtained by soaking the seeds at least 1 day, using sand as a substrate with capsule parts scattered over the sand, providing an illumination of about 900 ft-c, and supplying abundant moisture. Combining the two treatments which gave the best germination, “addition of placenta" and *209C-light- abundant moisture," was tried, but no increase in percentage of germination resulted. Although germination of Shortia seeds in nature outside the capsule has never been observed, the results of the labo- ratory experiment indicate that it undoubtedly occurs. Fac- tors which were found to be important for good germination in the laboratory are present in the natural habitat. Micro- environmental measurements made in several different stands of Shortia show that the air temperature during the growing season ranges from 13°C to 32°C, that the relative humidity never falls below 50%, and that the light intensity ranges from 20 to 1000 ft-c (Vivian, 1965). Decaying capsule parts and leaf litter are present on a sandy soil. In 1936, Ross thought the answer to the extreme locali- zation of Shortia was the fact, that the seeds are held tightly in the open capsule and must germinate there, thus limiting the spread by seed. Although Ross noted that some seeds fall from the capsules, he did not mention how many. The author has observed that 50% or more of the seeds may be lost within a month after the capsules open. Rain is proba- bly the primary agent in releasing the seeds, which could easily be carried far from the mature plants by rivulets or streams. It does not seem that lack of seed dispersal can be used to explain why Shortia fails to colonize new, apparently suitable habitats. Low seedling survival may be much more important in limiting the spread of Shortia by seed. Obser- vations suggest that the seedlings grow very slowly in nature and are quite subject to winter kill (Vivian, 1965). Seeds germinated in the laboratory produced seedlings about 154 Rhodora [Vol. 68 2 cm high after 8 months of growth. The amount of competi. tion that seedlings can tolerate is probably small. It seems that the problem of the restricted and discontinuous distri- bution of Shortia galacifolia has not been solved, but that requirements of mature plants, competition, and seedling survival, perhaps along with a low percentage of germi- nation, are more significant factors than lack of seed dispersal, DEPARTMENT OF BOTANY, DUKE UNIVERSITY, DURHAM, NORTH CAROLINA. LITERATURE CITED CRANDALL, DoRoTHY L. 1956. Shortia galacifolia in Gray’s Manual range. Rhodora 58: 38-40. Kersey, H. P. 1902. Shortia, p. 1664. In L. H. Bailey, Cyclopedia of American Horticulture. Macmillan Company, New York, 4 York Bot, Gard. 37: 208-211. VIVIAN, V. E. 1965. Department of Biology, Glassboro State College, vols. Ross, M. N. 1936. Seed reproduction of Shortia galacifolia. J. New Glassboro, New Jersey. Personal communication. THE APPLICATION OF THE LINNAEAN NAMES OF SOME NEW WORLD SPECIES OF EUPHORBIA SUBGENUS CHAMAESYCE DEREK BURCH In 1960 Wheeler resurrected a subject which most workers in the Euphorbiaceae must have thought long since settled. His proposal, in 1939, to change an established and well understood usage of the name Euphorbia maculata L. stirred up a controversy which had died down in print by 1948 (Svenson 1945; Fosberg 1946, 1947; Croizat 1947, 1948), and which Fosberg's rejustification of his earlier stand failed to revive in 1953. In 1962 Croizat answered Wheeler's case with what should have been the final word on the subject, but in the course of a revision of the genus Chamaesyce in the Caribbean it has been brought home to me that confusion stil exists in American herbaria, and that some collectors are following Wheeler’s suggestion even though the case for the change was never proved. If this were the only Linnaean name whose application to a New World species of Euphorbia sensu lato had been questioned the matter would best be allowed to lie, but there are other situations to clarify and, since some of them de- pend on the use of E. maculata, I have reluctantly decided to drag the old bones of the argument out once more. The other areas which must be discussed are the names for a group of erect, large-leaved plants including E. hyperi- cifolia (the name which Wheeler refers to as a *bandora's box" — presumably with some knowledge since he was the last to force its lid closed); the confusion between the Caribbean E. prostrata and E. chamaesyce of the Old World, and finally a note on the application of the name E. thymi- folia L. In the interests of brevity, only the relevant parts of earlier papers will be discussed. Most of the confusion in applying the names proposed by Linnaeus stems from a difference in interpretation of the relative importance of the parts of the material on which 155 156 Rhodora [Vol. 68 his description was based. Both Svenson (1945) and Stearn (1957) have published very lucid accounts of his descriptive method, pointing out the importance of the phrase-name, not only as a partial characterization of the plant but also as a key to the genus when taken in conjunction with the polynomials of the other species. These phrases were re- vised to maintain this differentiating function as new material or information was received, and the superseded phrases were put into synonymy. A mistake commonly made in deciding the application of a Linnaean name is to confuse the need to establish the concept of species which was in the author's mind with the present-day wish to have a single specimen as a representa- tion of the name given to the species. Linnaeus did not work with "type specimens", but based his species on an aggre- gate of earlier descriptions, illustrations, and actual ma- terial. It must be left to a worker familiar with the group, and having a clear understanding of the Linnaean method, to decide just what elements were included by Linnaeus in composing his polynomial phrase-name. Only when this has been done should the specimens now in the herbarium, which may or may not have been present when the differ- entiae were drawn up, be considered in the light of the International Code for designation as types of the names. 1. The application of Euphorbia maculata L. The species to which Linnaeus gave the trivial name “maculata” is number 21 in Species Plantarum edition 1. The full entry is made up of the usual polynomial phrase- name and references to earlier works (in this case a descrip- tion and illustration by Plukenet), an indication of habitat, and also a supplementary description including characters such as color which Dandy suggests to mean that it was taken from fresh material (personal comm. 1964). The author’s concept, then, we know to be based on a plant which Plukenet described and illustrated, with the phrase-name rewritten to be comparable with others in the genus, and probably to include information taken from a specimen. 1966] Euphorbia subgenus Chamaesyce — Burch 157 These are the facts in the case, and from this point we move into conjecture. For many years the name E. maculata has been applied to a prostrate plant, a specimen of which is in the Linnaean Herbarium on sheet number 630-11. The sheet is marked “21 maculata", which agrees with the entry in Species Plantarum. Most authors have taken the number to be one written on the sheet by Linnaeus, although Savage (1945), in his invaluable catalogue of the herbarium, ascribes it to Sir J. E. Smith. When I examined the speci- mens last summer I was unable to decide between these two viewpoints, but, in any case, there is general agreement that the epithet “maculata” on the sheet was added by Smith. There is another sheet, 630-4, on which Linnaeus wrote “17 Euphorbia maculata". In Species Plantarum number 17 is the species to which the trivial-name “hypericifolia” was given, and the specimen on the sheet is of one of a group of superficially similar erect species to which Linnaeus applied this name. It is clear that the number and the name on this sheet are contradictory, and that Linnaeus was mistaken in writing one or the other. The choice of which is in error is the first matter for conjecture, but in reaching a decision it should be borne in mind that the number was part of a major work which was in progress or had recently been completed (for it is not certain when the annotation on the sheet was made), while the name was, literally, a trivial name — a new, “shorthand” way of referring to a species, the use of which may still have been unfamiliar to Linnaeus. Wheeler, however, chooses to regard the trivial name as correct, the number as wrong, and on this builds the whole case for his application of the name. In effect, he first selected his type specimen, and then showed that it is feasible that it might have fitted Linnaeus’ concept of the species. Let us instead examine the entry in Species Plantarum, keeping in mind the two entities which have been offered as candidates for the name. The polynomial specific name, which was designed to separate the species from all others 158 Rhodora [Vol. 68 rather than as a description, contains no features which positively identify or disqualify either plant. Wheeler was correct in omitting this as a factor in his argument. Con- sider, however, the position of the species in the genus. Linnaeus was a methodical worker, and his purpose was to distinguish between the discrete groups that he considered species. What better way is there of doing this than by arranging similar things together, and then pointing out the features in which they differ? It is not suggested that Linnaeus followed this plan with complete consistency, but Croizat (1962) has shown in convincing detail that his genus Euphorbia is laid out in this fashion. The succulent species fall in the early part, followed by the shrubby types, while in the section Dichotomae, species 16-26, the arrange- ment is even more systematic. Species 16 through 19 are of erect or ascending habit with serrate leaves, species 20 through 24 are prostrate and have various leaf margins, and species 25 and 26 are erect with entire leaves. The placement of 21 maculata between 20 thymifolia and 22 prostrata, is strong evidence that the traditional application of the name to the prostrate plant is the correct one to follow. The descriptive phrase of Plukenet is noncommittal for our purpose of deciding between the two species, and I would agree with Wheeler that the illustration must also be re- jected as a decisive factor. However, I feel that he is rash in seizing on features of the italicized description that follows the note on the habitat in Species Plantarum. to support his choice of a type specimen on which to base the name. In his most recent article (1960) he persists in assuming that one of the two now in the Linnaean Her- barium must be treated as a type specimen in the modern sense. Even if these sheets were in Linnaeus' possession at the time of writing Species Plantarum, Wheeler's choice of one over the other because it agrees with parts of the description may be countered by reference to other charac- ters which would support the opposite choice. The shape of the leaves, and the description *subpilosa", for example, 1966] Euphorbia subgenus Chamaesyce — Burch 159 are each more appropriate to the specimen of the prostrate plant on sheet 630-11 than to the other. Nothing in the original entry in Species Plantarum, then, gives grounds for changing the application of the name, and there is sufficient doubt about Linnaeus' view of the speci- mens now in the herbarium for them to form an unreliable basis for such a drastic step. On the other hand it has been shown that the layout of Species Plantarum gives a positive reason for not making such a change, and the application to the prostrate species, affirmed by Jacquin's plate in 1772 and adopted by the vast majority of subsequent authors, should be maintained. There is one further point that must be discussed before this matter may finally be laid to rest. Both Wheeler (1939, 1960) and Fosberg (1946, 1953), who correctly apply the name E. hypericifolia to an erect plant, take the statement “Euphorbia maculata similis E. hypericifoliae" by Linnaeus in the Mantissa Altera (1771) to be selection of the erect element from the two which they feel he had earlier com- bined under the name E. maculata. Croizat's dismissal of this as an irrelevancy (1947, 1948) was not satisfactory, but it was not until I had the opportunity to examine some of the books in the Linnaean Library that I found a reason- able explanation. One of the first books to use differential phrase-names in the manner of Linnaeus was Patrick Browne's Civil and Natural History of Jamaica (1756). Stearn (1957) reports that this delighted Linnaeus, who went through the book providing (for his own use in later publications) the nomina trivialia that Browne had thought it unnecessary to give. Browne’s “Euphorbia 2. Minima reclinata, foliolis ovatis denticulatis . . ." gives as synonym the same “Tithymalus erectus acris . . ., Sloane Cat. 82, & H. t. 126" that Linnaeus included in his 17. hypericifolia. In recognition of this, “E. hypericifolia?” has been written in the margin of this entry, probably by Linnaeus himself. Browne goes on to describe his plant *. . . it is a slender weakly creeper and seldom runs above three or four inches from the root; its branches are smooth and slender and the 160 Rhodora [Vol. 68 leaves small and oval". Here, then, is the basis for the statement in the Mantissa. Linnaeus knew “maculata” to be a prostrate plant, and, on reading Browne's description of the plant based on the same element as his “hypericifolia” as a creeper, he simply drew attention to this similarity of habit. 