SIDA ‘Susie™ Volume 6 1975 — 1976 SIDA Contributions to Botany volume 6 (in 4 numbers) Copyright 1975, 1976 by Wm. I*. Mahler SMU Herbarium V1 Dallas, Texas 7527! DATES OF PUBLICATION No. 1, pp. 1—62: 4 June 1975 No. 2, pp. 683—122: 11 Nov. 1975 No. 3, pp. 123—242: 14 May 1976 No. 4, pp. 2483—331: 29 Sept. 1976 Index, pp. 333—341 p. 332 blank For contents, see the unnumbered pages forming the front cover of the separate issues. SIDA ‘suse VOLUME 6 NUMBER 1 JUNE 1975 CONTENTS Sidus Sidarum. Paul A. Fryxell. ] Notes preliminary to an account of Cassia in the Chihuahuan Desert. H. S. Irwin and R. C. Barneby. rf An SEM study of leaf surface pubescence in the southeastern taxa of Persea. B. Eugene Wofford and Ronald W. Pearman. 19 A bibliography of manuals and checklists of aquatic vascular plants for regions and states in the conterminous United States. Ronald L. Stuckey. 24 Typification and distribution of the varieties of Gnaphalium helleri Britton (Compositae—Inuleae). Wm. F. Mahler. Studies in the Ranunculaceae of the southeastern United States. Ill. Clematis L. Carl S. Keener. 33 Notes on the flora of the Mogollon Mountains, New Mexico. William J. Hess. 48 Additions and corrections to the Bahama flora—ll. William T. Gillis and George R. Proctor. 52 US ISSN 0036-1488 SIDA, CONTRIBUTIONS TO BOTANY Founded by Lloyd H. Shinners, 1962 Editor & Publisher Wm. F. Mahler SMU Herbarium Dallas, Texas, 75275 Associate Editor John W. Thieret Northern Kentucky State College Highland Heights, Kentucky, 41076 Contributors to Sida should refer to the latest issues for style, follow the CBE Style Manual (3rd ed.), and use the ‘adopted’ journal abbreviations given in Botanico-Periodicum-Huntianum., Subscriptions: Libraries—$6.00 (U.S.) per year; individuals—$8.00 (U.S.) per volume; numbers issued Sida, Contributions to twice a year. (©) Botany, Volume 6, Number 1, pages 1-62 Copyright 1975 by Wm. F. Mahler SIDUS SIDARUM' PAUL A. FRYXELL Agricultural Research Service United States Department of Agriculture Texas A&M University College Station, Texas 77843 In 1962 Lloyd H. Shinners launched the journal Sida: Contributions to Bot- any as a private venture. His choice of a name for the journal (using the name of the well-known and large malvaceous genus Sida L.) was commented upon by Mahler (1973) in the issue of the journal devoted to Shinners’ mem- ory. I find it curious, however, that in 12 years of publication of Sida, the genus of the same name has been mentioned in its pages only incidentally on two or three occasions. I think it appropriate that studies of the genus Sida should appear in the journal Sida. Few genera have that option (see Appen- dix). I offer the present notes as a beginning. NOTE 1. TYPIFICATION OF SIDA SECTION MALACROIDEAE G. DON This taxon, a natural group within Sida, was recently revised by Clement (1957), who followed Schumann (1891) in characterizing it. Clement correctly noted the priority of Don’s epithet (Malacroideae) over that of Schumann (Pseudomalachra) even though Schumann’s name had been used almost ex- clusively before Clement’s study. However, Clement erred in indicating S. ciliaris L. to be the type species of the section, since Don did not include this species in his section. A lectotype should therefore be chosen from among the three species included by Don: S. fulva St.-Hil., S. anomala St.-Hil., and S. plumosa Cav. I therefore choose S. anomala St.-Hil., Fl. Bras. Mer. 1:177. 1827 to lecto- typify Sida sect. Malacroideae G. Don, Gen. Hist. 1: 498. 1831. NOTE 2. A NEW SPECIES OF SIDA SECT. MALACROIDEAE G. DON SIDA meridiana Fryxell, sp. nov., sectionis Malacroidearum (Fig. Caules plures, ramosi, procumbentes, pallidi, dense stellato-pubescentes, pilis exalbidis, 0.5-2.0 mm longis, demum glabrescentes. Folia late elliptica, foliis plus amplis 15-18 mm longis, 10-13 mm latis, basaliter truncata vel leniter cordata, ad apicem obtusa, marginibus in dimidiis proximalibus integris, in dimidiis distalibus foliorum bene evolutorum manifeste 10-20 dentatis; pube- scentia foliorum dense stellata, exalbida, utrinque similiter praeter supra antrorsa, marginibus ciliatis, pilis usque ad 2 mm longis. Petioli 4-6 mm longi, manifeste hirsuti, pili saepe simplicibus. Stipulae 4-8 mm longae, manifeste ciliatae, spatulatae, laminis acutis 1-3(-4) mm latis, in stipibus infra angu- ' Under this title (meaning ‘‘a constellation of Sidas”) I propose to publish occasional notes on nthe genus Sida. SIDA 6. 6(1): 1-6. 1975. 1, SRELGPYP& Deter: & Fryxell is" Photograph of the holotype of Sida meridiana Fryxell. UB HERBARIO PLS ey statis vel (praesertim in inflorescentiis apicalis) anguste lineares, congestae, et usque ad 12 mm longae. Flores in inflorescentia pauciflora, pseudoinvolu- cellata et apicali congesti, pedicellis 2-3 mm longis. Calyces 5-6 mm longi, basaliter luteoli, ad medium divisi, 5 lobis acutis, intus et extus dense hir- sutis. Petala alba basibus flavidis, ca. 1 cm longa, in marginibus minute ciliatis, ceterum glabris. Columna staminalis glabra, pallida, ca. 3 mm longa; filamenta pallida; antherae et pollen aureae. Styli 5-6, androecium exce- dentes, pallidi; stigmata capitata, pallida. Fructus oblati, per calyces inclusi; mericarpia ca. 6, ad apicem breviter birostrata, in dorso rugosa et glandu- loso-puberula (saltem ubi immatura), in lateribus grosse reticulata. TYPE: BRAZIL: Minas Gerais: Rio Pandeiros, ca. 52 km by road W of Januaria near road to Serra das Araras; elev. 520 m; sandy beach. Decum- bent from perennial base; flowers white, open in afternoon, base of petals and stamens yellow. 21 Apr 1973, William R. Anderson et al. 9271 (holotype, UB; isotypes, NY, pf,” and to be distributed from NY). The new species was found growing sympatrically with S. ciliaris (Ander- son et al. 9265; NY, UB, pf,’ and elsewhere). The two species were distinct not only in gross morphology but also in floral biology. Since the plants were immediately adjacent to our site of encampment, we could observe them closely over a period of time. The flowers of S. ciliaris, yellow-orange, undergo anthesis near dawn and remain open through the morning but are withered by noon. Flowers of S. meridiana, on the other hand, remain in tight buds through the morning, open shortly after noon, and wither 2 or 3 hours later. Flowers of S. meridiana are white with a yellow center, the yellow being characteristic of the anthers and of the bases of the petals. Although the two species are sympatric to the extent that individual plants of cne species grow within a few feet of those of the other species, they are completely isolated temporally in their floral behavior. The choice of the spe- cific epithet of the new species is in reference to this midday flowering pat- tern. In S. meridiana the flower buds and the developing fruits are concealed amcng the congested leaves at the apices of the branches and are therefore inconspicuous; the spent corollas of late afternoon are equally inconspicuous. Thus, the plants appear, superficially, to be vegetative except during the few midday hours when the flowers are open. Perhaps because of this relatively cryptic reproductive pattern, this species has not previously been recognized. Sida meridiana may be distinguished from the other species of section Malacroideae by its broader leaves; by its more numerous teeth in the mar- gins of the leaves; by its tendency to have leafy stems rather than bare stems with leaves crowded apically; by its broader stipules; by its dense whitish stellate indumentum, especially on the stems where, in other species (e.g., S. anomala), the hairs are sometimes submalpighiaceous and on the * Personal herbarium of Paul A. Fryxell. upper leaf surfaces where, in other species, the hairs are either sparse or absent; and by the distinctive flowering pattern described previously. NOTE 3. SIDA COLLINA IN MEXICO The species S. collina Schlecht. is common in Mexico. However, it has not been well understood because it has often been confused with the more wide- spread and better-known S. rhombifolia L. When it has been distinguished from the latter species, it has been treated under the name S. corymbosa R. I. Fries. Sida collina is closely allied to S. woronowiti Ulbr. but is distinct enough in general aspect not to be confused with it. Both S. collina and S. costata have previously been considered as doubtful or unidentified species (Standley, 1923; Kearney, 1954), but the opportunity to examine the Schlechtendal types cited below has enabled me to establish this synonymy. This note further provides the occasion to add observations that more clearly distinguish S. collina from S. rhombifolia. I am grateful to Dr. K. Werner for making the Schlechtendal specimens available for examination, SIDA COLLINA Schlechtendal, Linnaea 11: 364. 1837. Type: MEXICO: [ Veracruz: ] prope Hacienda de la Laguna, Jul 1829, Schiede s.n. (holotype: HAL-36722!); idem loc., Sep 1829, Schiede s.n. (paratype: HAL-36721! ). Sida costata Schlechtendal, Linnaea 11: 365. 1837. Type: MEXICO: | Veracruz: | prope Hacienda de la Laguna, Jul 1829, Schiede s.n. (holo- type: HAL-34056-1!; isotype: HAL-34056-2! ). Sida corymbosa R. E. Fries, Bull. Herb. Boiss., ser. 2, 7: 998. 1907. Type: MEXICO: |Veracruz:| region d’Orizaba, 17 Aug 1866, Bourgeau 2863 (G, K). The type of S. costata is a more densely pubescent specimen than is the type of S. collina, but they appear to differ in no other significant respect. This difference in pubescence was the principal distinction noted by Schlech- tendal. The types of all three names are from Veracruz, but the species is also well known from western Mexico in the states of Jalisco, Nayarit, and Sinaloa. Both S. rhombifolia and S. collina are variable species. Instead of describ- ing the particulars of that variation or its range, I only note those characters that consistently differentiate the two species, especially since currently available keys (Standley, 1923: Kearney, 1954) distinguish them imperfectly. Pubescence. The herbage of S. collina is more or less pubescent with stel- late hairs 0.5-1.0 mm long, although the hairs on the upper (adaxial) leaf sur- face are commonly simple and antrorsely oriented. By contrast, the herbage of S. rhombifolia is minutely puberulent with stellate hairs no more than 0.1 mm long. The nature of the indumentum in the latter species is such that the leaves are generally discolorous. Stipules. The stipules of S. collina are considerably more prominent (7-12 mm long, 0.6-1.0 mm wide, often 3-nerved, ciliate) than are those of S. rhom- bifolia (4-6 mm long, 0.2-0.5 mm wide, l-nerved, puberulent). This promin- ence is especially noticeable if one compares a pair of stipules with the flow- er borne in the same axil. In S. collina the stipules usually exceed the ped- uncle and often are subequal to the calyx; in S. rhombifolia the stipules are only a fraction of the length of the peduncle. Leaf form. The leaves of S. collina tend to be narrower, more nearly lan- ceolate-elliptic, and more sharply serrate than those of 8. rkombifolia, which by contrast tend to be broader, more nearly rhomboid, and less prominently serrate. Inflorescences, Although individual flowers are solitary in the axils in both species, they tend to form crowded, apical, corymbiform inflorescences in S. collina, whereas in S. rhombifolia the flowers generally are widely scat- tered and not crowded into apical aggregations. This difference is accentu- ated by a difference in peduncle length. The peduncles of S. collina are only rarely 1 cm long; they are more commonly subsessile. In S. rhombifolia, on the other hand, the peduncles are commonly 1-3 cm long and are sometimes subequal to the subtending leaf. Flowers. In flowering specimens I have examined, the flowers of S. collina consistently have a large purple spot with radiating purple veins at the base of each petal. Flowers of S. rhombifolia lack basal petal spots. Flowers of S. collina are generally larger (calyx length: 8-10 mm) than those of S. rkombi- folia (calyx length: 5-7 mm) Carpel number. The number of mericarps per fruit or the number of styles and stigmas per flower is 7-9 in S. collina, 10-14 in S. rhombifolia. Schlech- tendal’s description ‘‘carpidiis 5” for S. collina is not supported by examina- tion of type material. REFERENCES CLEMENT, I. D. 1957. Studies in Sida (Malvaceae). Contr. Gray Herb. 180: 1-91. (Sect. Malacroideae, 16-30 KEARNEY, T. H. 1954. A tentative key to the North American species of Sida L. Leafl. ot. 7: 138-150. MAHLER, W. F. 1973. By . other name. Sida 5: 180-181. SCHUMANN, C. 1891. Malvaceae I. Pages 253-456 in C. F. P. Martius, Fl. Bras. 12(3). (Sida sect. Sasol 80- 287) STANDLEY, P. C. . Sida Pages 761-767 in Trees and Shrubs of Mexico. Contr. U.S, Natl. Herb. 2 APPENDIX An unimportant digression on the titles of botanical journals Many botanical journals have one-word titles that are coincident with the name of a genus. A quick tabulation provides me with a list of 81 such names, listed below. About one-third of these journals are no longer published, and it is difficult to be certain why most of these journals were named as they were. However, in 57, marked below with an asterisk, the name of the journal was chosen presumably to honor a noted botanist, and the same name was incidentally and independently chosen for a genus as another honor for the same botanist. Thus, the identity of the names of these journals and of the genera is only coincidental. Ten, marked with a dagger, are specialty journals, mostly for horticultural enthusiasts. The titles of these are simply descriptive of the contents of the publication. The titles of three journals, marked with a double dagger, are only acci- dentally identical to the names of genera. Three more, marked with a ques- tion mark, have names whose probable origins I am unable to judge. Only the 10 journal names marked with a section mark (§) were chosen to commemorate genera. Three of these, Betula, Erythea, and Nemophila, are no longer published. The remaining names are Araucaria, Castanea, Ceiba, Kalmia (?), Rhodora, Sarracenia (?), and Sida. The brevity of this list is the basis for my statement (in the introduction to ‘‘Sidus Sidarum’’) that few genera can be written about in journals named in their honor. Adansonia* Caldasia* Hedwigia* Regelia*® Addisonia* Candollea* Herbertia*® Reinwardtia* Allionia* Castaneas Iris Rhododendron* Araucarias Cavanillesia* Irmischia* Rhodora’ Arnoldia Ceiba Kalmia* or § Rodriguezia* Baileya* Cinchonat Kudoa? Sargentia* Bartonia* Citrus? Lensi Sarracenia® or § Bauhinia* Claytonia®* Lindenia* Sellowia* Belmontia? Dahlia* Linnaea* Sida§ Betulas Darwiniana* Lloydia* Sieboldia* Billctia* Delphinium* Molina* Sylviat Blumea* Delpinoa* Malpighia* Teysmannia* Blyttia* Dodonaea* Muhlenbergia* Torreya* Boissiera® Drudea* Mutisia®* Uraniat Bonplandia* Dusenia* Nemophilas$ Vellozia* Borbasia* Erythea$ Notarisia* Viola? Bragantia* Friesia* Orchis+ Vitis? Brebissonia* Gladiolus? Persoonia* Watsonia* Brittonia®* Gorteria* Pittonia* Webbia* Cactus? Grevillea* Presha* Willdenowia* Wrightia* * Name commemorates a botanist. + Name. describes i # Name accidentally coincides with a generic name. ? Name is of uncertain origin. § Name commemorates a plant genus. content of a specialty (often horticultural) journal. NOTES PRELIMINARY TO AN ACCOUNT OF CASSIA IN THE CHIHUAHUAN DESERT H. S. IRWIN & R. C. BARNEBY New York Botanical Garden Bronx, N. Y. 10458 ABSTRACT. Descriptive and interpretive taxonomy of CASSIA (Leguminosae: Caesalpinioideae) of the Chihuahuan Desert, involving 6 nomenclatural trans- fers: CASSIA sect. Earleocassia (Britt.), C. sect. EARLEOCASSIA ser. Tharpia (Britt. & Rose), C. CROTALARIOIDES Kunth var. vogeliana (Schlechtd.), C. pilosior (Macbr.), C. WISLIZENI Gray var. villosa (Britt.) and var. painteri (Britt.); 5 new taxa: C. DEMISSA Rose var. radicans (Coahuila), C. mensicola (Coahuila to Hidalgo), C. DURANGENSIS Rose var. iselyi (s. Tex. to San Luis Potosi), C. ripleyana (trans-Pecos Texas to n. Zacatecas), C. monozyx (Coahuila); and 1 new epithet: C. parralensis (= Chamaecrista goldmanii Britt. & Rose). In the course of revising Cassia for the Chihuahuan Desert Flora in prepa- ration, we have encountred several undescribed taxa and the need for a few nomenclatural changes. These are presented in the framework of Bentham’s classification of the genus (1871). I. Sect. EARLEOCASSIA (Britt.) Irwin & Barneby CASSIA subgen. SENNA sect. Earleocassia (Britton) Irwin & Barneby, stat. nov. Earleocassia Britton in N. Amer. FI. 24 (4): 247. 1930, pro gen.— Generitypus: E. roemerana (Scheele) Britt. = Cassia roemerana Scheele. Cassia subgen. Senna sect. Chamacefistula ser. Brachycarpae Bentham in Trans. Linn. Soc. (London) 27: 529. 1871. — Sp. lectotypica: C. roemerana Scheele. — Non Cassia sect. Chamaesenna ser. Brachycarpae Bentham, op. A well defined group of xerophytic cassias native to the Mexican Plateau, the Sonoran and Chihuahuan deserts, floristically related arid grasslands of southwestern United States, and tropical desert enclaves in south-central Mexico. The centers of diversification and abundance coincide within the Chihuahuan Desert, where 12 of the 18 described species are native and 5 en- demic. The Brazilian C. nana Benth., still indifferently known, is provision- ally excluded. Characters common to all (or almost all) members of the section are: roots black in age; herbaceous stems dying back yearly usually to the ground; flaccid attenuate stipules; a complex vesture composed of short hairs, min- ute yellowish clavate ones, and stiffer longer setae; fusiform, often stipitate interfoliolar glands; a turbinate hypanthium heavily glandular about the torus; subequal obtuse sepals; petals expanded for 1 day, then wilting, fading SIDA 6(1): 7—18. 1975. whitish when dried but coarsely arborescent-veined; and a pod, usually more or less turgid and internally septate, dehiscent downward along both sutures, the tips of the valves diverging to passively release the pear- or paddle- shaped seeds. Bentham considered his ser. Brachycarpae, of which he knew (disregarding C. nana and the probably also extraneous C. mexicana Jacq.) seven members, to form a transition between sect. Chamaesenna ser. Corym- bosae and sect. Oncolobium, being close to the last in form, but not dehis- cence, of the pod; different from both in the homomorphic sepals and from Oncolobium more decisively by want of the basal petiolar gland and lack of malodorous oils. The syndrome of characters listed, combined with a natural collective range of dispersal, characterizes a closely coherent and obviously natural group of species entitled to sectional rank within subgen. Senna. The chromosome number n = 14 (in tetraploid C. crotalarioides Kunth, n = 28) determined for 8 members of the section (Irwin & Turner 1960, p. 315) is the base number of subgen. Senna. In view of the fact that Bentham employed the epithet Brachycarpae for series in two different subgenera of Cassia, it seems preferable to take up at sectional level Earleocassia, proposed as a genus for an almost exactly conterminous group. The monotypic genus Tharpia Britt. & Rose, consisting of the peculiar little subacaulescent C. pumilio Gray included in ser. Brachycarpae by Ben- tham, was said to differ from related segregates of Cassia, in particular from Earleocassia, in the want of a petiolar gland combined with subglobose, indehiscent pod. In reality, a petiolar gland, sometimes small and concealed between the narrow wings of the leafstalk, is consistently present, and the pod, now known to vary from subglobose to oblong, is tardily dehiscent down- ward from the tip, just as in other Earleocassias. The flowers and seeds of C. pumilio are essentially identical with those of Earleocassia, and in C. ripleyana, described below, we now have a species almost as dwarf, resem- bling C. pumilio in the scapiform, one-flowered peduncles and small turgid pod, Characters that remain peculiar to C. pumilio in context of its relatives are the tuberously thickened roots, the firm texture of the stipules and foli- age, the thickened margins of the leaflets, and in particular the modification of the leaflets from asymmetrically oblong or obovate, cordate at base on the proximal side, to subsymmetrically linear-lanceolate, at base cuneate on both sides. We evaluate Tharpia as a monotypic series of sect. Earleocas- sia: CASSIA subgen. SENNA sect. EARLEOCASSIA ser. Tharpia (Britton & Rose) Irwin & Barneby, stat. nov. Tharpia Britton & Rose in N. Amer. FI, 23 (4): 246. 19380. — Sp. unica: C. pumilio Gray. CASSIA CROTALARIOIDES Kunth var. vegeliana (Schlechtd.) Irwin & Bar- neby, stat. nov. C. vogeliana Schlechtd. in Linnaea 12: 342. 1838. Earleocas- sia vogeliana (Schlechtd.) Britt. ex Rose in N. Amer. Fl. 23 (4): 248. 1930. Bentham (1871, p. 530) maintained C. vogeliana as akin to C. crotalarioides but different in the long spreading pubescence of stems and leafstalks com- bined with supposedly larger flowers. Britton & Rose (1 c.) relied for differ- ential characters solely on vesture, and this indeed seems to be the only sub- stantial difference. The collective range of the two entities encircles the south- ern margins of the Mexican Plateau from central Durango to Aguas Calientes and Guanajuato, thence east to Hidalgo, thence north along Sierra Madre Ori- ental to central Nuevo Ledn. Along the segment of this great arc between Durango and Hidalgo the stem-pubescence is appressed or subappressed and composed of hairs mostly less than 1 mm long. From Hidalgo northward it becomes pilose and in extreme form shaggy-pilose with fine lustrous setae up to 3 mm long. Both types of vesture occur in Hidalgo (whence the type of C. vogeliana), and plants with occasional long spreading hairs scattered among the appressed ones form passage between the two, CASSIA DEMISSA Rose var. radicans Irwin & Barneby, var. nov., a var. demissa caudice diffuse ramoso hine inde radicanti, caulibus annotinis elatioribus (1-3.5 nec 0.8 usque dm longis), cauliumque pube hirsutula recedens. — MEXICO. Coahuila: s.-w. end of Sa. de la Fragua, 1-2 km n. of Puerto Colorado, 2. [X. 1941 (fl, fr), I. M. Johnston 8774.—Holotypus, TEX. Hillsides and canyons, in izotal and chaparral, upward into the pinyon wocdland, 1400-2100 m, strongly calciphile, scattered in the mts. of centr. and s. Coahuila from Sa. de la Madera s. through Sas. de Fragua and de Organos to Sa. de Parras and mts. s.-e. of Saltillo; apparently disjunct in extreme s.- w. Tamps. Material seen: Coahuila: Sa. de Madera, I. M. Johnston 9082 (TEX); ibid., Canon de la Hacienda, Hendrickson 11874, 13567 (TEX). Sa. de Organos, 62 air miles w.-s.-w. of Cuatro Ciénegas, Henrickson 12156 (TEX). s. of Parras, Stanford & al. 229 (NY); S. S. José del Refugio, Chiang & al. 8267B (NY); Saltillo, Gregg 244 (NY); 17 mi. s.-e. of Saltillo, Bierner & Turner 86 (NY, TEX) Britton & Rose (1930, p. 284) knew C. demissa only from the plants collect- ed by Pringle in 1889 near (probably east of) Carneros Pass in extreme south- ern Coahuila. Precisely comparable material has been encountered since then only within a narrow circumference of the type station: once close by at Fraile (Stanford & al. 253); once in adjoining Zacatecas (Chiang & al. 7938); and once in the Sierra Madre Oriental near Pablillo (M. Taylor 153). North- ward in Coahuila, from a line connecting Sierra de Parras and Saltillo by way of General Cepeda northward to Sierras de Madera and La Fragua, the species is represented by the coarser, taller form described as var. radicans, different from var. diffusa in its openly branched, often adventitiously root- ing caudex, hirsutulous pubescence of the stems, and usually ampler, always more coarsely pilose foliage. The oldest known collection of var. radicans, 10 Gregg 244 from near Saltillo in 1848/9, was mistakenly referred by Bentham (1871, p. 530) and by Britton (in hb.) to C. vogeliana Schlechtd., a species similar in gross aspect but with 4-5, not 2 pairs of leaflets. One flowering specimen from the southwestern corner of Tamaulipas (Stanford & al. 829, NY) is provisionally referred to var. radicans but may yet prove to represent a distinct form. It resembles the Coahuilan plants in general habit, but is notable for some very long spiral hairs on the upper stems, suggesting the shaggy pilose vesture of the genuine C. vogeliana, vicariant southward in the Sierra Madre. CASSIA pilosior (Macbride) Irwin & Barneby, stat. nov. C. bauhinioides var. pilosior Robinson ex Macbride in Cont. Gray Herb. n. ser. 59: 27. 1919. — “TEXAS: Bofecillos Mts., Sept. 1883, Harvard, no. a . MEXICO: Tor- reon, Coahuila, Oct. 13-20, 1898, Palmer, no. 455; 75 qiles southwest of Parras, Coahuila, May, 1880, Palmer, no. 2132.’ Lectoholotypus (Isely, Leg. U.S. II. Subfam. Caesalpinioideae, ms.): Havard 14, GH. Cassia durangensis sensu B. L. Turner, Leg. Tex. 74, map 35. 1959; sensu Correll & Johnston, Man. Vasc. Pl. Tex. 788. 1970 Confusion of C. pilosior with genuine C. durangensis Rose started with Britton’s account of Marleocassia (Britt. & Rose, 1930, p. 248) and has been perpetuated in the Texan literature. The two species are certainly close al- lies, but upon examination of material taken throughout their ranges are found to be differentiated by the following syndrome of characters: a. Weak spiral hairs up to 2-4 mm present on upper stem and leafstalks; sepals marcescent about the forming pod; style filiform, 2.5-3.2 mm long, 0.2-0.25 mm diam; fl. small, the petals 8.5-10 mm; ripe pod strong- ly compressed, the septa between seeds very narrow (to 0.3 mm wide); seeds smooth; trans-Pecos Texas to n.-e. Durango and s. Coahu- ila... . ‘, pilostor a. Longer fais mostly. aes fia 2 mm, sélifoxm one gas Ps ones in distantly allopatric var. iselyi); sepals promptly deciduous at anthesis; style 1-2 mm long, 0.3-0.4 mm diam; fl. variable in size, but where near- ly allopatric with C. pilosior larger, the petals 10-13 mm; pod turgid, the septa broader; seeds strongly rugulose; bicentric in dispersal, var. durangensis local around margins of Mapimi depression in s.-e. Chihua- hua and e. Durango, var. iselyi on the Gulf Coastal Plain in extreme s. Texas and Tamaulipas : . C. durangensis The known range of C. siesta eta on the Rio Grande valley in Brewster and Presidio counties, Texas, south to the periphery of the Mapimi depression in southwestern Coahuila, eastern Chihuahua, and immediately ad- joining Durango, In the lower Nazas valley southwest of Torre6n C. pilosior and C. durangensis var. durangensis have been collected within a radius of a few kilometers of Cuencamé (Irwin 1231, 1234) but despite essential sym- patry are there sharply marked. 