2. The application of the name Euphorbia hypericifolia L., and the names for the group of morphologically similar species. This group, with strong, spreading to erect stems and leaves which are usually toothed, includes at least four species in America. They are distinguishable on pubescence, size of capsule, and on stipule and seed-coat characters, and have somewhat distinct geographical distributions. The names under which they have been known may be sum- marized thus: Entity A — northern plant with pubescence only in lines on stem; cyathia never glomerulate; capsules large; seed coat rippled . . . Euphorbia nutans Lag.; E. preslii Guss.; E. hypericifolia L. sensu Michaux, Torrey; E. maculata L. sensu Wheeler. Entity B — plant of the southern United States to South America; rarely and sparingly pubescent; cyathia never glomerulate; capsules large; seed coat transversely ridged . . . E. hyssopifolia L.; E. brasi- liensis Lam.; Chamaesyce nirurioides Millsp.; C. jenningsii Millsp. Entity C — tropical or subtropical plant; rarely pubescent; cyathia glomerulate; capsules small; seed coat wrinkled . . . E. hypericifolia L.; E. pilulifera L.; E. glomerifera (Millsp.) L. C. Wheeler. Entity D — tropical plant with all herbage close-pubescent; cyathia never glomerulate; eapsules large; seed coat transversely ridged ... E. lasiocarpa Klotzsch; E. hypericifolia L. sensu Wheeler. The central problem here is the application of the name E. hypericifolia L., which has been widely used for both Entity A and Entity C, and which Wheeler proposed in 1939 for Entity D. Entity B need not be involved in this discussion. It was described by Linnaeus in Systema Naturae edition 10 (1759) with the trivial name hyssopifolia, and this has been con- sistently applied by all authors except those who divided the taxon on the basis of leaf shape. Examination of a long series of this highly variable species has so far failed to 1966] Euphorbia subgenus Chamaesyce — Burch 161 substantiate any such division, and the name £F. hyssopifolia L. should probably be applied throughout. Wheeler’s case for the application of the name E. hyperi- cifolia to the pubescent West Indian species, Entity D, was made at a time when neither the Linnaean specimens nor Savage’s catalogue was available for study. He examined photographs of the two sheets in the herbarium labelled “hypericifolia,” and, having disposed of sheet 630-4 as the type of E. maculata L. in the fashion discussed above, he selected the second sheet, 630-3, as the type of the epithet hypericifolia. He correctly identified the specimen on this sheet as an example of Entity D, and proposed that the epithet be used for this species. The sheet, however, is marked “Br,” indicating that it came from Patrick Browne, most of whose collections were not added to the herbarium until 1758 (Stearn, 1957), and, thus, could not have been considered by Linnaeus when he described his E. hyperici- folia. Wheeler's proposal may be rejected on these grounds alone. Entity D was first described by Klotzsch in 1843 under the specific name E. lasiocarpa. Grisebach, in his Flora of the British West Indian Islands (1864), reduced the taxon to varietal level under E. hypericifolia L., but there has been general agreement that it merits specific rank, and that the name used by Klotzsch should be maintained. The question of the application of *hypericifolia" should also be approached by considering the entry in Species Plantarum. The polynomial phrase-name is taken without alteration from Wiman's Euphorbia, ejusque Historia . . ., a dissertation published under the supervision of Linnaeus in 1752. The entry here includes even more synonyms than are carried over into Species Plantarum, and also a descrip- tion which makes it clear that a variety a, with almost soli- tary “flowers”, and another 8, with “flowers” collected into heads, were recognized and included in the concept of the species. Variety « is probably Entity A, which appears to be the plant described in most of the earlier works of Lin- naeus given here as synonyms. Variety B is easily recog- 162 Rhodora [Vol. 68 nisable as Entity C, and this was the plant described and illustrated by Sloane as “Tithymalus erectus acris . . ." and by Commelin as “Tithymalus Africanus, seu Peplis major Brasiliensis . . ." (fide Tab. 10, “Tithymalus Ameri- canus, flosculis albis") to which reference is made in the synonymy. ]t is clear that Linnaeus held a broad concept of E. hypericifolia at the time of publication of Wiman's work and Species Plantarum, since the polynomial phrase- name of his earlier works was expanded in these to include features of Sloane's plant. In a case such as this, where the elements included by Linnaeus are now considered to form more than one species, Stearn (1957) suggests that the choice as to which is to bear the name should take into account not only all parts of the type material, but also the author's intent, and, if possible, subsequent and current usage. We have seen that both Entity A and Entity C are represented in the synonymy, and that while Entity A was the first to be described, this is balanced by the later broadening of the description to include Entity C. There is little in the entry in Species Plantarum to guide the choice, and no indication of the author's intent. The usage of the names, however, is another matter. Sir J. E. Smith, who acquired the herbarium in 1783, favored Entity A (Torrey, 1843), and his example was followed by Michaux (1803), Torrey, and a few early American authors. With these exceptions there has been an overwhelming use of E. hypericifolia as the name for Entity C, including Crantz' Institutiones Rei Herbariae (1766), Miller's Gardener's Dictionary edition 8 (1768), Aublet's Histoire des Plantes de la Guiane Francaise (1775), Humboldt's Nova Genera... (1817), Hooker's Exotic Flora (1823), and Boissier in DC. Prodromus (1862). In fact, the usage is so firmly established that it would be very much against the interests of stability of nomenclature to consider making any change, and the name E. hypericifolia L. should continue to be applied to this subtropical, erect plant with small capsules usually borne in glomerules. Sloane's illus- tration, cited by Linnaeus in Species Plantarum, is an ap- 1966] Euphorbia subgenus Chamaesyce — Burch 163 propriate lectotype for the name used in this fashion, and the specimen on which the illustration was based, if it still exists in the Sloane Herbarium in the British Museum, would constitute what Dandy has referred to as a “typo- type" (Stearn, 1957). The name for Entity A remains to be decided. The varietal status accorded it by Linnaeus was not questioned for a number of years, but in 1816 Lagasca described a taxon under the specific name E. nutans. Wheeler (1941) reports that he examined a specimen which probably repre- sents the plant that Lagasca described, and concluded that the name was apparently based on unusually vestite plants of Entity A. On these grounds the name E. nutans Lag. will be adopted for this species, with the more commonly used but later name E. preslii Guss. passing into synonymy. In summary, then, the names for this group of species are as follows: Entity A (Northern plant with large capsules) — E. nutans Lag. Entity B (Southern plant with large capsules) — E. hyssopifolia L. Entity C (Tropical plant with small capsules) — E. hypericifolia L. Entity D (Tropical plant which is densely pubescent) — E. lasio- carpa, Klotzsch. 3. Euphorbia prostrata Ait. and E. chamaesyce L. Wheeler suggested in 1941 that these two species are con- specific, basing his case on the fact that one of the two sheets labelled *chamaesyce" in the Linnean Herbarium (sheet 630-17) is actually a specimen of the plant which has been known as E. prostrata Ait. This sheet, however, is another of Brown's specimens, and was not in Linnaeus' possession until long after his description of E. chamaesyce in Species Plantarum. The other sheet (630-15) is a good example of the European species to which the name has usually been applied, and nothing in the original entry sug- gests that Linnaeus intended the limits of his species to be set broad enough to include the West Indian plant. The two species are, in fact, quite distinct, and the anno- tation by Linnaeus on sheet 630-17 is a simple misidentifi- cation. Croizat (1945) has documented the consistent use of the two names for the distinct entities, and there seems 164 Rhodora [Vol. 68 little room for argument against maintaining E. chamaesyce L. as the name for a Mediterranean plant not so far found in the New World, and E. prostrata Ait. as the name for a weed of the Old and New World tropics. Wheeler accepts this usage in a supplement to the second edition of Kearney & Peebles Arizona Flora published in 1962, but his correc- tion has been overlooked by the authors of other floras which have appeared since that date. 4. Application of the name E. thymifolia L. Wheeler (1941) drew attention to a suggestion by A. Gray that the specimens 630-10 “Euphorbia 20 thymifolia" and 630-11 in the Linnaean Herbarium are of the same entity. An examination of the sheets confirmed this, and that both are the prostrate E. maculata L. There is no specimen in the Herbarium of the plant which has tradi- tionally been called E. thymifolia L., and the protologue in Species Plantarum is of little direct help in deciding the application of the name. The polynomial phrase-name and descriptions by Burmann and Plukenet indicate that the plant is prostrate and pubescent, with serrate leaves and "flowers" somewhat collected into groups, and the illustra- tions cited fit these descriptions but show no characters which are diagnostic. Burmann, however, includes in his synonymy *'Cha- maesyce, Sloane Cat. 83". This was based on a Specimen from St. Jago de la Vega, now in the Sloane Herbarium, which is cited by Fawcett and Rendle in Flora of Jamaica (1920) as an example of E. thymifolia L. in the sense in which the name is used by most authors. This gives a posi- tive link between the name and a well-understood taxon, and since no other such link can be made, and since nothing in the protologue contradicts this usage, it seems reasonable to accept it even though Sloane’s plant was probably not seen by Linnaeus (Stearn, 1957). The plant to which this name has been applied is found in most tropical regions, and is distinguished from similar species by the shortness of the stalk of the female flower, which forces the cyathium to split down one side as the 1966] Euphorbia subgenus Chamaesyce — Burch 165 capsule develops. Bernard, Icones Bogorienses, 4, t. 315. 1910, illustrates this character well. Grisebach, in his Flora of the British West Indian Islands (1864), treats the taxon as a variety thymifolia of E. maculata L., but Boissier in DC. Prodromus (1862), Drury in Handbook of the Indian Flora (1869), Urban in Symb. Antill. (1910), and most subsequent authors consider that it merits specific rank. This usage is now so widespread and so well understood that nomenclatural stability will be best served by maintaining it, unless some contrary intention on the part of Linnaeus can be shown. MISSOURI BOTANICAL GARDEN, ST. LOUIS, MISSOURI LITERATURE CITED CRoIzAT, L. 1945. “Euphorbia chamaesyce" in the United States. Bull. Torr. Bot. Club 72: 312-318. 1947. Euphorbia maculata L. Bull. Torr. Bot. Club 74: 153-155. 1948. Euphorbia maculata: A rejoinder. Bull. Torr. Bot. Club 75: 188. 1962. Typification of Euphorbia maculata L. A Re- statement and a Conclusion. Webbia 17: 187-205. FosBERG, F. R. 1946. Application of the name Euphorbia maculata L. Rhodora 48: 197-200. 1947. Euphorbia maculata again. Bull. Torr. Bot. Club 74: 332-355. 1953. Typification of Euphorbia maculata L. Rhodora 55: 241-243. JacQUIN, N. J. 1772. Hortus Botanicus Vindobonensis 2: 87, t. 186. Vienna. KLoTzsCcH, J. F. 1843. Nov. Act. Nat. Cur. 19: (supp. 1): 414. La4GASCA, M. 1816. Genera et Species Plantarum... 17. Madrid. LINNAEUS, C. 1753. Species Plantarum ed. 1. Stockholm. 1771. Mantissa plantarum altera. Stockholm. MicHaUX, F. A. 1803. Flora boreali-americana. Paris. SAVAGE, S. 1945. A catalogue of the Linnaean Herbarium. London. STEARN, W. T. 1957. An introduction to the Species Plantarum and cognate botanical works of Carl Linnaeus, (Introduction to Ray Society facsimile of Species Plantarum ed. 1). London. 166 Rhodora [Vol. 68 SVENSON, H. K. 1945. On the descriptive method of Linnaeus. Rhodora 47: 273-302, 363-388. TORREY, J. 1843. Flora of the State of New York, Albany. WHEELER, L. C. 1939. A miscellany of the New World Euphorbia- ceae — II. Contr. Gray Herb. 127: 48-78. 1941. Euphorbia subgenus Chamaesyce in Canada and the United States exclusive of southern Florida, Rhodora 43: 97-154, 168-205, 223-286. (Reprinted as Contr. Gray Herb. 136). 1960. Typification of Euphorbia maculata. Rhodora 62: 134-141. IDENTITY OF WITCHWEED IN THE SOUTHEASTERN UNITED STATES In 1956 witchweed was found to be parasitizing field plantings of corn (Zea mays L.) in North Carolina and South Carolina. Because potential loss from this infestation is severe, immediate steps were effectively taken to prevent the spread of the parasite. The original Specimens were tentatively identified as Striga asiatica (L.) O. Kuntze. The botanical nomenclature of this genus has been subject to interpretation by the various botanists who have dealt with it and only recently has an effective review been written (Saldanha, C. J. 1963). In late 1964, Dr. V. P. Rao, a visitor from India, ques- tioned the identification of the witchweed in the United States. He kindly furnished a copy of the Saldanha paper and a suite of specimens identified by Saldanha with which the local Striga could be compared. No doubt remains that this plant is Striga lutea Lour., which includes one of the elements of S. asiatica, a name now abandoned as a source of confusion. However, Rao writes (in a personal communication) * ... the species found in the United States has red flowers which I have never seen in the common witchweeds which we come across in this country [India]. While it is true that the bulk of the population in the United States has very bright chinese-red flowers, individuals with lemon yellow flowers occasionally appear. Rarely an individual plant is found in which the flowers are yellow with red marks or stripes. Perhaps the flower color is due to the original introduction of red-flowered plants from Kottayam, India, where a population is known, or from South Africa, where the bulk of the plants of this species are said to be red-flowered. C. EARLE SMITH, JR. BOTANIST, CROPS RESEARCH DIVISION, AGRICULTURAL RESEARCH SERVICE, U.S. DEPARTMENT OF AGRICULTURE, BELTSVILLE, MARYLAND. 167 BROMUS MOLLIS AND ALLIES IN NEW ENGLAND FRANK C. SEYMOUR A problem like that in many other species is presented by Bromus mollis L., in which confusion has arisen because of unrecognized species, varieties, or forms. Apparently this species has been interpreted as including all specimens with short peduncles and hairy lemmas. In its form with glabrous lemmas, it has often been confused with B. race- mosus L., and the form of B. commutatus Schrader with hairy lemmas has often been mistaken for typical B. mollis. To Mr. Peter S. Green of the Arnold Arboretum I am indebted for calling my attention to the presence, in New England, of more than one species commonly included in Bromus mollis, for lending me a key which set apart ne- glected or little recognized taxa, and for making very val- uable suggestions regarding this article. To him I express my hearty thanks. I thank Dr. Richard A. Howard, Direc- tor of the Arnold Arboretum, and Dr. Reed C. Rollins, Di- rector of the Gray Herbarium, and Mr. Richard J. Eaton, Curator of the Herbarium of the New England Botanical Club, for the privilege of referring to specimens deposited in their institutions. I appreciate very much the help of Dr. Robert C. Foster, of the Gray Herbarium, and Dr. Lily M. Perry and Mr. Green, of the Arnold Arboretum, in finding references for original descriptions. I thank Mrs. Lazella M. Schwarten for the facilities of the Harvard Herbarium Library, and Dr. C. E. Hubbard of the Royal Botanic Gardens, Kew, England, for the reference for B. commutatus Schrader var. pubens Wats. Appreciation is expressed also to Dr. Howard W. Pfeifer, of the University of Connecticut and Mr. Leopold A. Charette, until recently of Yale University, for citations of specimens in their re- spective institutions. In 1833, Hardouin (Congr. Sci. France 1: 59. 1833) pro- posed a new name, Bromus Thominii, for a grass which had previously been described and named B. arenarius 168 1966] Bromus mollis — Seymour 169 Thomine-Desmasures (Mem. Soc. Linn. Calvados 1: 40. 