11 CASSIA DURANGENSIS Rose var. iselyi Irwin & Barneby, var. nov., a var. durangensi vix nisi pedunculis pauci (1-3, nec 2-5)-floris necnon glandula petiolari sessili glaberrima (nec stipite villosulo elevata) recedens sed patria valde remota separanda. — MEXICO. Tamaulipas: San Fernando a Jimenez [Santander], X.1830 (fr) J. L. Berlandier 840. — Holotypus, NY; isotypus, K. Mesquite thickets and along roadsides and canal banks of the Gulf Coastal Plain below 300 m, in sandy or sandy loam soils, lower Rio Grande valley in extreme s. Texas (Zapata, Starr, Hidalgo, Cameron cos), s. through Tamaulipas just into San Luis Potosi. Representative: UNITED STATES. Texas, Zapata: Zapata, 14.X1.61 (fr), Cabrera & Munoz 50 (TEX). Starr: road to McCook, 9.X.54 (fr), Tharp & M. C. Johnston 541888 (TEX). Hidalgo: Lomita Alta, 18.V1I.37 (fl), C. E. R. Cam- eron 30 (TEX). Cameron: Olmito, 15.X.36 (fl), R. Runyon 2992 (TEX). MEX- ICO, Tamaulipas: Rio Purificacién, 7 mi. n. of Padilla, Hagee & Kerr 1059 (SMU); Rio Blanco to Victoria, Karwinski in 1842 (LE); Victoria, Berlandier 2260 (K, NY). San Luis Potosi: Salto de Agua, mun. Cd. del Maiz, Shott & Drewe 54 (SMU). While one of the earliest collections (Berlandier 2260, NY) of var. iselyi was correctly referred by Britton (in hb.) to C. durangensis, modern collec- ions, especially those from southern Texas, have passed, surprisingly, as C. bauhinioides, although fundamentally different in the broad leaflets, cylin- dric style, and straight pod. It differs, in fact, only very slightly from gen- uine C. durangensis of the Chihuahuan Desert in its glabrous sessile petiolar gland and usually 1-2, rarely 3-flowered peduncles. The dispersal of the two varieties is remarkable, var. durangensis being known from the upper forks of the Conchos and Nazas rivers in southeastern Chihuahua (n.-e. of Parral) and eastern Durango at elevations of some 1400-1570 m, whereas var. iselyi is confined to the Gulf Coastal Plain below 300m. A hispidulous ecotype of C. bauhinioides, distinguished by its very small, sometimes cleistogamous flow- ers and prevailing 1-flowered peduncles is the only Earleocassia known to enter the edge of the range of var. iselyi, reaching Starr and Zapata coun- ties in the Rio Grande valley (Runyon 1719, 2605, TEX) and northern Nuevo Leon (Ripley & Barneby 13252, NY). This, however, with the narrow leaflets and dilated style of its species, is probably not truly sympatric with var. iselyi, occurring in Texas on gravelly hilltops and in adjacent Mexico on the calcareous first bench of Sierra Madre, apparently never straying out onto the sandy and loam soils of the Coastal Plain proper. The variety is named for our friend Duane Isely, distinguished student of Leguminosae, who first recognized the distinct character of the taxon in re- lation to the polymorphic C. bauhinioides. CASSIA mensicola Irwin & Barneby, sp. nov. inter C. bauhinioidem A. Gray et C. roemeranam Scheele quasi intermedia, Ulam habitu simulans sed floribus majusculis (petalis 11-18 nec 6-10 mm longis), stylo exacte cylin- drico 0.2-0.3 mm tantum diam, necnon seminibus laevibus cito extricanda, ab hac, quoad florum magnitudinem stylique forma praesimili, imprimis statura minori (caulibus 0.5-2, rarius 2.5 dm usque, nec 2-7 dm longis) foliolis minoribus oblongo-ellipticis ‘cae pro rata latioribus (2-3 nec 4-9- plo longioribus quam latis), patriaque procul aliena abstans. — MEXICO. San Luis Potosi: Charcas, VU-VUI. 1934 (fl, fr), C. L. Lundell 5345. — Holo- typus, CAS; istotypus, LL, TEX. Loosely cespitose perennial herbs, in habit resembling C. bauhinioides, the mostly simple, incurved-ascending stems 0.5-2 (2.5) dm, these and the leaf- stalks pilosulous with short subretrorse and minute yellow clavate hairs mix- ed with a few coarser ascending setae to 0.4-1 mm, the leaflets (dry) ashen beneath, golden-green above, oo. both sides, the axillary, few-flowered racemes short-exserted. Leaves (2-) 2.5-5 (-5.5) em; stipules lance-caudate 4-8 mm; petioles 1-2.7 cm; intrafoliolar een slenderly fusiform acute, including the often puberulent stipe 0.8-1.5 mm; leaflets 1 pair, narrowly oblong or oblong-elliptic obtuse, minutely mucronulate, (10-)12-37 X 3-11(-12) mm, most- ly 2-3 times longer than wide; peduncles 2-4(-4.5) em; racemes (1-)2 (-4)- flowered, the axis 0-11 (-15) mm; bracts 2.5-4.5 mm; pedicels 4-9 (-12) mm; sepals narrowly elliptic to elliptic-obovate obtuse, (5-) 5.5-8.5 (-9.5) X 2.2-3.6 mm; petals yellow drying whitish, broadly oblanceolate to obovate-cuneate, 11-16.5 (-18) X 6-11 (-12) mm, arborescently 3-veined from claw; androecium of C. bauhinioides; style slenderly cylindric 1.4-2.8 X 0.2-0.3 mm; ovary pilo- sulous: ovules 34-42. Pod linear-oblong, usually a trifle incurved, 2.3-3.5 X (.35-0.55 cm, moderately turgid, the valves at once densely minutely puberu- lent and setose with ascending hairs to 0.9-1.2 mm, the cavity narrowly sep- tate; seeds paddle-shaped, 2.2-2.5 mm, the testa bluntly folded and ridged but otherwise smooth or almost so, dark lustrous brown. rravelly clay flats and hillsides, in desert or desert grassland, sometimes on roadsides, mostly, perhaps exclusively calciphile, 990-2300 m, Meseta Cen- tral of n. Mexico from upper Moctezuma valley in n, Hidalgo (Ixmiquilpan) n. through San Luis Potosi and adjacent Zacatecas to s.-e. Coahuila. Representative: COAHUILA: 15 km w. of Concepcion del Oro, 19.VI11.41 (fl), Stanford, Wetherford & Northcraft 500 (NY); s. of Saltillo, 29.VII.29 (fl), R. Runyon 1388 (NY, TEX). Zacatecas: w.-n.-w. of Tecolotes on road _ to Coapa, 17.V1.72 (fl), Chiang, Wendt & Johnston 7895 (NY); e. of Troncoso 6.1V.70 (fl), Mahler 5757 (SMU). SAN LUIS POTOSI: San Luis Potosi and vicinity, 2.1X.02 (fr), Pringle 9728 (NY), in 1879, Perry & Palmer 206 (NY), Schaffner 624/831 (NY). HIDALGO: Ixmiquilpan, VII.05 (fl), Rose Painter & Rose 8999 (NY, US), Purpus 1370 (NY) Five collections of C. mensicola known to Britton in 1930 were referred by him to C. bauhinioides, and modern ones have generally been so identified, following the key to Marleocassia in North America Flora. The two species, although widely separated geographically, are indeed similar in stature and foliage, but C. mensicola differs in the conspicuously larger flower, the syI- 13 indric style, and the smooth seeds. Its flower is essentially like that of the probably more closely related C. roemerana, but this differs in the greater stature (stems mostly 2.5-7, not 2-2.5 dm), in the longer, proportionately nar- rower and more acute leaflets (4.5-9.5 not 2-3 times longer than wide), which are in consequence of their length, at least in all upper leaves, more and not less than twice as long as their petiole. The range of C. mensicola, a very na- tural one extending over the southern part of the Chihuahuan Desert prov- ince southward from the Coahuila-Zacatecas boundary at approximately 25° N. into Hidalgo, is fully disjunct from that of C. roemerana, a species com- mon over Edwards Plateau, trans-Pecos Texas, and desert segments of the Pecos and Rio Grande valleys in New Mexico, which extends south only feebly into Mexico, in Coahuila to near 28° and along the east piedmont of Sierra Madre Oriental to near 26° N. in Nuevo Leon. CASSIA ripleyana Irwin & Barneby, sp. nov., inter affines sect. Earleocas- siae foliolis 1-jugis signatas legumine breviusculo turgido inter semina late septato C. bauhinioidem A. Gray spectans, sed habitu pumilo caespitoso, pedunculis (plerisque 1-, paucis raris 2-floris) scapiformibus, styloque elongato filiformi apice haud dilatato, necnon seminibus laevissimis (nec rugulosis) recedens. MEXICO, Chihuahua: arid calcareous flats, + 1400m, 18 km w. of Jiménez (27° 5’ N, 105° 10’ W), 2.X.1965 (fr), H. D. Ripley & R. C. Barneby 13,904.—Holotypus, NY. Dwarf, laxly cespitose, subacaulescent perennial herbs from woody tap- root and (when adult) a pluricipital caudex, pilosulous throughout with coars- er, forwardly incurved-ascending hairs less than 1 mm and a few weak spreading filiform ones to + 2-2.5 mm, the 1 pair of leaflets ashen beneath, greener above, silky-pilosulous both sides. Stipules erect, narrowly linear- caudate (4-) 6-10 mm, strongly 1-nerved, the filiform glabrescent tips often recurved. Leaves 1-5 cm; petiole stiff but subfiliform, 1-2.5 (-3.2) cm; intra- foliolar gland linear-claviform 0.7-2.4 mm; leaflets oblong- to obliquely obo- vate-elliptic, obtuse mucronulate or subapiculate, (6-) 8-20 X 3 -10 mm, at base broadly cordate on proximal and cuneate on distal side. Peduncles stiff- ly erect or in age declined, 2.5 cm, 1 (-2)-flowered; bracts triangular-subu- late 1-1.5 mm; pedicels 5-8 mm; buds nodding, obtuse, pilosulous. Sepals 5.9-6.5 mm, the outer elliptic-oblanceolate, the inner obovate, all membran- ous-margined; petals yellow (whitish when dry) not widely expanding, all spatulate-oblanceolate, short-clawed, 8.5-9.5 X 5-6 mm, arborescently 3-vein- ed from claw; androecium of C. bauhinioides; ovary densely pilose: style 3-3.5 X + 0.38 mm, filiform-cylindric; ovules 18-28. Pod in outline oblong- oblanceolate, straight or slightly incurved, (13-) 15-24 XK 5-8 mm, abruptly subulate-beaked, turgid, the ripe valves papery, brown, hispidulous with coarse, basally dilated ascending setae up to 0.7-1 mm, dehiscent downward through both sutures, the valves narrowly gaping to release the seeds. Seeds paddle-shaped, + 2-3.5 mm, the testa olivaceous or pinkish-brown, smooth, 14 lustrous.—F ig. 1. Gravelly hilltops and flats in arid grassland and Larrea desert, + 1400-1660 m, local within the borders of the Chihuahuan Desert from trans-Pecos Tex- as (Glass Mts., Brewster Co.) to s.-e. Chihuahua near Bols6n de Mapimi (w. of Jiménez; Rancho La Gloria) and extreme n. Zacatecas (Cedros), not yet collected but to be expected in n. and w. Coahuila. Material seen: UNITED STATES. Texas. Brewster: Glass Mts., 13.VII.40 (fl, fr), Warnock W-44 (TEX); s. 1., 9.X.36, Tharp s. n. (CAS, in part, mixed with C. bauhinioides). MEXICO. Chihuahua: Sa. del Diablo 14 km s.-e. of Rancho La Gloria, 29.VIII.72 (fr), Chiang & al. 9000d (TEX). Zacatecas: n. of Cedros, 22.1X.73 (fl, fr), Reveal & Atwood 3357 (NY, US) A neat little cassia related to C. bauhinioides but readily distinguished by the tufted habit, the wiry scapiform peduncles, and the lustrously smooth olivaceous to pinkish-brown seeds. The subacaulescent growth-form and rela- tively stiff stipules recall C. pumilio Gray, but the oblong-obovate, basally semicordate leaflets are entirely different. Named in memory of Harry Dwight Dillon Ripley, 1908-1973, an avid col- lector and eclectic devotee of small and rare desert plants, who in the field noted of the type-collection: ‘‘near C. bauhinioides, but not the same.”’ II. Section CHAMAESENNA Bentham The genus Palme? t Britt. (1930, p. 253), based on Cassia wislizeni A. Gray, consisted in ie first instance of five species, endemic to desert re- gions of Mexico and immediately adjoining United States. They are twiggy, subspinescent, microphyllous shrubs and treelets with glandless leafstalks, large flowers, the androecium and tardily dehiscent, compressed pod of sect. Chamaesenna Benth., forming in Bentham’s scheme of classification a xerophytic offshoot of the Eglandulosae series. An interesting feature of the group is the dimorphic foliage, an adaptation to desert climate achieved by many Leguminosae of different groups. The leaves of the current season are solitary, but subtend in their axil a compressed-conical bud which, lying dormant the first year, develops during the second into a brachyblast, taking the form of a fascicle of small leaves intermixed with setiform stipules. Three of Britton’s five Palmerocassias are almost or quite confined to the Chihuahuan Desert, but are so closely related that they may best be inter- preted as geographic varieties of a widespread C. wislizeni; their differen- tial characters appear in the key following. The perhaps also closely related C. pringlei Rose, differing in the often more flexuous branches and greatly elongated pod (18-29 not 8-16 cm long) represents this complex in the Balsas Depression and low valleys of Oaxaca; while C. wnijuga Rose, locally en- demic in southern Puebla, is well marked by the combination of villous pube- scence and unijugate leaflets with revolute margins. A parallel reduction of the leaflets to one pair has occurred also in the Chihuahua Desert, described below as C. monozyx. The members of the group occurring in our range are Fig. 1. Cassia ripleyana Irwin & Barneby. Habit X1; sepals + petal (ventral view) X2; ovary and seed X5. 16 separable as follows (Fig. 2): 1. Leaflets 2-6 pairs (C. wislizeni, sens. lat.) 2. Pubescence of young branchlets, leafstalks, and (if present) of leaflets appressed or forwardly subappressed. Stipules of primary leaves persistent; leaflets of primary leaves most- ly 3 pairs, the largest 3-10 mm long, on lower (or both) faces coarsely penninerved; s.-e. Chihuahua (Sas. del Diablo, Hechiceros, Encinillas) and Chihuahua Desert n. from middle Conchos valley to Rio Grande valley in trans-Pecos Texas (Hudspeth, Jeff Davis, Presidio cos), the adjoining s. corners of New Mexico and ie and thence w. to the Magdalena valley in n. Sonor .C. wislizeni var wislizeni Stipules of primary leaves aie ie eae of primary leaves 3-5 (-6) pairs, the larger ones 5-15 mm long, the venation of the lower face not or scarcely elevated, commonly immersed; Chihuahua Desert, s. and s.-e. from n.-w. Zacatecas and extreme s. Coahuila to n. Querétaro, e. feebly into s. Nuevo Leén and extreme s.-w. Tamauli- c. C. wislizeni var. painteri w w pas. 2. punexeenee of young antics eset. and leaflets spreading, villo- sulous; leaflets in number, size, and venation as in var. wislizeni; local around edge of Mapimi Depression in extreme s.-w. Coahuila and ad- jacent Durango. . fo. hl. AD. C. wislizent var. villosa 1. Leaflets 1 pair; w. spite and s.-centr, Coahuila. . . . 2. C. monozyx CASSIA WISLIZENI A. Gray var. villosa (Britton) Irwin & Barneby, stat nov. Palmerocassia villosa Britton in N. Amer. FI. 23 (4): 254. 1930. CASSIA WISLIZENI A. Gray var. painteri (Britton) Irwin & Barneby, stat. nov. Palmerocassia painteri Britton in N. Amer. FI. 23 (4):254, 1930. CASSIA monozyx Irwin & Barneby, sp. nov., C. wislizeni Gray arcte affinis imprimis foliolis 1 (nec 2-3)-jugis primo intuitu absimilis, a C. unijuga Rose pube sparsa appressissima vel subnulla (nec copiosa molli patula) pat- riaque aliena distat—MEXICO. Coahuila: limestone ridge, 1800 m, Valle Seco in Sierra da la Paila, mun. General Cepeda, 7.VI.1944, G. Hinton 16565.—Holotypus, NY. Erect, stiffly branched, subspinescent shrubs to 1 m, the old trunks fus- cous glabrous, the young branchlets strigulose with appressed hairs to 0.25 mm, early glabrate, the foliage glabrous or the leaflets thinly strigulose dor- sally. Leaves solitary only on young branchlets, mostly fasciculate from knotty brachyblasts, 5-15 mm, stipules setiform, 1-4 mm, persistent; petiole (1-) 1.5-5 mm; leaflets 1 pair obovate or cuneate-obovate, 3-9 (10) mm, at apex commonly apiculate, rarely retuse and mucronulate, the blades stiffly chartaceous, yellow-green or glaucescent, the midrib and 3-5 (-6) pairs of secondary veins prominulous both faces. Racemes loosely 1-4-flowered, axil- lary to reduced upper leaves, forming a small panicle, the axis often shorte ! \ ’ aa i ~ \ < Besos N ! A x ) = a J a rm m™ \ Texa a aay any a \ a 4 \ L re 4 \ s ‘ om j Tey | H \ f a ae é ‘ a ‘ _~ | | a | 1 uy - Ww \ S } oo a ee a , \ ; af ' ff ( ws Lf f | : fe vA ’ é fy | 7 f ae | re e N ‘ nH ~. qe e e : \ J Zi ! ce iS D4 fl) od : oy S vv / | + Y h i] v | XQ / v \ 7 Y ¥ | \ ~~ \ « w \ Cassia wislizeni ee v v var. wislizeni a \- ( v Vy 4 var. villosa x var. painteri ¥ Cassia monozyx ) | / 4 Vv Fig. 2. Distribution of Cassia monozyx and the varieties of Cassia wislizent. than the flexuous pedicels; buds globose, except for the minutely ciliolate sepals glabrous; sepals oblong-elliptic to suborbicular obtuse, 4.5-7 mm, petaloid-margined; petals yellow, the largest + 18-20 mm; ovary glabrous; ovules + 10-15; pod linear, 6-10 X 0.65-0.8 cm, contracted at base into a short stipe, the purplish-nigrescent lustrous valves expressed over the rhombic- compressed, dull brown seeds. Gravelly clay slopes of bajadas and canyons at and near base of calcareous mountains, 750-1500 m, w.-centr. to s.-centr. Coahuila (Sa. Mojada e. to Sa. de la Madera, s. to Sa. de la Paila). Material seen: COAHUILA. Sa. Planchada n. of Esmeralda, 16.VIII.40 (1lvs), Johnston & Muller 343 (TEX). Sa. Mojada s. of Esmeralda, 1.1X.72 (fl, fr), Chiang & al. 9066 (NY). Canon de Calabazas in Sa. Mojada, 6.v.73 (fl), M. C. Johnston & al. 10883 (NY). Sa. de la Madera 8 km w. of Cuatro Ciénegas, 5.VIII.73 (fl), M. C. Johnston 12084F (NY). 138 km s.-w. of Cuatro Ciénegas, 11. VI. 72 (fl), Chiang & al. 7633 (NY). Cuatro Ciénegas, 23.VIII.39 (fl), E. M. Marsh 2060 (TEX). 7 mi. w. of Sacramento on road from Mon- 18 clova to Cuatro Ciénegas, 12.VI.55 (fl), M. C. Johnston 2584A (TEX). Sa. de Organos 62 km w.-s.w. of Cuatro Ciénegas, 8.VIII.73 (fl, fr), Hendrickson 12170b (TEX). n.-w. of Las Delicias, 29. VIII.71 (f1), Henrickson 6098b (TEX). Sa. de los Alamitos 11 mi. n. of Australia, 13.VI.72 (fl, fr), Chiang & al. 7723 (TEX). Closely akin to C. wislizeni Gray and C. unijuga Rose, in habit and details of flower and fruit resembling both, but different from the first in the con- stantly bifoliolate leaves, those of the primary branchlets and secondary short-shoots being essentially alike, and from the latter in the sparse appress- ed, not copious and villous pubescence of the young stems and foliage, and in the smaller, firmer, more veiny, non-revolute leaflets. The two bifoliolate species, C. unijuga and C. monozyx, are widely separated geographically, the former being endemic to the Tehuacan Desert in Puebla, the latter to a small segment of the Chihuahuan Desert in Coahuila, distant 900 km apart. The accompanying map shows the vicariant dispersal of C. monozyx and C. wislizeni with its varieties which together form a replacement series occu- pying virtually the whole floor of the Chihuahuan Desert. Ill. Sectio CHAMAECRISTA Bentham emend. Irwin CASSIA parralensis Irwin & Barneby, nom. nov. Chamaecrista goldmanii Britton & Rose in N. Amer. FI. 23 (5): 285. 19380. Non Cassia goldmanii Rose (1906). REFERENCES BENTHAM, G. 1871. Revision of the genus Cassia. In Trans. Linn. Soc. (London) 27: 3-591, t. 60-63. IRWIN, H. S. & B. L. TURNER. 1960. Chromosomal relationships and taxonomic con- Am N, H. S. 8 siderations in the genus Cassia. In Am. J. Bot. 47 (4): 309-318. AN SEM STUDY OF LEAF SURFACE PUBESCENCE IN THE SOUTHEASTERN TAXA OF PERSEA’ B. EUGENE WOFFORD AND RONALD W. PEARMAN Department of Botany University of Tennessee Knoxville, Tennessee Scanning electron microscope (SEM) studies of anatomical and morpholo- gical structures are rapidly appearing in the systematic literature. Although most of these studies have dealt with spore or pollen grain architecture, Faust and Jones (1973), in Vernonia, and Mulligan (1971), in Draba, recently used the SEM for systematic purposes in examining leaf trichomes. For the most part, however, SEM studies of anatomical structures associated with leaf surfaces, i.e., distribution and morphology of hairs, glands, stomata, etc., have been generally ignored by systematists. SEM studies of leaf pubescence were included as part of a biosystematic study of the genus Persea (Lauraceae) in the southeastern United States (Wofford, 1973). On the basis of these and other data, P. palustris (Raf.) Sarg., P. humilis Nash, and P. borbonia (L.) Spreng, are recognized as dis- tinct species and P. littoralis Small is considered to be nothing more than a coastal dune ecotype of P. borbonia. According to Small (1903), the leaf blades of P. borbonia are ‘‘finely reticulated beneath and mainly over thrice as long as wide’”’ while the leaf blades of P. littoralis are ‘‘not reticulated and mainly about twice as long as wide.’’ These differences, however, are incon- sistent and clearly represent morphological adaptations to the environmental conditions operative on coastal dunes. METHODS Lower leaf surfaces (ca. 5 mm”) were mounted with plastic cement to standard 34 inch aluminum studs. The mounted specimens were coated with vaporized carbon and gold in vacuo using a Denton vacuum coater equipped with a random rotating head. The specimens were viewed on an AMR model 900 Scanning Electron Microscope provided by the Department of Chemical and Metallurgical Engineering, The University of Tennessee, Knoxville. DISCUSSION The relatively small, unicellular, cupreous, filiform leaf trichomes in the southeastern taxa of Persea, provide the most reliable set of characters for readily distinguishing species. Several characters elucidated using the SEM, 1 Contributions from the Botanical Laboratory, University of Tennessee, Series Number 449, SIDA 6(1): 19—23. 1975. 21 e.g., trichome length and density, are not adequately resolved with the dissecting microscope. Persea palustris (Fig. 1), the most widespread and commonly encountered species, occurs throughout the Coastal Plain from eastern Virginia south to Florida and west to eastern Texas. This species possesses lanate trichomes that easily separate it from the sericeous type of P. humilis (Fig. 2) and the strigose type of P. borbonia (Fig. 3). The trichomes of P. palustris are mod- erately distributed over the lower leaf surface. Persea humilis (Fig. 2), a central Florida endemic, is restricted to the evergreen oak-sand pine-scrub community. The moderately falcate, sericeous hairs of this species resemble those of P. borbonia (Fig. 3) with the excep- tions of being denser and much longer. Also, the trichomes of P. humilis are about half as long as those of P. palustris. The shiny, cupreous trichomes of this species provide a reliable field character and make the common name ‘silk bay’’ an understandable one. As Nash (1895) stated, ‘‘Persea humilis, a most beautiful little shrub, makes itself very conspicuous by its bright brown silky pubescence, which is noticeable a long way off.” Persea borbonia (Fig. 3) has a distribution similar to that of P. palustris except that it is absent from Virginia and is restricted primarily to the outer Coastal Plain. The strigose hairs of this species, as previously stated, resem- ble those of P. humilis but are shorter and much less dense. Persea littoralis, sensu Small (Fig. 4), now considered only a sand dune variant of P. borbonia, was found to show no differences in pubescence type, length, and density from that of P. borbonia. In a statistical analysis of trichome length, using the Duncan’s Multiple Range Test, Wofford (1973) found significant differences among all three species but no differences between P. borbonia and its sand dune ecotype (P. littoralis). Mean values (mm.) for each taxon are: P. littoralis, 0.14; P. borbonia, 0.16; P. humilis, 0.26; and P. palustris, 0.58. Intraspecific varia- tions in trichome density were observed but are considered to be a function of age rather than being genetically based. The pubescence type of P. palustris is quite distinct from that of the other taxa examined in being much longer and held in an upright position. The hairs of P. borbonia (including P. littoralis) and P. humilis are similar in type but are denser and much longer in P. humilis. It is not the intent of this paper to suggest one-character taxonomy but rather to point to the taxonomic and systematic potential of SEM leaf surface studies since the specimen can be more critically examined and the SEM micrographs are far superior to other methods of illustration. Fig. 1-2. Scanning electron micrographs of Persea pubescence. Fig. 1, P. palustris. Fig. 2, P. humilis. REFERENCES eee W. Z. and S. B. cee JR. 1973. The systematic value of ere complements a North aa ans group of Vernonia (Compositae). Rhodora 75: -528. MULLIGAN, G. 1971. Cee ag es i the aoe roe ee cana, D. cinera and D. pea ioe in Canada and Alas nad. J. 89-93 NASH, G. V. 1895. Notes on some Florida ee Bull. ees ee Club 22: 141-161. SMALL, J. - 1903. Flora of the southeastern United States. Published by the author, New York. 1370 p. WOFFORD, . E. 1973. A biosystematic study of the genus Persea (Lauraceae) in = southeastern United States. Doctoral Dissertation, University of Tennessee, Knoxvi Fig. 3-4. Scanning electron micrographs of Persea pubescence. Fig. 3, P. borbonia. Fig. 4, P. littoralis, sensu Small. A BIBLIOGRAPHY OF MANUALS AND CHECKLISTS OF AQUATIC VASCULAR PLANTS FOR REGIONS AND STATES IN THE CONTERMINOUS UNITED STATES RONALD L. STUCKEY Department of Botany, College of Biological Sciences, The Ohio State University, Columbus, Ohio 43210 In the conterminous United States considerable interest has been shown in the identification and cataloguing of aquatic vascular plants, especially since 1940. This interest has undoubtedly been fostered because of the development of various areas of the aquatic sciences, such as aquatic botany, wildlife and fisheries management, wildlife conservation, and environmental protection agencies. The need for basic information on identification manuals and checklists of aquatic vascular plants for all areas of the United States has therefore become increasingly more frequent. This paper provides a ready reference list of all those known manuals and checklists of aquatic vascular plants covering the following geographical categories: (1) the conterminous United States, (2) five different regions of the United States, and (3) 16 individual states. A citation of a bibliography on aquatic vascular plants is included for Ohio. The five regions that have aquatic vascular plant manuals essentially cover the entire United States, but not all of these publications are of equal scope with similar objectives, and some are older than others and in need of revision. Certainly more work at the state level is necessary, even in some of those states that already have manuals and checklists. The bibliography presented here gives some idea of the extent and state of knowledge in this science and hopefully will provide a guide for future basic and applied research on aquatic vascular plants. Grateful appreciation will be given for calling overlooked references to my attention. UNITED STATES Hotchkiss, Neil. 1967. Underwater and Floating-leaved Plants of the United States and Canada. Bureau of Sport Fisheries and Wildlife, Washington, D.C. Resource Publication 44. 124 pp. (Reprinted 1972, in combination with the following work, by Dover Publications, Inc., New York, under the title, ‘Common Marsh, Underwater and Floating-leaved Plants of the United States and Canada.’’) Hotchkiss, Neil. 1970. Common Marsh Plants of the United States and Can- Contribution from the Department of Botany (Paper No. 871) and the Herbarium of The Ohio State University, Columbus, Ohio 43210. SIDA 6(1): 24—29. 1975. 25 ada. Bureau of Sport Fisheries and Wildlife, Washington, D.C. Resource Publication 93. 99 pp. (Reprinted 1972, in combination with the previous work, by Dover Publications, Inc., New York, under the title, “Common Marsh, Underwater and Floating-leaved Plants of the United States and Canada.’’) Muenscher, Walter Conrad. 1944. Aquatic Plants of the United States. Com- stock Publishing Company, Ithaca, New York. 374 pp. (Reprinted 1948, 1964, 1967.) Prescott, G. W. 1969. How to Know the Aquatic Plants. Wm. C. Brown Pub- lishers, Dubuque, Iowa. 171 pp. Weldon, L. W., R. D. Blackburn, and D. S. Harrison. 1969. Common Aquatic Weeds. United States Department of Agriculture, Agricultural Research Service, Washington, D.C. Agricultural Handbook No. 352. 43 pp. (Re- printed 1973 by Dover Publications, Inc., New York.) REGIONAL CENTRAL UNITED STATES Lindstrom, Lester Emerald. 1968. The Aquatic and Marsh Plants of the Great Plains of Central North America. Ph.D. Dissertation, Kansas State Uni- versity, Manhattan. 