1824) because Labillardiére had used the name B. arenarius for another species in 1804. However, many botanists have considered this species not sufficiently distinct from B. mol- lis; and in America, where it has been introduced, it has commonly been overlooked or ignored to the present time. In 1924, another segregate of B. mollis was recognized at specific rank, namely B. lepidus Holmberg (Botaniska Noti- ser 1924: 326.). Plants differing from the typical by having, or lacking, pubescence on the lemmas occur in these species and have often been treated as varieties. Here, the example of Fern- ald and Holmberg is followed by reducing them to forms. When a taxon differs from the typical variant in only one character and does not seem to have a distinct geographic range, it seems more appropriate to recognize it as a form rather than a variety. Much of the confusion in these species arises from the fact that all three of them and also Bromus commutatus Schrader and B. secalinus L. may have either glabrous or hairy lemmas. Consequently, pubescence on the lemmas, or lack of it, does not separate these species. In an effort to clarify this group and some others to which they are closely related, the following key is presented, dealing with all the New England species in the sections Bromium Dum. in the Manual of Grasses of the United States by A. S. Hitchcock, ed. 2 by Agnes Chase, 1951, and Zeobromus Griseb. in Gray’s Manual, ed. 8, by M. L. Fernald, 1950. A. Branches of panicle short; peduncles all or mostly shorter than spikelets; panicle erect; margins of lemmas overlapping the lemma next to them. B. Lower lemmas 8-11 mm long; lower glume (6-)7-9 mm long; upper glume (7-) 8-9 mm long. ...m 1. B. mollis. C. Lemmas hairy. «HH 1a. B. mollis f. mollis. C. Lemmas glabrous. ......... 1b. B. mollis f. leiostachys. B. Lower lemmas 5.5-7.5 mm long; lower glume 4.5-6 mm long; upper glume 5.5-6.6 mm long. D. Lower lemmas 6.5-7.5 mm long; lower glume 4.5-6 mm 170 Rhodora [Vol. 68 long; upper glume 6-6.6 mm long; grain shorter than palea. —— HERRERA 2. B. Thominii. E. Lemmas glabrous, except minutely scabrous margins. —— RE 2a. B. Thominii f. Thoméanii E. Lemmas hairy. .................. 2b. B. Thominii f. hirsutus. D. Lower lemmas 5.5-6.5 mm long; lower glume 4.5-5 mm long; upper glume 5.5-6 mm long; grain longer than DEI MM 3. B. lepidus. F. Lemmas glabrous, .................. 3a. B. lepidus f. lepidus. F. Lemmas hairy. .................... 3b. B. lepidus f. lasiolepis. A. Branches of panicle longer; peduncles all or mostly longer than spikelets; panicle erect or drooping; margins of lemmas not always overlapping the lemma next to them. G. Margins of lemmas inrolled at maturity, leaving axis visible; lemmas 6.5-9 mm long. ......... eee 4. B. secalinus. H. Lower lemmas 7-9 mm long; sheaths glabrous. ................-- RM 4a. B. secalinus var. secalinus. I. Lemmas awnless. .....cccccsssecscesesssssesessssceeccescsececccecsenscesceceete MEME 4c. B. secalinus var. secalinus f. submuticus. I. Lemmas with awns 1-6 mm long. J. Lemmas glabrous. .......... eese — 4b. B. secalinus var. secalinus f. secalinus. J. Lemmas hairy. ...... sese 4d. B. secalinus var. secalinus f. hirtus. H. Lower lemmas 6.5-7 mm long, glabrous; lower sheaths hairy. ...eeeeseeeeeeteee nre 4e. B. secalinus var, hirsutus. G. Margins of lemmas concealing axis, not inrolled; lower lemmas 6.5-12 mm long. K. Lower lemmas 6.5-8(-9) mm long. L. Anthers 3-4.5 mm long; upper lemmas about as long as lower; panicle open, its branches spreading, lower branches more than two at a node, each bearing two or more spikelets; lower sheaths hairy, upper sheaths glabrous. .......eeeeeeree eee eere 7. B, arvensis. L. Anthers 2-2.5 mm long; upper lemmas distinctly shorter than lower; panicle rather dense. its branches erect or ascending, lower branches 1-2 at a node, each bearing 1 (rarely 2) spikelets; sheaths all hairy. ...... RR 5. B. racemosus. K. Lower lemmas 9-12 mm long. M. Spikelets drooping, 0.8-1.3 em wide, awnless or nearly so; anthers 0.8-1 mm long; branches of panicle usually bearing 1 spikelet; lower lemmas about as wide as long; palea much shorter than lemma, about 6 mm long. .......eeeeeeeeene 8. B. brizaeformis Fischer & Meyer 1966] Bromus mollis — Seymour 171 M. Spikelets erect or ascending, 4.5-6 mm wide; awns usually 4-10 mm long; anthers 1.5-2 mm long; branches of panicle usually bearing 2 or more spikelets; lower (and other) lemmas much longer than wide; palea about as long as lemma. ..............- 6. B. commutatus. N. Lemmas glabrous except minutely scabrous mar- EIE esie 6a. B. commutatus f. commutatus. N. Lemmas hairy. ........ 6b. B. commutatus f. pubens. To assist further in identifying these species, a few rep- resentative specimens are cited. Bromus mollis is too familiar to require citations; likewise the typical forms of some other species. It will be evident from the list that in America B. Thominii f. hirsutus also is frequent whereas the other species and forms most closely related to them are less common. 1. Bromus mollis L. Sp. Pl. ed. 2. 112. 1763. 1a. Bromus mollis f. mollis. 1b. Bromus mollis f. leiostachys (Hartman) Fernald, Rhodora 35: 316. 1933. In this article Fernald makes clear the reasons why this infraspecific epithet must be used instead of leptostachys of Persoon. REPRESENTATIVE SPECIMENS. MASSACHUSETTS: Roxbury, 1878, E. Faxon (YT); Lancaster, 1 July 1943, F. C. Seymour 7240 (NEBC). CONNECTICUT: Fairfield, 12 June 1914, E. H. Eames 8753 (GH). NEW JERSEY: Camden, 3 June 1877, Parker (GH). DISTRICT OF COLUMBIA : Washington, 31 May 1905, A. S. Hitchcock, Am. Grasses 891 (GH), distributed as B. racemosus L., see Rhodora 35: 316. 1933. MICHIGAN: near Lansing, 22 June 1886, L. H. Bailey, Jr. (GH). IDAHO: New Plymouth, 21 May 1910, J. F. Macbride 94 (GH). OREGON: without precise locality, Elihu Hall 1871 (GH). 2. Bromus Thominii Hardouin, Congr. Sci. France 1: 59. 1833. 2a. Bromus Thominii f. Thominii. REPRESENTATIVE SPECIMENS. MASSACHUSETTS: Stockbridge, 25 July 1914, R. Hoffmann (NEBC). RHODE ISLAND: Block Island, 12 Aug. 1919, C. B. Graves (NEBC). CONNECTICUT: Bridgeport, 24 June 1915, E. H. Eames 9022 (NEBC). BRITISH COLUMBIA: Vancouver Island, 20 June 1907, C. O. Rosendahl 1887 (GH). 2b. Bromus Thominii f. hirsutus (Holmberg) Hylander, Uppsala Univ. Arskr 1945 (7): 85. 1945. REPRESENTATIVE SPECIMENS. MAINE: Seal Harbor, Mt. Desert Island, 9 July 1928, G. L. Stebbins, Jr. 327 (NEBOC) ; N. Berwick, 4 July 1892, 172 Rhodora [Vol. 68 J. C. Parlin (GH, NEBC). MASSACHUSETTS: Salem, 26 May 1890, W. P. Conant (NEBC) ; Cambridge, 10 June 1891, M. L. Fernald (NEBC) ; Lincoln, 16 June 1913, M. L. Fernald (NEBC) ; Boston, Parker's Hill, June 1877, C. E, Faxon (Gu); Centerville, 17 June 1896, E. F. Wil- liams (NEBC) ; Nonamesset, 29 May 1928, J. M. Fogg, Jr. 3394 (NEBC) ; Penikese, 3 July 1928, J. M. Fogg, Jr. 3412 (NEBC); Chilmark, 23 Aug. 1895, Sydney Harris (NEBC); Nantucket, 15 June 1909, Mrs. N. F. Flynn (VT). RHODE ISLAND: Block Island, 30 June 1938, K. P. Jansson (CONN). CONNECTICUT: New London, 1891, C. B. Graves (GH); 16 June 1932, K. P. Jansson (CONN); Hartford, 2 June 1901, A. W. Driggs (GH). NEW YORK, Long Island, Orient Point, 26 July 1940, K. P. Jansson (CONN). MICHIGAN: Meridan Twp. T4N, RIW sec. 7, Ingham Co., G. W, Parmelee 231 (VT). ILLINOIS: Peoria, July 1904, F. E. McDonald (GH). NEVADA: King's Canon, C. F. Baker 1104 (VT). CALIFORNIA: Mather, 30 May 1931, D. D. Keck 1126 (GH); Redding, Shasta Co., 4 July 1914, L. E. Smith (GH); Webb Ranch, San Mateo Co., 5 May 1960, D. M. Porter 678 (GH). WASHINGTON ; Pullman, 8 July 1950, L. C. Hulbert 499 (CONN). 3. Bromus lepidus Holmberg, Bot, Not. 1924: 326. 1924. 3a. Bromus lepidus f. lepidus. REPRESENTATIVE SPECIMENS. MASSACHUSETTS: Stockbridge, 21 June 1915, R. Hoffmann (NEBC). CONNECTICUT: Bridgeport, 24 June 1915, E. H. Eames (vT). 3b. Bromus lepidus f. lasiolepis Holmberg, Bot. Not. 1924: 326. 1924. Syn. f. micromollis (Krósche) Holmberg in Hartman, Handb. Skand. Fl. ed. Holmberg, 265. 1926. REPRESENTATIVE SPECIMEN. MASSACHUSETTS: Jamaica Plain, 19 June 1963, P. S. Green 49/63 (NEBC). 4. Bromus secalinus L. Sp. Pl. 76. 1753. 4a. Bromus secalinus var. secalinus. 4b. Bromus secalinus var. secalinus f. secalinus 4c. Bromus secalinus var. secalinus f. submuticus (Reichenb.) Ascher- son & Graebner, Syn. Mittelleurop. Fl. 2(1): 604. 1901. REPRESENTATIVE SPECIMENS. MAINE: E. Auburn, 1 Sept. 1897, ex. herb. E. B. Chamberlain (NEBC); N. Berwick, 26 June 1896, J. C. Parlin (NEBC). MASSACHUSETTS: Jamaica Plain, 6 Aug. 1925, E. J. Palmer 28164 (NEBC); New Bedford, s.d., E. W. Hervey (NEBC). CONNECTICUT: West Haven, 30 June 1908, R. W. Woodward (NEBC). 4d. Bromus secalinus var. secalinus f. hirtus (F. Schultz) Seymour, stat. nov. Basionym: B. mutabilis F. Schultz var. hirtus F. Schultz, Flora 32: 234. 1849. Synonym: B. secalimus L. var. hirtus (F. Schultz) Ascherson & Graebner, Syn. Mitteleurop. Fl. 2(1): 604. 1901. 1966] Bromus mollis — Seymour 173 REPRESENTATIVE SPECIMENS. MAINE: Bethel, 25 June 1938, L. A. Wheeler (NEBC). CONNECTICUT: Southington, 26 June 1903, C. H. Bissell (NEBC). 4e. Bromus secalinus var. hirsutus Kindberg, Bot. Not. 1887: 32. 1887. REPRESENTATIVE SPECIMENS: VERMONT: Townshend, 23 June 1912, L. A. Wheeler (NEBC). MASSACHUSETTS: Revere, 1878, H. A. Young (NEBC); Needham, 30 July 1883, T. O. Fuller (NEBC) ; Milton, 15 July 1918, N. T. Kidder (NEBC); Hingham, 11 July 1926, C. H. Knowlton (NEBC). RHODE ISLAND: Edgewood, 26 June 1913, T. Hope (NEBC). CONNECTICUT: Waterbury, 14 July 1910, A. E. Blewitt 160 (NEBC). 5. Bromus racemosus L. Sp. Pl. ed 2. 114. REPRESENTATIVE SPECIMENS: MAINE: Foxcroft, 25 June 1894, M. L. Fernald 565 (NEBC) ; Seal Harbor, 9 July 1928, G. L. Stebbins, Jr. 326 (NEBC) ; 29 June 1899, E. L. Rand (NEBC). MASSACHUSETTS: Charles- town, 19 Aug. 1879, C. E. Perkins (NEBC) ; Brookline, 14 June 1883, C. W. Jenks (NEBC); Nantucket, 21 June 1910, E. P. Bicknell 9773 (NEBC). 6. Bromus commutatus Schrader, Fl. Germ. 1: 353. 1806. Synonym: B. pratensis Erhart ex Hoffm. Deutschl. Fl. ed. 2. 2: 52. 1800, non Lamarck 1783. 6a. Bromus commutatus f. commutatus. 6b. Bromus commutatus f. pubens (Wats.) Seymour, stat. nov. Basionym: B. commutatus Schrader var. pubens Wats., Phytologist 1: 1062. 1844. REPRESENTATIVE SPECIMENS: MAINE: N. Berwick, 27 June 1897, J. C. Parlin 849 (NEBC). VERMONT: Ely, 27 June 1935, A. A. Beetle 37 (NEBC). MASSACHUSETTS: Milton, 8 June 1924, N. T. Kidder (NEBC). CONNECTICUT: North Haven, 30 June 1962, R. J. Eaton 2 June 1909 & 18 June 1910, E. P. Bicknell (NEBC). 7. Bromus arvensis L. Sp. Pl. 77. 1753. REPRESENTATIVE SPECIMENS: MASSACHUSETTS: Arnold Arboretum, 8 June 1942, E. J. Palmer 46002 (NEBC) ; Grafton, 30 June 1951, B. N. Gates 27059 (NEBC); Worcester, 10 July 1946, B. N. Gates 20499 (NEBC). CONNECTICUT: North Haven, 30 June 1962, R. J. Eaton (NEBC) ; Bridgeport, 1 July 1912, E. H. Eames 8543 (NEBC) ; Bridge- port, 10 July 1912, E. H. Eames 8568 (NEBC). 8. Bromus brizaeformis Fischer & Meyer, Ind. Sem. Hort. Petrop. 3: 30. 1837. REPRESENTATIVE SPECIMENS: VERMONT: Charlotte, 27 June 1880, 174 Rhodora [Vol. 68 C. G. Pringle (VT). MASSACHUSETTS: South Boston, 28 May 1879 & 8 June 1879, C. E. Perkins (NEBC) ; Charlestown, 23 July 1881, Herb. C. W. Swan (yu); New Bedford, 15 July 1903, Herb. E. W. Hervey (NEBC). CONNECTICUT: East Haddam, 5 July 1903, H. S. Clark (NEBC); New Haven, 1884, R. W. Woodward (NEBC). WASHINGTON: Pullman, May 1898, A. D. E. Elmer 1330 (vT); Whitman Co., Karniak Mt., Willamette Mer. 20 July 1951, L. C. Hulbert 2326 (CONN). ONTARIO: comte d'Haldimand, 2 July 1936, F. Marie-Victorin, F. Rolland-Germain, F. Dominique 46200 (CONN). PRINGLE HERBARIUM UNIVERSITY OF VERMONT, BURLINGTON CONTRIBUTION TO THE FUNGUS FLORA OF NORTHEASTERN NORTH AMERICA. IV" HowARD E. BIGELOW AND MARGARET E. BARR? The following continues a report of unusual or rare species of higher fungi that we have collected in New Eng- land during recent years. In the summer of 1963 our field work was done in the White Mountains of New Hampshire, and during the summer and fall of 1964 we collected in the Mt. Mansfield area of Vermont. Although both years were generally dry in New England, sufficient moisture did ac- cumulate at these two areas to produce a number of interest- ing species. The senior author is responsible for the information on agarics, while the junior author has provided the data on ascomycetes. The colors within quotation marks are from Ridgway, R., 1912. Color Standards and Color Nomencla- ture. Washington, D. C. The collections are deposited in the Herbarium of the University of Massachusetts, Amherst, unless otherwise noted. Recent research has been possible through the support of National Science Foundation Grant G-19534. We are grate- ful also to Mr. Stanley J. Smith, Senior Curator of Botany, New York State Museum, and Dr. I. MacKenzie Lamb of the Farlow Herbarium, for the opportunity to examine the type specimens cited. Mr. R. Goodrich, District Ranger, White Mountains National Forest, New Hampshire, and Mr. Perry Merrill, Commissioner, Department of Forests and Parks, State of Vermont, kindly extended permission for our col- lecting in New Hampshire and Vermont. We also wish to acknowledge the Research Council, University of Massa- chusetts for financial assistance with the cost of plates and reprints of this article. T. Rhodora 62: 186-198. 1960. II. Rhodora 64: 126-187. 1962. III. Rhodora 65: 289-309. 1963, Contribution from the Department of Botany, University of Massachusetts, Amherst. "Mrs. Howard E. Bigelow. 175 176 Rhodora [Vol. 68 * Plate 1331 Callistosporium luteoolivaceum (Berk. & Curt.) Singer. x 1. BASIDIOMYCETES Callistosporium luteoolivaceum (Berk. & Curt.) Singer, Lloydia 9: 117. 1946. Plate 1331. Agaricus luteoolivaceus Berk. & Curt. Ann. Mag, Nat. Hist. III, 4: 286. 1859. Agaricus coloreus Peck, Buffalo Soc. Nat. Sci. Bull. 1: 46. 1873. Agaricus rubescentifolius Peck, N. Y. State Mus. Rep. 39: 38. 1886. Collybia luteoolivacea (Berk. & Curt.) Sacce. Syll. Fung. 5: 215. 1887. Collybia colorea (Peck) Sace. Syll. Fung, 5: 230. 1887. Collybia rubescentifolia Peck, N. Y. State Mus. Rep. 41: 83. 1888. Tricholoma rubescentifolium (Peck) Sace. Syll. Fung, 9: 15. 1891. Collybia colorea var. rubescentifolia (Peck) Peck, N. Y. State Mus. Rep. 49: 51. 1896. Callistosporium psilocybe Murr. & Singer, Mycologia 36: 363. 1944. Pileus 1-4 cm. broad, convex at first with an inrolled and slightly incurved margin, expanding to broadly convex, finally plane, margin slightly inrolled for some time, exceeding gills at times, sometimes sulcate but not pellucid-striate, surface glabrous when moist, matted 1966] Fungus Flora — Bigelow and Barr TUE fibrillose or innately radiate-fibrillose when faded, dull yellow at first becoming cinnamon to rufous brown (“orange cinnamon”, “Hays russet”, “brick red”, “Brussels brown”, “cinnamon rufous”), hygroph- anous, fading to yellowish or buff (near “antimony yellow”, “deep colonial buff”, “light ochraceous salmon”, “light ochraceous buff”), disc darkest; flesh thin, rather brittle, watery yellowish fading to pale yellowish or whitish, no odor and taste or taste rarely bitter. Lamellae adnate or adnexed, finally emarginate then seceding, close, broad (up to 7 mm.), not forked or intervenose, arched at times, yellow (“empire yellow", “yellow ochre”, “antimony yellow”, “cha- mois”) not fading, edges crenate or merely uneven, at times marginate with reddish brown in age. Stipe 2.5-5 em. long 1.5-5 mm. thick at apex, equal or the base en- larged (up to 7 mm.), and tapering upward, solid at first (interior concolorous with pileus flesh), becoming hollow, at times eccentric, often curved, surface innately fibrillose-striate, concolorous with lamel- lae, darkening to dingy “ochraceous buff", “cinnamon”, or “dresden brown”, base with white tomentum and sometimes a few white rhizoids. Spores 4.5-7 X 2.5-4 n, elliptical, sometimes elliptic-oblong, smooth, not amyloid, often with vinaceous droplets or contents in KOH, white in mass; basidia 16-28 X 4.5-7 u, 2- or 4-spored, often with vinaceous droplets or granules in KOH; pileus tissue: fresh material yellowish in water mount, revived material pale vinaceous in KOH, pigment in cell contents and often punctate-encrusted, surface at times with protruding cystidioid end cells, cuticular hyphae cylindrical, 2.5-5.5 & in diameter, tramal hyphae cylindrical to slightly inflated, 2-8 u in diameter, clamp connections absent; gill trama subparallel, broad, hyphae cylindrical to slightly inflated, 2.5-7.5(-9) & in diameter. Scattered to gregarious or occasionally cespitose, On conifer logs and stumps. Material examined: Bigelow 10788, near Katahdin Stream Camp- ground, Baxter State Park, Maine, August 6, 1962; 7309, 7310, 7311, Amherst, Massachusetts, August 4, 1958; 7411, Amherst, August 9, 1958; 7515, Amherst, August 17, 1958; 7600, Mt. Toby, Sunderland, August 27 1958; 7739, Mt. Toby, September 15, 1958; 7802, Mt. Toby, September 20, 1958; 8264, Amherst, July 25, 1959; 8956, Wahconah Falls, Dalton, July 26, 1960; 9007, D.A.R. State Forest, Goshen, August 2, 1960; 9234, D.A.R. State Forest, September 1, 1960; 9624, Pittsfield State Forest, Pittsfield, July 26, 1961; 9683, New Salem, August 2, 1961; 11990, near Sawyer Rock, White Mts. Nat, Forest, New Hampshire, July 27, 1963; 12246, Oliverian Brook Trail, White Mts. Nat. Forest, August 13, 1963; C. H. Peck, North Greenbush, New York (type of Agaricus rubescentifolius, nys), August; C. H. Peck, Croghan (type of Agaricus coloreus pro parte, NYS). This agaric can be very confusing in the field as the 178 Rhodora [Vol. 68 colors of the pileus and stipe vary markedly according to the age of the carpophore or its moisture content. In my field work the brown stage of the pileus seemed to be encountered much more frequently than the yellow stage. I have not had the opportunity to study other species in this genus, but the carpophores of C. luteoolivaceum at least turn a rather bright vinaceous red when dilute KOH solu- tion is applied to fresh material. In drying, specimens be- come a dark vinaceous color over all. Clitocybe albimontana Bigelow, nomen novum Omphalina oreades Singer, Pap. Mich. Acad. Sci., Arts & Letters 32: 123, 1946. Pileus (2.5-)4-10 mm. broad, convex with an incurved margin at first, expanding to broadly convex, dise rarely slightly depressed, glabrous and striate moist, disc brown, striations dark brown (“bister”), paler in between striations (near “snuff brown”), hygro- phanous, fading to sordid alutaceous, fibrillose and opaque faded, fibrils often erect near margin; flesh thin, pliant, concolorous with pileus moist or faded, no odor, taste mild. Lamellae decurrent, usually evenly and forming a collar on the stipe apex, narrow to moderately broad (+ 1 mm.), distant, rarely forked, not intervenose, arched, color a pale “buffy brown", edges even. Stipe 6-9 mm. long, 0.5-1.5 mm. thick, equal or tapering downward from a slightly enlarged apex, pruinose at first becoming glabrous, solid, rather soft, darker brown than lamellae ("buffy brown", *Sac- cardo's umber”). Spores 7.5-10 X 3-4 a, shape variable: elliptical to obovate or sub- cylindrical in face view, sublacrymoid or lacrymoid in side view, some- times slightly bent at either end, smooth, not amyloid; basidia 24-31 AX 5.5-7 u, 4-spored, sterigmata up to 7 u long, basidioles somewhat irregular in shape at times; cystidia not differentiated; pileus tissue: cuticle brown in KOH, pigment finely but distinctly encrusted, hyphae cylindrical or inflated, 3-8.5 u in diameter, trama pale brown, hyphae smooth or encrusted, cylindrical or inflated, 3-11 4 in diameter, clamp connections absent; eill trama interwoven, pale brown, hyphae smooth or encrusted, cylindrical or inflated, 1.5-7(-11) 4 in diameter. Scattered or gregarious. On moss or sandy soil in alpine zone. Material examined: Bigelow 11918, 11919, Lake-of-the-Clouds Trail from peak, Mt. Washington, New Hampshire, July 23, 1963; 12339, 12340, trail to Alpine Garden from peak, Mt. Washington, August 19, 1963; R. Singer, Mt. Washington (type of Omphalina oreades, F). 1966] Fungus Flora — Bigelow and Barr 179 O% Plate 1332 Hygrophorus lilacinus (Laest.) M. Lange. X L The preceding records are apparently the first of this species since Singer found it twenty years ago in the same locality. C. albimontana is easily overlooked because of its small size and a tendency to grow under overhanging moss where water or paths have cut down into the soil. A number of carpophores were found on both east and west sides of the mountain below the peak. In transferring Omphalina oreades to Cline rape. a new species epithet is necessary because of Clitocybe oreades Murrill. Clitocybe asterospora (Lange) Moser apud Gams, Kleine Kryptogamen Flora II, p. 52. 1953. Gregarious. On wet moss, in depression between rocks. Material examined: Bigelow 14223, top of Mt. Mansfield, Vermont, September 24, 1964. This agaric appears to be rare in North America. I know of fruit- ings only in Michigan and Ontario besides the specimens found on Mt. Mansfield. Hygrophorus lilacinus (Laest.) M. Lange, Meddelelser om Gr¢nland 148(2): 68. 1957. Plate 1332. 180 Rhodora Plate 1583 Hygrophorus monticola Hesler & Smith. X 1. 1966] Fungus Flora — Bigelow and Barr 181 Agaricus (Omphalia) lilacinus Laestadius, Lapp. Torn. p. 45, n. 156. 1860. Hygrophorus violeipes M. Lange, Meddelelser om Grénland 147 (11) : 18. 1955. Material examined: Bigelow 13656, Mt. Mansfield, Vermont, August 29, 1964. A small group of carpophores was found in a mossy de- pression near the “upper lip” on top of Mt. Mansfield. The bright luteous colors of cap and gills and the violet stipe make this species easy to recognize in the field. In micro- scopic characters as well as all field characters, my speci- mens agree exactly with Lange’s (1955) description of H. violeiipes. Apparently Bigelow 13656 is the first record of H. lilacinus in North America. In 1957, Lange placed Omphalia luteolilacina Favre in synonymy with H. lilacinus. While there is certainly a resemblance between the two species, I do not agree with Lange’s treatment. As described by Favre (1955), O. luteo- lilacina lacks clamp connections, has elliptic-cylindric spores and a hispid stipe. These characters are unlike H. lilacinus and certainly sufficient to maintain O. luteolilacina as a separate taxon. Hygrophorus monticola Hesler & Smith, North American Species of Hygrophorus, p. 359. 1963. Plate 1333. Pileus 1.5-16 em. broad, at first convex with an incurved and inrolled margin, not striate, slowly expanding to plane with a decurved margin, elevated in age, surface dry and unpolished, finely matted-fibrillose on disc under a lens, glabrous at the margin, cracked at times in age or dry weather, color dingy "cinnamon" to *vinaceous cinnamon” or “sayal brown", margin whitish at times, “russet vinaceous" over all when dry; flesh thick on dise, tapering abruptly to margin, whitish (near “cartridge buff"), firm, brittle, odor strong, sweet like that of cherries, taste mild. Lamellae decurrent (evenly at times and forming a collar on the stipe apex), distant, thickened, broad (3-11 mm.), occasionally forked, sides venose at times, arched, color “cream buff" to “ochraceous buff", edges even. Stipe 2-12 em, long, apex 0.8-4.2 cm. thick, equal in smaller car- pophores but tapering downward strongly in larger ones, solid (white 182 Rhodora [Vol. 68 within), no veil, surface dry and unpolished, velvety to pubescent under a lens, appressed in age, whitish. Spores 9-11.5(-13) X 6-7.5 4, elliptical, smooth, not amyloid; basidia 46-68 X 5.5-8 u, 4-spored; cystidia not differentiated; pileus tissue: cuticle pale yellowish in KOH, pigment dilute in the contents of some surface hyphae, cuticle an ixocutis, hyphae cylindrical, 2.5-4 4 in diameter, walls thin, tramal hyphae mostly cylindrical, 3-9 (-14) u in diameter, walls often slightly thickened, clamp connections present; gill trama: divergent in young carpophores but appearing more interwoven in large caps, hyphae cylindrical, 4-7.5 u in diameter. Gregarious to cespitose, In soil and needle beds under spruce. Material examined: Bigelow 13917, 13984, 14124, 14204, 14278, 14323, 14350, 14373, Lamoille and Washington counties, Vermont, September and October. In Hesler and Smith's (1963) monograph of Hygro- phorus, H. monticola is described from specimens collected in Idaho and eastern Canada. The pilei measured 2-5 cm. broad, and the stipes 3-6 em. X 3-25 mm. In view of these dimensions, the Vermont specimens are most impressive in size. The odor of cherries was very pronounced as well and could be detected without crushing the flesh, sometimes from several feet away. Of further note, and perhaps worthy of emphasis in diagnosis, was the tendency of the pileus to become more intense in color with drying. Hygrophorus nitiosus Blytt, Norges Hymenomycetes. Videnskab Selskabets Skrifter. Math. Naturv. Kl. 1904, No. 6. 1905. Gregarious. On soil of dirt road in hardwoods. Material examined: Bigelow 13132, Stowe, Lamoille Co., Vermont, July 28, 1964. Hygrophorus tahquamenonensis Smith & Hesler, Sydowia 8: 831. 1954. Gregarious to subcespitose, on moss by stream, near hardwoods. Material examined: Bigelow 13242, Bingham Falls, near Stowe, Lamoille Co., Vermont, August 3, 1964. Mycena atkinsoniana A. H. Smith, North American Species of Mycena, p. 144, 1947. On leaves and humus under beech. Material examined: Bigelow 13936, Bingham Falls, near Stowe, Lamoille Co., Vermont, September 9, 1964; 14059, near Lake Mans- field, Lamoille Co., September 14, 1964. 1966] Fungus Flora — Bigelow and Barr 183 Paxillus curtisii Berkeley, Ann. Mag. Nat. Hist. II, 12: 423. 1853. On hemlock logs. Material examined: Bigelow 11906, South Conway, Carroll Co., New Hampshire, July 22, 1963; 12092, same locality, August 3, 1963; 12398, 12399, same locality, September 4, 1963. Phaeocollybia kauffmanii (Smith) Singer, Rev. Mycol. 5: 11. 1940. Naucoria kauffmanii Smith, Mycologia 29: 52. 1937. Pileus 1.3-5.5 em. broad, obtuse-conie or convex at first with mar- gin slightly incurved and narrowly inrolled, not striate, expanding somewhat to broadly convex with margin remaining incurved and inrolled, sometimes lobed, surface glutinous or viscid, smooth, gla- brous, color watery “cinnamon” to “sayal brown", fading to “cinnamon buff", at times with a few dark reddish flecks; flesh thick, firm, whitish, odor and taste farinaceous. Lamellae emarginate (notch small and abrupt, superficially ap- pearing '"free"), close, narrow to medium broad (2-7 mm.), forked at times, white at first then “vinaceous buff" or “pale pinkish buff", edges slightly uneven. Stipe 4-13 cm. long, apex 5-11 mm. thick, often subventricose (up to 16 mm. thick), attenuated and radicating (about half of stipe above soil), at times twisted, usually curved below, surface innately fibrillose, moist but not viscid, solid (cortex thick and brown in sec- tion, interior firm, whitish to 'pale ochraceous buff"), color watery *cinnamon buff" above, rufous stained below, especially on lower por- tion embedded in soil, Spores 7.5-9 X 5.5-6 uw, ovate in face view, inequilateral in side view, with a minute apical pore but no conspicuous beak, rugulose- roughened, fulvous in KOH; basidia 23-31 X 7.5-10 u, 4-spored, pro- truding beyond the hymenium at maturity; cystidia: cheilocystidia abundant, pleurocystidia rare, both basidioid to clavate in shape, sometimes rather irregular and with broad lump near apex, rarely 2-celled, often with dilute brownish contents, smooth, thin-walled; pileus tissue: surface layer a distinct gelatinous pellicle in KOH, hyaline, hyphae 1.5-2.5 u in diameter, trama brown in KOH, pigment finely but distinctly encrusted on some hyphae in subpellicular zone, in smooth and thickened walls below, hyphae cylindrical to somewhat inflated, 4-13 u in diameter, clamp connections absent; gill trama parallel, brownish in KOH, pigment in thickened but smooth walls, hyphae cylindrical, 3-8 u in diameter. Seattered or subcespitose. On soil of river bank near hemlock. Material examined: Bigelow 13570, Bingham Falls, near Stowe, Vermont, August 20, 1964. P. kauffmanii is well known in the Pacific northwest 184 Rhodora [Vol. 68 although apparently not common. Hesler (1949) has re- ported its occurrence in Tennessee, but I have not found any other record from the east. My specimens from Ver- mont fit Smith’s (1957a) description except in dimensions of the carpophore. Western specimens of P. kauffmanii are considerably larger in size. Pleurotus elongatipes Peck, Jour. Myc. 14: 1. 1908. Pileus 2.5-14 cm. broad, convex to broadly convex or nearly plane, margin narrowly inrolled and incurved for some time, slightly pubes- cent when young, sometimes with short suleations, surface glabrous, mottled about the dise with small watery-appearing spots, white to pale avellaneous or pale buff ("cartridge buff", *light buff"); flesh usually rather thick on dise, firm, brittle, odor often Spicy and fra- grant when flesh crushed, taste not distinctive, Lamellae at first adnexed or adnate, becoming sinuate or emar- ginate, seceding at times, moderately broad (5-11 mm.) near stipe but narrowed at margin of pileus, usually close but sometimes crowded, not forked or intervenose, white or whitish, edges uneven and brittle in age, Stipe 3-10 em. long, apex 4-20 mm. thick, base usually somewhat enlarged and tapering upward, sometimes ventricose, solid (interior white and soft), cavernous in age, often curved, eccentric at times, surface with a thin fibrillose coating, becoming innately fibrillose- striate, apex pruinose to pubescent at times, white or nearly so. Spores 4-5.5 X 4-4.5 u, globose to subglobose, smooth, not amyloid, white to slightly cream tinged in deposit; basidia 16-24 X 4-6 u, 4-spored, carminophilous granules absent; cystidia not differentiated; pileus tissue: surface subgelatinous to gelatinous in KOH, cuticular hyphae cylindrical, 1.5-3 u in diameter, tramal hyphae cylindrical to slightly inflated, 2.5-13 u in diameter, clamp connections present, walls thin or slightly thickened and refractive; gill trama regular to subparallel, broad, hyphae mostly cylindrical, (1.5-)4-9(-11) 4 in diameter, wall thin or slightly thickened and refractive. Solitary or cespitose. On hardwood logs and stumps, sometimes in wounds of living maple. Material examined: Bigelow 8665, North Pond, Savoy Mt. State Forest, Florida, Massachusetts, September 22, 1959; 9362, Beartown State Forest, Monterey, September 27, 1960; 10509, Katahdin Stream Campground, Baxter State Park, Maine, August 6, 1962; 11146, near Norcross, Penobscot Co., August 18, 1962; 11259, Cedar Lake, Penob- scot Co., August 24, 1962; D. E. Fischer, Detroit, Michigan, October 5, 1905 (type, Nys); Bigelow 12524, Oliverian Brook Trail, White Mts. Nat. Forest, New Hampshire, September 7, 1963; 13842, near John- son, Lamoille Co., Vermont, September 2, 1964; 13937, Bingham 1966] Fungus Flora — Bigelow and Barr 185 Falls, near Stowe, Lamoille Co., September 9, 1964; 14048, Stowe, September 13, 1964; 14110, Ranch Brook Trail, Mt. Mansfield State Forest, September 16, 1964; 14116, Monroe State Park, Washington Co., September 17, 1964. Pleurotus elongatipes has proven to be quite common in New England. I have found it growing on birch and poplar debris as well as on dead or living maple. A description based upon specimens from Mt. Rainer, Washington, has been published by A. H. Smith (1957b). This species is a Pleurotus in the traditional sense of the genus, but the globose spores exclude it from modern inter- pretations. According to Singer's (1962) classification, the closest affinities may be with Hypsizygus, but some emenda- tion of this genus would be necessary to admit P. elonga- tipes. ASCOMYCETES Keissleriella caudata (Müller) Corbaz, Phytopath. Zeitschr. 