247 pp. NORTHEASTERN UNITED STATES Fassett, Norman C. 1940. A Manual of Aquatic Plants. McGraw Hill Book Company, ie York. 382 p Fassett, Norman C. 1957. A ae of Aquatic Plants. With Revision Appen- dix by Eugene C. Ogden. University of Wisconsin Press, Madison. 405 pp. (Reprinted 1960, 1966, 1969, 1972.) RTHWESTERN UNITED STATES Steward, Albert N., LaRea J. Dennis, and Helen M. Gilkey. 1960. Aquatic Plants of the Pacific Northwest With Vegetative Keys. Oregon State Col- lege, Studies in Botany, Number 11. Corvallis. 184 pp. Steward, Albert N., LaRea J. Dennis, and Helen M. Gilkey. 1963. Aquatic Plants of the Pacific Northwest With Vegetative Keys. 2nd ed. Oregon State University Press, Corvallis. 261 pp. SOUTHEASTERN UNITED STATES yles, Don E., J. Lynne Robertson, Jr., and Garnet W. Jex. 1944. A Guide and Key to the Aquatic Plants of the Southeastern United States. United States Public Health Bulletin No. 286. 151 pp. (Reprinted 1963 by Bureau of Sport Fisheries and Wildlife, Washington, D.C. Circular 158.) SOUTHWESTERN UNITED STATES Correll, Donovan S., and Helen B. Correll. 1972. Aquatic and Wetland Plants of Southwestern United States. Water Pollution Control Research Series, Environmental Protection Agency. United States Government Printing ice, Washington, D.C. THE STATES CALIFORNIA Mason, Herbert L. 1957. A Flora of the Marshes of California. University of California Press, Berkeley and Los Angeles. 878 pp. COLORADO Matsumura, Y., and H. D. Harrington. 1955. The True Aquatic Vascular Plants of Colorado. Colorado Agricultural Experiment Station, Technical Bulletin 57. 130 pp. ILLINOIS Winterringer, Glen S., and Alvin C. Lopinot. 1966. Aquatic Plants of Illinois. Department of Registration and Education, Illinois State Museum Divi- sion and Department of Conservation, Division of Fisheries. Illinois State Museum Popular Science Series Vol. VI. 142 pp. IOWA Beal, Ernest O, 1952. The Distribution of Aquatic Monocotyledons in lowa. M.S. Thesis, University of Iowa, Iowa City. Beal, Ernest O. 1954. Aquatic Monocotyledons of Iowa. Proceedings Iowa Academy of Science (1953) 60: 89-91. Beal, Ernest O., and Paul H. Monson. 1954. Marsh and Aquatic Angiosperms of Iowa. State University of Iowa Studies in Natural History 19(5): 1-95. Monson, Paul Herman. 1952. The Marsh and Aquatic Dicotyledonous Flora of Iowa. M.S. Thesis, lowa State College [University], Ames. Monson, Paul H. 1954. A Preliminary Report on the Marsh and Aquatic Dico- tyledonous Flora of Iowa. Proceedings Iowa Academy of Science (1953) 60: 200-201 KANSAS Wilson, James S., and Robert J. Boles. 1967. Common Aquatic Weeds of Kan- sas Ponds and Lakes. The Emporia State Research Studies 15(3): 1-36. LOUISIANA Thieret, John W. 1972. Aquatic and Marsh Plants of Louisiana: i akc Louisiana Society for Horticultural Research Journal 13(1): MICHIGAN Voss, Edward G. 1966. Checklist of Aquatic Vascular Plants of Michigan. Her- barium, The University of Michigan, Ann Arbor. Mimeographed. 7 pp. Voss, Edward G. 1967. A Vegetative Key to the Genera of Submersed and Floating Aquatic Vascular Plants of Michigan. The Michigan Botanist 6: 35 MINNESOTA Carlson, Richard A., and John B. Moyle. 1968. Key to the Common Aquatic Plants of Minnesota. Minnesota Department of Conservation, Division of Game and Fish Section of Technical Services. Special Publication No. 27 53. 64 pp. Moyle, John B., and Neil Hotchkiss. 1945. The Aquatic and Marsh Vegetation of Minnesota and its Value to Waterfowl. Minnesota Department of Con- servation, Technical Bulletin No. 3. 122 pp. NEW JERSEY Essback, Alban R., Donald N. Riemer, and Donald A. Schallock. 1965. Aqua- tic Vegetation of New Jersey. Part II. Problems and Methods of Control. College of Agriculture, Rutgers—The State University, New Brunswick, Extension Bulletin 382. 98-107. Fairbrothers, David E., and Edwin T. Moul. 1965. Aquatic Vegetation of New Jersey. Part I. Ecology and Identification. College of Agriculture, Rut- gers—The State University, New Brunswick, Extension Bulletin 382. 1-97 pp. NEW YORK Clausen, R. T. 1940. Aquatic Vegetation of the Lake Ontario Watershed, pp. 167-187, annotated list, pp. 180-187. In A Biological Survey of the Lake Ontario Watershed Supplemental to Twenty-ninth Annual Report, 1939. Conservation Department, State of New York. Biological Survey No. 16 (1939). 261 pp. + maps. McVaugh, Rogers. 1938. Aquatic Vegetation of the Allegheny and Chemung Watersheds, pp. 176-195, annotated list, pp. 189-195. In A Biological Survey of the Allegheny and Chemung Watersheds Supplemental to Twenty-sev- enth Annual Report, 1937. Conservation Department, State of New York. Biological Survey No. 12 (1937). 287 pp. + maps. Muenscher, W. C. 1927. Vegetation of Silver Lake and Conesus Lake, pp. 66- 67, list of the larger plants, p. 86. In A Biological Survey of the Genesee River System Supplemental to Sixteenth Annual Report, 1926. Conserva- tion Department, State of New York. [Biological Survey No. 1 (1926)]. 100 pp. + maps. Muenscher, W. C. 1928. Vegetation of Cayuga and Seneca Lakes, pp. 243-248, list, pp. 246-248. In A Biological Survey of the Oswego River System Sup- plemental to Seventeenth Annual Report, 1927. Conservation Department, State of New York. [Biological Survey No. 2 (1927)]. 248 pp. -+ maps. Muenscher, W. C. 1929. Vegetation of the Niagara River and the Eastern End of Lake Erie, pp. 189-197, list, pp. 195-197. In A Biological Survey of the Erie-Niagara System Supplemental to Eighteenth Annual Report, 1928. Conservation Department, State of New York. [Biological Survey No. 3 (1928)]. 244 pp. + maps. Muenscher, W. C. 1930. Aquatic Vegetation of the Lake Champlain Watershed, pp. 164-185, list, pp. 178-185. In A Biological Survey of the Champlain Wa- tershed Supplemental to Nineteenth Annual Report, 1929. Conservation Department, State of New York. [Biological Survey No. 4 (1929)]. 321 pp. -+ maps. 28 Muenscher, W. C. 1931. Aquatic Vegetation of the St. Lawrence Watershed including the Grass, St. Regis, Salmon, Chateaugay Systems and the St. Lawrence between Ogdensburg and the International Boundary, pp. 121- 144, list, pp. 143-144. In A Biological Survey of the St. Lawrence Water- shed Supplemental to Twentieth Annual Report, 1930. Conservation De- partment, State of New York. Biological Survey No. 5 (19380). 261 pp. + maps Muenscher, W. C. 1932. Aquatic Vegetation of the Oswegatchie and Black River Watersheds (Including the eastern end of Lake Ontario and the upper stretch of the St. Lawrence River), pp. 199-221, annotated list, pp. 213-221. In A Biological Survey of the Oswegatchie and Black River Sys- tems Supplemental to Twenty-first Annual Report, 1931. Conservation Department, State of New York. Biological Survey No. 6 (1981). 344 pp. + maps. Muenscher, W. C. 1933. Aquatic Vegetation of the Upper Hudson Watershed, pp. 216-238, annotated list, pp. 231-238. In A Biological Survey of the Up- per Hudson Watershed Supplemental to Twenty-second Annual Report, 1932. Conservation Department, State of New York. Biological Survey No. 7 (1982). 341 pp. + maps. Muenscher, W. C. 1934. Aquatic Vegetation of the Raquette River Watershed, pp. 209-221, annotated list, pp. 216-221. In A Biological Survey of the Ra- quette Watershed Supplemental to Twenty-third Annual Report, 1933. Conservation Department, State of New York. Biological Survey No. 8 (1933). 301 pp. + maps. Muenscher, W. C. 1985. Aquatic Vegetation of the Mohawk Watershed, pp. 228-249, annotated list, pp. 242-249. In A Biological Survey of the Mohawk- Hudson Watershed Supplemental to Twenty-fourth Annual Report, 1934. Conservation Department, State of New York. Biological Survey No. 9 (1934). 379 pp. + maps. Muenscher, W. C. 1936. Aquatic Vegetation of the Susquehanna and Delaware Areas, pp. 205-221, annotated list, pp. 216-221. In A Biological Survey of the Delaware and Susquehanna Watersheds Supplemental to Twenty-fifth Annual Report, 1935. Conservation Department, State of New York. Bio- logical Survey No. 10 (1935). 356 pp. + maps. Muenscher, W. C. 1937. Aquatic Vegetation of the Lower Hudson Area, pp. 231- 248, annotated list, pp. 240-248. In A Biological Survey of the Lower Hud- son Watershed Supplemental to Twenty-sixth Annual Report, 1936. Con- servation pekeey State of New York. Biological Survey No. 11 (1936). 373 pp. -- maps. Muenscher, W. C. 1939. Aquatic Vegetation of Long Island Waters, pp. 88-101, annotated list, pp. 95-101. In A Biological Survey of the Fresh Waters of Long Island Supplemental to Twenty-eighth Annual Report, 1938. Conser- vation Department, State of New York. Biological Survey No. 13 (1938). 128 pp. + maps. OHIO Stuckey, Ronald L. 1973. Bibliography of Aquatic Flowering Plants of Ohio. Ohio Biological Survey Informative Circular No. 2. 11 pp. OKLAHOMA DeGruchy, James H. B. 1938. A Preliminary Study [and Checklist] of the Larger Aquatic Plants of Oklahoma with Special Reference to Their Value in Fish Culture. Oklahoma Agricultural and Mechanical College Agricultural Experiment Station, Stillwater, Oklahoma. Technical Bulletin No. 4. 31 pp. including 68 figures + one insert [which contains the check- list]. PENNSYLVANIA Cole, Herbert, Jr., R. G. Wingard, R. L. Butler, and A. D. Bradford. 1967. Aquatic Plants: Management and Control in Pennsylvania. The Pennsyl- vania State University, University Park. Natural Resource Series Spe- cial Circular 79. 32 pp. SOUTH DAKOTA Martin, John H. 1965. The Marsh and Aquatic Monocotyledons of South Da- kota. Proceedings South Dakota Academy of Science 44: 180-184. TENNESSEE Robinson, Franklin D., and Royal E. Shanks. 1959. Checklist of Vascular Aquatic Plants of Tennessee. Journal Tennessee Academy of Science 34: 58-65 WEST VIRGINIA Clovis, Jesse F. Without Date. Plants of West Virginia waters. West Virginia Department of Agriculture, Charleston. 15 pp. TYPIFICATION AND DISTRIBUTION F THE VARIETIES OF GNAPHALIUM HELLERI BRITTON (COMPOSITAE-INULEAE) WM. F. MAHLER Herbarium, aaa Methodist University Dallas, Texas 75275 A study was initiated to elucidate the status of a collection—Mahler 6850 (SMU)—of Gnaphalium helleri Britton taken during a botanical survey of a lake site and strip mining area in northeastern Texas (Mahler, 1973; 1974). I am indebted to the curators of the following herbaria: FSU, GA, GH, LL, MICH, MO, NCU, NY, TENN, TEX, VDB, Northeastern Louisiana Univer- sity, Monroe, and University of Southern Mississippi, Hattiesburg. Gratitude is also extended to R. C. Barneby. KEY TO TAXA 1. Stems white, tomentum loose to dense, obscuring glandular pubescence... . ... G, obtusifolium 1. Stems greenish, not tomentose, visibly gienduiar . 2. . . GG. helleri 2. Stems glandular-villous; plants of the southeastern United States . . 2...) .) .) svar. helleri 2. Stems glandular-puberulent; plants of the northeastern United States . 2. 2. 2.) 2) svar. micradenium yNAPHALIUM OBTUSIFOLIUM L., Sp. P1.2: 851. 1753. G. polycephalum Michx. The white stems of this species possess hairs matted parallel to the stem. Upon removal of the white mat, the glandular-pubescence is usually promi- nent. GNAPHALIUM HELLERI Britton Bull. Torr. Bot. Club 20:280, 1893. LECTOTYPE selected: Northwest, Nor- folk County, Virginia, A. A. Heller s.n. 23 Sep. 1892, annotated by Britton (NY). SYNTYPES: Northwest, Norfolk County, Virginia, A. A. Heller s.n. 23 Sep. 1892 (NY, MO); Southern United States, Georgia, Boykin s.n., an- notated by Gray and Britton (NY). G. polycephalum Michx. var. 8B, Torr. & Gray, Fl. N. Amer. 2:427, 1843. G. polycephalum Michx. var. helleri (Britt.) Fernald, Rhodora 10: 94. 1908. G. obtusifolium L. var. helleri (Britt.) Blake, Rhodora 20: 72. 1918. Annuals, aromatic; stems erect, green, pubescence of short, erect, stipitate, glandular hairs or of long, spreading, white, jointed, gland-tipped hairs; SIDA 6(1): 30—32. 1975. Fig. 1. Documented distribution of Gnaphalium helleri Britton: var. helleri (squares), var. micradenium (stars). leaves sessile, green and stipitate glandular above, white tomentose to gla- brate below, narrowly elliptic, margins entire to undulately toothed; inflore- scence of heads in wide to narrow corymbs; heads peduncled, not glomerate at maturity; involucres 7-8 mm long; phyllaries white, rounded to erose apic- ally; achenes brown, glabrous; pappus bristles individually deciduous. Aug.- Nov. Openings and clearings, pine or pine-hardwood forest. 32 The absence of the white tomentum was regarded as being only of varietal value by Fernald (1908) and Blake (1918). However, the distribution patterns, absence of intermediates, and the conspicuous pubescence patterns result in the recognition of the glandular-viscid taxa as one species. GNAPHALIUM HELLERI Britton var. HELLERI Illustrations (photographs); Fernald, Rhodora 38: Plate 434. Figs. 8, 9. 1936. Distribution Map: Fig Main stems and lower floral branches greenish and glandular-villous with spreading, white, jointed, gland-tipped hairs. This variety is well documented in Arkansas (Delzie Demaree collections). Louisiana (R. D. Thomas collec- tions), and the Carolinas (Radford, Ahles, & Bell, 1968). It is probable that this is a result of ‘“‘disjunct collections’ instead of disjunct ranges (Demaree, pers. comm.). GNAPHALIUM HELLERI Britton var. micradenium (Weatherby) Mahler, comb. nov. G. obtusifolium L. var. micradenium Weatherby, Rhodora 25: 22. 1923. TYPE: Dry, sandy openings among scrub oaks, Barnstable, Mass., 7 Oct 1917, Fernald 15870 (GH). lllustrations (photographs): Fernald, Rhodora 38: Plate 434. Figs. 6, 7. 1936. Distribution Map: Fig. 1 Stem, pubescence of short, straight, gland-tipped, spreading hairs, not con- spicuously whitish or flattened basally. Two specimens examined from Lou- isiana and western New York did not have specific localities and are plotted adjacent to their respective states in Figure 1. REFERENCES F. 1918. Notes on the Clayton Herbarium. Rhodora 20; 21-28, 48-54, 65-73. BLé ey Rhodora 10: 46-55 FE RNAL Db, M. L. 1908. Notes on some plants of northeastern America. ‘ena R, W. FL 1973. Botanical survey of the Lake Monticello area. SMU = Contrib. Anewop: No. 9. L974. Guaphalinin cei Britton (Compositae - Inuleae) in the Texas The Southwestern Nat. 19: RADFORD, A. H. E. AHLES, me Cc. R. BELL. 1968. Manual of the vascular flora of the re University of North Carolina Press, Chapel Hill. STUDIES IN THE RANUNCULACEAE OF THE SOUTHEASTERN UNITED STATES. Ill. CLEMATIS L CARL S. KEENER 202 Buckhout Laboratory The Pennsylvania State University University Park, Pennsylvania 16802 While I was completing a treatment of the Ranunculaceae for the forthcom- ing Vascular Flora of the Southeastern United States, certain nomenclatural and taxonomic decisions were made; these call for additional clarification. This treatment of Clematis is conservative in that I am following a species concept allowing for considerable intraspecific population variability. Unless populations show discrete geographic patterns correlated with morphological discontinuities, I see no compelling reason to describe subspecies, varieties, etc. Furthermore, unless breeding studies suggest intraspecific relationships, I prefer to recognize morphologically-defined species. The aim throughout is to provide a rationale for certain taxonomic and nomenclatural conclusions, a key to the species, and data on their distribution in southeastern United States. Clematis can be distinguished from other genera of Ranunculaceae by its opposite leaves, elongate and generally plumose styles, apetalous actinomor- phic flowers, and usually four showy valvate petaloid sepals (Tamura, 1967). I am recognizing 16 species within the southeastern United States (as com- pared to about 300 species for the world; Buchheim, 1964), an area bounded by and including Louisiana, Arkansas, Kentucky, West Virginia, Maryland, and Delaware. Aside from the taxonomic treatments by James (1883), Kuntze (1885), and Gray (1895), there have been no recent comprehensive treatments of Clem- atis of eastern North America. On the other hand, in recent years several groups represented in this region have been studied from a critical taxonom- ic standpoint, e.g., section Viorna (Erickson, 1943), the C. ochroleuca com- plex (Wherry, 1931; Fernald, 1943; Keener, 1967), and section Atragene in North America (Pringle, 1971). In addition, Fish (1970) studied megagameto- genesis in 26 species of Clematis (including seven occurring within the south- eastern United States) with a view toward examining certain taxonomic and sed on a manuscript and notes compiled for the forthcoming Vascular of the Reb eine United States. In general, the format follows Radford ef al. (196 Any suggestions for improving this treatment should be sent to me so that necessary corrections and additions can be made before the Vascular Flora is in’ pr > Contribution No, 121 from the Department of Biology, The Pennsylvania State Uni- versity. SIDA 6(1): 33—47. 1975 34 phylogenetic implications from his data. Nevertheless, there is still need for comparative biosystematic studies of the C. virginiana group in North Amer- ica, C. viorna and its allies, and C. crispa. Moreover, although there are chromosome counts for 10 species (Table 1), counts for six species native to the southeastern United States are still lacking. Table I. Chromosome Numbers of the Species of Clematis in the Southeastern t United States Chromosome Speci ies Number Reference C. addisonii 2n = 16 wregory (1941) C..albicoma 2n = 16 Keener (1966) C. baldwinit 2n — 16 Keener (1974) C. catesbyana no report C. coactilis 2n = 16 Keener (1966) C. crispa 2n = 16 Gregory (1941), Smith (1965) C. glaucophylla no report C. maximowieziana 2n — 16 yregory (1941) (= C. paniculata Thunb., 2n = 48 Meurman & Therman (1939) C. dioscoretfolia Levl. n = 32 Gregory (1941) & Vaniot) C. occidentalis 2n = 16 Kurita (1956) (= C. verticillaris DC.) C. ochroleuca 2n — 16 Gregory (1941), Keener (1966) C. pitcheri no report C. reticulata no report C. versicolor no report C. viorna no report C. virginiana In — 16 Lindsay (1929), Gregory (1941), Kurita (1958), Bostick (1965) C. viticaulis 2n = 16 Keener (1966) KEY TO SPECIES la. F’owers perfect or imperfect, numerous, in cymose-paniculate inflore- scences; sepals whitish, spreading; filaments glabrous ae Ib. IF aspers perfect, usually solitary; sepals variously nulexed: erect to somewhat spreading; filaments pubescent . 4 2a. Flowers perfect; leaflets entire, seldom cleft, suey coriaceous; anth- ers 1.5 mm or more long. ; . 1. C. maximowicziana 2b. Flowers imperfect; leaflets speteels toothed or lobed, usually mem- branous; anthers less than 1.5 mm long oe ew we se we = oo 3a. Leaves 3-foliolate; leaflets usually coarsely serrate, but seldom conspicuously 3-lobed; achenes light to dark brown or greenish 3b. Leaves bitemnately compouna or pinnately 5. ‘liclane eatiets often co spicucusly 38-lobed or toothed, the lobes entire to coarsely ae achenes reddish-brown to dark blackish-purple . . 38. C. catesbyana .o ist) a) > 39 . Sepals thin, slightly spreading, not connivent; leaves ternate; petaloid staminodia sometimes present. . : 4. C. occidentalis . Sepals thick, usually erect and more or less eonnivent: leaves simple to pinnately compound; petaloid staminodia absent. . Plants erect herbs; leaves usually simple, entire to ane cleft, ee oin 6 all pinnate. . Plants aimbine vines or ececanunne nes Het eiecte Teawes. siamouly 11 pinnately compoun . Leaves green and ‘iStaliiy ube eeu panedihi the uppermost entire, 7 lobed or cleft, neither pinnate nor tendril-bearing. Leaves glaucous and glabrous beneath, the uppermost eomubaly pin- nate and tendril-bearing. . . . 16. C. addisonii . Leaves usually simple and entire, dances to jens ovate, the larg- est at least 2 cm wide; mature styles whitish to deep reddish-brown, the longest usually less than 5 cm long; mature peduncles usually less than 20 cm long. 8 . Leaves usually eaieass divided, or, if cane linear to cameo ovate, the largest segment or leaf usually less than 2 cm wide; mature styles light yellow, the longest more than 5 cm long; mature peduncles usually more than 20 cm long. . . . 9. C. baldwinti . Leaves of flowering material soft eibesect beneath, the largest 3-9 cm wide, hypostomatic; leaves of fruiting material usually light green with secondary and tertiary veins forming prominent reticulations on the upper surface. 9 Leaves of flowering material seationed sesok to sien beneath, the largest 2-5 cm wide, amphistomatic; leaves of fruiting material often dark green with the secondary and tertiary veins not forming prominent reticulations on the upper surface. . Stems and leaves villous; sepal backs medenaias sericeous- eiloas: mature styles yellowish-white to deep tawny, loosely spreading- recurved. . . . . 9d. C, ochroleuca N) nd_ leaves cca: ence halescre cone: sepal backs pen holosericeous; mature styles white to pale yellow, sharply recurved and flexuous. . 6. C. coactilis . Sepal backs villous; pubescence on er margins apes mature styles white to pale yellow, sharply recurved and flexous. 7. C. albicoma . Sepal backs finely puberulent; pubescence on carpel margins closely appressed-ascending; mature styles tawny to deep reddish-brown, loosely spreading-recurved. .. : . 8. C. viticaulis . Leaflets generally coriaceous, the Sedona and ne veins forming prominent reticulations on the upper surface. . Leaflets thin, membranous, the secondary and oe veins forming faint or indistinct reticulations on the upper surface. . Leaflets green and often soft-pubescent beneath; stem nodes and rachis joints finely to sparsely pubescent; sepal backs usually sericeous-canescent. 3 12b. Leaflets glaucous and ebro peneathe- stem wee nd raehis joints usually glabrous and glaucous; sepal backs usually glabrous. 12. C. versicolor 13a. Mature styles fess Gian 3 cm oe aapioits or rath appressed to spread- ing villous hairs, not evenly plumose throughout; leaflets generally subglabrous beneath, rarely densely soft-pubescent, more or less coarse- ly reticulate above... . 10. C. pitcheri 13b. Mature styles more than 3 cm iene. aes einennes with pilose hairs; leaflets soft-pubescent beneath, rarely glabrate, finely and prominently reticulate above. . 2... . . . LLC. reticulata lda. Sepals usually (2-)3-5 em long, ‘dilated distally into thin broad crisped margins up to 6 mm _ wide; mature styles sericeous, finely appressed-pubescent, rarely evenly plumose; peduncles us- — ually ebracteate. —. . . 13. C. crispa 14b. Sepals 1.2-2.5 cm long, he mareins iepenisee Bue not dilated distally; mature styles plumose; peduncles usually bracteate. . . . . 15 loa. Leaflets greenish and pubescent beneath; stem nodes and rachis joints pubescent; sepal backs finely pubescent. . . . . . 14. C. viorna 15b. Leaflets glaucous and glabrous beneath; stem nodes and rachis joints glabrous and glaucous; sepal backs usually glabrous... 16 16a. Plants climbing vines; leaves seldom simple, often 6-10 foliolate. fo a mee ©: em». & & « so 15. C. glaucophylla 16b. Plants erect or ascending; lower leaves simple, upper simple to 2-6 foliolate. . Boe ee ee 16. CL addisonii 1. C. MAXIMOWICZIANA Franch. & Savat. Thickets, fence rows, and waste places; widely scattered in pied. and cp., rarely mts. Ala., Ark., Ky., Md., Miss., N.C., S.C., Tenn., Va., W.Va. [Tex., Ill., Pa., N.J.]. C. dioscoreifolia Levl. & Vaniot—Fernald (1950), Glea- son and Cronquist (1963), Radford (1968); C. paniculata Thunberg—S (1933); incl. C. d. var. robusta (Carr.) Rehder—Fernald (1950). Cultivated escape; native to Japan. Even though two widely used names for this species appear in current manuals, nevertheless C. maximowicziana, published in 1879 by A. R. Fran- chet and P. A. L. Savatier (fnum. Pl. Japon. 2: 261), clearly has priority over C. dioscoreifolia Levl. & Vaniot (Fedde, Repert. Nov. Sp. 7: 339. 1939; cf. Ohwi, 1965). Furthermore, C. paniculata Thunberg (Trans. Linn. Soc. 2: 337. 1794) is preempted by C. paniculata J. F. Gmelin (Syst. 873. 1791), a completely different species. 2. C. VIRGINIANA L., Virgin’s-bower Low woods, stream banks, thickets, and waste places; all prov. SE except Del. and Fla. [ALL except Okla. and N.J.]. 37 A comprehensive comparative study of the population variability and breeding behavior of the short-styled virgin’s-bowers is sorely needed. Pend- ing such studies, I am recognizing three species, two in eastern North Amer- ica (C. catesbyana, C. virginiana) and one in western North America (C. ligusticifolia Nutt.) outside the range covered by this paper. The three species are apparently closely related (Table II; also cf. Fish, 1970) but usually with mature fruiting material they can be distinguished. Neverthe- less, intergradient forms occur, especially in Missouri, and, to date, the tax- onomy of this complex is not completely clear. 3. C. CATESBYANA Pursh Low sandy woods, river banks and upland woods; all prov. Ala. mts., Ark. mts., Fla., Ga. ep., Ky. IP., Miss. cp., N.C. cp. and pied., Va. mts. [Mo.|. Incl. C. micrantha Small—Small (1933), a small-flowered ex- treme; incl. virginiana L.—Radford (1968), in part; C. ligusticifolia sensu Radford, not Nuttall—Radford (1968). On the basis of leaf and achene characters this species appears distinct (see Table II), even though herbarium material frequently passes as C. li- gusticifolia or C. virginiana, A cytotaxonomic study of this species would be most instructive, especially in respect to its relationship with C. virginiana. Clematis micrantha Small, erected to accommodate small-flowered plants occurring in west peninsular Florida, does not appear to be sufficiently dis- tinct to warrant recognition. 4. C. OCCIDENTALIS (Hornemann) DC., Purple Clematis Open weods, banks, rocky calcareous slopes and cliffs, rare; chiefly mts. Del. ep., N.C., Va., [Ohio, Pa.]. Atragene americana Sims—Small (1933): C. verticillaris DC.—Fernald (1950), Gleason and Cronquist (1963), Radford (1968); incl. C. v. var. cacuminis Fern.—Fernald (1950). Pringle (1971) in working through the taxonomy of Clematis section Atra- gene in North America critically reviewed the nomenclature and persuasive- ly argued that the correct name for this distinctive species is C. occidentalis (Hornemann) DC. Pringle also showed that variety cacuminis, described by Fernald (1947) to accommodate small-flowering material collected from the Peaks of Otter, Bedford County, Virginia, was actually erected on the basis of immature material. Inasmuch as mature specimens from the same local- ity do not appear to differ significantly from collections made elsewhere, Pringle therefore regarded var. cacuminis as synonymous with C. occidental- is var. occidentalis (i.e., the wide-ranging eastern North American taxon). 5. C. OCHROLEUCA Aiton, Leatherflower, Curly-heads Shaded woods, dry open banks and clearings; chiefly cp. and pied. Ga., N.C., S.C., Va. Viorna ochroleuca (Aiton) Small—Small (1933). The eastern short-stemmed leatherflowers (nos. 5, 6, 7, 8; Table III) have been critically studied by Wherry (1931), Fernald (1943), and Keener (1967). Table Il. Comparison of Species in the C. virginiana Complex. Character C. virginiana C. catesbyana C. ligusticifolia Leaves 3-foliolate )-foliolate to biternate Pinnately (3-)5-7 foliolate; rarely biternate Leaflets: color Usually green to dark green length (cm) 2-10 3-6 shape Ovate to subcordate Subcordate to + lanceolate margin Usually coarsely serrate Usually 3-lobed to entire texture Membranous pubescence Glabrous to villous on veins beneath, plus scattered to abundant pubescence over entire abaxial surface Pale to yellowish-green (2-)3-6(-8) Lanceolate to lance-ovate or ovate Usually coarsely few-toothed or lobed Membranous to coriaceous Sparingly strigose on veins yeas otherwise usually glabro stomates Lower leaflet surface Upper and lower leaflet surfaces Peduncles: length (cm) 2-6 (avg. :3.9) 3.5-8 (avg. :4-7) 4-7 (avg. 15.3) pubescence Glabrous to tomentose; hairs Tomentose; hairs short Villous; hairs both long short (0.3-0.5 mm) (ca. 0.3 mm) (0.5-0.7 mm) and short Table II. Continued. Character C. virginiana C. catesbyana C. ligusticifolia Sepals: length Often longer than the fertile stamens Often shorter than the fertile stamens shape Oval to oblong-spatulate Linear to cbovate Narrowly oblong-lanceolate Achenes: color Light to dark brown Reddish or purplish-brown to Light brown to deep reddish-brown or greenish brown dark blackish-purple pubescence Sparingly short pilose to villous hirsute Densely pilose to crisp hairy Chromosome number 2n = 16 No report 2n — 16 Range Eastern North America Va. to Ky., south to Ark. and Fla. Western North America 40 As a group they are erect herbs (2-6 dm tall) with simple usually entire leaves and with mature fruits having plumose styles less than 5 cm long. Although these plants appear to be closely related, their evolutionary rela- tionships remain relatively obscure (Keener, 1967, 1970). Still, with mature fruiting material, the four species can be determined readily (See Table IIL). 6. C. COACTILIS (Fernald) Keener Shale barrens and thin limestone woods, rare; RV. West-central Va. C. albicoma Wherry var. coactilis Fern.—Fernald (1950), Gleason and Cron- quist (1963). This species combines the features of both C. ochroleuca (as to vegetative habit and leaf morphology) and C. albicoma (fruits) but differs by the pro- nounced pubescence on the stem, leaves, and sepals (Table III; Keener, 1967). Clematis coactilis is a rare species, occurring only in western Virginia, and it may represent a hybrid derivative of C. ochroleuca and C. albicoma (Keener, 1967, 1970). Until careful breeding studies involving all three taxa are undertaken, the evolutionary relationships of C. coactilis will remain obscure. 7. C. ALBICOMA Wherry, Leatherflower Shale barrens, rare; RV. Western Va. and adjacent W.Va. This is one of several strict shale barren endemics occurring only on ex- posed south-facing outcrops of Upper Devonian shales in Virginia and ad- jacent West Virginia (Keener, 1970). Although C. albicoma is relatively dis- tinct from C. ochroleuca and C. coactilis, it appears to be closest to the ex- tremely rare C. viticaulis (Seener, 1967). 8. C. VITICAULIS Steele Shale barrens, very rare; RV. Western Va. This species, restricted to only a few shale barrens in Bath and Rockbridge counties in western Virginia, is the rarest Clematis in the southeastern United States. It is closest to C. albicoma but differs from that species by its finely puberulent sepal backs and usually deep reddish-brown styles (‘‘achene tails’). 9. C. BALDWINIL Torrey & Gray Sandy flat pine woods; cp. Fla. Viorna baldwinii (Torrey & Gray) Small—Small (1933). This distinctive species is restricted to peninsular Florida. Long (1970) pointed out that population studies of C. baldwinii would be most desirable, especially to assess leaf shape and sepal size throughout the range. Although he recognized the larger flowered and broader leaved plants as var. latius- cula, there appears to be no geographic correlation with respect to the two varieties. Therefore, pending additional field study, I prefer to regard C. baldwinti as a unitary polymorphic species. 10. C. PITCHERI Torrey & Gray Low grounds, woods and thickets; all prov. Ark., Ky. ME. and IP., Tenn. IP. [Tex., Okla., Mo., Ill., Ind.]. Viorna pitcheri (Torrey & Gray) Small —Small (1933). Clematis pitcheri is a wide-ranging mid-central North American species having considerable morphological variability (Erickson, 1943). It enters the western portion of the range, occurring in Arkansas and in western Kentucky and Tennessee. Frequently C. pitcheri is confused with C. reticulata but the two species are generally readily separable (Table IV). 11. C. RETICULATA Walter Dry sandy woods and thickets; chiefly cp. and pied. Ala., Ark., Fla., Ga., La., Miss., $.C., Tenn. IP. [Tex.]. Viorna reticulata (Walter) Small— Small (1933); inel. V. subreticulata Harbison ex Small, V. beadlei Small— Small (1933). This variable species occurs throughout the southeastern United States; a thorough population study is needed. There is some variation in the degree of reticulation of the venation pattern on the adaxial surfaces of the leaves, leading Small (1933) to segregate additional species (as Viorna subreticulata Harbison ex Small, Viorna beadlei Small). However, pending further analy- sis, it seems best to regard C. reticulata as a widespread polymorphic spe- cies. Except for its viny habit, C. reticulata appears similar to C. ochroleuca. Furthermore, C. reticulata is confused with C. pitcheri, but the two species can usually be distinguished especially with mature material (Table IV). In any case, their ranges generally do not overlap (cf. Erickson, 1943). 12. C. VERSICOLOR Small ex Rydberg Dry rocky woods, bluffs and barrens; all prov. Ark., Ky. IP., Tenn. ME. and IP. [Okla., Mo.]. Clematis versicolor is one of three species of viny Clematis having gener- ally pronounced adaxial leaf reticulations. It differs from C. pitcheri and C. reticulata, however, by having both glabrous and glaucous abaxial leaf sur- faces, leaf rachises, and stem nodes (Table IV). Aside from its prominent venation pattern, C. versicolor apparently is closely related to C. glaucophyl- la and C. addisonii. Both C. pitcheri and C. versicolor should be carefully studied from a population standpoint, especially in Arkansas. 13. C. CRISPA L., Leatherflower, Blue-jasmine Low woods, swamps and creek bottoms; chiefly cp. and pied. SE ex- cept Del., Md., and W.Va. [Tex., Okla., Mo., Lll.]. Viorna crispa (L.) Small— Small (1933); incl. V. obliqua Small—Small (1933). Clematis crispa frequently occurs on the coastal plain and piedmont throughout the southeastern United States. It is readily distinguished by its thin leaves, typically large crisped-margined sepals, sericeous styles, and Table III Character tem pubescence . Comparison of Species in the C C. ochroleuca Silky pilose . ochroleuca Complex.’ C. coactilis C. albicoma C. viticaulis Axillary branches Seldom overtopping the central axis —— ae uiting spec color shape width x enon (cm) pubescence texture venation stomates epal pubescence ial surface) Achene rim pubescence Densely holosericeous Tomentose to loosely pilose Usually + Light green Narrowly to broadly ovate 3-8 x 4-12 Soft pubescent beneath + Strongly yellowish-reticulate on upper surface Moderately sericeous- pilose Strongly essed- ascendin Narrowly 4-9 x 6-11 2-5 x 4-8 + Coriaceous Lower surface Densely holosericeous Villous + Spreading Usually finely hirtellous overtopping the central axis + Dark green ovate to elliptic-lanceolate 2-3.5 X 4-7 Scattered pilose to glabrate beneath Membranous to coriaceous Searcely yellowish-reticulate on upper surface Both surfaces Finely puberulent Closely appressed- ascending Table III. Continued. Character C. ochroleuca C. coactilis C. albicoma C. viticaulis Mature styles: color disposition Habitat Range Yellowish-white to deep tawny + Loosely seems recu pea ds spheric re woods, ae open s and cleat ngs; often. in ee dericcd from basi Staten Island, N.Y., Va., N.C., 8.C., Ga. White to pale yellow Usually sharply recurved and flexuous; heads compact, + flattened ane barrens; rarely Shale barrens calcareous soils Four counties in west- Western Va. and central Va. eastern W.Va. Tawny to deep reddish-brown Loosely spreading- recurved; heads + spherical Shale barrens | Bath and Rockbridge counties, Va. > * Based on Keener (1967). Table IV. Comparison of C. pitcheri, G. reticulata, and C. versicolor.’ Character C. pitcheri C. reticulata C. versicolor Similar features Node pubescence Leaflets: color Green beneath Glaucous beneath pubescence Glabrous to soft pubescent Soft pilose to glabrate beneath Glabrous beneath eneath shape Lance-ovate to broadly Ovate to elliptical or subcordate Ovate-oblong to cordate-ovate cordate-ovate apex + Acute-tipped Rounded to acute venation + Coarsely reticulate Usually i and closely + Closely reticulate (adaxial reticulat surface) Sepals: abaxial Usually sericeous-canescent Nearly glabrous surface color Dull purple to dark red Purplish-red or bluish to pink- Dull purple to blue lavender lavender Mature styles: length (cm) 1-3 (often broken) 4-6 5-6 pubescence Nearly glabrous to a ae Plumose with pilose hairs silky or villous, not plumos Range Ind. to Iowa, s. to w. Ky., S.C. to Tenn. and Ark., to Ky. to Mo. and se. Okla., s. to w. Tenn., Ark., and Tex. Fla., Ala., Miss., La., and Tex, ‘Tenn. and Ark. Vines with + coriaceous leaflets having a prominent reticulate venation on the adaxial surface Finely to sparsely pubescent Usually glabrous and glaucous * Based in part on Erickson (1943). 45 ebractate peduncles. This species shows considerable variability, reflected by the segregate taxa of earlier workers, e.g. C. crispa var. walteri (Pursh) Gray, a morph with linear-lanceolate leaflets, and Viorna obliqua Small, scribed as having smaller, scarcely crisped-margined sepals. Although care- ful study of population variability might reveal definable geographic variants, I am currently treating C. crispa as a single polymorphic species. 14. C. VIORNA L., Leatherflower Rich wocded banks and thickets; all prov. SE except Fla. and La. [Mo., Ill., Ind., Ohio, Pa.]. Viorna viorna (L.) Small—Small (1933); incl. V. gattingeri Small and V. flaccida Small—Small (1933) and C. v. var. flaccida (Small) Erickson—Fernald (1950). The wide-ranging C. viorna can be distinguished by its viny habit, mem- branous soft-pubescent leaves, pubescent nodes and leaf rachises, bracteate peduncles, finely pubescent sepal backs, and plumose styles. It is somewhat variable, especially with respect to leaf size and pubescence and sepal size, a fact reflected in two segregate taxa described by Small. Viorna (= Cle- matis) gattingert Small, collected from the banks of the Cumberland River near Nashville, Tennessee, was erected on the basis of its smaller flowers and leaflets and glandular pubescence (an incorrect observation by Small; cf. Erickson, 1943). A second species, Viorna (= Clematis) flaccida Small (= C. viorna var. flaccida (Small) Erickson), was based on plants originally collected in Warren County, Kentucky and that were characterized by their large membranous entire velvety-pubescent leaves. Both species do not ap- pear sufficiently distinctive to warrant taxonomic recognition. Nevertheless, critical population studies of C. viorna, especially in central Tennessee and adjacent Kentucky, would be instructive and might reveal more precise tax- onomically definable topogamodemes. 15. C. GLAUCOPHYLLA Small Rich woods and river banks; all prov. Ala., Ark., Fla., Ga., Ky., Miss., N.C., $.C., Tenn. [Okla.]. Vierna glaucophylla (Small) Small—Small (1933). Even though C. glaucophylla and C. addisonii share the membranous leaf- let character of C. viorna, they can be distinguished from that species by their glabrous and glaucous leaves, rachis joints, and stem nodes and by their usually glabrous sepal backs. Clematis cleacophulld is a rare, morphol- ogically variable species in need of critical biosystematic study, especially in relation to C. versicolor and C. addisonii. Particularly helpful in this regard would be population studies of the glaucous-leaved leatherflowers from cen- tral Tennessee and adjacent Kentucky (cf. Erickson 19438, p. 21). 16. C. ADDISONII Britton Dry limestone hillsides and wooded banks, rare; RV. West-Central Va. Viorna addisonii (Britton) Small—Small (1933). This very rare species is closest to C. glawcophylla. It is quite variable in 46 habit; e.g., in the same population some plants may be small and erect with only simple leaves, whereas other plants may be more or less viny, up to 1 m long, and with both simple lower leaves and upper 2-6 foliolate pinnately compound ones (Keener, unpubl. data). A critical comparative study of C. addisonii with C. glaucophylla is required before its taxonomic status can be fully ascertained (cf. Gleason and Cronquist, 1963, p. 319). Pending such analysis, I am retaining C. addisonii as a species restricted to dry calcareous hills and banks in a few counties in west-central Virginia. ACKNOWLEDGMENTS Special thanks are due Dr. R. A. Pursell, Monte Manuel, and Paul Roth- rock for criticizing an earlier draft of this paper. Any remaining errors of fact or judgment are my responsibility. REFERENCES BOSTICK, P. FE. 1965. Documented plant chromosome numbers 65: 2. Sida 2: 165-168. BUCHHEIM, G. 1964. Ranunculaceae, p. 133-137. In H. Melchior (ed.) Engler’s Syllabus der Pflanzenfamilien, Band UH, Auflage 12. Gebriider Borntraeger, Berlin- Nikolas ERIC KSON, R. O. 1943. Taxonomy of Clematis section Viorna. Ann. Missouri Bot. Gard. 30: 1-62 FE RNALD, M. L. 1943. Morphological differentiation of Clematis ochrolenca and Rhodora 45: 401-412. 1947, A new Clematis from the Peaks of Otter. Rhodora 49: 219-220. . 1950. Gray’s manual of botany, 8th ed. American Book Co., New York FISH, R. K. 1970. Megagametogenesis in to nafis and its taxonomic and pl plications. Phytomorphology ae 317-: GLEASON, H. A. and A. CRC NOUIST. 1963. Manual of vascular plants of northeastern United States and adjacent en Van es GRAY, A. 1895. Ranunculaceae, p. . ee In North America, a GREGORY, W. C. . Phylogenetic ea cy ae studies in ie Ranunculaceae. Trans. Amer, Phitas, ae cc S. 31 (pt. 41- JAMES, J. F. 1883. oe of the genus Cle ae of the United States. J. Cincinnati Soc. Nat. Hist. 6: 118-1 KEENER, C. S. 1966. allies. ly logenetic im- Princeton, N.. .. Robinson Nis cel flora of - pe. I. American Boo : Co., New York teen plant chromosome numbers 196 s 35. . 1967. A biosystematic study ae watis subsection oe Referee laceae). . Elisha Mitchell Sei. Soc. 83: 1-4 . The natural history of deh ve Appalachian shale barren flora, p. 215-248. In PLC. ce a5, The distributional history of the biota of the anne Apealaeltns Pe. OU: — Virginia Polytechnic Inst. and State Univ. Res. Div. Monog. 2 . 1974, IOPB chromosome number reports XLV. Taxon 23: 19-624. KUNTZE, - 1885. Monographie der Gattung Clematis. Verh. Bot. Vereins Prov. Branden- burg 26: 83-202 KURITA, M. 1956. Cytological studies in Ranunculaceae. X. Further notes on ae iis types of pela Cimicifuga, and Clematis. Bot. Mag. (Tokyo) 69: 239-2 _ oN 958, Chromosome stuidics in Ranunculaceae. XI. Karyotype of seven genera. Mem. - ae aa Sect. 2, Nat. Sci., Ser. B, 3 (1): 13 LINDSAY, R. . 1929, 7 Ge haaainan: of some — angiosperms. Proc. Natl. Acad. eeu - 611-613. LONG, R. W. 1970. Additions and nomenclatural changes in tl Rhodora 72: 17-46. MEURMAN, O. and E. THERMAN. 1939. Studies on sae ee morphology and structural SS in the genus Clematis. Cytologia 1-14. Se 19 PRINGI re flora of southern Florida. Ni Cc Taxonomy and distribution of on pidiie sect. Afragene (Ranun- ulaceac), in North America. Brittonia 23: 361- RADFORD, A. E. 1968. Ranunculaceae, 452- a on A. E. Radford, H. E. Ah and Bell, Manual of the vascular flora of the Carolinas. Univ. of North Gaccling. Daw and R. L. WILBUR. Dept. Cc; ee Hill, . R. BELL, J. W. HARDIN, 1967. Contributor’s guide for the vascular flora oi the southeastern United States. of Botany, Univ f F , Chapel H . . Manual - a Southeastern flora. Publ. by the author, New York. SMITH, E. B. 1 f . In Documented chromosome numbers of plants. Madrofio 18: 122-126. 1967. Morphology, ecology 4 phylogeny of the Ranunculaceae. VII. Sci. S. ‘Coll i 6 -43, ou E. The eastern short-stemmed leatherflowers. J. Wash. Acad. Sci. 21: 194-198, NOTES ON THE FLORA OF THE MOGOLLON MOUNTAINS, NEW MEXICO WILLIAM J. HESS The Morton Arboretum, Lisle, Illinois, 60532 This brief paper begins another ‘‘series’’ of comments on the flora in the southwest, particularly in New Mexico. Voucher specimens have been or will be distributed to various herbaria in the United States. The following comments are about five species of plants in southwestern New Mexico that I have collected and observed during different seasons since 1966. In the Mogollon Mountains of New Mexico at an elevation of 2760-2470 meters, Indian Creek drains from west to east and flows into Gilita Creek. Only 0.5 km south, Willow Creek converges with Gilita Creek to form the Middle Fork of the Gila River, a major wilderness river of the southwest. The south facing slope of the canyon is forested by Pinus ponderosa while on the north facing slope is a mixed coniferous woodland with Pinus flevilis, Picea pungens, P. engelmannii, Pseudotsuga menziesti, and Abies concolor its major constitutents. Quaking aspen ( Populus tremuloides) and Quercus gambelii become conspicuous angiospermous components of the mixed conif- erous woodland. Numerous small side canyons empty into the main canyon adding to the diversity of habitats that are common in the mountains of the southwest. The canyon bottom (100 meters or more broad in some places) is primarily grassland interspersed with scattered stands of Pinus ponderosa and occa- sional thickets of Salix sp. and Sambucus neomexicana. The following four species of plants that occur in this canyon are of interest. ARCTOSTAPHYLOS UVA-URSI (L) Spreng. Up Indian Creek canyon 3 km, and 20 km east of Mogollon on State Highway 78, Catron Co., New Mexico. Hess 2143 (July 16, 1968). Locally common on either side of the canyon in dense shade under Pinus ponderosa or Picea pungens, P. engelmannii, and Pseudotsuga menziesit. The collection of Arctostaphylos uva-ursi from this canyon extends the range by over 250 km from previous known regions in northern New Mexico. Most likely, other locations of this plant exist in the Mogollon Mountains. It has not been reported from the White Mountains of Arizona and one would expect to find it there. This distribution pattern illustrates two points of in- terest in southwestern phytogeography. The first would be the absence of data from field investigations on which widely distributed taxa are recorded for either the Mogollon Mountains or the White Mountains respectively, but not from both of the mountain ranges. This often is indicative of our poor knowledge concerning the flora of the region, but it may be these taxa simply SIDA 6(1): 48—S51, 1975, 49 don’t occur in both mountain ranges. The second is the southern extension of some northern Rocky Mountain plants as far south as the Mogollon Moun- tains and the northern extension of some northern Sierra Occidental plants from Mexico as far north as the Mogollon Range. The only place in Arizona that A. uva-ursi is known to occur is in the Chuska Mountains (McDougal, 1973). These mountains occur in the north- western and northeastern corners of New Mexico and Arizona, respectively. It has been known from these mountains in New Mexico. In addition, I have examined in the Field Museum, Chicago, a herbarium specimen of A. uva-ursi collected by E. A. Goldman in August of 1917 from the Arizona side. ALLIUM GOODDINGII Ownbey. Up Indian Creek Canyon 3 km, and 20 km east of Mogollon on State Highway 78, Catron Co., New Mexico. Hess 2181 (July 18, 1968). Locally common in damp and densely shaded floors of some of the north side canyons near their openings into the main canyon. Previously, Allium gooddingii was known only from the White Mountains and Mogollon Rim in Arizona (Dr. Charles Mason, Jr., pers. comm.). Its occurrence in the Mogollon Mountains of New Mexico could be expected. In 1968 at the University of Arizona, I examined the herbarium specimens and labels, and mcre recently, Dr. Mason has forwarded all the label information on their plants identifed as A. gooddingit. From the label information only three main localities in Arizona are indicated, and on one label, concern was expressed by the collector (Phillips 3600) that ‘“‘road re-alignment may destroy this locality.’”’ When Ownbey (1947) described this plant, it was known only from the type locality. It now appears that A. gooddingii has a wider distribution with the likelihood of finding other sites and consequently a greater assurance that it will be preserved. SISYRINCHIUM ELMERI Greene. Up Indian Creek Canyon 3 km, and 20 km east of Mogollon on State Highway 78, Catron Co., New Mexico. Hess 2165 (July 18, 1968). Uncommon, along the creek margin with Sisyrinchium demis- sum, Veronica americana, Ranunculus hydrocharoides, Deschampsia. caespi- tosa, and others. In the Arizona Flora (Kearney & Peebles, 1960), this little plant is called Sisyrinchium longipes, and its distribution is recorded as ranging from the San Francisco and White Mountains south into northern Mexico. Bicknell (1900) monographed the yellow flowered Sisyrinchium (Hydastylus) and our plants correspond well to what is described as H. elmeri. The latter taxon was said to occur in the southern California mountains north in the Sierra Nevadas to northern California. Rickett (1970) states that S. longipes is equivalent to S. elmeri. The range for S. elmeri is extended from the moun- tains of California into Arizona and northern Mexico and now into the Mogol- lon Mountains of New Mexico. PEDICULARIS ANGUSTIFOLIA Benth. Pl. Hartw. 22, 1839. P. angustissima Greene. Leaflts. Bot. Obs. & Crit. 1:151. 1905. I first collected Pedicularis angustifolia during August of 1966 in the Gila 30 Wilderness on a trail from Willow Creek to Whitewater Baldy and about 20 km east of Mogollon on State Highway 78, Catron Co., New Mexico. Hess 930. In the Flora of New Mexico (Wooten & Standley, 1915), the plants keyed out to P. angustissima. Upon checking the Bebb Herbarium specimens of Pedicu- laris at the University of Oklahoma, none called P. angustissima could be found. Further checking uncovered specimens identified as P. angustifolia, and it appeared that the plants from New Mexico were the same taxon. I have further checked plants labeled either P. angustifolia or P. angustissima in various herbaria (US, NY, ARIZ, MICH, F) with the same results, that is, all of the collections from Mexico are labeled P. angustifolia and the plants from New Mexico are called P. angustissima. Morphologically, the two taxa are indistinguishable and their variation completely overlaps. Green (1905) describes his specimens as ‘‘flowers crowded” and “leaves narrowly lin- ear, 2 in. long, callous-crenulate’? and Bentham (1839) writes ‘‘spicis densis paucifloris” and ‘‘foliis anguste linearibus integris margine minute carti- laginco-serrulatis.’”’ There seems to be little doubt that the two named taxa are the same, and that the plant collected and named by Greene (1. c.), P. angustissima, should be placed in synonymy under P. angustifolia. Pedicularis angustifolia is widespread in the Mogollon Mountains and found mostly in association with Picea engelmannii. In these areas at an ele- vation of approximately 3200 meters, it makes up a dominant portion of the understory. In localities, such as Indian Creek, it is also found on the hill- sides in shaded regions of mixed coniferous woodland. This is another plant species that has the major part of its distribution in the mountains of north- ern Mexico just extending north into the Mogollon Mountains. As far as I know, P. angustifolia hasn’t been found in Arizona and one would expect to find it in either the White or Chiricahua Mountains. JUNIPERUS COMMUNIS L. var. DEPRESSA Pursh. On trail from White- water Baldy to Mogollon Baldy Mountain, approximately 19 km due east of Glenwocd, Catron Co., New Mexico. Hess 2268 (Aug. 22, 1968). Although the range extension is not great (only 50 km southeast of the White Mountains of Arizona), Dr. Marion T. Hall (pers. comm.) did think it a significant factor to find it in the Mogollon Mountains. It occurs at an ele- vation of 3,225 meters under Picea engelmannii. This is still another example of the phytogeographic richness of the Mogollon Mountains with respect to the distributional overlap of northern Rocky Mountain plants and northern Sierra Occidental plants of Mexico. I appreciate the suggestions of Dr. Marion T. Hall, Dr. George Ware, and Mr. Floyd Swink concerning this manuscript. REFERENCES BENTHAM, G. 1839. Plantae Hartwegianae. G. Pamplin, Londor p. 393. BICKNELL, E. P. 1900. Studies in Sisyrinchinm VIL: Siigrinchinm californicum and. re- lated species of neglected genus Hydasfylus. Bull. . Club 27: 373-38 1a GREENE, i. 905. New plants from ha cnoedte mountains. Leaflets Bot. Obs. & Crit. 1: pny 51 KEARNEY, T. H. and R. H. PEEBLES. 1960. Ne flora with supplement. Univ. of Calif. a ae & = Angelos, Calif. p. MCDOUG W. 1973. Seed plants of northern eas The Museum of Northern Arizona, eee p. 594. 1947, The genus Allium in Arizona. Research Studies State College Wash. 15: 211-232. RICKETT, H. W. 1970. Wild flowers of the United States: V. 4: The southeastern states, pt. 1. McGraw-Hill Book Co., p. 241. WOOTTON, E. O., and P. C. STANDLEY. LOLS. Flora of New Mexico. Contrib. U. S. Nat. Herb. 19: 1-794. ADDITIONS AND CORRECTIONS TO THE BAHAMA FLORA—II WILLIAM T. GILLIS' Arnold Arboretum of Harvard University, 22 Divinity Ave., Cambridge, MA 02138 GEORGE R. PROCTOR Science Museum, Institute of Jamaica, Kingston, Jamaica Earlier, we presented a compilation of species new to the Bahama flora since the publication (1920) of Britton and Millspaugh’s work (Gillis, Howard, and Procter, 1973), a summary of some nomenclatural changes in the flora from usage in Britton and Millspaugh (Gillis, 1974a), and a union of species resulting from a broader approach to species concepts than employed ear- lier (Gillis, 1974b). Meanwhile, Correll (1974) and Hill (1974) have reported further new species to the flora. We present here, in the process of prepar- ing a new vascular flora of the Bahama Islands and Turks and Caicos Islands, a compilation of still more additions to the flora, and additional changes in nomenclature from that in Britton and Millspaugh. Voucher specimens for plants reported here for the first time are deposited in the herbaria of the Arnold Arboretum and the Gray Herbarium, and at the Institute of Jamaica. ADDITIONS POLY PODIACEAE NEPHROLEPIS MULTIFLORA (Roxb.) Jarrett ex Morton. This Old World fern has become established in the West Indies and South Florida. It differs from N. exaltata by its scaly leaves and stem. Moreover, it does not have the ragged frond margins of N. hirsutula by which name American populations of it have been called in recent years. Both of these alternative species lack the short hairs on the upper side of the midribs of the pinnae as found in N. multiflora. The species has been found near the abandoned town site of the village replaced by Freeport (Philippi) on Grand Bahama Island (Gillis 7833), from Eleuthera (Proctor 19129) and Andros (Proctor 20977). It has also been found for the first time in South Florida on Key Biscayne at Cape Florida State Park (Gillis 10856), although it is probably much more common in the south- ern part of Florida but has been confused with the other two species mention- ed above. The recent publication of the combination for this fern is in the summary of Roxburgh fern types by Morton (1974). Although Morton claims that New World material identified as N. multi- flora is actually N. exaltata, the first author has seen the two species grow- "Current address: Department of Biology, Hope College, Holland, MI 49423. SIDA 6(1): 52—62. 