28: 411. 1957. Figure 8. Material examined: Barr 2896, on cultivated rose stems, Conway, Massachusetts, May 7, 1961. K. caudata was described from Triticum stalks in Europe. The collection on rose stems agrees with the description of the species, although the ascospores are slightly longer, 15.5-17.5 X 3.5-4.5 y rather than 10-13 x 4-4.5 y. This species differs from K. cladophila (Niessl) Corbaz by smaller, thinner-walled ascostromata, lack of a clypeus, and narrower ascospores. Keissleriella is included in the Pleosporaceae. Lasiosphaeria chrysentera Carroll & Munk, Mycologia 56: 83. 1964. Material examined: Barr 3338, Abol Field, near Baxter State Park, Maine, July 6, 1962. The species, described from Costa Rica, is closely related to L. ovina (Pers. ex Fr.) Ces. & de Not. L. chrysentera differs in golden-yellow pigmentation of the hymenium, bright yellow globule at apex of ascus, and yellowish asco- spores. One additional collection has been examined from 186 Rhodora [Vol. 68 Idaho, A. W. Slipp, U.LF.P. Herb. 3834, on decorticated Alnus tenuifolia, 1943. Lasiosphaeria is a member of the Sordariaceae. Leiosphaerella falcata Barr, sp. nov., figures 1-4. Perithecia 180-300 yu diametro, 130-250 wu alta, immersa, gregaria, ostiola eccentrica vel lateralia, 60-120 u diametro, 70-150 & alta, per- iphysata; peridium 13-16.5 x crassum, brunneum, hyphis brunneis vestita. Asci 55-100 X 11-14.5 u, elliptica, unitunicata, ad apicem poro amyloideo, paraphysati. Ascosporae 46-86 X 3-5 4, fusiformae vel faleatae, hyalinae, uniseptatae, fasciculatae in ascis. Specimen typicum in caulis Aceris pensylvanicum L., a M.E. Barr lectus, n. 3460, prope “Katahdin Stream Campground, Baxter State Park, Maine, July 23, 1962", in herbario Univ. Mass. depositum. Perithecia 180-300 & in diameter, 130-250 u high, depressed globose, immersed beneath epidermis and raising it slightly, thickly scattered, ostiole eccentric to lateral, 60-120 u in diameter, 70-150 u high, pierc- ing the covering layer of substrate, canal periphysate, perithecial wall 13-16.5 & wide, of polygonal, brown, compressed layers of cells, inner layers hyaline, with brown hyphae surrounding perithecium and penetrating substrate tissues. Asci 55-100 X 11-14.5 à, elliptical, apex rounded or truncate, narrowed abruptly below spores into rounded base, wall single, thickened above with refractive pore, amy- loid in Melzer's, paraphyses deliquescent and forming an amorphous mass around asci. Ascospores 46-86 X 3-5 u, greenish hyaline, fusi- form to faleate, tapered to pointed ends, 1-septate in or near the middle, not constricted at the septum, wall thin, smooth, contents homogeneous or globular, in two overlapping fascicles in the ascus. In small dead branches of Acer pensylvanicum and A. spicatum. Material examined: Barr 3267, Norcross, Penobscot Co., Maine, June 30, 1962; 3460 (type), Katahdin Stream Campground, Baxter State Park, July 23, 1962; 3644, Norcross, August 18, 1962; 3995A, Stony Brook Trail, near Gorham, New Hampshire, July 26, 1963; 2098, Lake Munroe, Mt. Tremblant Park, Quebec, July 24, 1957; 4179, Underhill Camp Area, Mt. Mansfield State Forest, Vermont, June 30, 1964; 4334, Mt. Elmore State Park, July 16, 1964. L. falcata is closely related to L. praeclara. (Rehm) von Hóhnel, the type of the genus, and L. phoenicis (Chona & Munjal) Miiller & Ahmad in spore shape. The other mem- bers of the genus have shorter, relatively broad spores. The eccentric to lateral position of the beak of L. falcata and the longer spores distinguish it from the other species. This fungus shows affinities to members of Ceriospora but differs by lacking spore appendages and by simpler apical appara- 1966] Fungus Flora — Bigelow and Barr 187 Legend for Figures 1-8 Figs. 1-4. Leiosphaerella falcata Barr, sp. nov.: 1, perithecia be- neath bark, 2, ascus, 3, ascus showing amyloid apical ring, 4, asco- spores. Figs. 5-T. Mycosphaerella linnaeae Barr, sp. nov.: 5, asco- stroma, 6, ascus, 7, ascospores. Fig. 8. Keissleriella caudata (Müller) Corbaz: ascospores. Figs. 1, 5, X 80; figs. 2-4, 6-8, X 750. tus of the ascus. In the classification of Müller & von Arx (1962), Leiosphaerella is considered to be a member of the Amphisphaeriace2e. Mycosphaerella linnaeae Barr, sp. nov., figures 5-7. Ascostromata 80-100 & diametro, 72-100 & alta, immersa, gregaria, epiphylla; peridium 11-14.5 4 crassum, brunneum, hyphis brunneis. Asci 26-45 X 7.5-9 p, oblonga, bitunicata, aparaphysata. Ascosporae 16.5-21 X 2-3 y, fusiformae, hyalinae, uniseptatae, in duobus fasci- culatae in ascis. Specimen typicum in folia Linnaeae borealis L. var. americanae (Forbes) Rehd., a M.E. Barr lectus, n. 3831, prope “Millbrook Trail, Jefferson, White Mts. Nat. Forest, New Hampshire, July 6, 1963", in herbario Univ. Mass. depositum. Ascostromata 80-100 u in diameter, 72-110 » high, globose, slightly depressed or conical, immersed in leaf tissue, with broad blunt apex 188 Rhodora [Vol. 68 erumpent through upper epidermis, scattered thickly and evenly over most of leaf, wall thin, 11-14.5 u wide, two to three layers of dark brown, polygonal cells, compressed hyaline cell layers toward interior, brown hyphae copious in leaf tissues, Asci 36-45 X 7.5-9 u, oblong, apex rounded, base foot-like, wall double, thickened above, numerous in a fascicle, aparaphysate. Ascospores 16.5-21 X 2-3 u, greenish hyaline, fusiform, ends pointed, straight to curved, one-septate in mid- dle, not constricted at septum, wall thin and smooth, contents guttulate or with two globules in each cell, overlapping in two fascicles in the ascus. On overwintered leaves of Linnaea borealis var. americana. Material examined: Barr 3831 (type), Millbrook Trail, Jefferson, White Mts. Nat. Forest, New Hampshire, July 6, 1963; 1906, woods near chalet, Mt. Albert, Gaspe Prov. Park, Quebec, July 6, 1957. A species of Cercospora is intimately associated with M. linnaeae, and I suggest that it may form part of the life cycle of this species. This Cercospora has conidiophores short, brown, septate, erumpent through upper leaf epider- mis, often arising from apex of young ascostromata ; conidia 24-54 X 2 m, greenish hyaline, cylindrical, ends rounded, several-septate, contents guttulate. I have found no records of Cercospora on Linnaea (Chupp, 1953). At least twenty-four species of M ycosphaerella have been described on members of the Caprifoliaceae, but only two were reported on Linnaea. M. minor (Karst.) Johans. has smaller ascostromata, broadly saccate asci and clavate as- cospores 11-12 X 4 u. No measurements have been given for Sphaerella leightonii (Berk. & Br.) Cooke, but ascospore shape and septation is quite different from that of M. lin- naeae. As described in the literature, the other species of Mycosphaerella on members of the Caprifoliaceae do not appear similar to Barr 3831 and 1906. M. crepidophora (Mont.) Rehm and M. ebulina Petrak have elongate asco- spores and seem most closely related to M. linnaeae. M. crepidophora develops in spots in living leaves of Viburnum: the ascospores are cylindrical or oblong and are somewhat broader than those of M. linnaeae. M. ebulina occurs on dead leaves of Sambucus. The ascostromata are densely gregarious to form angular black spots, asci are larger, and ascospores broader than in M. linnaeae. The broad apex of 1966] Fungus Flora — Bigelow and Barr 189 the ascostromata and elongate, narrow fusoid ascospores are distinguishing characters of M. linnaeae. Plagiostigme petrakii Müller, Sydowia 18: 90. 1965. Plagiophiale petrakii (Müller) Petrak, Sydowia 18: 387, 1965. Material examined: Barr 4094, on overwintered leaves of Polygo- num viviparum, Alpine Gardens, Mt. Washington, New Hampshire, August 19, 1963. This rare and distinctive fungus has been illustrated and described recently by Müller (1965), who made several col- lections of it on Polygonum viviparum from the Swiss and Italian Alps. Petrak (1955) was first to collect P. petrakii, but thought it to be identical with Sphaerella eucarpa Karst. At that time he erected the genus Plagiophiale for the species. However, Sphaerella eucarpa Karst. is a species of Wettsteinina (Barr 1959, Müller and von Arx 1962), and therefore Plagiophiale Petrak is a nomenclatural synonym of Wettsteinina. Miiller (1965) has discussed in detail the generic position of the species on Polygonum. The North American material appears to be identical in all respects with that described from Europe. Plectosphaera salicis (Fckl.) v. Arx & Müller, Beitr. Kryptogamenfl. Schweiz 11(1): 204. 1954. Material examined: Barr 3928, Salix twigs, South Conway, New Hampshire, July 18, 1963. This appears to be the first record of P. salicis in North America. According to the literature, this pyrenomycete has been found previously only in Europe. My collection agrees with the description and figure in von Arx and Müller (1954). Podosphaera clandestina (Wallr. ex Fr.) Lév. Ann. Sci. Nat. Bot. III, 15: 86. 1851, var. clandestina Material examined: Barr 4578, Vaccinium angustifolium, top of Mt. Mansfield, Verment, August 28, 1964; 4583, V. uliginosum, data as above; 4585, Ledum groenlandicum, data as above, This minute powdery mildew has been recorded only rarely on members of the Ericaceae, chiefly species of Vac- cinium, and never, as far as I could determine, on Ledum. 190 Rhodora [Vol. 68 The specimens are in close accord with the description of P. oxyacanthae var. oxyacanthae in Salmon (1900). Thaxteria fusca (Fckl.) Booth, The Naturalist, London, p. 90. 1958. Sphaeria phaeostromoides Peck, N. Y. State Mus. Rep. 28: 77. 1876. Chaetosphaeria phaeostromoides (Peck) Sacce. Syll. Fung. 2: 93, 1883. Chaetosphaeria phaeostroma var. phaeostromoides (Peck) Ell, & Everh. North Amer. Pyreno., p. 160. 1892. Material examined: Barr 4041, on fallen beech, near Sawyer Rock Picnic Area, White Mts. Nat. Forest, New Hampshire, August 10, 1963; C. H. Peck, on decorticated Acer, North Greenbush, New York, September 1874 (type of Sphaeria phaeostromoides Peck, NYS). The collections cited above agree with Booth’s (1958) description and figures of T. fusca, and bear both conidial and perithecial stages. Ellis and Everhart (1892) cited a collection from Missouri, but otherwise the fungus appears to be collected rarely. The genus Thaxteria is a member of the Coronophoraceae. LITERATURE CITED ARX, A. J, VON and E. MÜLLER. 1954. Die Gattungen der amero- sporen Pyrenomyceten. Beitr. Kryptgamenfl. Schweiz 11(1): 1-434. Barr, M. E. 1959. Northern Pyrenomycetes I, Canadian Eastern Arctic. Contrib. Inst. Bot. Univ. Montreal 73: 1-101. BoorH, C. 1958. The genera Chaetosphaeria and Thaxteria in Bri- tain. The Naturalist, 83-90. CHUPP, C. 1953. A monograph of the fungus genus Cercospora. Ithaca, N. Y. 667 pp. ELLIis, J. B. and B. M. EVERHART. 1892. North American Pyreno- mycetes, Newfield, N. J. FAVRE, J. 1955. Les Champignons Supérieurs de la Zone Alpine du Pare National Suisse. Rés. tech, scient. entr. Pare Nat. Suisse 5: 1-212. Hester, L. R. 1949. Notes on Southern Appalachian Fungi, VIII. Jour. Tenn. Acad. Sci. 24: 81-93. HESLER, L. R. and A. H. SMITH. 1963. North American Species of Hygrophorus. Knoxville, Tenn. 416 pp. 1966] Fungus Flora — Bigelow and Barr 191 LANGE, M. 1955. Macromycetes Part II. Meddelelser om Grénland 147 (11) : 1-69. LANGE, M. 1957. Macromycetes Part III. Meddelelser om Grénland 148(2): 1-125. MÜLLER, E. 1965. Beobachtungen an Ascomyceten. Sydowia 18: 86-105. MÜLLER, E. and A. J. von ARX. 1962. Die Gattungen der didymo- sporen Pyrenomyceten. Beitr. Krypt. Fl. Schweiz, 11(2) : 1-922. PETRAK, F. 1955. Plagiophiale nov. gen., eine neue Gattung der sphaerialen Ascomyceten. Sydowia 9: 585-587. SALMON, E. S. 1900. A monograph of the Erysiphaceae. Torrey Bot. Club Mem. 9: 1-292. SMiTH, A. H. 1957a. A contribution toward a monograph of Phaeo- collybia, Brittonia 9: 196-217. 1957b. Additional new or unusual North American agaries, Sydowia, Beih. 1 (Festschr. f. Franz Petrak): 46-61. SINGER, R. 1962. The Agaricales in Modern Taxonomy, 2nd ed. Weinheim. 915 pp. NOTES ON RAFINESQUE'S SPECIES OF LECHEA (CISTACEAE) ROBERT L. WILBUR" It was suggested in an earlier paper (Wilbur & Daoud, 1961) that some of the twenty species of Lechea published by Rafinesque (1836) would eventually prove identifiable and that their names would replace binomials then cur- rent. This may have seemed an unwarranted assumption since three careful students of this technical genus had diligently studied Rafinesque’s monograph and concluded that the descriptions were not adequate to identify the species intended with any degree of confidence. Leggett (1878) republished much of Rafinesque’s monograph and suggested the possible identity of several of the taxa. Brit- ton (1894) indicated that he had worked closely with Leg- gett in attempting to unravel the problems of the genus and together they had “pored over Rafinesque’s mono- graph”. They finally concluded that “unless type specimens could be found, there could be no certainty” in applying any of the names. Hodgdon (1938:30) suggested that Rafinesque’s “perplexing” monograph left the genus in a “deplorable state" and claimed to have been able to identify only four of Rafinesque’s twenty new binomials. I believe however that the majority of Rafinesque’s species of Lechea can be identified with considerable cer- tainty even in the absence of types by comparing his descrip- tions with the species known to occur in the areas mentioned in the original publication. The results of this analysis are presented in the following paragraphs presented in the same order as Rafinesque’s monograph (1836) and followed chronologically by his other papers dealing with the genus. Unfortunately, for stability’s sake, three of Rafinesque’s ‘Grateful acknowledgement is made to the National Research Foun- dation for a grant of research funds to Duke University (NSF- G18799) which were used in part for this study. 192 1966] Lechea — Wilbur 193 binomials must replace those of species long known by later names. 1. Lechea pulchella Raf., New Fl. N. Am. 1: 91. 1836. Rafinesque definitely mentioned L. pulchella only from the pine barrens of New Jersey although suggesting that its range probably extended to the south. The species was described as being "quite smooth" and, as a member of Rafinesque's subgenus Menandra, it can be assumed that its outer sepals were noticeably shorter than the inner, à fea- ture which helps to narrow the list of possible species to four. Of the seven species of Lechea known from New Jersey, only L. maritima Legg., L. intermedia Legg. ex Britt., L. leggettii Britt. & Holl. and usually L. racemulosa Michx. have external sepals exceeded by the inner. Lechea maritima may be safely discounted as a suspect as it could never be described as “quite smooth” even upon the most superficial observation and its cauline leaves tend to be whorled and not “scattered.” Rafinesque described the leaves of L. pulchella as “scattered long linear" and its capsules as *obovate." These features make it possible to exclude L. racemulosa from consideration for its principal or lower and median cauline leaves are commonly opposite or even whorled and elliptic to lanceolate or even oblanceo- late and perhaps more significantly it possesses a barrel- shaped or cylindrical capsule. Rafinesque’s subgenus Menandra was described as having seeds 1-3 in number which agrees with the number characteristic of L. leggettü but not with the 4-6 seeds characteristic of L. intermedia. Furthermore L. leggettii is known from the New Jersey pine barrens while L. intermedia apparently is not. Leggett (1878: 247) questioningly suggested that Rafinesque's L. pulchella was the same as the species then called “L. minor Lam.” which is what has recently been called L. leggetti. Leggett (1878: 250) twice encountered in the pine barrens *an abnormally smooth Lechea" which pretty well matched Rafinesque's description, These smooth plants were what is now called L. leggettii. Accompanying the original publica-. 