1975, 03 ing together in South Florida where they can be seen to be distinct. Further- more, most of our Bahama material has been annotated by Dr. Jarrett. CYPERACEAE FIMBRISTYLIS ANNUA (All.) R. & S. Populations of this sedge are re- ported from Andros and New Providence by Kral (1971). GRAMINEAE BOTHRIOCHLOA PERTUSA (L.) Willd. This grass was found for the first time in the Bahamas by Dunbar (from Inagua) and reported by Howard and Dunbar (1964) as Andropogon pertusus. Since that time, we have collected it on New Providence (Gillis 11907) and Grand Turk (Gillis 11852). It appears to be spreading rapidly throughout the archipelago. It was not known from the Bahamas at the time of Britton and Millspaugh’s writing, nor at the time of Hitchcock’s treatment of West Indian grasses (1936). It can easily be dis- tinguished from its congeners by the pit in the middle of the back of the first glume, seen readily by the naked eye. It has been called an Andropogon by some (including the second author, GRP) but Gould (1967) preferred placing it in Bothriochloa. LEPTOCHLOA UNINERVIA (Presl) Hitchc. & Chase. We report this spe- cies from Inagua (Proctor & Gillis 33930). ORCHIDACEAE Twelve species of orchids are mentioned in Luer’s treatment of Florida orchids (1972) as having ranges in the Bahamas. Although we respect the careful work of Luer in drawing up the ranges of his species, we have not been able to find specimens to verify the presence of the following orchids in the Bahamas. Triphora gentianoides (Sw.) Ames & Schlechter in Ames Ponthieva racemosa (Walter) Mohr var. racemosa Cranichis muscosa Sw Spiranthes reneeneides (Griseb.) Cogn. in Urban Spiranthes elata (Sw.) L. C. Rich. Tropidia polystachya (Sw.) Ames Tetramicra canaliculata (Aubl.) Urban Cyrtopodium punctatum (L.) Lindl. Oncidium floridanum Ames Oncidium luridum Lindl. Macradenia lutescens R. Br. Polyrrhiza lindenti (Lindl.) Cogn. EUPHORBIACEAE DALECHAMPIA SCANDENS L. This common species of the Greater An- tilles and Central America has now been found on North Caicos, where it in- habits roadsides to the east of Bottle Creek (Proctor and Gillis 34039). From a4 the stinging hairs often present on the plant, especially the fruits, comes the local name ‘‘itchy bus ARGYTHAMNIA CANDICANS Sw. Ingram (1967) has found that Britton and Millspaugh misdetermined collections of this species and generally placed them in A. lwcayana. A native of the Greater Antilles, this species is found throughout the southern islands in the Bahama group, and is evidently migrating northward. ” ZYGOPHYLLACEAE TRIBULUS TERRESTRIS L. This Old World weed, common in eastern United States, has now been found in the village of Conch Bar on Middle Caicos (Gillis 12298, Proctor 34064). This collection represents a wide disjunc- tion from populations in the United States. MALVACEAE MALACHRA CAPITATA (L.) L. This species was growing at the Horse Pond or ‘‘waterworks’”’ near Matthew Town, Inagua (Gillis & Proctor 11713). The nomenclature of Borssum Waalkes (1966) has been followed here. Popu- lations known as M. alceifolia are occasionally segregated from M. capitata, depending upon one’s species concept in this group. LACHRA URENS Poit. Despite the fact that the genus Malachra had not previously been reported from the Bahamas, we now have two species represented in the flora. Malachra urens, a weedy species, has been found in vacant lots in Nassau, New Providence (Gillis 11893). MALVAVISCUS ARBOREUS var. CUBENSIS Schlecht. Instead of var. mexicanus which had been reported from the Islands, Schery (1942) main- tained the Bahama populations were this variety of M. arboreus. SIDA CILIARIS var. INVOLUCRATA (A. Rich.) Clement. In his treatment of Sida, Clement (1957) has recognized two varieties of this species in the Bahamas. This report of var. involucrata is new to the Islands. STERCULIACEAE AYENIA TENUICAULIS Urban. Cristobal (1960) indicated that the south- ern Bahamas have this species in addition to A. insulicola. This species dif- fers from A. insulicola by its much smaller and nearly orbicular leaves and pubescent ovary and fruits, in addition to shorter (‘‘mamiform’’) projections from the fruits. MELOCHIA TOMENTOSA var. FRUTESCENS (Jacg.) DC. In his treat- ment of Melochia, Goldberg (1967) recognized a small-leaved form of Mel- ochia tomentosa in the southern islands. This differs in a number of charac- ters from typical var. tomentosa. The authors have found it on Mayaguana, Inagua, and in the Caicos Islands. CUSUTACEAE CUSCUTA GLOBULOSUS Benth. Yuncker (1965) reported the presence of Hb) this additional species of dodder in the Bahamas on the Cay Sal Bank. This remote cluster of islands at the eastern end of the archipelago has an inter- esting assemblage of Cuban species in its flora. : A UMBELLATA Kunth. A Proctor collection (No. 8819) from South Caicos was determined by Yuncker to be this species. We have since found it on Grand Turk (Proctor and Gillis 34044). CONVOLVULACEAE IPOMOEA NIL (L.) Roth. We suspect that this species becomes introduced to the Bahamas periodically and manages to be collected from time to time and then dies out. It has not been found consistently on any island. An old collection of Wight (No. 130) from New Providence was found at the Gray Herbarium, apparently not discovered by Britton and Millspaugh when they wrote their flora. Proctor found it once again (No. 30783) on Andros in a farm area south of San Andros airfield. BORAGINACEAE TOURNEFORTIA STENOPHYLLA Urban. This Cuban species has turned up several times on Inagua where it appears to be well established. It was common with T. volubilis at Devil’s Point (Proctor 33905), and was a primary invader on Maroon Hill (Gillis 12116) which had been stripped of its soil and upper layers of rock for use as fill. LENTIBULARIACEAE UTRICULARIA GIBBA L. Peter Taylor has examined many of the early twentieth century collections upon which B&M based their treatment of Utricularia in the flora. He has realigned several of the species names in- volved. Although true U. foliosa is known from the Bahamas (from Grand Bahama Island, B&M 2679) as indicated in the flora, most of the species orig- inally named as U. foliosa by B&M should instead be called U. gibba. COMPOSITAE SPILANTHES IODISCAEA A. H. Moore. This tiny composite was found in the dry muck at the bottom of Smith’s Thatch Pond on Inagua during the dry season (Gillis 11738). It had been found earlier in a pond near Salt Pond Hill on the same island by Dunbar, but unreported in Howard and Dunbar (1964). It is represented by a Dunbar collection at the Arnold Arboretum. RON BELLIOIDES DC. Also from Inagua, near Salt Pond Hill, comes this small plant, found by Dunbar (No. 318), but previously unreport- ed. This, too, is represented by a specimen at the Arnold Arboretum. CORRECTIONS For ease of reference this section will follow the order of species presented in Britton and Millspaugh’s Flora, which will be Monocotyledons first, fol- lowed by the Dicotyledons, and then the Pteridophytes. Again for brevity, 06 the previous flora is referred to by the expression B&M in the text herein. The number preceding the names refers to the page in B&M on which the tax- on in question is discussed. The name in capital letters is the name considered to be correct. These names are not necessarily nomenclatural or taxonomic equivalents, hence the reason for not employing an equals sign (=). In some instances, B&M misidentified the plant or used a binomial incorrectly. GRAMINEAE 16 Syntherisma sanguinalis—DIGITARIA CILIARIS (Retz.) Koeler, As Adams (1972) has pointed out, Digitaria sanguinalis (l.) Seop., based on Panicum sanguinale L., is not the plant of the West Indies. The two plants differ in a number of ways, however subtle. Ebinger (1962), Gould (1963), and Blake (1969) have noted differences in chromosome number, length of second glume in relation to the length of the spikelet, mean pollen size, presence or absence of spicules on the sterile lemma, and presence or ab- sence of papillose-based hairs on the foliage as the differentiating charac- ters. The West Indian species has long been called Digitaria adscendens (HBK.) Henrard, based on Panicum adscendens HBK., dating from 1816. Pani- cum ciliare Retz., however is the earliest name for this species (1786) and should be taken up. When placed in Digitaria, the combination is D. ciliaris (Retz.) Koeler, Descr. Gram.: 27 (1802) 24 In the earlier list of corrections in the flora (Gillis, 1974a), a change was not made properly due to a typographical error, It should have read Panicum coerulescens—PANICUM CAERULESCENS Hack. ex. Hitche. PALMAE 58 Thrinax microcarpa—THRINAX MORRISII H. Wendl. Read (1974) has indicated that this small thatch palm must undergo a name change as the result of proper typtification. 09 Coccothrinax argentea—COCCOTHRINAX ARGENTATA (Jacq.) L. H Bailey. In the earlier treatment of name changes (Gillis, 1974a) a misplaced line of type made nonsense of the attempt to explain this name change. True C. argentea may exist in the southern Bahamas, but this has yet to be dem- onstrated. It is now treated as being indigenous to Hispaniola. The Bahama plant is C. argentata. RANUNCULACEAE Clematis bahamica—CLEMATIS DIOICA L. There appears to be no significant difference in these plants. Therefore, the earlier name must pre- vail. LEGUMINOSAE 178 Sophora tomentosa—SOPHORA TOMENTOSA subsp. BAHAMENSIS Yakovlev. Earlier, one of us (Gillis, 1974a) indicated that all Bahama plants were typical subspecies and we did not recognize the infraspecific taxa of a7 Yakovlev (1967a and b). We now find that we should follow his treatment, at least in part, in recognizing that the Bahama plants all represent subsp. bahamensis instead of the typical subspecies. In this choice, we follow Rudd (1972). 184 In an earlier updating of nomenclature (Gillis, 1974a), an incorrect author citation was attributed to Desmodium tortuosum. It should be D. tor- tuosum (Sw.) DC., based on Hedysarum tortuosum Sw., Prodr. 1788. EUPHORBIACEAE 220 Phyllanthus niruritPHYLLANTHUS AMARUS Schum. & Thonn. Web- ster (1957) has shown that the widespread plant of ruderal sites in South Florida and the West Indies is not P. niruri as often named, but rather P. amarus. TAMARICACEAE 279 Tamarix gallica—TAMARIX CANARIENSIS Willd. Upon examination of all specimens of this species available to Britton and Millspaugh from the Turks and Caicos Islands, we find that all they had was sterile material! We have been fortunate enough to find a precocious flowering specimen of the population on Grand Turk for specialist Dr. Bernard Baum to study. Our collection ( Gillis 12340; Proctor 34094) was made in June; full flowering in the Turks and Caicos is in July. (Proctor also found it some years ago on South Caicos in flower: Proctor 8907). Dr. Baum determined our material to be T. canariensis on the basis of papillae on the perianth. The latitude of the Turks Islands is more compatible with T. canariensis than T. gallica. We are grateful to Dr. Baum for his courtesy in making this determination for us. FLACOURTIACEAE 285 Zuelania guidonia—CASEARIA GUIDONIA (Sw.) Lundell. Lundell (1974) has merged this species of Zuelania with Casearia. In absence of broadly described generic lines in this difficult family, it seems appropriate to accept this name change until someone performs cons‘derable in-depth studies in the whole family. HALORAGIDACEAE 311 Proserpinaca platycarpa Small—PROSERPINACA PALUSTRIS L. var. PALUSTRIS. It appears that Small’s name for a slightly different popu- lation in South Florida influenced Britton in his compilation of the Bahama flora. Small’s plant is now considered to be quite typical P. palustris. UMBELLIFERAE 313 Foeniculum foeniculum— FOENICULUM VULGARE Mill. In making the name change for this species, Gillis (1974a) used the correct new name, but attributed it to the wrong author. The earliest use of this binomial is by Miller in ed. 8 of the Gardener’s Dictionary, Art. Foeniculum, No. 1, 20 years 08 earlier than Gaertner’s use. 313 Centella asiatica—CENTELLA ERECTA (L.f.) Fern. Fernald (1940) recognized that the Old World Centella was different from the common, New World species and based his new combination on Hydrocotyle erecta L.f., Suppl. 177. 1781. See also Mathias and Constance, 1944. 314 Helosciadium ammi—APIUM LEPTOPHYLLUM (Pers.) F. Muell. Benth. et Muell. The weedy and cultivated plant of the West Indies and aon of North America must have its name based on Pimpinella leptophylla Pers. (Syn. Pl. 1: 324. 1805) as shown by Mathias and Constance (1944) SAPOTACEAE 321 MASTICHODENDRON' FOETIDISSIMUM (Jaeq.) H. J. Lam. In assigning author citations to M. foetidissimum in an earlier paper (Gillis, 1974a), the author of the combination was given as Cronquist. The combina- tion, however, was made by H. J. Lam in 1939. APOCYNACEAE 338 Cerbera thevetia L.—THEVETIA PERUVIANA (Pers.) K. Schum. When this commonly cultivated plant is moved from Cerbera to Thevetia, the transfer of the Linnaean epithet would result in a tautonym,. Hence, the use of a later name in Thevetia. SOLANACEAE 382 Solanum aculeatissimum—SOLANUM CILIATUM Lam. The plant commonly found in South Florida and the West Indies has the latter name, and not the former. These two species are closely related, but separated by the fact that S. aculeatissimum Jacq. is an Old World species with brown fruits and smaller seeds instead of the red fruits and seeds about 4 mm. in diameter of S. ciliatum. In fact, Tjaden (1970) has made a strong case for replacing S. aculeatissimum Jacq. by the still earlier (1773) name, S. capsicoides Allioni. 383 Solanum nigrum—SOLANUM AMERICANUM var. NODIFLORUM (Jacq.) Edmonds. Edmonds (in Stearn, 1971) has discussed the break-up of the aggregate species Solanum nigrum into smaller species realities. Solanum nigrum appears to be a temperate species now fairly widely distributed in both New and Old World. The Solanum americanum complex is also widely distributed, chiefly—but not exclusively—in the tropics, with classic Miller and Jacquin specimens having originated from such diverse places as West Virginia and Mauritius. Differences among members of the Solanum nigrum complex are subtle, but consistent, and include polyploidy. The following couplet summarizes differences discussed by Edmonds in making distinctions between S. nigrum and S. americanum: Corolla 10-18 mm. across; anthers 1.5-2.7 mm. long; fruit often nodding; berries black or green, ovoid, 6-10 mm. across: seeds 1.7-2.4 mm. long; 09 pollen 27-384; sepals usually appressed to fruits; hexaploid (2n = 72) or as ee We ew S. nigrum. Corolla 5-9 mm. across; anthers 1-2 mm. eae fruit erect neniiee black, d-7 mm. across; seeds 1.0-1.5 mm. long; pollen 19- one sepals well reflexed in fruit; diploid (2n = 24) . . ...5. americanum In S. americanum, West Indian and eastern South American plants are usually glabrescent and are interpreted by Edmonds as var. nodiflorum. Western South American plants are fairly pubescent and are interpreted by Edmonds as var. americanum. Both varieties may be typified by specimens at BM. Our population is S. americanum var. nodiflorum which is also found in Australia (Henderson, 1974) and New Zealand (Baylis, 1958). SCROPHULARIACEAE 392 Agalinus spiciformis—AGALINUS MARITIMA (Raf.) Raf. According to Pennell’s treatment (1929), the plants from the Bahamas fall well within the circumscription of A. maritima, an earlier name than A. spiciformis. He recognized a large flowered form, var. grandiflora, for the Bahama popula- tions. RUBIACEAE 423 Spermacoce tenuior—SPERMACOCE CONFUSA Rendle ex Gillis. Gil- lis (1974c) validated the publication of Rendle’s invalid binomial. Rendle (1936) had neglected to provide a Latin diagnosis which was required at the time of his publication. All populations previously reported from the Baha- mas as S. tenuior appear to be S. confusa. An interpretation by Bacigalupo (1972) would treat S. tenuior and S. confusa as synonyms. 424 Galium bermudense—GALIUM HISPIDULUM Michx. Galium bermu- dense L. is difficult to interpret inasmuch as there is no specimen at the Linnaean Herbarium as is suggested by the way in which Linnaeus wrote the protologue. The two pre-Linnaean names cited in the original description are different species. The Gronovius reference is to G. pilosum Aiton, while the Plukenet reference is apparently to the species which we have in our flora. Although one might make a case for taking up one or another of these names and typifying it, tradition since the time of Asa Gray (1878) has been to consider G. bermudense to be a nomen dubium and eliminate it from con- sideration. Until and unless someone chooses to resurrect G. bermudense L., our plant is best named by the next available binomial, G. hispidulum Michx. We wish to thank Univ. Prof. Dr. F. Ehrendorfer of the Botanisches Insti- tut und Botanischer Garten der Universitat Wien for his interpretation of this matter, and for his consent to publish his heretofore unpublished analysis herein. Incidentally, another Galium species from Bermuda which might have been confused with Linnaeus’s G. bermudense is G. aparine, represent- ing a binomial of the same date of publication as G. bermudense. COMPOSITAE 435 Vernonia obcordata—VERNONIA ARBUSCULA Less. 436 Vernonia bahamensis—VERNONIA ARBUSCULA Less. 436 Vernonia insularis—VERNONIA BLODGETTI Small. As has been our inclination with Aster (See Gillis, 1974b), we have inter- preted the northwestern populations in the islands to be merely extensions of the ranges of southeastern U.S. species, and not endemics to the Bahamas. Thus, we wish to reduce the number of Vernonia species in the archipelago. Veronia insularis Gleason is synonymous with V. blodgettii Small, the latter being the earlier name. Vernonia blodgettii occurs in South Florida and the so-called V. insularis is simply the Bahama extension of its range. (Our judgment is supported by personal communication with Dr. S. B. Jones and Mrs. S. C. Keeley who have Vernonia under study.) Furthermore, we believe that the only woody Vernonia in the Bahamas is V. arbuscula. The leaf shape variations which have been used to separate this species into two others in addition to V. arbuscula are likely environ- mentally induced, and are highly variable. Typical V. arbuscula as seen in the pinelands of New Providence Island has leaves that are chiefly elliptic to ovate. Farther south in the island chain, especially south of the Crooked Island Passage, the leaves become decidedly spatulate. Then, on Little Inagua (and known from only one collection—the type of V. obcordata), the leaves appear larger, nearly obcordate. Whether this specimen represents a single individual variation or is typical of the population on Little Inagua cannot be answered at this juncture for lack of material. Our judgment that these several ‘‘species’? should be merged into V. arbuscula is supported by the observations of Mrs. Sterling Keeley (personal communication) who is presently studying the Vernonias of the West Indies. The treatment of Vernonia herein eliminates two supposed ‘‘endemic”’ spe- cies from the treatment of B&M. It is considerably different from the treat- ment of Vernonia in the Bahamas by Gleason (1906). KEY TO VERNONIA IN THE BAHAMAS 1. Annual. . . .) ee Vernonia cinerea (L) Less. 1. Perennial. 2. Shrub; heads scorpioid, in axils of upper leaves; foliage canescent or tomentose; leaves ovate, elliptic to obovate, spatulate, or obcordate. Ce ke ee) 6 Vernonia arbuscula Less, 2. Herb; heads cymose, long-peduncled; foliage glabrous or puberulent; leaves narrow, elongated, broadly linear. . Vernonia blodgettii Small. SUMMARY We report herein 20 plants new to the Bahama flora, and explain the change of name for an additional 22 species. REFERENCES ADAMS, C. D, 1972. Flowering plants of Jamaica. University of the West Indies. Mona, Jamaica. 848 BACIGALUPO, ‘NELINDA M. 1972. Observaciones sobre algunas especies de los géneros Spermacoce L. y oo O.K (ubiacese). Darwiniana 17: 341-357. BAYLIS, G.T.S. 1958. A cytogenetical study = New Zealand forms of Solanum nigruni Ls S. ia i a. and S. gracile Otto. Trans. Roy. Soc. New Zealand 85: 37 ee 383, BLAKE, . Taxonomic and nomenclatural studies in hig Gramineae, No. 2. Proc. Roy. Paes 81: 1-26. BORSSUM ee apres J. ee 1966. Malesian Malvaceae revised. Blumea 14: ee CLEMENT, I. D. Studies in Sida (Malvaceae). 1. A review of the genus an graph of the Pe nee Maleceoidesa, Physalodes, Pseud [ , Incanifolia, pe drac, Pseudonapaea, Hookeria, and Steninda, Contr. Gray 5-91, erb. 180: CORRELL, DONOVAN 5S. ae ee of the Bahama eet angte additions. Fairchild Tro ard. Bull. 29: 11-12; CRISTOBAL, CARMEN L. oh Revision del genero “Ayenia’? (Sterculiaceae). Opera tlloa Tucuman, Argentina. pp. 5-230 PBINGER Jor IN E. 1962. Walidity of he: grass species Digitaria adscendens. Brittonia oe ee R_ K. 1971. Solanaceae in William T. Stearn: Taxonomic and_no- menclatural notes on Pe nee gamosepalous plants. J. "Arnol d pees 52: 634-635. L. 19 FERNALD, M. . Spermatophytes of eastern North America. - Rhodors 42: 281-302. GILLIS, WILLIAM T. 1974a. Name changes for the seed plants in the Bahama flora. Rhodora er 67-13 97 4b. eee in the flora of the Bahamas. Phytologia 29: 154-166. aa oe The confused a oce. Phytologia 29: 185-187. Richard A. Howard, woe R. Proctor. 1973. Additions to the Bahama flora : since Britton and Millsaush— Rhodora 75: 411-425. ee oe A. 1906. The genus Vernonia in the Bahamas. Bull. Torrey Bot. Club 33; o. DBERG, nO The genus Melochia L. (Sterculiaceae). Contr. U.S. Natl. Herb. coo FRANK oe 1963. Cytotaxonomy of Digitaria sanguinalis and D. adscendens. Brittonia 1 ie 19 The: grass genus Andropogon in - United States. ee 19: 70-76. k. sRAY, ASA. He Synoptical flora of North America. Vol. 2. New Yor ee en R. oa 1974, Solanum nigrum L. (slices) and related species in Aus- ralia. nn Herb. No. 16: 1-78. HILL, STEVEN | Range extensions and new records for the Bahama flora. Rhodora A os R MARY a 1941. A revision of the subgenus Mariscus found in the United States. Contr. Biol. Lab, Catholic Univ. Amer. 33: 1-1 HOWARD, RICHARD A. and HENRY F. DUNBAR. 1964. Additions to tha: “flora of Inagua, the Bahamas. Rhodora 66: 6-15. INGRAM, JOHN. 1967. A revisional study of Argyfhamnia, subgenus Argythamnia (Eu- phorbiaceae). Gentes Herb. 10: 1-38. KRAL, ROBERT. 1971. A treatment of es Bulbostylis, and Fimbristylis (Cy- peraceae) for North America. Sida 4: 57-22 KUKENTHAL, G. 1935-36. Ce ee: Cypereae. Pflanzenreich iv. 20: 1-671. ee aac a aa The native orchids of Florida. New York Botanical Garden. rk, N.Y LUNDEL as TON ORE 1974. Studies of American plants -vi. Wrightia 5: Penne Te a and LINCOLN CONSTANCE. 1944. Umbelliferae. No. Amer. FI. ‘ 28B, Pr. 1: MORTON, C. V. 1974. wiles Roxburgh’s fern types. Contr. U.S. Natl. Herb. 38 (7): 283-39 PENNELL, FRANCIS W. 1929. Agalinus and allies in North America -II. Proc. Acad, Nat. Sci. Philadelphia 81: 111-249 62 READ, ROBERT W. 1974. The ecology of the pane Principes oe 39-50. RENDLE, B. 1936. Spermacoce remota La J. Bot. 74: 10- R ae i EK. 1972. Leguminosae - F shotdese - eee No. Amer. FI., Ser. II, E> § Pt 3. SCHE ne ROBERT WALTER. 1942. Monograph of Malvaviscus. Ann. Missouri Bot. Gard, 29: 183-244 + 17 pl. STEARN, WILLIAM T. 1971. Taxonomic and nomenclatural notes on Jamaican gamo- sepalous plants. J. Arnold Arbor. 52: 614-648. TJADEN, W. L. 1970. Carlo Alliont and his Auctarinm Ad Synopsim Methodicam Stir- binm Horti Reg. Taurinensis. Vaxon 19: 61 WEBSTER, GRADY L. 1957. A monographic ore of the West Indian species of Phyl- lanthus. m a —— 38: 295-373. 96 YAKOVL G. 7a. Systematical and geographical studies of the genus Sophora L. and se genera. Prob. Pharmacog. 21: 42-62 (in Russian). 1967b. Some characteristics of structure of seeds of Sophora L. and allied genera in connection with its systematics and sieleceia: Repr. Proc. Leningrad Chem. Pharm. Inst. 21: 90-98. (in Russian YUNCKER, TRUMAN G. 1965. Cue ie. No. Amer. Fl., Ser. HII, Pe. 4: 1-51. SIDA susie" VOLUME 6 NUMBER 2 NOVEMBER 1975 CONTENTS Sex and the angiosperms—another proposition. W. H. Wagner, Jr. 63 Revision of the Rhamnus serrata complex. Laverne A. Johnston. 67 Taxonomy of the Lythraceae in the southeastern United States, Shirley A. Graham. 80 The importance of revisionary studies in plant systematics. Tod F. Stuessy. 104 NOTES. Additions and corrections to the flora of Texas. 114—-Schizachyrium stoloniferum Nash var. wolfei DeSelm, var. nov. (Gramineae). 115—Hemigraphis reptans (Acanthaceae), a greenhouse weed in Louisiana. 115—Brazoria (Labiatae) in Mexico. 116—Fatoua villosa (Moraceae): additional notes on distribution in the southeastern United States. 116——-New 117 and verified additions to the Nebraska vascular flora. 117. REVIEWS. Flowering Plants: Evolution Above fhe Species Level. 118—List of Flowering Plants in Hawaii. 120 APPEAL. 122. US ISSN 0036-1488 SIDA, CONTRIBUTIONS TO BOTANY Founded by Lloyd H. Shinners, 1962 Editor & Publisher Wm. F. Mahler SMU Herbarium Dallas, Texas, 75275 Associate Editor John W. Thieret Northern Kentucky State College Highland Heights, Kentucky, 41076 Contributors to Sida should refer to the latest issues for style, follow the CBE Style Manual (3rd ed.), and use the “adopted”? journal abbreviations given in Botanico-Periodicum-Huntianum. Subscriptions: Libraries $6.00 (U.S.) per year; individuals—$8.00 (U.S.) per volume; numbers issued twice a year. © Sida, Contributions to Botany, Volume 6, Number 2, pages 63-122 Copyright 1975 by Wm. F. Mahler SEX AND THE ANGIOSPERMS— ANOTHER PROPOSITION W. H. WAGNER, JR. Department of Botany The University of Michigan Ann Arbor, Michigan 48104 When Dr. John W. Thieret published his thoughtful critique of sex and the angiosperms, I found it stimulating. I have always had an active interest in sex. According to Thieret (1973) we are in error to apply sexual termi- nology to sporophytic structures of flowering plants. Such terms as ‘‘male,”’ “female,’’ and ‘“‘bisexual flower,’ for example, are inadmissible because they are ‘misleading, ”* “inconsistent,” ‘‘inaccurate,’’ and ‘‘superfluous.”’ Instead we should use “‘staminate’’ for ‘“‘male,”’ ‘‘carpellate’’ for ‘‘female,”’ and ‘‘perfect” for ‘‘bisexual.’’ It was my impression that wherever there are structures and HUIGULONS related to production of sperms or male gametes, we used the term ‘‘male,’ and of eggs or female gametes, the term ‘female.’ I am not quite clear as to what is wrong with this procedure as it applies to angiosperms (or, for that matter, any land plants, or fungi or algae) whether the stage in- volved is sporophytic or gametophytic. In vernacular the words ‘‘male’’ and ‘‘female’’ have been commonly asso- ciated with forcefulness vs. passiveness, hard vs. soft, yin vs. yang, hus- band vs. wife, leader vs. follower, inflexible vs. willowy, and so on—largely supposed cultural attributes of man and woman, related to certain kinds of family organization and to formerly ees traditions for behavior, now, fortunately, being more or less eroded away. ology, however, the words have a more scientific and precise ene < pertain to the process of sexual reproduction, which comes down to making sperms and making eggs, as well as caring for embryos and juvenile diploid stages. In people the last is baby care; in higher plants, providing seed coats and endosperm. As soon as ‘‘syndromes’’ of structures and (or) functions become differen- tiated as part of these basic processes, then they become biologically male and female respectively. Below are listed some examples of such differen- tiation in the land plants, using two taxonomic groups as illustrations: a SIDA 6(2): 63-66. 1975. 64 MALE FEMALE FERN Microsporophyll Megasporophyll Microsporangium Megasporangium Microspore Megaspore Microgametophyte Megagametophyte Antheridium Archegonium Microgamete Macrogamete FLOWERING PLANT Staminate tree Pistillate tree Staminate flower Pistillate flower Stamen Carpel Pollen sac Nucellus Connective Ovary wall and style Pollen tube Embryo sac Tube nucleus Polar nuclei When one is concerned with functional aspects of biology, precise homol- ogy of participating parts is traditionally thrown overboard—in my opinion, wisely—in favor of clear expression of roles. The study of analogous struc- tures is one of the most fascinating phases of modern ecological morphology. Thus we have nouns like arm, beak, eye, grazer, herb, leg, predator, tree— words in constant use in biology in writing and speaking but involving all proportions of homology and analogy. If we rid ourselves of the adjectives “male” and “female’’ in reference to higher plants, we should also include in our house cleaning such familiar adjectives as cortical, carnivorous, embryonic, glandular, pedicellate, pinnate, sessile, vascular, winged, and the like. Should we delete all the words based upon the sexual roots ‘‘andr, ‘oam,” ‘gyn,’ and ‘“ov?’’ Is it better that we use “perfect” from some quaint antiquarian concept, or ‘‘pistillate,’’ derived etymologically from the word for pestle (strong biological meaning here), or “‘staminate’’ from the word for hair or thread (ditto)? Figure 1 is a diagrammatic representation of the evolutionary advance- ment of sexual differentiation in vascular plants. The most primitive condi- tion is that in which one kind of spore and one kind of gametophyte is pro- duced. From the time it forms in the sporangium, the spore embodies the ability to produce sex cells of both types, and the entire sexual generation is thus hermaphroditic. In heterosporous pteridophytes, on the other hand, the sex organs are borne upon differentiated gametophytes. The spores pro- ducing them are different because of divergent processes operating in the ” micro- and megasporangia, so that the microsporangia produce only male spores and the megasporangia female. In seed plants, the differences be- come more and more profound and involve more and more structures. Fin- ally, genetic changes arise that change the sporophytes themselves. Now EVOLUTION of SEXUALITY in VASCULAR PLANTS 2n Melosis n S P E R UNISEXUAL UNISEX UAL M A T O P UNI-BISEXUAL UNISEXUAL H Y T E S BISEXUAL UNISEXUAL BISEXUAL UNISEXUAL P T E R BISEXUAL UNI-BISEXUAL , D Oo P H BISEXUAL BISEXUAL ¥ T E (©) SPORANGIUM O GAMETANGIUM S } | SPOROPHYTE Vv GAMETOPHYTE Fig. 1. Diagrammatic representation of the evolutionary advancement of sexual differentiation in vascular plants. 66 the sexual differentiation has moved into the 2n generation, and the stami- nate plant can produce only sperms and the pistillate only eggs, each utiliz- ing its own now very distinctive structures. In the course of these basic transformations in evolution, we see not only repeated parallels to conditions known in fungi and algae (but surely not homologous, even though they are clearly sexual phenomena), but—what impresses me particularly—beautiful convergences of higher seed plants toward higher animals. It is a remark- able fact that no lower vascular plants are yet known to have evolved sexual differentiation of the sporophytic plant. In terms of biological strategy, just as the 2x scrotum is a male structure, the 2n microsporangium is a male structure. Just as ejaculation is a male function in mammals, pollen release is a male function in angiosperms whether or not the organs involved are homologous. I propose that both words “male” and ‘female’ continue to be used, where appropriate, in connection with vascular plants, and that they be defined as follows: MALE. Adjective applied to all character-states—diploid or haploid, spo- rophytic or gametophytic, primary or secondary—adapted to or directly or indirectly correlated with the production and transport of sperms. FEMALE, Adjective applied to all character-states—diploid or haploid, ete.—adapted to or directly or indirectly correlated with the production and functioning of eggs. Also the structures and functions specifically modified for the protection and nourishment of zygote, embryo, and germling. As I see them, the advantages of utilizing ‘male’? and ‘‘female’” in our biological (as opposed to vernacular) terminology are not only in (a) cutting down on polysyllabic verbosity, (b) taking advantage of their obvious mean- ings, but also—the most important point scientifically—(c) to bring out the remarkable adaptive parallelisms and the universality of “sporophytic”’ maleness and femaleness, as has evolved separately in higher animals and in higher plants. To recapitulate, in all probability much of the differentiation into ‘“‘male- ness’ and ‘‘femaleness’’ in such diverse groups of organisms as fungi, animals, red algae, and vascular plants constitutes analogy. Nevertheless, the biological functions involved are closely similar in their methods of ac- complishment and in their results. Perhaps indeed in some cases the simi- larities are not so superficial as to some they may seem, for the possibility should be entertained that we are observing homologous genetic potentiali- ties originating in a flagellate common ancestry long since lost in antiquity. In these terms, a staminate tree of Salix fragilis—in spite of its soft and willowy appearance—is just as ‘‘male’’ as a bull, a stallion, or a billy-goat. ” REFERENCE THIERET, J. W. 1973. Sex and the angiosperms. Sida 5: 59-60. REVISION OF THE RHAMNUS SERRATA COMPLEX LAVERNE A. JOHNSTON Herbarium, The University of Texas, Austin, Texas 78712 While I was working on a revision of Rkamnus as found in Mexico and southward, my interest was captured by a small group of closely-related species that have narrowish shiny leaves, bud scales, four-merous flowers that are solitary or in axillary clusters of two or three (rarely four) and that tend to be unisexual, and glabrous two-seeded fruits. I call this group informally the Rkhamnus serrata complex, and I find it especially interesting and exasperating because of the puzzling populations in northern Mexico and western Texas. These are ambiguous, suggesting relationships, on the one hand, to the Coloradan and New Mexican plants called R. smithii Greene and, on the other, to R. serrata Schult. of the central Mexican highlands. The latest monographer of the group, C. B. Wolf (The North American Spe- cies of Rhamnus, Rancho Santa Ana Bot. Gard. Monogr., Bot. Ser. 1: 1-136. 1938) called the plants of Colorado and northern New Mexico R. smithii ssp. typica and those of southern New Mexico and western Texas R. smithii ssp. fasciculata (Greene) Wolf. That he did not consider the latter taxon more nearly related to R. serrata is apparently a function of the real or fancied geographic proximity of the populations of southern New Mexico to those of northern New Mexico and the fact that R. serrata had at that time not been found north of San Luis Potosi. The few poor collections available to Wolf did not reveal a taxonomic problem to him. In the years after Wolf’s monographic work, botanical exploration began to show the occurrence of plants of this affinity in several areas of northern Mexico. Collections have been made from a number of mountain masses that, though still leaving some small gaps, very nearly ‘‘connect”’ the central Mexican area with the Rocky Mountain area and reveal the ambiguous taxonomic position of the plants between them (Fig. 1). Examination of about five times as many collections of this complex as were studied by Wolf now suggests that the Mexican R. serrata extends in some of its numerous local forms to southern New Mexico and to two localities in Arizona (thus em- bracing the R. fasciculata Greene), but that the northern New Mexican and Coloradan plants are still best treated as a distinct species, R. smithii. The latter, in fact, on the basis of length of leaves and petioles, color of leaves beneath, characteristics of the leaf margin, and time of anthesis, seems more closely related to R. lanceolata Pursh, a species of lowland forests of eastern United States (Fig. 1). As will be evident from the following key, and from Wolf’s key, the dis- tinctions are tenuous and difficult to put down on paper. One valid alter- SIDA 6(2): 67-79. 1975. 68 native to my recognition of several species is to submerge all the members of this complex as varieties of R. lanceolata. This would create a wide- ranging and somewhat diverse species such as has been done for the Prunus serotina complex (R. MeVaugh, Brittonia 7: 279-315. 1951), for the Ptelea trifoliata complex (V. L. Bailey, Brittonia 14: 1-45, 1962), and for the Cercis canadensis complex (M. Hopkins, Rhodora 44: 193-211. 1942). But on the other hand the species that I recognize here are as well marked as those in other sections of Rkamnus, and there appears to be little to gain from the more inclusive approach. An ancillary problem is presented by R. standleyana, based by Wolf on a single collection (Purpus 4968) from a seldom-visited mountain range in Coahuila. The general habit of growth and the very small leaves, some- times fascicled, suggest that this is indeed a distinct taxon. Lending support to this thesis is another specimen from another range in Coahuila (Johnston 9031), which seems to me to belong to this category. In other characters (e.g., moderate number of very short almost invisible hairs on both leaf surfaces, yellowish color of undersurface of leaf, and prominent. sharp- pointed marginal serrations) these plants appear to be so similar to other members of R. serrata that they could be considered a variety thereof. However, until field study and other techniques can be used to clarify the relationships, I do not feel justified in making a new combination for these two specimens and will follow Wolf’s nomenclature, as in the following key. om la. Shrub usually more than 1 m high, the branchlets averaging longer than 1.5 cm and leaves averaging longer than 1.2 cm and not fascicled 2 1b. Small shrubs with branchlets averaging about 1.5 cm long and leaves all under 1.2 cm long, fascicled or tending toward fasciculate. R. standleyana 2a. Stipules quickly deciduous or semewiae persistent but only rarely at leafless nodes, averaging more than 3 mm long; below 3000 m. . . 3 2b. Stipules abundant, persistent even at leafless oe averaging less than 3 mm long; Guatemala above 3000 1 2 a % Ob: R. serrata var. guatemalensis 3a. Branchlets often long, averaging 5 cm or more in length; eastern United States below 1000 m. . . . . . . 1. R. lanceolata 3b. Branchlets shorter; at higher elevations, ant in caste United States oe 4 4a. Upper surface of leaf gray or olive-green, lower surface paler, rarely yellowish; bud-scales thin, pale; southern Colorado and northern New Mexico . . m4 2. R. smithit 4b. Upper Sines of leat green, ewes siace vallowish to Scone, rarely paler green; bud-scales coriaceous, dark reddish; Chiapas northwest- ward through western Texas, southern New Mexico, and Arizona. 3a. R. serrata var. serrata 69 Fig. 1. Documented distribution of members of the Rhamnus serrata com- plex. A, R. standleyana; G, R. serrata var. guatemalensis; R, R. serrata var. serrata; small dots, R. lanceolata, incomplete distribu- tion, the species occurring as far east as Pennsylvania, Maryland, and Virginia. — . RHAMNUS LANCEOLATA Pursh, Fl. Amer. Sept. 1: 166. 1814. Sageretia lanceolata (Pursh) G. Don, Gen. Syst. 2: 29. 1832. Rhamnus shortit Nuttall, J. Acad. Nat. Sci. Philadelphia 7: 91. 1834. Rhamnus parvifolius Torrey & Gray, Fl. North Amer, 1: 262. 1838. Cardiolepis nigra Raf., First Cat. Bot. Gard. Transylv. Univ. 18. 1824, nom.; Neogen. 1. 1825; Sylv. Tellur. 28. 1838. Cardiolepis rubra Raf., Neogen, 1. 1825; Sylv. Tellur. 28. 1838. Sarcomphalus shortianus Raf., Sylv. Tellur. 29. 1838. Shrub usually about 2 m high but at times larger, deciduous; bark green- ish and sparsely to moderately beset with short hairs when young, reddish and glabrous when older with a thin broken or peeling light-colored outer layer; branchlets frequently long and whip-like with distal leaves larger than proximal and not crowded; bud-scales 1-1.75 mm long and 1-2 mm broad, deltoid, dark reddish, tough and brittle when dried, apically acute or obtuse, with marginal hairs at apex only. Leaves mostly more than 3 cm long when mature, 2—-3.5 times as long as broad, at apex acuminate on the large distal leaves through acute on medium-sized leaves to obtuse on small proximal leaves of a branchlet, beneath somewhat paler green than or occasionally the same as above, above and beneath glabrous or with a few hairs which are sometimes visible only when the leaf is at an angle and either short on both surfaces or longer beneath than above; nerves usually glabrous above and with some hairs beneath, and beneath either dark- or light-colored and conspicuous or not; marginally plane (rarely revolute), crenate or occasionally serrate, the teeth (6—)8-18 per em and with or without dark-colored apexes; leaves thin and easily flexible; petioles (2—)3-6(-8) mm long (longer on type specimen) and usually with hairs only on the margin of an anterior furrow; stipules lorate, 3—-5(—6.5) mm long, 0.5 mm broad, pale yellowish, thin and flexible and quickly deciduous. Flowers unisexual, associated with immature leaves, in clusters of 1-3 in the axils mainly of small proximal leaves; pedicels (1-)2-3 mm long, gla- brous; floral cup (1—)1.5-2.5 mm deep, 1-1.5 (—-2) mm broad, deeper than broad, campanulate or funnelform (rarely hemispheric); sepals 4, ca. (1.25—)1.5-2 mm long, 1-1.25 mm broad, longer than broad: petals 4, ca. 1 mm long and broad (smaller in pistillate flowers), the broad portion gradually narrowing to form a slaw in staminate flowers, broadly notched, uniformly colored; style of pistillate flowers ca. 2 mm long, exserted, bi- furcate a fourth its length: filament of staminate flowers ca. 1 mm long, broadly based, the anther small and entirely enclosed by petal in pistillate flowers and partially exposed in staminate ones. Fruit 1-2 usually in axils of small proximal leaves of a branchlet, g abrous, 2-seeded (a few with 3 carpels at an early stage but one aborting), 4—5 mm in diameter (to 7 mm fide Wolf), either globose or somewhat prolate, black; pedicels up to 4 mm long, about 1 mm longer than at anthesis; floral cup glabrous and with margin entire or with 2 notches. 71 In moist wooded areas at less than 1000 m. The map (Fig. 1) shows a rather continuous distribution in the midwestern states but rather large dis- junctions in the South. The most striking hiatus, apparently real and not simply the result of inadequate collecting, is from eastern Texas to Ala- bama and northeastern Oklahoma or northwestern Arkansas, a gap of well over 500 km. The species is unknown in Louisiana and in the southern four- fifths of Arkansas. Rhamnus lanceolata var. glabrata Gleason (Phytologia 2: 288. 1947) seems to be no more than a trivial pubescence form. TYPE: United States: Tennessee: Lyon s.n. (PH!). REPRESENTATIVE SPECIMENS: United States: Arkansas: Moore s.n. (TEX, WIS). Hlinois: Umbach 8349 (TEX, WIS). Indiana: Friesner 18795 (TEX). Iowa: Davidson 546 (TEX). Kansas: McGregor E375 (TEX, WIS). Kentucky: Short s.n. (WIS); Wharton 10844 (NA). Maryland: Lyon s.n. (PH). Mississippi: Ray 8226 (WIS). Missouri: Hardin 666 (TEX). Ne- braska: Osborn 772R (MO). Pennsylvania: Davis 7061 (MO). Tennessee: Quarterman 4010 (NA). Texas: Palmer 5180 (US); Lindheimer 161 = exs, II 228 (MO). Virginia: Lyon s.n. (PH). 2. RHAMNUS SMITHII Greene, Pittonia 3: 17. 1896. Rhamnus smithii subsp. typica C. B. Wolf, Rancho Santa Ana Bot. Gard. Monoer., Bot. Ser. 1: 56. 1938. Shrub 1-2 m tall, deciduous; branchlets (1.5—)2-4(-6) ecm long with 6-8(-11) leaves and internodes averaging about 3 mm in length with mod- erate numbers of short hairs; bark greenish, smooth, with moderate num- bers of short hairs when young to grayish and rough and glabrous or sparse- ly hairy when older and often with a thin, white exterior layer; bud-scales 2.5-6 mm long, pale, flexible, with fimbriate margin and acute or obtuse or notched apex. Leaves J-1.5 cm long, 0.5-0.6 cm broad on flowering spe- cimens (immature), 4-7.5(-8) cm long, 1.38-2.7(-3.8) cm broad when ma- ture, (2.2—)2.4-2.7(-3.5) times longer than broad, at apex obtuse (usually on smaller leaves) or acute (usually on larger leaves), above gray- or olive-green, beneath paler than or occasionally the same as above or rarely yellowish or yellowish brown on older leaves, on both surfaces glabrous or apparently glabrous with scattered short erect hairs that are invisible when leaf surface is viewed at right angles; nervature above and bencath glabrous or with a few short hairs, beneath dark colored with rare excep- tions and conspicuous; margins plane or occasionally slightly revolute, crenate or crenulate with (7—)9-15(-17) usually remote teeth that are apical- ly amber colored to colorless; texture thick or semi-thick and barely flex- ible; petioles (2—-)3-6(-8) mm long with short to medium-length erect hairs usually more heavily concentrated along a_ ventral furrow; — stipules (2.5-)3-4(-5.5) mm long, 0.5 mm broad, pale brownish, membranous, with scattered marginal hairs, quickly deciduous. Flowers associated with imma- ture leaves, 1-3(—4) per axil, either unisexual or protandrous (only appar- 72 ently staminate flowers seen), pedicels 1.5-2.5 mm long, with sparse to moderate amounts of short, erect hairs; floral cup (1—)1.5-2 mm _ long, (0.75-)1-1.75 mm _ broad, longer than broad, campanulate to funnelform, glabrous, dark-colored; sepals 4, 1.25-1.5(-2) mm long, ca. 1 mm broad, longer than broad, glabrous, pale colored; petals 4, ca. 1 mm long, 0.75 mm bread, the broad part gradually narrowing to form a claw, narrowly and acutely notched, mostly uniformly colored or a few with darker-colored upper portion; style bifurcate about half its length: stamens 4, the filaments broad based, each enclosed by a petal, the anthers ca. 0.5 mm long, from half to completely exposed above petal, uniformly yellow; ovary glabrous. Fruit 1-2 per axil of small proximal leaves usually, glabrous, 2-seeded, (3—)4-6(-8) mm long and either globose or somewhat longer than broad, appearing mostly dark brown when dry, the floral parts either absent or a very short portion of the style retained, the floral cup shallow with two notches or less frequently entire and sparsely to moderately beset with short, erect hairs, and pedicels ca. 1 mm longer than at anthesis. On dry hillsides or along streams usually between 2000 and 2500 m but rarely up to 3000 m. It is known in Colorado and New Mexico (Fig. 1). The report of its occurrence in Arizona (T. H. Kearney and R. H. Peebles, Arizona Flora, p. 532, 1951) is based on a specimen now referred to R. serrata. ab] TYPE: United States: Colorado: Pagosa Springs, Smith s.n. (PH!) REPRESENTATIVE SPECIMENS: United States, Colorado: Andrews s§.n. (PH, US); Baker 458 (POM, US); Bethel, Willey, & Clokey 4197 (PH, POM, US); Brandegee s.n. (PH); Cary 172 (US); Eastwood 17 (POM, US): Eg- gleston 14182 (US); Graham 9709 (US); Payson 113-7 (TEX); Tucker 3226 UCLA); Vasey s.n. (US); Waterjail 15146 & 11710 (RSA): Weber & Living- ston 6333 (RSA, TEX); Wolf 3082 (RSA, UCLA, US); Wooton 2899 (US). New Mexico: Eggleston 6483 & 6566 (US): Standley 6835 (US). 3. RHAMNUS SERRATA Humb. & Bonpl. ex Schult. in R. & S., Syst. 5: 5. 181! Two varieties can be discerned, distinguished by those characters noted in the key; each of the varieties is described and treated below. 3a, RHAMNUS SERRATA Schult. var. see Rhamnus serrulata H.B.K., Nov. Gen. et Sp. 7: 51. . 1825. Rhamnus fasciculata Greene, Leafl. Bot. ae ae = 63. 1904. Rhamnus smithii subsp. fasciculata oe C. B. Wolf, Rancho Santa Ana Bot. Gard. Monogr., Bot. Ser 58. 1938 Rhamnus smithii subsp. fesesilinibe var. mulleri Fosberg, Lloydia 4: 286. 1941 Shrub from 1-3(-5) m tall, deciduous or evergreen; bark greenish or when young reddish, grayish or brownish when older with hairs short. to long (long only in Texas and most Coahuila collections), mostly erect and straight, of usually moderate amount sometimes glabrate with age (bark 73 observed to be rough and scaly on a collection from western Texas and several from Coahuila and some from uevo Le6én); branchlets (0.6-)0.9-9.5(-12) cm long (the average being about 1-2.5 cm), (4—)6-8(-14) leaves per branchlet; bud-scales 1-2 mm long, coriaceous, dark reddish, with apex broadly obtuse and with hairs dorsally (collections from New Mexico, Texas, and Coahuila) or marginally only (from farther south). Leaves (0.3-)0.5-4(-8.4) cm long with the shorter leaves usually proximal on the longer branchlets or throughout the shorter ones, (0.1—-)0.4-1.5(-2) cm broad, from broadly elliptic and ovate (the small leaves, usually) to oblong and lanceolate, at base cuneate to rounded, at apex mostly acute or obtuse (some apexes acuminate in the Chiapas collection; some retuse in a few collections from farther north), above green and in most cases either actually or apparently glabrous (hairs, when present, usually visible only if leaf surface is at an angle to the lens, short and mostly erect), beneath yellowish to brownish (a few New Mexico collections show simply paler green undersurfaces) and either actually or apparently glabrous or with moderate numbers of usually short, erect hairs; nervature above incon- spicuous or conspicuously pallid and either glabrous or with variable num- bers of mostly short, erect hairs and with midvein hairy and secondary nerves with the same amount of hair or less (rarely glabrous), the hairs erect and of variable lengths; margins revolute or plane (revolute more commonly to the South and plane more commonly to the North), usually serrate but sometimes crenate with (6—)10—-15(-17) teeth per cm having apexes all colorless or amber colored or dark and sometimes prominently large; texture from thin and flexible to firm and moderately thick; petioles 1-4(-5) mm long with moderate numbers of short to moderately long erect hairs which may be confined to or concentrated on the margin of a ventral furrow; stipules 2-5(-6) mm long and about 0.5 mm broad, pale-colored (usually reddish) and either thin or moderately thick with sparse to mod- erate numbers of hairs which may be exclusively on the margin or midvein and either quickly deciduous or somewhat papa — at leafless nodes). Flowers associated with mature leaves, monoecious or polygamod- ioecious or protandrous, 1-3 (rarely 4 or 5) per axil; ul ae absent; ped- icels usually 3 mm long or shorter (rarely up to 6 mm); floral cups varying n shape from funnelform through urceolate and campanulate to hemispheric ae latter uncommon), (0.6~-)1-1.5(-2) mm deep (usually shallower in pis- tillate than in staminate flowers), 1-2 mm broad, usually broader than deep, glabrous or with sparse to moderate numbers of short, erect hairs; sepals 4, ca. 1-2 mm long, 0.6-1.5 mm broad (usually smaller in pistillate than in staminate flowers), glabrous or with sparse to moderate pubescence; petals 4, ca. 0.5-1.1 mm long, 0.5-1 mm broad at widest part (smaller in pistillate than in staminate flowers), claw apparently more distinct in pistillate flow- ers, with a notch a fourth to a third the length of the petal, uniformly col- ored or with the distal portion darker; anther 0.5-0.75 mm long (shorter in pistillate flowers), enclosed by petal partially in staminate flowers and com- 74 pletely in pistillate flowers. Fruit 1-2 per axil, glabrous, 2-seeded, (3—)4-6(-8) mm _ broad, either globose or somewhat prolate, pedicels (1.5-)2-4(-6) mm long and with usually moderate pubescence; floral cup with margin circular or split into two parts and glabrous or with short hairs; style sometimes retained. Found from 1200 to 3000 m. TYPE: Mexico: Humboldt & Bonpland s.n. (B-WILLD No. 4637!) REPRESENTATIVE SPIECIMISNS: United States: Arizona: Hevly s.n. (ARIZ); Schroeder s.n. (ARIZ). New Mexico: Bailey 454 (US); Barlow s.n. (UCLA); Goodding 6574 (ARIZ): Eggleston 14361 & 14366 (US); Hinckley 6567 (ARIZ): Rehder 427 & 431 (US): Standley 40687 (US); Vaughan 1798 & 1799 (ARIZ); Wolf 2784 (RSA, UCLA); Wolf 2824 (POM, UCLA, Wooton s.n. (ARIZ, POM, US): Wooton 203 (POM, US). Texas, Correll 29751 (TIEX): Correll & Hanson 29821 (TEX); Ferris & Duncan 2528 (CA, DS); Havard 129 (US); Hinckley s.n. & 985a (TEX); Marsh s.n. & 354 (TEX); McVaugh 7445 (TEX); Moore & Steyermark 3161 (CA, UC, US); Palmer 30919 & 31992 & 34337 (US); Tharp, Warnock, & Hinckley s.n. (TEX); Warnock 1023 & 6843 & 12713 (TEX): Warnock & Tharp 5100 & 6843 (TIX). Mexico: Coahuila: Graber 135 (TEX); Johnston, Chiang, Wendt, & Ris- kind 11859 (TEX); Johnston et al. 9393 & 11240 & IIS86 & 11924 (TEX); Marsh 1938 (F, TEX): Marsh 1389 (F, GH, TEX): Johnston & Muller 597 (MICH, MIEXU, TEX); Muller 3188 (GH, NA); Pringle 2810 (GTI); War- nock 11604 (TEX). Chihuahua: Johnston et al. 10754 (TEX). Tamaulipas: Sullivan 590 (ENCB, TEX): Webster 134 & 146 (TEX). Nuevo Leon: John- ston et al. 11051 (TEX): Muller 399 & 455 & 846 (A, ', MICH, TEX); Muller 1034 (A, TEX); Muller 2872 (GH, NA): Taylor 98 (ARIZ, F, THX); White 1485 (ARIZ, GH, MICH, US). San Luis Potosi: McVaugh 10454 & 22161 (MICH); Palmer 57 & 214 (F, GH, US): Parry & Palmer 116 (GH, NA, US); Pennell 17810 (US); Purpus 5830 (F, GH, MEXU, US); Rzedowski 4504 & 8050 (ENCB); Rzedowski 6818 (ENCB, TEX): Rzedowski 7741 (ENCB, MICH, TIX). Hidalgo: Chiang ea al. 8128 (TEX): Rzedowski 22190 & 22937 (ISNCB, MICH). State of Mexico: Ehrenberg 979.6 (US); Paray 361 (NCB); Rzedowski 22021 CENCB, F, MICH, TEX); Rzedowski 24180 & 28525 CGU&NCB); Rzedowski 28825 (IaNCB, MICH); Uhde 1163 & 1169 (US). Distrito Federal: Bourgeau 698 (US); Espinosa 710 (NCB, TEX); Pringle 8055 (A, F, MEXU, MICH, POM, US); Pringle 11947 (F, GH, mixed with Rhamnus mucronata Schlecht.); Paray 1907 (IENCB); Reiche s.n. (ME ; US); Rose & Hough 4233 (US); Rzedowski 981 & 2266 (ENCB); Rzedowski 27245 (EANCB, MICH, TEX). Morelos: Lyonnet 745. Chiapas: Laughlin 1784 (TEX). ~ : 3b. RHAMNUS SERRATA Schult. var. guatemalensis L. A. Johnst., va nov. Frutex 0.3-3 m altus sempervirens cortice modice pubescenti pilis bre- vibus; ramuli (1.5-)3-9(-16) saepissime ca. 4.7 cm longi arcuati saepe 75 modice laxi; internodia (0.1—)0.2-1(-1.5) saepissime 0.2-0.3 em _ longa; squamae gemmae 1-2 mm longae ferruginosae coriaceae apice late obtusae nonnunquamque incisuratae marginae pubescentes. Folia (0.6—)1-4.2 cm longa (0.3—-)0.4-1.2(-1.7) cm lata longissima saepissime 3.07 cm longa 1.32 cm lata parvissima saepissime 1.33 cm longa 0.51 cm lata saepe duplo nonnunquam 3.3-plo raro 3.7-plo longiora quam lata crassiuscula supra nitida pilis modice abundis brevissimis fere invisibilibus raro absentibus subtus fere concoloria glabra vel rarissime pilis brevissimis margine revo- luta obtuse serrata crenatave dentibus saepissime introrsis mucronibus praesentibus vel nullis; nervi supra glabri subtus glabri vel pubescentes; costa subtus flavescens; petioli 1-2.5(-3.5) mm longi; stipulae tenues pal- lidae 2-3.5 mm longae 0.5 mm latae pilis sparsis multis persistentibus. Flores axillares; petalae omnes unicolores. Fructi glabri 2-spermi, pedicelli 2.5-4 mm longi pilis modice abundis brevissimis erecti. Shrubs 0.3-3 m tall, evergreen (leaves present in December and Januar collections), bark with moderate amount of short pubescence; eee (1.5—)3-9(-16) cm long (averaging about 4.7 cm), with curving habit, gen- erally not very crowded (exception: Molina 21222); internodes (0.1—)0.2—1 (-1.5) cm long, most frequently ca. 0.2-0.3 cm long; bud-scales 1-2 mm long, at apex broadly obtuse and occasionally notched, dark rust-red, cori- aceous, with marginal hairs. Leaves (0.6—)1-4.2 cm long, (0.3-)0.4-1.2(-1.7) cm broad, the largest averaging 3.07 cm long, 1.32 cm broad, the smallest averaging 1.33 cm long, 0.51 em broad, mostly twice as long as broad, at times to 3.3 times and rarely 3.7 times as long as broad, thickish, at apex acute, above lustrous with a moderate amount of very short almost invisible pubescence or rarely glabrous, beneath about the same color as above and glabrous or very rarely with short hairs, at margin revolute, bluntly serrate (the teeth introrse, with rare exceptions) or crenate and with mucros pres- ent or absent; petioles 1—-2.5(-3.5) mm long, with dense pubescence along the ventral suture, sparser pubescence elsewhere with short hairs; stipules thin, pallid, 2-3.5 mm long, 0.5 mm broad, persistent, with scattered hairs. Flowers (only one seen) axillary; petals uniformly colored. Fruits glabrous, 2-seeded (rarely 3-seeded on an otherwise 2-seeded specimen); stigma re- tained, birfurcate or trifurcate; floral cup retained either circular marginal- ly or notched on two sides: pedicels 2.5-4 mm long with moderate amount of very short, erect pubescence. In Juniperus woods, on rocky slopes and limestone bluffs at 3300 m most frequently but up to 3900 m at times. TYPE: Guatemala: Huehuetenango: Sierra de los Cuchumatanes, vicinity of Tunima, Steyermark 48377 (holotype, A!, isotype, F!). OTHER SPECIMENS SEEN: Guatemala: Huehuetenango: Johnston 1775 (F); Molina 21222 (F); Molina, Burger &, Wallenta 16477 (F); Standley 81741 & 81769 (F) 76 4, RHAMNUS STANDLEYANA C. B. Wolf, Rancho Santa Ana Bot. Gard. Monoer., Bot. Ser. 1: 51. 