194 Rhodora [Vol. 68 tion of L. leggettii the name var. pulchella was appended with L. pulchella Raf. questioningly listed in synonymy indicating Britton’s earlier similar suspicions. The evidence seems sufficiently convincing that I do not hesitate to take up Rafinesque’s binomial for the species most recently known as L. leggettii, New combinations for the varieties recognized by Hodgdon are not made for I am not convinced that these tendencies represent biological varieties or subspecies. Rafinesque (1836: 91) named three varieties based upon the luxuriousness of the plants and the Shape of the inflorescence. These varieties were var. minor Raf., var. elegans Raf. and var. pyramidalis Raf. 2. Lechea cinerea Raf., New Fl. N. Am. 1: 91. 1836. Leggett (1878: 250) felt that Rafinesque's description was a "pretty correct" one for the species he had previously distributed as L. torreyi. He noted however that there were 3-seeded and 6-seeded forms of the species and hoped that his proposed L. torreyi would prove distinct. Hodgdon (1938: 106-107) discussed Rafinesque's description and concluded that it was apparently based upon two or more species. The original description includes nothing that would suggest to me that Rafinesque had both L. torreyi and L. patula as seemed likely to Hodgdon but the point is academic. Rafinesque parenthetically placed L. thymifolia Michx. in synonymy of his own L. cinerea and again in closing mentioned the possibility that it was “the real thymifolia.” Rafinesque’s binomial hence is illegitimate as nomenclaturally superfluous when published according to the International Code (Article 63). The fact that Michaux’s binomial is actually a synonym of L. minor L. does not alter the fact that Rafinesque first placed Michaux’s name unequivocally in the synonymy of his own species. 3. Lechea ternifolia Raf., New Fl. N. Am. 1: 91. 1836. Rafinesque stated that his L. ternifolia from “Virginia, New Jersey and probably elsewhere” “must be the real 1966] Lechea — Wilbur 195 L. racemulosa” of Michaux. Both Rafinesque and Michaux quote the same Gronovian polynomial. And Rafinesque's description itself (“stem . . . adpressed pubescent, leaves and branches mostly ternate cuneate acute ciliate nearly smooth; racemes paniculate lax nearly naked, peduncles elongate, capsules oblong’’) indicates that Lechea ternifolia Raf. can be placed with confidence in the synonymy of L. racemulosa Michx. Leggett (1878: 247) with some ques- tion and Grosser (1894: 135) unequivocally placed Rafi- nesque’s binomial in the synonymy of Michaux’s species while Hodgdon (1938) did not mention it. 4. Lechea furfuracea Raf., New Fl. N. Am, 1: 92. 1836. Leggett stated (1878: 250) that he had “never seen any- thing to which this description would apply.” No one else has been able to suggest what species Rafinesque had from “Kentucky and Illinois” with “mealy pubescence and globose capsules,” very short external calyx-lobes, and rounded internal and “narrow linear scattered nearly obtuse" leaves. According to the most authoritative listings of Lechea in Illinois (Jones & Fuller, 1955) and in Kentucky (Braun, 1943) there are in fact no species known to be common to both states possessing external calyx-lobes shorter than the inner. Lechea racemulosa Michx. is the only species from Kentucky whose calyx approximates Rafinesque’s descrip- tion and its capsule and foliage would definitely exclude it as a possible synonym of L. furfuracea. The three species from Illinois with appreciably shorter external than inner sepals are L. stricta, Legg., L. intermedia Legg. and L. leg- gettii Britt. & Holl. Rafinesque's description is not detailed enough to permit me to determine which species he had. 5. Lechea laxiflora Raf., New Fl. N. Am. 1: 92. 1856. Rafinesque's description seems certainly to apply to the species long known as L. leggettii. Leggett (1878: 247) thought it probably was “L. minor Lam." which was the name by which L. leggettii was known in the later nine- 196 Rhodora [Vol. 68 teenth century. Rafinesque found this species in “the New Jersey Pine Woods with L. pulchella” to which he felt it closely related. The more significant points in Rafinesque's description are as follows: “smooth ... branches lax, leaves scattered linear cuneate acuminate ciliate; racemes scattered lax, flowers remote naked, pedicels elongate, calix and capsules ovate.” I recommend (see No. 1 above) that L. pulchella be adopted as the earliest name for this species with L. laxiflora placed in its synonymy. Rafinesque also distinguished a var. brevifolia with reddish rather than greenish flowers and shorter, more cuneate, less ciliate leaves. Rafinesque indicated that the principal difference between this species and his L. pulchella was chiefly in the leaves which in L. pulchella were “scattered long linear acute” and by inference eciliate in contrast to the acuminate, ciliate leaves of L. laxiflora. 6. Lechea tenuifolia Michx., Fl. Bor.-Am. 1: 77. 1803. Rafinesque’s account is largely a translation of Michaux’s original description. It is placed by Rafinesque with those species with short external sepals while the species to which we apply the name has external calyx lobes equaling or more commonly longer than the inner. Rafinesque cited no locali- ties other than the Santee River which had been cited in Michaeux’s original account so it is not certain whether Rafinesque had actually seen specimens of this species in spite of his statement that he had “specimens of all of the described species.” Perhaps Rafinesque placed it with the species of Lechea with short calyx-lobes since Elliott (Sk. Bot. S. C. & Ga. 1: 185. 1816.) implied that this species had no external calyx and this statement was included by Rafinesque. Elliott in turn had been misled by Walter (FI. Car. 83. 1788) who reported there was no outer calyx in his L. juncifolia, a species which has never been satisfacto- rily identified. 6[a]. Lechea verna Raf., New Fl. N. Am. 1: 92. 1836. This binomial was published as a substitute name (*6. L. 1966] Lechea — Wilbur 197 tenuifolia Mx. Elliot or L. verna Raf, L. juncifolia Walter.” ) and is therefore illegitimate. 7. Lechea recurvata Raf. New Fl. N. Am. 1: 93. 1856. This species was placed by Rafinesque in his subgenus Menandra possessing short external calyx-lobes. It was said to possess globose capsules and to occur "in Carolina and Virginia." I believe that this is the species referred to in the most recent monographs of Britton (1894) and Hodgdon (1938) as L. leggettii. If my interpretations of Rafinesque's species are correct, Lechea pulchella, L. laxiflora and L. recurvata are all prior names for the species now called L. leggettii. There are only two Lecheas common to both Virginia and “Carolina” possessing external calyx-lobes shorter than the inner; these two are L. racemulosa and L. leggettii. The described globose capsules of L. recurvata would seem to exclude L. racemulosa as a possibility al- though, judging by the following features mentioned *branches opposite and ternate, leaves broad oblong acute at both ends ciliate," it might seem a better match for Rafinesque's description than L. leggettii. 8. Lechea mucronata Raf., Précis des Découvertes Somiol. 37. 1814, To my knowledge only House (New York St. Mus. Bull. 254: 497. 1924) has taken up this name although Robinson (Syn. Fl. N. Am. 1: 192. 1895) and Grosser (1903: 135) both placed it unquestioningly in the synonymy of L. major in the sense of Michaux but not of Linnaeus. Lechea major sensu Michaux, Robinson and Grosser is synonymous with L. villosa Ell. Hodgdon, the most recent monographer of the genus, concluded that there was not enough evidence presented in Rafinesque’s original description to warrant the displacement of Elliott’s L. villosa, a “well-characterized and clearly typified name.” Rafinesque's original description is presented below in full so that others may form their own conclusions independent of my own interpretation. I am 198 Rhodora [Vol. 68 indebted to Mrs. Lazella Schwarten, Librarian of the Har- vard Herbaria, for making a copy of it available. 120. Lechea mucronata. Poilue, tige droite et simple, feuilles oblongues-cunéiformes mucronées, fleurs en grappe composée, bractéolées, bractées oblon- gues aigues. Dans les bois du N. Jersey. Hodgdon stated (1938: 56) that the "only possible diag- nostic features of Rafinesque's description are “feuilles oblongues-cunéiformes mucronées, fleur en grappe com- posée." He concluded that the statement concerning the arrangement of tne flowers “might apply to most members of the genus" and that the described leaf shape would apply "equally well to L. minor as to L. villosa." The description's most definitive word in my opinion is “poilue,” and in addi- tion the habitat and geographical location certainly con- siderably narrow the number of possibilities. "Poilue" meaning “shaggy, hairy or pilose” best describes the species recently called L. villosa. The features men- tioned in the brief description single out the species long incorrectly known as L. major and more recently known as L. villosa. There are seven species of Lechea known from New Jersey: L. minor, L. tenuifolia, L. racemulosa, L. mari- tima, L. leggettii, L. intermedia and L. villosa. Of these seven the last mentioned alone seems indicated by Rafines- que’s description of 1814. Lechea maritima, L. leggettii and L. intermedia all possess external sepals conspicuously shorter than the internal which would exclude them from consideration. Of the seven species of Lechea mentioned above found in New Jersey the term "poilue" would seem to me to exclude all but L. villosa and perhaps L. maritima. The description of the leaves, although not definitive, would seem to exclude L. maritima but include L. villosa whose leaves are often mucronate. It is true that the leaves of L. minor sometimes would fit Rafinesque’s description as indicated by Hodgdon but certainly “poilue” would scarcely be used to describe the subappressed pubescence of that species. Rafinesque (1836: 93) more fully described his L. mucro- 1966] Lechea — Wilbur 199 nata indicating it to be the same as the L. villosa of Elliott from whom he had received specimens. Included are state- ments puzzling even to a casual student of the group (e.g. "capsules ovate villose") but sufficient, I believe, to convince one that Rafinesque was indeed describing, although perhaps poorly, the species called L. villosa by Elliott. In passing it may be noted that the capsules of all species of Lechea are glabrous. There actually seems to be nothing other than prejudice why we should not take up Rafinesque's name for this species since it has two years priority over Elliott’s L. villosa. Rafinesque (1836: 93) briefly described three varie- ties all of which came from the extensive area of Hodgdon's var. typica. The three varieties were var. simplex, var. ramosa and var. sessiliflora. New combinations have not been made for Hodgdon's var. macrotheca ranging from central Nebraska south into Texas or his var. schaffneri from northeastern Mexico since I have not seen enough material to form an opinion as to the merits of these pro- posed taxa. 9. Lechea heterophylla Raf., New Fl. N. Am. 1: 93. 1836. Rafinesque described this species from Kentucky and Illinois as having an appressed pubescent stem and acute, carinate inner sepals about equaling the external. Neither Leggett (1878: 248) nor Grosser (1903: 140) would hazard a guess as to what it might be. The three Species common to Kentucky and Illinois possessing external sepals at least occasionally equaling the inner are Lechea villosa Ell., L. tenuifolia Michx. and L. minor L. The shaggy pubescence of the stem of L. villosa would exclude it from further consideration. The leaves were described as “commonly ternate, lower obovate, subsessile, upper cuneate and linear petiolate acute smooth.” I þelieve this a better approxima- tion of the greater variation shown between the lower and upper leaves of L. minor than the comparatively uniform condition exhibited by L. tenuifolia. The acute inner sepals on the other hand are a better fit for the “subacute” condi- 200 Rhodora [Vol. 68 tion of L. tenuifolia than the “obtuse” inner sepals of L. minor. I suspect Rafinesque was describing L. minor since the conspicuous difference in the foliage was less subject to error by Rafinesque than the shape of the apex of the inner sepals. Rafinesque placed “L. minor Smith” in the Synonymy of his L. heterophylla but his description would exclude the villous species that was apparently considered L. minor by J. E. Smith. Britton (1895: 248) has demon- strated that Smith’s L. minor was the species which Elliott called L. villosa. Rafinesque published the following three varieties of this species: var. major Raf., op. cit. 94. 1836, var. minor Raf., l.c. and var. parviflora Raf., Lc. and only the last of these was even briefly described. 10. Lechea glomerata Raf., New Fl. N. Am. 1: 94. 1836. Rafinesque's description of this species from the " Apla- chian mts." strongly points to L. mucronata Raf. (= L. vil- losa) except for his characterization of it as “smooth.” The carinate sepals, short pedicellate, glomerate flowers, linear external sepals about equaling the internal, the trig- onous calyx surrounding the subglobose capsule and the mucronate, petiolate leaves are all indicative of L. villosa but that species is apparently always densely villous to spreading pilose. Leggett (1878), Grosser (1903) and Hodgdon (1938) have all found it impossible to identify Rafinesque’s description and I claim no greater success. ll. Lechea corymbosa Raf., New Fl. N. Am. 1: 94. 1836. The description of this species from the “Alleghany” Mountains, belonging to Rafinesque's subgenus Lechea, a group characterized by the external and internal sepals equaling one another, is most suggestive of L. villosa. EL., a later synonym of L. mucronata Raf. (see No. 8 above). The original description of L. corybosa (“stem . . . pilose . , leaves petiolate broad oblong nearly obtuse, . . flowers corymbose, pedicels equal to flowers, sepals round concave, ext. linear, capsules globose .. . 6 to 12 inches high 1966] Lechea — Wilbur 201 ... capsules large with 3 to 6 seeds... ") is not especially definitive but I believe it is most suggestive of L. mucronata of the species occurring in the Alleghenies. 12. Lechea surculosa Raf., New Fl. N. Am. 1: 94. 