1938 Small shrub (probably less than 1 m high); bark reddish and fairly smooth, with short hairs; primary branches long; secondary branchlets short, 0.3-0.4 cm long, averaging about 0.6 em, with (8—)4-12 leaves per branchlet; internodes 0.1-0.6 mm long or leaves tending to be fascicled; bud-seales minute and brown. Leaves 0.38-1.2 em long, 0.2-0.5(-9.7) em bread, about 1.5 to 2.5 times as long as broad, thickish, at apex obtuse or tending to be obtuse, above green, beneath yellowish, above and beneath with moderate amounts of very short pubescence that is almost invisible when leaf is flat, at margin plane and serrate with teeth that are antrorse and slender and apically sharp-pointed and colorless; nervature with short hairs; petioles 1-2 mm long with moderate amount of short pubescence; stipules quickly deciduous. Flowers unknown. Fruit solitary in the leaf axils, glabrous, 2-seeded, ca. 6 mm in diameter, the retained portion of the floral cup with seattered short hairs and entire margin; pedicels 2-4 mm long with scattered short hairs. Found from about 25090 to 3000 m in limestone sierras. TYPE: Mexico: Coahuila: Sierra de la Paila, Purpus 4968 (F, GH, US) OTHER SPECIMEN SEEN: Mexico: Coahuila: Johnston 9031 (GH). — DISCUSSION In a species with as large a range as R. serrata, morphological and other diversity can be expected. I have delimited only two varieties, but I shall discuss certain geographic patterns and show affinities with the other species treated in this paper. In the first place R. smithii and both varieties of R. serrata comprise shrubs usually 1-3 m tall, but shrubs up to 4-5 m have been recorded by collectors in San Luis Potosi and Hidalgo, although this great stature does not seem to correlate with any other distinctive features. Rhamnus stand- leyana is evidently shorter than 1 m. It appears that plants from Texas northward are deciduous and from Mexico southward are evergreen. Ob- servations suggesting the former are (a) the lack of winter or early spring collections from Texas, New Mexico, and Colorado, (b) a collection (Correll 29751) in mid-June at a high elevation in western Texas (‘‘base of ledge below Kmory Peak’) with flowers all in bud and leaves immature both in texture and size, and (c) the presence of only very young immature leaves on early spring collections of R. lanceolata. Evidences of the evergreen condition are (a) a spring collection (Webster 146), a March collection (Rze- dowski 22021) and a January collection from the Federal District (Pringle 8055) all with large leaves along a primary axis (presumably from the pre- vious year’s growth) and secondary axes, shorter and with smaller leaves arising from its axils (presumably from the current year’s growth), (b) De- cember and January collections from Guatemala with intact leaves. I found only one exception to this pattern, viz. a February collection from the state 77 of Mexico (Rzedowski 28825) having only small (immature?) leaves along with flowers that are almost all in bud stage. Wolf (Rancho Santa Ana Bot. Gard Monogr., Bot. Ser. 1: 55. 1938) attributed the deciduous character to R. smithii (sensu latiore, including R. fasciculata, which I now place under R. serrata); he claimed that everything he referred to R. serrata was ever- green. However, because we are dealing with plants some of which are found in the temperate zone where strong sasonal changes may affect this character and others are found in tropical mountains Where the major seasonal change is from wet to dry, I am unwilling to base taxonomic dis- tinctions on the character expression of deciduous versus everereen. Based on averages and appearances, R. serrata has longer branchlets at the southern end of its range and shorter branchlets at the northern end; but taken individually, the collections do not always fall into neat categories, For instance, most Coahuila and Texas collections have branchlets not ex- ceeding 5 cm in length, but in five eases out of 18, there are branchlets ex- ceeding 7 cm in length and even up to 11 cm in one case. At the southern end of the range, in San Luis Potosi, branchlets of 5 em are quite common and roughly half the collections have branchlets of 8 and 9 cm and even up to 12 cm in length; but in the Federal District and state of Mexico, al- though 5 cm is a rather common length, only one collection has branchlets exceeding 6 cm in length (Espinosa 710 has branchlets ca. 15 and 30 em in length and therefore in this respect resembles R. lanceolata more than it oes Rh. serrata). Therefore, although a tendency can be seen in branchliet length, it is not correlated sufficiently with other characters or localized enough geographically to permit its use in the infraspecific taxonomy. Because of the superficial appearance of greater crowding of leaves in the two collections (Hinckley s.n.) from Mt. Livermore in the Davis Moun- tains of western Texas and in a group of Coahuila collections, as compared with otherwise similar collections from the Big Bend National Park, Texas, internode length was measured, but the findings were inconclusive. Through- out the range from Chiapas to New Mexico, the most numerous measure- ments were 1-4 mm, the next most numerous were 5-9 mm and the least numerous were 10-20 mm. The longest internode lengths, as expected, are found in general on the longest branchlets. Leaf sizes of collections from Mexico, Texas, and New Mexico reveal that Muller 3188 and 2872 (described by Fosberg as R. smithii subsp. fasciculata var. mulleri), two Rzedowski collections from Hidalgo, and Espinosa 710 from the Federal District are the only ones with any leaves exceeding 5 cm in length. In this respect they are like R. smithii of Colorado and R. lance- olata but not like the smaller-leaved plants of Texas and southern New Mexico that Wolf calls R. smithii subsp. fasciculata. But leaf-size data alone and as a whole do not support segregation of these collections from R. ser- rata. The two Rzedowski collections from Hidalgo have leaves as much as five times as long as wide but the others (notably Muller 3188 and 2872) have leaves about 2.5 times as long as wide, a common proportion for 78 medium-sized leaves throughout the range. Also common is that leaves which measure less than 1 em are usually 1-1.5 times as long as broad (rarely up to 2.8 times) and are usually proximal on longer branchlets and exclusive on shorter ones. In general, therefore, for R, serrata as a whole, there is a gradual decrease in ratios of length to width from upper to lower extremes in length of leaf. A specimen from New Mexico (Goodding 6574) and one from Arizona (Schroeder s.n.), which belong to R. serrata on the basis of geographic lo- cality, leaf size, and dark-tipped petals, lack the yellowish or brownish coloration of the leaf undersurface. In this respect, they are more like R. smithii and R. lanceolata. In the original description of R. serrata, the leaves are said to be glabrous. Actually, hairs in variable densities and lengths can sometimes be found, and some degree of geographic correlation can be seen. The undersurface beneath is glabrous in Chiapas, state of Mexico, Hidalgo, Tamaulipas, most Nuevo Leon, and some Coahuila specimens. Pubescence is sparse to mod- erate in a few collections from Nuevo Leén and in most from Coahuila, Chihuahua, Texas, and New Mexico. Long and fairly abundant hairs are common on both surfaces in Texas collections. Such hairs are otherwise found only on the undersurfaces of a few collections from northern Mexico On the uppersurface, short hairs are common in New Mexico, Nuevo Leon, and Coahuila collections and in the solitary collection from Chiapas, while hairs are absent commonly in Tamaulipas, San Luis Potosi, the Federal District, and Morelos. There is some evidence of a geographic gradient in pubescence of nerva- ture of the leaf upper surface. Most collections from the state of Mexico have glabrous nerves. Both glabrous and hairy nerves can be found in San Luis Potosi and Nuevo Leon, and only hairy ones in Coahuila, Texas, and New Mexico. Tamaulipas and Chihuahua collections apparently fall into the latter category but they are too few for certainty Apexes of marginal teeth are exclusively colorless in all collections from the Federal District, state of Mexico (except from Espinosa 710), Hidalgo, and San Luis Potosi (except Pennell 17810) and in one collection from Ta- maulipas, one from Nuevo Le6én, and one from Coahuila. They are amber colored in most New Mexico collections, in that ped! resembling R. smithii. Dark-colored apexes are seen in all other collections, including the one from Chiapas. In general, stipule length averages shorter in the northern part of the range and longer from Nuevo Leén southward, but great variability and overlap occurs, and the character is considered insignificant taxonomically. Pedicel length of 1-3 mm at anthesis is essentially uniform for all four species with only two exceptions, Sullivan 590 from Tamaulipas and [ze- dowski 7741 from San Luis Potosi, which have pedicels measuring 4-6 mm. There is no constant distinction between staminate and pistillate flowers in this character. Because I am defining R. smithii and R. serrata differently from Wolf, 79 something must be said about the floral characters that he tried to use to distinguish these two species. According to Wolf, the petals of R. smithii are uniformly colored whereas in R. serrata they have dark tips. I find ex- clusively unicolored petals only in R. lanceolata and in R. serrata var. gua- temalensis. Otherwise the findings are as follows. Gooddings 6574 from New Mexico has dark-tipped petals. Schroeder s.n. from Arizona has some dark- tipped and some unicolored petals on the same branch. Among the Texas collections, Palmer 34337 (pistillate) and Correll 29751 (staminate) have unicolored petals, Moore & Steyermark 3161 (staminate; cited by Wolf) has dark tips gradually fading into paler claws, and Palmer 30919 (stami- nate) has the blade somewhat darker than the claw. Among the Mexican collections, some have both unicolored and dark-tipped petals on the same plant with some petals mottled, including Johnston et al. 11240 (staminate), Sullivan 590 (staminate), Rzedowski 7741 (staminate), Pringle 8055 (stam1- nate and pistillate; cited by Wolf), Rzedowski 2266 (staminate), Paray 1907 (staminate), and Ehrenberg 979.6 (staminate, cited by Wolf). Other Mexican collections (e.g., Johnston et al. 10754) have petals unicolored or with the tip only ohtly. darker than the claw; and still others have all unicolored petals, including Graber 135 (pistillate), Muller 455 (staminate), Purpus 5330 (pistillate, difficult to interpret; cited by Wolf), and Rzedowski 22190 (staminate and pistillate branches). There is slight evidence for a sex cor- relation from the foregoing data. Female flowers may always have unicol- ored petals, but male flowers may have either unicolored or dark-tipped petals. The small size of the sample precludes a definitive decision at this time. Another of Wolf's statements bears rather full discussion. He attributed only unisexual flowers to R. smithii (including the New Mexico and Texas populations), and ‘‘apparently perfect’? ones to R. serrata. My own observa- tions suggest that the flowers of all these plants are unisexual. It is possi- ble that they may be ea but the following observations (based on all the flowering specimens of R. smithii and R. serrata seen) suggest other- wise to me. The ay staminate flower (judging from the pollen pres- ent on the anthers and the small size of the pistil) has the following charac- ter-expressions: (a) anther only partially enclosed by the petal, (b) anther 0.50.75 mm long, (c) filament 0.5-1 mm long, (d) floral cup 1-2 mm deep, (e) sepals as much as 2 mm long and 1.5 mm broad at base, (f) petals more often dark-tipped than in pistillate flowers, (g) petals 0.6-1.1 mm long and wide, and (h) petal claw broad and indistinct. On the other hand, the ap- parently pistillate flowers (judging from the well developed pistil and very small anthers) show the following: (a) anther completely enclosed by petal, (b) anther only 0.25 mm long, (c) filament about 0.2 mm long, (d) floral cup 0.6-0.75 mm deep, (e) sepals not exceeding 1.5 mm long and 1.2 mm broad at base, (f) petals more often than not uniform in color, (g) petals only 0.5-0.6 mm long and wide, and (h) claw narrow and distinct. The sample may not have been large enough to serve as a basis for firm conclusions, but the evidence is certainly suggestive. TAXONOMY OF THE LYTHRACEAE IN THE SOUTHEASTERN UNITED STATES SHIRLEY A. GRAHAM ce of Biological Sciences t State University po Ohio 44240 ABSTRACT. Keys, distributions, and taxonomic and nomenclatural notes are given for seven genera and 16 species of Lythraceae occurring in south- eastern United States. Changes in the treatment of southeastern Lythraceae include recognition of the adventive ROTALA INDICA (Willd.) Koehne in Louisiana, uniting of AMMANNIA TERES Raf. of the United States with the Caribbean A. LATIFOLIA L., acceptance of the monotypic genus DIDI- PLIS, and presentation of evidence for uniting LYTHRUM ALATUM Pursh and L. LANCEOLATUM Ell. as LYTHRUM ALATUM var. LANCEOLATUM (Ell.) T. & G. ex Rothrock, a previously overlooked combination, The chromosome number of CUPHEA GLUTINOSA Cham & Schlect. is reported for the first time. INTRODUCTION This paper is an outgrowth of a floristic treatment of the Lythraceae writ- ten for a proposed Vascular Flora of the Southeastern United States, edited by A. E. Radford. Although the Lythraceae are represented in southeastern United States by only seven genera and 16 species, Ammannia, Didiplis, and Lythrum presented taxonomic problems requiring investigation beyond the geographical boundaries of the flora. The treatment of these genera therefore is more detailed than that of the rest of the family. The curators of the following herbaria are gratefully acknowledged for loan of the material indicated: all Lythraceae, ALU, LAF, LTU, TENN, UNC, MISS, VDB, WVA; Lythrum, A, GA, GH, MO, TEX, and US; Am- mannia, A, GH, MO, TEX, US, and WIS. I also wish to thank the many botanists who shared freely their special knowledge or gathered living ma- terial for me. These include Loran Anderson, Delsie Demaree, Joseph Ewan, Samuel Jones, Elizabeth Shaw, John W. Thieret, and Daniel Ward. I am in- debted to Robert Ornduff for his helpful comments on the manuscript. 1 Based on manuseript compiled for the Vascular bide of a southeastern United States. The area covered is bounded by and includes Delay Maryland, West Virginia, Kentucky, Arkansas, and Louisiana. The ae) of the neater and method of citing distribution and synonymy basically follows Radford cf al. (1967). Letters Following synonyms cited at the end of species seen refer to the following recent floras of the eastern United States in which those names are used: S == Small, J.K., Masaal of he sontheastern flora, 1933; F - Fernald, M.L. ae ‘ mre of botany, 8th ed., 195 G = Gleason, H.A. and A. Cronquist, Mania of the vascular plants of the er ere United States and a Canada, 1963: BR Radford, A.E., ELE. Ahles, and C.R. Bell, Manual of the f 4 ea vascular flora of the € es een 196 SIDA 6(2): 80-103. 1975. LYTHRACEAE Loosestrife Family Herbs, shrubs, or small trees with quadrangulate or terete stems. Leaves simple, entire, decussately opposite, seldom alternate or whorled, exstipu- late. Flowers perfect, regular or irregular; bracteoles 2, opposite on the pedicels; floral tube campanulate to cylindrical, persistent; calyx lobes 4-6, generally alternating with appendages; petals 0-6, crumpled, decidu- ous, inserted on the inner surface of the floral tube between the calyx lobes; stamens as many as or twice as many as the petals or numerous, inserted on the inner surface of the floral tube below the petals, anthers versatile; gynoecium often subtended at the base by a disc, stigma capitate or bi- lobed, style filiform, ovary superior, free in the floral tube, 2—4-locular; fruit a membranaceous capsule enclosed by the persistent floral tube, sep- ticidally, septifragally, or loculicidally dehiscent, or indehiscent, splitting irregularly, Seeds 3 to many, the embryo straight, endosperm scant or none. 1. Floral tube campanulate to globose, about as long as wide. 2. Plant woody, perennial, over 1 m tall: petals showy, 8-20 mm long. wo . Flowers in axillary dichasia: petals 8-15 mm long; stamens 8-10; shrubs of shallow water... . . . . . 4. Decodon ew) . Flowers in terminal panicles; petals 12-20 mm long; stamens nu- merous, more than 10; terrestrial trees or shrubs . 2. I 4 . © AUGETSUPOE TUL 2. Plant herbaceous, annual or perennial, generally less than 50 cm tall; petals to 4 mm long or none. 4. Appendages present in the sinuses of the calyx lobes, or if absent then the bracteoles nearly equal to the length of the floral tube; flow- ers 1 to many in the axils of leaves: petals 4 (rarely 0); capsule de- hiscent. 9. Middle and upper leaves attenuate at base; flowers solitary in the axils of leaves; capsule dehiscing septicidally, the outer wall of the capsule finely and densely transversely striate, . . 3. Rotala or . Middle and upper leaves cordate to auriculate at base; flowers (1-)3 to many in the axils of leaves; capsule dehiscing irregularly, the outer wall of the capsule neoth: not striate . . 4. Ammannia 4. Appendages absent in the sinuses of the calyx lobes; bracteoles mi- nute on the pedicel; flowers solitary in the axils of leaves: petals 0; Capsule lndehiscent «. 4 << « «4» &€ <6 «aa & & Didiplis 1. Floral tube cylindrical, elongate, about twice as long as wide. 6. Flowers regular; floral tube entire in fruit; capsule dehiscing septicidal- ly from the apex, the placenta included; seeds numerous, mostly more than 20 oe : : : e i Je Ube 6. Flowers irregular; floral tube and capsule splitting longitudinally along the upper side in fruit, the placenta then exserted: seeds 3—20.7. Cuphea Lythrum 1. DECODON J. F. Gmelin 1.D. VERTICILLATUS (L.) EIll., Swamp Loosestrife—Perennial shrubs, glabrous to velutinous, spreading by arching stems rooting at tips and with submerged stems basally thickened by spongy aerenchyma. Leaves opposite or whorled, lanceolate, shortly petiolate, 3-20 cm long, 0.5-5 cm wide. In- florescence of 1-3 shortly pedunculate axillary dichasia at a node. Floral tube campanulate, greenish; flowers trimorphic with styles and stamens of 3 lengths; calyx lobes 4 or 5(-7), alternating with and mostly } the length of narrowly triangular, thickened appendages; petals 4 or 5(-7), rose purple, 8-15 mm long, about twice the length of the floral tube; stamens 8-10, the filaments of 3 possible lengths, 2 of the 3 lengths occurring in any one flower; capsule globose, loculicidal, 3-5-locular. Seeds 20-30, small, inverted- pyramidal. (n=16) Summer. Swamps, edges of ponds and lakes in shallow water; all prov. SE /All/ A monotypic eastern North American genus dis- tributed from southeastern Canada and Maine to Minnesota, south to Louisi- ana and Florida. nie Glabrous specimens said to occur mostly inland (Gleason and Cronquist, 1963) have been called var. laevigatus T. & G. On the basis of admittedly limited herbarium material from beyond the southeastern states it appears that there are in fact relatively few entirely glabrous specimens and that, although these occur primarily inland, glabrous plants occur near coastal areas as well, e.g., Cameron Parish, La., Jones Co., N.C., Washington, D.C. Conversely, plants with the more common velutinous condition, although mostly coastal or in the Mississippi Valley, have also been collected inland, e.g., Noble Co., Ind. and Jefferson Co., N.Y. Distinctive geographical limits for the varieties based on degree of vestiture seem to be lacking, and in the absence of other discriminatory characters I prefer not to recognize vari- etal rank within the species. 2. LAGERSTROEMIA LL. 1. L. INDICA L., Crape-Myrtle—Long-lived shrubs or small trees. Leaves + alternate, obovate or oblong-elliptic, sessile or shortly petiolate, 15-60 mm long, 5-40 mm wide. Inflorescence of showy multi-flowered terminal panicles. Floral tube subglobose, 10 mm long, on pedicels 3-15 mm long; calyx lobes 5-7, erect, shorter than the floral tube, appendages lacking; petals 5-7, purple, rose, or white, long-clawed, 12-20 mm long; stamens numerous; capsule subglobose, loculicidal, (3—)6—locular. Seeds flattened, winged, 3 or 4 per locule. (n= 24, 25) Spring—Fall. Native to Asia, widely cultivated throughout the SE and possibly naturalized to a limited extent. 3. ROTALA L. Annual glabrous herbs, often much branched, 0.5-4 dm tall. Leaves op- posite or whorled, linear to spathulate, attenuate at the base, mostly sessile. Flowers solitary, axillary, sessile or subsessile, 4(-6)—merous; pedicels bi- 83 bracteolate. Floral tube campanulate to globose or urceolate, 2-5 mm long, calyx lobes 4(-6), appendages present or wanting between the lobes; petals 4(-6), white to pink; stamens 4(-6); capsule with microscopic dense trans- verse striations on the outer wall, septicidally dehiscent, 2-4-locular. Seeds many, minute, ovoid, plano-convex. 1. Leaves linear to oblanceolate; floral tube bearing appendages between the calyx lobes; bracteoles less than 4% the length of the floral tube. 1. R. ramosior 1. Leaves obovate to spathulate; floral tube lacking appendages; bracteoles + equalling the length of the floral tube . . . . . 2. R. indica 1. R. RAMOSIOR (L.) Koehne, Tooth-Cup—Herb 0.5-4 dm tall. Leaves oppo- site or whorled, linear to oblanceolate, 10-50 mm long, 2-12 mm wide. Flow- ers axillary, sessile, bracteoles at the base of the floral tube less than % the length of the floral tube. Floral tube 2-5 mm long; calyx lobes alternating with triangular appendages of size nearly equal to the lobes; petals scarcely exceeding the calyx lobes, deciduous; capsule 2—4-locular. (n=16) Mid- summer-Fall. Muddy or sandy shores or damp depressions; all prov. SE /All/ Also from Mass § to Fla, E to E Tex, US west coast, Cent. Am, S. Am, Antilles. The species bears a superficial resemblance to both Ammannia coccinea Rottb. and Ludwigia palustris (L.) Ell. (Onagraceae), with which it is some- times confused. The obvious characters in Rotala—i.e., upper leaf bases tapering, solitary flowers in the axils, and minute transverse striations on the capsule—easily distinguish it from Ammannia, which has upper leaf bases auriculate, generally 3 or more flowers in axillary cymes, and a mi- croscopically smooth capsule wall. Easily observed characters of Ludwigia palustris that distinguish it from R. ramosior are the absence of appendages between the lobes and the presence on the floral tube below each calyx lobe of a thickened green band through which the outline of seeds may be seen in the dry state. More robust plants of R. ramosior occurring inland from the coastal plain have been recognized as var. interior Fern. & Griscom on the basis of but slightly greater size of leaves, bracteoles, and capsules. 2. R. INDICA (Willd.) Koehne—Herb 0.5—-3 dm tall. Leaves opposite, obovate to spathulate, 4-17 mm long, 1-8 mm wide, the margins thickened, trans- lucent, the apex minutely mucronate on young leaves. Flowers subsessile on foliaceous branchlets; bracteoles + equalling floral tube in length, narrowly linear, acute. Floral tube 2-3 mm long; calyx lobes narrowly triangular mm long, lacking alternating appendages; petals much shorter than the calyx lobes, persistent; capsule 2-locular. Summer-Fall. Rice fields; Acadia Parish, La/Calif/ Native to S Asia. A native of southern Asia, this species is thought to be an introduction through rice culture in California (Mason, 1957). It has recently also been found in rice fields in Louisiana (Thieret, 1972, and M. Piehl, 1974, pers. 84 comm.) According to Thieret (pers. comm., 1972) the species has persisted and spread in its second year in the area where it was first collected. For this reason and because it might be expected to appear in other rice grow- ing areas of the southeast it has been included in this treatment. 