1856. Leggett (1878: 248) questioningly suggested that this description might be Lechea minor [in the sense of] Lam. which is the species most recently called L. leggettii Britt. & Hollick. This seems an unlikely identification since L. leggettii is characterized by its external sepals being pro- nouncedly shorter than the inner while L. surculosa, as a member of Rafinesque's subgenus Lechea, would be expected to have the inner and outer sepals of approximately equal length. It is certainly true that one can not place the species of Lechea into three groups depending on whether the ex- ternal sepals are shorter, equaling, or longer than the in- ternal since there is too much variation in the relative length of the sepals in several species but the external sepals of L. leggettii are always clearly appreciably shorter than the inner. Rafinesque's species was described as possessing “acute” inner sepals while those of L. leggettii are clearly obtuse. There are other discrepancies, but the two men- tioned are sufficient to discount completely Leggett's sug- gestion. Rafinesque knew this plant from *Pennsylvania on dry hills" and indicated it was the L. minor of some authors. I believe the description of L. surculosa best matches the characteristics of L. racemulosa, a species whose external sepals range from obviously shorter than the inner sepals to equaling them. The inner sepals of L. racemulosa, accord- ing to Hodgdon's description, vary from “subacute to obtuse" while Rafinesque described those of L. surculosa as “acute.” The capsule of L. racemulosa was described in Hodgdon's monograph as “slenderly ellipsoid to slenderly ovoid" while Rafinesque said those of L. surculosa were “ovate.” Rafinesque described the sterile stems of L. surcu- losa as “pilose with leaves ternate ovatoblong acute ciliate" 202 Rhodora [Vol. 68 which is a reasonably close approximation of Hodgdon’s description of the basal shoot of L. racemulosa: ‘basal leaves frequently verticillate, . . . oblong-ovate to elliptic- lanceolate, acute to slightly mucronulate . . . prominately Spreading-pilose to villous on the margins." The other species of Lechea from Pennsylvania with external lobes approximately equaling the inner at least occasionally (L. villosa, L. tenuifolia and L. minor) have characters which compare unfavorably with those delineated by Rafinesque. The “smooth” branches and “linear” leaves of Rafinesque’s L. surculosa are sufficient to discount it as a possible syno- nym of L. villosa (= L. mucronata). The basal leaves of L. tenuifolia were described by Hodgdon (1938: 87) as “narrowly lanceolate or oblanceolate to linear” which cer- tainly implies a much narrower leaf than does Rafinesque’s description and the cauline leaves of L. tenuifolia are described by Hodgdon as “scattered” while Rafinesque described those of L. surculosa as "opposite and alternate." I believe then that L. surculosa Raf. is an ecologically modi- fied form of L. racemulosa Michx. and that it should be placed in its synonymy. This binomial first appeared three years earlier in Rafi- nesque's Herb. Raf. 68. 1833 but without description. 18. Lechea revoluta Raf., New Fl. N. Am. 1: 94. 1836. Rafinesque stated that this species is the Lechea “minor of Lin. not of Smith nor other Authors” and that the ‘‘de- scription is taken from Linnaeus altogether, and agrees very well with some specimens [that Rafinesque had] from the Alleghenies of Pennsylvania." Nomenclaturally the bi- nomial is a synonym of Lechea minor L. Rafinesque pro- vided a new name since J. E. Smith had indicated that Linnaeus had confounded several species within his L. minor, 14. Lechea virgata Raf., New Fl. N. Am. 1: 95. 1836. This Alleghenian species was described as possessing’ 1966] Lechea — Wilbur 203 pubescent stems, opposite, whorled or even scattered, petio- late, linear, cuneate, ciliate, acuminate leaves with acute, carinate inner sepals, pedicels equaling the flower-length and ovate capsules. As a member of subgenus Lechea, it can be safely assumed that its external sepals equaled the inner. There are four species within the area of the Alle- gheny Mountains whose external sepals may commonly or occasionally equal the internal ones. These four species are Lechea minor L., L. racemulosa Michx., L. tenuifolia Michx. and what has been most recently referred as L. villosa Ell. The description is an unusually full one for Rafinesque and the above mentioned characters seems best to describe L. tenuifolia Michx. of the four species listed above from the Alleghenies whose external sepals commonly or occasionally equal the inner. The linear leaves, which are the only type mentioned for L. virgata, are sufficient to exclude L. villosa and L. minor from further consideration. The inner sepals of L. racemulosa are non-carinate while those of L. virgata were described as carinate. Two varieties were appended to this species with brief description: var. breviflora and var. bracteata. 15. Lechea floridana Raf., New Fl. N. Am. 1: 95. 1836. There are eight species of Lechea known from Florida (Hodgdon, 1938; Wilbur & Daoud, 1961). Five of these (L. deckertii Small, L. leggettii Britt. & Holl., a later synonym of L. pulchella Raf. as is pointed out in no. 1 above, L. tor- reyi Legg. ex Britt., L. divaricata Shuttlew. ex Britt. and L. cernua Small) possess external sepals conspicuously shorter than the internal. The three others (L. villosa Ell., a later synonym of L. mucronata Raf. as is discussed in no. 8 above, L. minor L. and L. patula Legg.) all possess external sepals conspicuously shorter than the internal, as was described by Rafinesque for his L. floridana. None of these three species is “smooth” as mentioned by Rafinesque but certainly the copious, spreading villosity of L. mucro- nata, (— L. villosa) together with its oblanceolate leaves ex- cludes it from consideration. The appressed pubescence of 204 Rhodora [Vol. 68 L. patula and L. minor could easily have been overlooked, especially if the specimens were aged, by an observer as hasty and superficial as Rafinesque often gives evidence as having been. The leaves of L. floridana were said to be "scattered, minute linear" which is more suggestive of L. patula than of L. minor. The external calyx of L. floridana was implied to equal the inner which is sometimes true of both L. patula and L. minor. The external sepals of L. patula most frequently are a little shorter than the inner but range up to slightly longer while those of L. minor rarely are as short as the inner and most frequently are conspicuously longer. The capsules of L. floridana were said to be ovate which is not a very apt description for the capsule of either species but, interpreted as “ovoid,” not impossibly bad for either. I suspect this to be a form of L. patula Leggett but due to the discrepancies mentioned would not feel its identity completely certain. 16. Lechea secundiflora Raf., New Fl. N. Am. 1: 95. 1836. This species came from “the glades of West Kentucky" making it impossible that its binomial could ever displace that of any of the four species definitely known from that state since each has an older name. The known Species are L. villosa Ell., a later synonym of L. mucronata Raf., L. minor L., L. tenuifolia Michx. and L. racemulosa Michx. Rafinesque’s description fits none well and there are men- tioned features that would exclude each of the geographical possibilities. Grosser (1903: 135) placed Rafinesque’s name in the synonymy of L. minor but his description of “leaves scattered, lax narrow linear, . . . flowers remote secund" would, as pointed out by Hodgdon (1938: 60), make that an unlikely choice. Hodgdon (1938: 88) placed Rafinesque's binomial in the synonymy of L. tenuifolia Michx. although the description of the capsules as "oblong" and the inner calyx as "lanceolate" causes considerable doubt as does Rafinesque's description of the plant as “smooth.” I am unable to identify the binomial and believe the mentioned discrepancies are of such magnitude as to prevent one from 1966] Lechea — Wilbur 205 assigning it with confidence to the synonymy of any species known from Kentucky. 17. Lechea pauciflora Raf., New Fl. N. Am. 1: 96. 1836. The identity of the plants described by Rafinesque under this binomial from “Near the Sea Shore in Long Island and New Jersey" was questioningly suggested by Leggett (1878: 250) to be L. thymifolia in the sense of Pursh which is L. maritima, Legg., à species whose external sepals are much shorter than the heavily appressed-pilose internal calyx. Rafinesque placed L. pauciflora in his subgenus Eudiexa, characterized by outer sepals exceeding the inner, and in the diagnosis of the species specifically stated “sepals smooth, external a little longer.” Grosser (1903: 135) assigned Rafinesque’s binomial to the synonomy of L. minor L., a disposition to which Hodgdon (1938: 60) objected, feeling the described 3-inch specimens to be an “ecological variant of some other species” whose identity was undecipherable because of the incompleteness of the original publication. The only two species apparently known from both Long Island and New Jersey with outer sepals longer than the inner and “adpressed pubescent” stems are L. tenuifolia Michx. and L. minor L. Rafinesque described the capsules of his L. pauciflora as “globose” which among other features mentioned in the description would eliminate L. minor. The leaves of Rafinesque’s species were described as “rather obtuse” which together with the “smooth” calyx makes it difficult to place this name in the synonymy of L. tenuifolia with which the description agrees in most other respects. Therefore I am unable to suggest the identity of this bi- nomial. 18. Lechea brevifolia Raf., New Fl. N. Am. 1: 96. 1836. Leggett (1879: 249) was "quite at a loss" as to what species Rafinesque's description might apply while Grosser (1903: 135) placed the binomial in the synonymy of L. minor L. without discussion, Hodgdon (1938: 60), while 206 Rhodora [Vol. 68 admitting that much of Rafinesque’s description applied to L. minor, felt that the diagnosis of “flowers secund, pedicels equal” indicated confusion on the author’s part. Hodgdon’s objection on these grounds are difficult to understand for the flowers and fruit are sometimes secund as his own description of L. minor L. admits (p. 59 “fruit mostly clustered, at times scattered or even secund") and the 1-2 mm. long pedicels are often equal to one another or to the flowers. The species came from the “Mts. Apalaches" (Ap- palachian Mts.) and I believe it most probable that this binomial belongs in the synonomy of L, minor. 19. Lechea uniflora Raf., New Fl. N. Am. 1: 96. 1836. Leggett (1878: 249) was unable to suggest what species Rafinesque was describing based upon 3-4 inch specimens from the mountains of Maryland possessing few leaves or flowers and longer outer sepals than inner and with a 6- seeded, globose capsule. Grosser (1903: 135) placed the binomial in the synonymy of L. minor L. but Hodgdon (1938: 61) disagreed with this placement concluding that it was “hopeless to try to identify this particular one of Rafinesque’s Lecheas.” I am in total agreement with Hodg- don for I too am unable to suggest the identity of L. uniflora. There are too many features mentioned in its original des- cription which are at complete variance with the character- istics of L. minor for us to accept Grosser’s identification. To my knowledge a “smooth” specimen of L. minor with globose 6-seeded capsules has not been reported. 20. Lechea stellata Raf., New Fl. N. Am. 1: 96. 1836 I am in complete agreement with Leggett’s decision (1878: 251) that this description is "undoubtedly the L. thymifolia" of Michaux which is a synonym of L. minor L. The more diagnostie portions of Rafinesque's description seem to be as follows: “Pilose . . leaves 3-4 nate petiolate elliptic mucronulate, upper leaves alternate lanceolate; racemes foliose, . . . ext. sepals very long, capsules ovate- 1966] Lechea — Wilbur 207 oblong.” Both Grosser (1903: 135) and Hodgdon (1938: 61) considered L. stellata to be synonymous with L. minor, although the latter fails to place it formally in the synonymy of that species. 21. Lechea sessiliflora Raf., New Fl. N. Am. 1: 97. 1836. Leggett (1878: 251) thought that the description ac- companying this binomial might “possibly be L. patula" but no species known to him had sessile mature flowers. In spite of this, Index Kewensis equated Rafinesque's binomial with L. patula Legg. and Grosser (1903: 138) unquestion- ingly placed it in the synonymy of that species. Hodgdon (1938: 63) felt that Rafinesque's description could not be positively identified but that its sessile flowers and 3-9- seeded capsules obviously excluded the pedicellate, 1(2)- seeded L. patula as a possibility. However, since no Lechea is actually sessile, it would perhaps be only reasonable to interpret Rafinesque's description as “subsessile” and con- sider short-pediceled species as possibilities. Also, since no known species of Lechea possesses more than 6 seeds, we had best treat the 3-9-seeded condition attributed by Rafines- que to his subgenus Hudiera as an obvious error. It is well to note that Rafinesque did not mention seed-number in his diagnosis of L. sessiliflora ; seed number was stated only in conjunction with the characterization of the subgenus. There are five species known to be common to both Florida and Alabama: Lechea villosa Ell., L. minor L., L. patula Legg., L. leggettii Britt. and L. torreyi Legg. ex Britt. The two last-mentioned of these can be discounted as possibilities since their external sepals are considerably shorter than the internal and the reverse is clearly described as the state of L. sessiliflora. Lechea villosa can be eliminated from con- sideration since its shaggy-villous stems could never be described as “adpressed pilose” while both L. minor and L. patula both fit that characteristic. Rafinesque's character- ization of the species “branches diffuse virgate, leaves scat- tered ... linear... ext. sepals linear not much longer..." all point most clearly to L. patula rather than to L. minor. 208 Rhodora [Vol. 68 I believe that Leggett's suspicion as to the identity of L. sessiliflora has been fully confirmed. There is hence no reason not to take up L. sessiliflora Raf. which has forty-two years priority over Leggett's L. patula. The remaining names proposed by Rafinesque are listed below: Lechea linifolia Raf., Atl. Jour. 1: 18. 1832. nom nud. Lechea paucifolia Raf., Atl. Jour. 1: 18. 1832. nom. nud. Lechea mexicana Raf., Sylva Tell. 133. 1838. Rafinesque provided this as a substitute name for Heli- anthemum tripetalum of Mexico no doubt because tripetalum was a pointless epithet in a genus characterized by three petals. The name is nomenclaturally superfluous and there- fore illegitimate (Art. 63). DEPARTMENT OF BOTANY, DUKE UNIVERSITY DURHAM, NORTH CAROLINA LITERATURE CITED BRAUN, E. Lucy. 1943. Lechea in An Annotated Catalog of Sperma- tophytes of Kentucky. p. 93. BRITTON, N. L. 1894. A Revision of the Genus Lechea. Bull, Torrey Club 21: 244-253. GROSSER, W. 1903. Lechea in Engler's Das Pflanzenreich 14 (IV, 193) : 133-140. HopGDoN, A. R. 1938. A Taxonomic Study of Lechea. Rhodora 40: 29-69, 87-131. (Reprinted as Contrib. Gray Herb. 121.) JONES, G. N. & G. D. FULLER. 1955. Lechea in Vascular Plants of Illinois. pp. 328-329. LEGGETT, W. H. 1878. Rafinesque's Lechea. Bull, Torrey Club 6: 246-252. RAFINESQUE, C. S. 1836. Monograph of Lechea, New Flora and Bot. of N. Am. pp. 89-98. WILBUR, R. L. & H. S. Daoup, 1961. The Genus Lechea (Cistaceae) in the Southeastern United States. Rhodora 63: 103-118. GENTIANA PUBERULENTA SP. NOV., A KNOWN BUT UNNAMED SPECIES OF THE NORTH AMERICAN PRAIRIES: JAMES S. PRINGLE The species which for the past century has been known as Gentiana puberula is one of the showiest of the North American gentians; Curtis (1959) called it “by all odds the most beautiful member of this famed genus in Wis- consin," and described its flowers as “unrivalled in the clarity and intensity of their deep blue color." It is a characteristic species of mesic and xeric prairie commu- nities from Ohio west to Manitoba and Arkansas, and has been collected as far east as western New York and Mary- land, as far south as Louisiana, and as far west as the Black Hills of South Dakota. Some authors have designated this species the “Prairie Gentian” ; others have used the translation “Downy Gentian”, although the puberulence is so minute that it is not a conspicuous feature to the casual observer. Gentiana puberula Michx. Fl. Bor.-Am. 1:76. 18083 was based on a specimen Michaux had collected at “Fort Chero- quis" (Fort Massac) in present-day Massac County, Illinois, where he had spent part of the autumn of 1795. This speci- men is preserved in the herbarium of the Laboratoire de Phanérogamie, Muséum National d'Histoire Naturelle, Paris, France. A photographic negative of this specimen was made available through the kindness of Prof. A. Aubré- ville and J. Leandri. This negative and enlarged prints have been placed in HAM. Duplicate prints have been deposited Tn addition to those persons whose contributions are noted in the text, I should like to express my appreciation to Dr. A. J. Sharp of the University of Tennessee and Dr. J. M. Gillett and Mr. W. J. Cody of the Plant Research Institute, Canada Department of Agriculture, for their reviews of the manuscript, and to Dr. A. H. Moser of the University of Tennessee, for his checking the Latin description. Contribution No. 5 from the Royal Botanical Gardens, Hamilton, Ontario, Canada. 209 [Vol. 68 Rhodora 210 —— —————^P—————É—————————————" 000 7 7 E '6'0 €» X ‘dA, 10g pojoo[os Ways uo doAO[ "p ‘DIA '&'0 BO X 'spoqe[ yya ‘pys umnrrieq -19u uo uaurreds ejo[duio)) *g ‘DIA Bung Djuojm4oqnd ^r) Jo uauitoods odÁ] 'p pu? g “SDIA ‘6'0 BOX ‘SLOMO[ ST Z 7014 gO €» X ‘Sfeqe, pue ueurroedg '[ D ‘XII Djm4oqnd Dupiuor) Jo sueunoeds ƏdAL 'g pue [ 'S5I4 me) emp, ty) 08 igh prse aluo gong: S ee fe 4 wy ag TID W jose ao Ao peat T Y7 fedt y sez ving ama "^5. 1 — "y megy i npo d + Hiker vpn eye pce wage TUVI NOU LEVEY. 1966] Gentiana puberulenta — Pringle 211 in DAO, GH, ILL, and WIS, and are reproduced here as Figures 1 and 2. The status of this specimen as the type is confirmed by the label in Michaux's handwriting on which both the collection data and the description correspond closely to those in Michaux's (1803) Flora Boreali-Americana. The type specimen of G. puberula Michx. is clearly not a specimen of the species described under that name in recent manuals (e.g. Fernald, 1950, and Gleason & Cronquist, 1963), but is, instead, a specimen of G. saponaria L. which falls well within the range of variation commonly found in populations of that species. Gentiana saponaria is usually described as having glabrous stems, but individuals with puberulent stems are occasionally encountered throughout most of its range. Table 1 lists some of the conspicuous differences between the type specimen of G. puberula Michx. and the species hitherto commonly known as G. puberula, here designated G. puberulenta. The characters selected for use in Table 1 are those readily discernible in the photo- graph and ones which have been found valuable in the dis- tinguishing of Gentiana species. No scientific name appears to be available for the Prairie Gentian. Authors of floras of North America and treatises on the North American gentians published prior to 1848 seem not to have known this species (with the possible exception of Rafinesque, as noted below). Grisebach (1838, 1839, 1845), in his works on the Gentianaceae, tentatively assigned the name G. puberula Michx. to the synonymy of his G. saponaria 3 linearis. (The latter name was based on G. linearis Froel., the name of a distinct species with which Grisebach had confused narrow-leaved specimens of G. saponaria.) In so doing, Grisebach was following Pursh (1814), but he (1838) added that this interpretation might be incorrect, and that “it is desirable that this question should be determined by the inspection of Michaux's her- barium." From labels on mid-nineteenth-century collections of the Prairie Gentian, it appears that the first botanists to collect it called it G. rubricaulis Schwein. (a name properly applied to a more northern species), and that it is the basis 212 Rhodora [Vol. 68 for the inclusion of “Gentiana rubricaulis" (without de- scription) in some plant lists of that period (see Jones and Fuller, 1955). Although many names for gentians were published by Rafinesque, it does not appear that any can be applied to this species. In certain parts the brief descriptions of Gentiana shortiana, G. torreyana, G. rigida, and G. gracilis (all in his Medical Flora, 1828) suggest the possibility that they were based on specimens of this species ; however, each of the descriptions, when considered in its entirety, appears much more probably to have been based on G. saponaria L. In Rafinesque’s (1837) Flora Telluriana these specific epi- thets appear in his genera Pneumonanthe and X olemia; the descriptions of both genera contain material which is in- applicable to the Prairie Gentian, although applicable to G. saponaria. Probably Rafinesque never encountered this species, which was rare in the area of his field work and had been collected by few botanists at the time of his pub- lishing. It is highly unlikely that any of these Rafinesque taxa will ever be typified by a specimen, since very few of the plant specimens collected by Rafinesque are extant. Dr. Ronald L. Stuckey, who has studied the herbarium (now in PH) of C. W. Short, to whom Rafinesque gave some speci- mens, reports that none of the specimens bear any of these names. John Torrey’s Gentiana specimens (now in NY) were annotated by Rafinesque, but I found none of these names among the annotations. Probably the earliest published description of the Prairie Gentian is that by Gray ( 1848) in the first edition of his Manual, in which it was designated G. saponaria B puberula (Michx.) Torr. & Gray. In later editions Gray recognized this taxon at the rank of species, as G. puberula Michx. Gray’s treatment has generally been followed by later authors. Thus, even though a number of descriptions of the Prairie Gentian have been published in the past, it is necessary to describe it here as a new species: 1966] Gentiana puberulenta — Pringle 213 Gentiana puberulenta Pringle, sp. nov. Herba perennis singulis vel pluribus caulibus suberectis longis 1-6 dm puberulentis subtiliter abundanterque. Folia lanceolato-oblonga interdum paene linearia prope basem latissima uninervia marginibus ciliatis saepe revolutis. Folia superiora longa 1.5-7 (-10.5) em lata 4-18 (-30) mm. Flores paene sessiles 2-bracteati aggregati apud sum- mum caulem et interdum de axillibus pedicellati. Tubus calycis longus 7-18 mm. Lobi calycis lineares longi 4-18 (-25) mm lati 1-3 (-4) mm. Corolla atrocyanea (raro violacea vel rubro-violacea) vasi- formis certe se aperiens longa 3-6 cm. Lobi corollae ovati patentes longi 6-14 mm lati 4-10 mm marginibus saepe erosis. Partes liberae appendicium corollae longae 2.5-6 mm bifidae praeterea segmentibus laciniatis. Antherae non cohaerentes, Semina alata. Type: Chase 1276, 8 mi. N.E. of Princeville, Peoria County, Illinois, 7 Oct. 1906 (ILL; isotypes DAO, ILL, KSC, MIN, NY, US). A more extensive description of this species, along with a discussion of its distribution and probable relationships to other species, will be included in a forthcoming mono- graph by the author on Gentiana, section Pneumonanthae, in eastern North America. No infraspecific taxa have been recognized within this species. The two named hybrids involving it are both typi- fied by specimens, so the applications of the names G. X bil- lingtonii Farw. (pro sp.) (Pap. Mich. Acad. 1:96. 1923) and G. X curtisii Pringle (Trans. Wis. Acad. 53: 277. 1964) remain unchanged. Their formulae, however, should be emended to G. puberulenta X G. andrewsii and G. puberul- enta X G. alba, respectively. LITERATURE CITED Curtis, J. T. 1959. The Vegetation of Wisconsin: An Ordination of Plant Communities. The Univ. of Wisconsin Press, Madison. FERNALD, M. L. 1950. Gray’s Manual of Botany. ed 8. American Book Co., New York. GLEASON, H. A., and A. CRONQUIST. 1963. Manual of Vascular Plants of Northeastern United States and Adjacent Canada. IDE Van Nostrand Co., Princeton. Gray, A. 1848. A Manual of the Botany of the Northern United States . . . James Monroe & Co., Boston. Rhodora [Vol. 68 214 GRISEBACH, A. H. R. 1838. Gentianeae Juss. In Hooker, W. J., Flora Boreali-Americana 2:54-71. Henry G. Bohn, London. [Re- printed 1960 as Hist. Nat. Class. 5. H. R. Engelmann (J. Cramer), Weinheim, and Wheldon & Wesley, Codicote.] 1839. Genera et Species Gentianearum. J. G. Cotta, Stuttgart and Tubingen. 1845. Gentianaceae. In Candolle, A. P. de, and A.L.P.P. de Candolle, Prodromus Systematis Naturalis Regni Vegetabilis . . . 9:38-141. Victor Masson, Paris. JONES, G. N., and H. J. FULLER. 1955. Vascular Plants of Illinois. The Univ. of Illinois Press, Urbana, and The Illinois State Mu- seum, Springfield, Sci. Ser. vol. 6. MICHAUX, A. (L. C. M. Richard, ed.)1803. Flora Boreali-Americana -.. Fratres Levrault, Paris PursH, F. T. 1814. Cochrane, & Co., London. RAFINESQUE, C. S. 1828-1830. Alexander, Philadelphia. 1836-1837. Flora Flora Americae Septentrionalis . . . Medical Flora and Strasburg. White, Atkinson & Telluriana, Publ. by the author, Philadelphia. (Photolith. 1946. Arnold Arboretum, Cambridge.) ROYAL BOTANICAL GARDENS, HAMILTON, ONTARIO, CANADA TABLE 1. THE TYPE SPECIMEN CONTRASTED WITH G., Type specimen of G. puberula Leaves few and widely spaced. Leaves widest at or above the middle. Calyx lobes oblong, widest above the middle. Corolla lobes comprising about 10% of the total corolla length. Corolla lobes widest at the base, less than twice as long as the appendages. OF GENTIANA PUBERULA MICHX, PUBERULENTA PRINGLE G. puberulenta Leaves many and crowded. Leaves widest near the base. Calyx lobes linear. Corolla lobes comprising 20-25% of the total corolla length. Corolla lobes ovate, more than twice as long as the appen- dages. NEW CHROMOSOME NUMBERS IN ZINNIA AND SANVITALIA ANDREW M. TORRES! In recent systematic papers concerning Zinnia ( Torres 1962, 1963) and Sanvitalia (Torres 1964) the author was unable to provide chromosome counts for 1 of the 17 species of Zinnia and for 3 of the 7 species of Sanvitalia. The Zinnia species, Z. purpusii, had been known only from the state of Chiapas in Mexico. Because the collection site, *Hacienda Monserrate", was not, and still is not, located and because the species was known only from the type col- lection, it was surprisingly pleasant to find a population on the hillsides of gravelly soil along Mexican Highway 190, 1.6 miles east of the Oaxaca-Chiapas border (1739) and again 12 miles east of the same border (1743). From the first collection the chromosome number of n = 12 was determined. Meiosis was regular. Its chromosome number, although predictable, was of interest because its nearest phenetic ally, Z. maritima, is also n = 12. Yet several other near allies are n = 11. The count is significant also since chromosome numbers of all species are now known. A few more collections of Sanvitalia tenuis had been made than of Z. purpusii yet its habitat too was relatively in- accessible. It apparently grows only on the eastern side of the Sierra Madre Occidental in western Chihuahua. Some 19 miles west of Mifiaca (near Ciudad Guerrero and La Junta) in Chihuahua near the edge of oak-pine hills, I chanced upon a population of some 20-25 young plants just beginning to flower. Interestingly, all my collections were from the base of young oak trees on gentle slopes in gravelly soil. During the rainy season when S. tenuis is in flower the area is accessible by logging roads and only in high- center vehicles. An anomalous chromosome number pattern had already been established for Sanvitalia because the species are n — 8 or 16, except for S. abertii which is TField work was supported by NSF grant GB-3842. 215 216 Rhodora [Vol. 68 ^ = 11. It therefore became of especial interest to obtain the number of S. tenuis to see if it formed a link in a pos- sible aneuploid series or if it was also n = 8, 16 or 11. Its chromosome number is n — 9; meiosis appeared regular. Perhaps one of the two species as yet uncounted, Mexican S. fruticosa and South American S. versicolor, will fill the n = 10 gap. Voucher specimens of both species will be housed at KANU with duplicates at Msc. DEPARTMENT OF BOTANY, UNIVERSITY OF KANSAS, LAWRENCE, KANSAS LITERATURE CITED TORRES, A. M. 1962. Cytotaxonomy of cespitose zinnias. Amer. Jour. Bot. 49: 1033-1037. 1968. Taxonomy of Zinnia. Brittonia. 15: 1-25. 1964. Revision of Sanvitalia (COMPOSITAE-HELIAN- THEAE). Brittonia. 16: 417-433. TWO NEW NAIADS FROM ILLINOIS AND DISTRIBUTIONAL RECORDS OF THE NAIADACEAE: PAUL L. FORE? AND ROBERT H. MOHLENBROCK? In the process of identifying Illinois species of Naias and in comparing fresh material with herbarium specimens, it was discovered that considerable misidentifications of speci- mens of this genus existed in Illinois. In order to present an accurate account of Naias in Ilinois, and to up-date ‘An additional species Najas marina L. is reported from Illinois by Winterringer in this number of Rhodora (p. 221). *Cooperative Fisheries Research Laboratory, Department of Zoology, Southern Illinois University, Carbondale. "Department of Botany, Southern Illinois University, Carbondale. 19661 Illinois Naiadaceae — Fore and Mohlenbrock 217 distributional data of the species, the first author undertook this study of Naiadaceae of Illinois. GEOGRAPHICAL CONSIDERATIONS Clausen (1936), summarizing the geographical distribu- tions of Naias, listed five species from the northern United States, two of which were known from Illinois. Likewise, Winterringer and Evers (1960) and Jones (1963) recorded only the occurrence of the same two species, N. flexilis (Willd.) Rostk. & Schmidt and N. guadalupensis (Spreng.) Magnus. The results of this investigation demonstrated that four species of Naias inhabit Illinois waters. Lawrence (1955) stated that a total of seven species were native to the continental United States out of a world-wide total of approximately forty species. Two Illinois species have a general distribution in eastern North America from Newfoundland to Minnesota, south to Missouri and Virginia. Included here are Navas flexilis and N. gracillima (A. Br.) Magnus. The remaining two species exhibit quite dissimilar ranges. Naias guadalupensis is known from Massachusetts to South Dakota, south to Texas and Florida, and Tropical America. Naias minor All., an introduced Eurasian species, is locally established in New York and West Virginia, according to Fernald (1950) ; this species is also reported by Muenscher (1944) to occur in Alabama and Tennessee. In addition to the above localities, Naias minor is now recorded from Illinois. In Illinois, Naias flexilis is widely distributed, although not abundant, throughout the state; N. guadalupensis is found mainly in the western and southern parts of the state. Naias gracillima, mostly distributed in southern Illinois, was recently rediscovered