4. AMMANNIA L. Annual glabrous herbs, 1-11 dm tall, with ascending branches. Leaves opposite, narrow, linear to lanceolate or oblanceolate, sessile, auriculate or cordate, rarely tapering to the base. Flowers small, 4-merous, in sessile or pedunculate axillary cymes, (1—)3—-10 flowers per cyme; pedicels bibrac- teolate. Floral tube campanulate to urceolate, 2-6 mm long, greenish to rose-colored, 4-angled at anthesis; calyx lobes 4, alternating with short, horn-shaped appendages; petals 4 or 0, small, purple to pink, early decidu- ous; stamens 4(—8); capsule globose, irregularly dehiscent, outer wall smooth, not striate. Seeds many, minute, ovoid. Style in fruit exserted, filiform, equal to or longer than the ovary; calyx 1. A. coccinea _ lobes triangular, the apices acute . . . . . . Style in fruit included, much shorter than the ovary; calyx lobes obtuse, . 2. A. latifolia —_ the apices often minutely mucronate nae 1. A. COCCINEA Rottb.—Leaves linear and attenuate to narrowly oblong, 2-11 cm long, 2-10 mm wide, the apex acute, the base auriculate- cordate, Inflorescence a closely-flowered, short-pedunculate cyme, flowers (1-)8-5(-10) per cyme, peduncles to 3 mm long. Petals 4; style filiform, equal to or longer than the ovary, exserted in fruiting calyx; calyx lobes triangular, the apex acute; capsule 3-5 mm in diameter, included to barely surpassing the calyx lobes. Summer-Fall. Wet places; all prov. SE (except Ala?) /Mo, Ill, Ind, Ohio, NJ/ Also sporadic to west coast and in Mex, Two long-styled species of Ammannia have been cited from southeastern United States (Fernald, 1950), A. coccinea, characterized by closely (1—)38—5(—10)-flowered sessile or nearly sessile cymes and A. auriculata Willd., with smaller, more-numerous flowered (3-15 flowers) long-peduncu- late cymes. Ammannia coccinea is frequently collected throughout eastern and central United States and ranges southward into South America via the Caribbean and Mexico. Ammannia auriculata is a species of world-wide distribution, occurring in Africa, Asia, Australia, Latin America and is infrequently collected in the United States from Texas north to South Da- kota and possibly also in California, but to my knowledge is not present within the boundaries of the Vascular Flora. Ammannia coccinea and A. auriculata are sympatric over a very wide area, 1.e., the central United States, Mexico, the Caribbean, and northern South America. Throughout the area of sympatry, specimens representing apparent hybrids or introgressants have been collected. These resemble A. coccinea most closely, but they display the elongate peduncle of A. 85 auriculata and are most often identified as that species. The possible intro- egressants are also found infrequently outside the range of A. auriculata in the southeastern states bordering the Mississippi River. Other characters such as flower number and capsule size, although statistically significant in separating the two species, may or may not be sufficiently distinct on any one individual to be a useful distinguishing feature. A biosystematic study of these species is currently in progress. Occasional specimens of A. coccinea from well outside the range of A. auriculata are found with the most mature cymes fully elongated into short branches, succeedingly younger cymes with slightly elongate to sessile pe- duncles (cf. Godfrey 64910 from Florida, LAF, TEX, UNC; and Thorne 5628 from Georgia, GH). In this case it would appear that elongate peduncles must be attributed to a genetic change not connected with hybridization. 2. A. LATIFOLIA L.—Leaves lanceolate to narrowly elliptic or spathulate, 5-14 cm long, mostly about 8 mm wide, the apex obtuse to subacute, the base of middle and upper leaves auriculate, the lowest cuneate. Inflorescence a closely flowered, short-pedunculate cyme, flowers 3-10 per cyme, peduncles to 3 mm long. Petals 4 or 0; style much shorter than the ovary, included in fruiting calyx; calyx lobes broad, rounding, the apex often minutely mu- cronate; capsule 46 mm in diameter, included. Summer-Fall. Brackish waters; cp. Md, Va, NC, GA, Fla, Miss, La/NJ, Tex/ ne Mex, S. Antilles. A. teres Raf.—F, G, R; A. koehnei Britt.—S. Two short-styled species of Ammannia are commonly recognized from the Atlantic and Gulf coasts, A. teres, distributed from New Jersey to Florida and westward to Texas, and A. 7 atifolia, primarily a Caribbean species reaching northward into southern peninsular Florida, The major difference between the two taxa is the purported presence of petals in A. teres and their absence in A. latifolia. Leaf shape and size and capsule size have also been used to separate the species. In the course of a revisionary study of the genus I have tabulated these characters for United States, Caribbean, and Latin American specimens. No geographically-limited character was found by which more than one species or even variety could be defined. Although petals occur on all plants examined from New Jersey to northern Florida, they also occur sporadically throughout the Caribbean on 30% of the specimens examined. At least four other species of Ammannia have petals that vary from 0-4, so that vari- ability of this character within a species is not unusual for the genus. Leaf size is not significantly different between the two species even at the northern and southern range extremes. The upper leaf shape is pri- marily lanceolate to spathulate in the north, becoming more commonly linear-lanceolate from North Carolina southward, with spathulate-leaved specimens occurring infrequently in the Caribbean. Capsule size is more variable in the north, but the difference in mean size at the geographic extremes is less than 0.5 mm. 86 Because of the lack of stable morphological characters which can be ge- ographically characterized, only one short-styled species of Ammannia, A. latifolia, 1s recognized in eastern United States. Although this species is variable, it is without definable infraspecific groups. Ammannia teres is relegated to synonymy. The complete synonymy and more extensive data supporting these changes will be presented in a revision of Ammannia in the Western Hemisphere (Graham, unpub.) Fernald’s A. teres var. exauriculaia is merely a form in which the basal auriculate lobes on some of the upper leaves are much reduced. This ap- parently sporadic character is known also from collections from the Ba- hamas (Small & Carter 8965, US) and from Yucatan (Lundell & Lundell 8188, MEXU). 5. DIDIPLIS Raf. 1. D. DIANDRA (DC) Wood, Water-Purslane—Delicate annual aquatic herbs, 1-4 dm long. Leaves thin, opposite, 5-30 mm long, 0.5-4 mm wide, narrowly linear or elliptic, the submersed leaves linear, truncate at the base, the emersed leaves narrowly elliptic, shorter with tapering base. Flowers minute, greenish, 4-merous, solitary, axillary; calyx lobes 4, broad- ly triangular, intersepalar appendages lacking; petals none; stamens 2-4, included; style short or wanting; capsule globose, indehiscent, irregularly splitting, 2-locular. Seeds many, small, spathulate, the distal end enlarged and curved. Late Spring-Summer. Margins or shallow waters of lakes, ponds, temporal pools; all prov? SE /All?/ Reported from Tex to Fla, north to Minn and Va, but SE herbaria contain specimens only from Ky, La, Miss, NC, and Va. A monotypic genus endemic to eastern North Amer- ica. Peplis diandra Nutt. ex DC.—F, R. The decision to recognize Didiplis as a monotypic genus endemic to North America versus placing it in Peplis (cf. Graham, 1964) is based on Webb’s (1967) convincing arguments for the merging of the European species of Peplis with Lythrum. According to Webb, morphological similarities or ac- tual overlap in characters of habit, petal and stamen number, shape of the floral tube, and type of capsule dehiscence, especially as evidenced in the little-known European Lythrum thesioides Bieb. and L. borysthenicum (Schrank) Litv. (Peplis erecta Req. ex Moris) erase the distinctions between the two genera. One apparent exception is Didiplis diandra, a species with 4-merous, rather than 6-merous flowers which lack the appendages (epi- calyx) found on all other species of Lythrum and Peplis. The flowers of Didiplis, often produced on totally submerged plants, are solitary, minute, and greenish, with both appendages and petals lacking, probably reduced in response to the aquatic environment. In Lythrum the 4-merous floral condition is present in a few species, such as L. thymifolia L. and L. tribracteatum Salzm. ex Sprengel and appendages may be conspicuously reduced (‘‘subnullae”’ fide Koehne, 1903, p. 59) or in- conspicuous, to 0.5 mm long or less, as in L. tribracteatum (Webb, 1968, 87 p. 302). The presence of such character states in Lythrum suggests that Didiplis is no more easily separated from Lythrum than is Peplis. Without studying the European species involved and lacking sufficient study ma- terial of Didiplis itself, I choose to follow Webb’s suggestion for the present, that Didiplis be treated as a monotypic American genus. Didiplis diandra is a delicate aquatic herb of shallow, often ephemeral waters. Though most floras record its range as covering much of eastern United States, there are only 12 herbarium specimens present in the south- eastern herbaria whose collections I studied. It is probably overlooked by collectors because of its inconspicuous habit and flowers. In view of its questionable taxonomic status, it is hoped botanists will be aware of the need for more extensive flowering and fruiting collections, from both sub- mersed and emersed populations. 6. LYTHRUM L. Perennial, rarely annual, herbs or small shrubs, stems often prominently 4-angled, 1-12 dm tall. Leaves glabrous or rarely pubescent, opposite, alternate, or whorled, ovate to linear, sessile or shortly petiolate, reduced in the inflorescence. Flowers regular or nearly so, homomorphic or hetero- morphic with styles and stamens of 2 or 3 lengths, 6-merous, axillary, 1 or 2 at a node or in terminal spikes; pedicels bibracteolate. Floral tube cylin- drical, greenish, 8-12-nerved, 4-8 mm long; calyx lobes alternating with appendages; petals 6, rose-purple, purple, pink, or white, deciduous; sta- mens 6 or 12; ovary 2-locular, with or without a thickened ring at the base; capsule septicidal or septifragal. Seeds many, small, ovoid. 1. Flowers solitary or paired in the axils; stamens usually 6. 9. Leaves mostly opposite along the entire stem, mostly shorter than the internode above 3. Leaves oblong, base obtuse; stems short, 1-4 dm tall, several from a creeping rhizome .. . .. . . LL. flagelilare 3. Leaves linear to narrowly lanceolate, base attenuate; stems tall, 3-15 dm, slender, wand-like . . . ne 2. L. lineare 2. Leaves mostly alternate, eancratiy only the lowermost opposite, the middle and upper ones alternate, mostly longer than the internode above 4. Branch leaves distinctly, abruptly reduced in size from stem leaves, openly spaced on branches; appendages and calyx lobes equal in length 3 curtissti 4. Branch leaves gradually reduced in size from stem leaves, crowded and overlapping in inflorescence; appendages 2x the length of the calyx lobes 4, L. alatum 1. Flowers numerous in showy terminal spikes; stamens usually 12. 5. L. salicaria 88 In the most recent treatment of Lythrum in the United States (Shinners, 1953) at least seven, possibly eight, species are reported to occur in south- eastern United States. My studies of the genus, including examination of type material not seen by Shinners, suggest that there are five species of Lythrum, one comprising two varieties, in the area covered by the proposed Vascular Flora. 1. L. FLAGELLARE Shuttl. ex Chapm.—Slender, herbaceous perennial with several erect to decumbent stems arising from a woody creeping rhizome, 1-4 dm tall; stems nearly terete, sparsely branched. Leaves strictly opposite, sessile or subsessile, oblong to obovate, 5-13 mm long, 2-6 mm _ wide, scarcely or not at all reduced in the inflorescence, the apex and base round- ed. Flowers solitary, dimorphic; pedicels 0-1.5 mm long. Floral tube 4-5 mm long, gradually widening toward the distal end; appendages about 2 the length of the calyx lobes; petals pale purple to purple, 3-4 mm long; hypogynous ring 0.5 mm tall and 0.5 mm wide. Spring. Wet places; cp. Fla. Infrequently collected endemic, known from Collier, Hendry, Lee, Manatee, Okeechobee, Orange, and Sarasota counties. This rare Florida endemic is distinguished by short stems from a creeping rhizome and small, oblong, strictly opposite leaves which are scarcely re- duced in the inflorescence, The species is very closely allied to the wide- spread Mexican L. acinifolium (DC) Koehne, differing in growth habit and size of stems and lower leaves, Both species are perennial from a creeping rhizome. Stems of L. acinifolium tend to be woody and erect with at least some stems up to 7 dm tall. Stems of L. flagellare are less woody, up to 4 dm tall and often decumbent. Lowermost leaves of L. acinifolium tend to be somewhat larger than in lL. flagellare. The species cannot be separated by floral morphology. A definitive statement of. their relationship awaits an understanding of the range of variability in the Mexican species. 2. L, LINEARE L.—Perennial with basal offshoots, the stems slender, wand- like, the upper portion of the plant much branched, 3-15 dm tall. Leaves of stems and inflorescence mostly opposite, narrowly linear to lanceolate, sessile, 5-40 mm long, 1-4 mm wide, mostly shorter than the internode above, slightly fleshy. Flowers paired in the axils. oriented nearly parallel to the stem, dimorphic; pedicels 0.5-1.5 mm long. Floral tube 3-5 mm long; calyx lobes and appendages subequal in length; petals pale purple or white, 3-4 mm long; basal stipe of ovary scarcely or not at all thickened into hypogynous ring, abaxial and adaxial sides of the stipe 0.5 mm tall. (n=10) Summer, Brackish marshes; ep. SE. The species is tall, denden with narrow, linear, opposite leaves which are mostly shorter than the internode above. The petals are pale purple to white and often shorter than the floral tube. The ring at the base of the ovary is not at all thickened, as it is in all other native species of Lythrum in the Southeast. Lythrum lineare has a more open, less coarse aspect than does L. alatum var, lanceolatum (Ell.) Rothrock, a species with which it 89 may grow and is occasionally confused. The species is very strictly distri- buted in brackish waters of the coastal plain from Delaware in the north (possibly north to Long Island, fide Fernald, 1950, though I have seen no specimens from there), south to southernmost Florida and west into Louisi- ana. It is not known from Texas, as reported by Fernald (1950), collections so labeled representing either L. alatuwm var. lanceolatum or L. californi- cum S. Wats. A possible hybrid with L. curtissii (see the following species) has been collected in Georgia. 3. L. CURTISSII Fernald—Slender, much branched perennial, 5-10 dm tall, stems 4-angled; branches numerous, short, ascending, flexuous, highly flori- ferous. Leaves thin, yellow-green, oblong to elliptic, the apex obtuse or acute, the base tapering, the lower leaves mostly opposite, the middle and upper ones mostly alternate, sessile or subsessile; stem leaves 20-75 mm long, 5-15 mm wide; branch leaves distinctly and abruptly much reduced, 3-15 mm long. Flowers mostly solitary, often oriented at a 45-90° angle with the stem, dimorphic; pedicels slender, 0.5-1.5 mm long; bracteoles present at the top of the pedicel. Floral tube 3-6 mm long, the distal end and nerves often purple-tinged; calyx lobes and appendages + equal in length; petals deep to pale purple with darker central vein, 2 mm long; hypogynous ring 0.5 mm tall. (n=10) Summer. Wet places, rare; cp. Fla, Ga. Endemic known from Calhoun and Miller Co., Ga. and Gadsen, Liberty, and Franklin counties, Fla. A species with slender, flexuous, pyramidal ascending branches and many small delicate flowers, it is a rare endemic of southwestern Georgia and adjacent counties in Florida. Besides its unique habit, it is distinguished by having main stem leaves distinctly larger than the branch leaves and pur- ple petals generally only 2 mm long. A specimen suggestive of both L. cur- tissiti and L. lineare has been collected at or near the type locality of L. curtissii (Kral 28630, VDB). It is a plant with linear leaves, alternate on the upper stem and gradually decreasing in size toward the summit, lead- ing it to be keyed with difficulty to L. lineare. The flowers tend to be held more erect than is typical of L. curtissii and are at the large end of the size range for that species with petals 4-5 mm long as opposed to 2 mm in typical L. curtissii. However, the petals are of the deep purple hue char- acteristic of L. curtissii rather than pale as in L. lineare, the appendages are longer than found in either species, and a hypogynous ring, normal in L. curtissii but lacking in L. lineare, is present. Lythrum lineare is not known to occur as far inland as Calhoun Co., Ga. Both species are known from Franklin Co., Fla., but no evidence of hybridization has been noted in collections from that county. 4. L. ALATUM Pursh—Perennial herb with basal offshoots, the stems slen- der to robust, up to 1 m tall, virgate, the upper part of the plant much branched. Leaves dark green above, green to gray-green beneath, sessile, ovate to oblong with rounded to subcordate base or lanceolate with tapering base, the lower stem leaves opposite to mostly subopposite, 10-76 mm long, 90 2-14 mm wide, the middle and upper leaves subopposite to mostly alternate, reduced in size, crowded and overlapping on the branches. Flowers solitary, axillary, held nearly erect, bibracteolate on pedicels 1-3 mm long, dimor- phic, with either style or stamens exserted. Floral tube 3-7 mm long, green; appendages ca. 2 the length of the calyx lobes: petals purple, 2—6.5 mm long; thickened hypogynous ring present at the base of the ovary, narrowed on the ventral (abaxial) side. (n=10)? Summer. Wet soils, ditches, thickets: all prov. Sle /All/ Common throughout eastern and central US. The more robust southern plants with lanceolate leaves and tapering leaf bases are referred to var. lanceolatum (Ill.) T. & G. ex Rothrock. L. lanceolatum ill.—I, R, S. Most widespread of the United States species of Lythrum, L. alaiwm has been divided by various authors into as many as four species. Shinners (1953) recognized as present in the Southeast L. alatum. L. lanceolatum, L. dacotanum Nieuwl., and provisionally L. cordifoliwm Nieuwl. He considered L. alatum and possibly L. cordifolium as rare southeastern endemics and felt that the epithet alatum had been incorrectly applied to the common midwestern plants that he referred to L. dacotanum. Floras treating south- eastern United States or adjacent areas have recognized L. alatum and L. lanceolatum as distinct species (Fernald, 1950; Small, 1933: Radford, Ahles, and Bell, 1964), as varieties of a single species (Torrey & Gray, 1840; Glea- son & Cronquist, 1963: Long and Lakela. 1971), or have followed Shinners’ taxonomic and nomenclatural interpretations (Blackwell, 1970: Correll and Correll, 1972). The taxa have been distinguished primarily on the basis of leaf shape, secondarily by quantitative (though overlapping) differenees in leaf size and floral tube and petal length, and by position of the bracteoles on the pedicel. Lythrum is considered a notoriously difficult genus due to the vegetative variability within a species and the superficial similarities among the spe- cies. In addition, introgression appears to occur in areas of sympatry, fur- ther blurring species distinctions. Analysis of the variability in this wide- spread complex centering around L. alafum was conducted as a basis for defining the number of species or infraspecifie taxa present in the Southeast. Approximately 200 specimens were selected from collections spanning the entire range of all species involved and were scored for leaf shape (by tracings of the longest leaf), and for leaf length and width, floral tube and petal length, and bracteole position. Type material of L. alatum, L. daco- fanum, L. cordifolium, and L. lanceolatum (using a photograph of the type) was included in the study To more easily observe any clinal changes that might be present, I divided the specimens into ten geographical areas. On the basis of the most con- spicuous and easily determined character, leaf base shape, the areal bound- 7A uid count of #w==5 (Smith, E, B., 1963 in Madrofio 17:117) has been cor- rected to #==10 based on material from the same locality (R. Ornduff, pers. comm., 1974) 91 aries were drawn to include all plants with round to subcordate-based leaves (L. alatum of most authors) in areas 1-5 and plants with tapering leaf bases (L. lanceolatum of several authors) in areas 6-10. These two major cate- gories were also compared with each other (Fig. 1; Tables 1 and 2). Although L. alatum is dimorphic in stamen and style length, no other gross morphological differences were correlated with this condition. Long-styled and short-styled plants both showed the same variability in the features measured and so were not considered separately in the survey. Fig. 1. Variability in leaf shape throughout the range of Lythrum alatum. Outline tracing of actual longest leaf on each plant examined with groups 1-10 selected to represent total diversity in shape for each area. Leaf tracings from type material of L. alatum, L. cordifolium, L. lanceolatum, and L. dacotanum are indicated by the letters A, 1 L, and D, respectively. Geographical limits of areas 1-10 are listed in Table 1 92 Table 1. Variability in selected characters of Lythrum alatum, arranged according to the following geographic areas. 1. Maine-W. Va, 2. Ohio, Mich, Ind, Hl, Wis, 3. Minn, N & S Dak, N. Okla, 4. Ia, Mo, N. Ark, 5. Ky, Tenn, N. Ala, N. Ga, 6. Va-S. Ga, 7. Fla, 8. S. Ala-Miss, 9. La-S. Ark, 10. Tex- S. Okla. n=number Character Leaf length/width Totals Totals Longest floral tube length, mm Totals Totals Longest petal length, mm Totals Totals of plants scored, Area n 1 3 , 6 3 10 4 6 5 7 1-5 32 6 8 7 7 8 6 9 8 10 5 6-10 34 1 6 2 19 3 19 4 12 7 5 17 ; 1-5 73 6 20 7 22 8 21 9 20 7 10 7 6-10 88 1 4 2 19 3 19 4 12 D 14 1-5 68 6 18 7 19 8 18 9 15 10 5 6-10 75 min-max mean 3.7 — 4.7 4.4 3.6 — 7.0 5.2 2.1-—7.3 3.9 2.8 — 6.7 4.2 2.6 — 5.3 7 4.0 2.1 = 73 4.3 2.5 -—8.4 5.9 4.5-—8.5 5.7 42-88 6.3 2.8 — 12.5 6.5 3.6 — 9.0 5.3 7 2.5 — 12.5 §.0 5.0 — 6.0 5.3 4.0 — 6.0 5.4 5.0 — 7.0 5.5 4.0 —6.0 5.0 4.5 — 6.0 5.1 4.0 — 7.0 5.3 4.0 — 6.0 4.7 4.0 — 6.0 4.8 3.0 — 7.0 5.2 4.0 — 6.0 4.7 4.0 — 6.0 _— 5.0 3.0 — 7.0 5.0 4.0 — 6.0 5.3 4.0-6.5 4.8 3.0 — 6.0 4.8 3.0 — 6.0 4.6 3.0 — 5.0 7 3.9 3.0 — 6.5 4.6 2.0 — 5.0 3.9 2.0 — 5.0 3.9 2.0 — 5.5 4] 3.0 —5.5 4.2 4.0 — 5.0 4.4 2.0 —5.5 4.0 standard deviation ee ae me wD mW Pena Do Bw rH eb OME CR CIT ICI! co ecccoe ai —_— oO Ss CO bo oo 93 Table 2. Bracteole position on pedicels of Lythrum alatum, arranged ac- cording to geographical areas listed in Table 1 No. of Plants with Bracteole Position at Area ase Mid Upper 1 7 3 4 2 15 0 3 3 9 2 8 4 9 0 3 5 7 3 4 47 8 _ 22 = is % of total 61% 10% 29% 6 4 2 17 7 3 1 17 8 7 3 28 9 6 9 28 10 1 1 . 5) a 21 16 7 95 = 132 7 % of total 6% 12% 72% The most obvious differences among the specimens lie in the leaves. After the leaf shape of the longest leaf of each specimen was traced the leaf tracings (Fig. 1) were selected to provide a picture of the leaf vari- ability within each area and a tracing representative of the extremes and middle of the geographic range for each area. Shinners’ distinctions, based on leaf length/width ratios between L. lanceolatum and L. dacotanum (4-15 as long as wide) and his L. alatum (1%-3% « as long as wide) are not maintained, these ratios also being found throughout most of the range of L. alatum (Fig. 1). Ovate to oblong leaves with rounded bases are typical of northern and midwestern plants while plants of the south have lanceolate leaves with tapering bases. No clinal differences can be recognized in this character since round-based ovate leaves broadest at the base become lanceolate in shape when there is a change to tapering base, the broadest width being transferred upward toward the middle of the leaf. There is con- siderable overlap in leaf length/width ratios of the two major areas, north and south (Table 1). The lanceolate leaves tend to be longer and narrower than ovate leaves but more variable, and there is no gradual shift to long- er, narrower leaves from north to south. For floral tube and petal length the maximum size on each specimen was recorded (Table 1). Floral tubes and petals of the northern and central plants tend to be slightly larger than those in the south, but the differences are scarcely significant taxonomically. Again, there is no clinal change in size of either character, Lythrum alatum is said to have bracteoles at the base of the pedicel and 94 L. lanceolatum to have bracteoles on the upper half of the pedicel. In the northern areas (1-5) 61% of the plants examined had bracteoles positioned at the base of the pedicel (Table 2). In southern areas (6-10) the position is reversed with 72% of the plants scored having bracteoles on the upper part of the pedicel just below the floral tube. Although there is a significant difference in this character between plants of the north and south, its prac- tical value in identification of individual specimens is limited by its vari- ability. Again, there is no obvious shift in bracteole position in a clinal man- ner from north to south. For the five characters examined only leaf shape provides a definitive taxonomic character. The other four characters are too similar to be utilized in identification and are certainly too similar to provide a basis for recog- nition of more than one species. yreenhouse-grown plants provided two additional taxonomic features not always evident on herbarium specimens. Plants representing populations of L. alatum from six states grown under uniform conditions showed marked differences in stature and production of winter shoots. The southern plants i.c., those from Mississippi, Arkansas, and Texas, were taller with sturdier, more substantial stems than those from Kansas, Michigan, and Ohio. Her- barium specimens of the southern plants most often have a single main stem while specimens from the northern and central states most often have several slender stems closely arising from an enlarged rootstock. Stature correlates with shape of the leaf base. The robust, tall plants all have taper- ing leaf bases, and the shorter, slender-stemmed plants have rounded to cordate-based leaves. A second feature, the production of prostrate, dark green, somewhat fleshy winter shoots, does not appear to be correlated with leaf base type. The shoots develop from the rootstock or from buds on the lower portions of older stems as the aerial stems die or become dormant. In the greenhouse, winter shoots are maintained into the spring when buds on the erect dor- mant stems begin growth. The shoots apparently do not become erect or persist into the summer but may give rise to new erect stems by rooting nodally and internodally where they contact the soil. Greenhouse grown plants from Kansas (with cordate-based leaves), Texas, Arkansas, and Mis- sissippi (all with lanceolate leaves) display winter shoots; those from Michi- gan and Ohio do not. Production of winter shoots may be a genetically-based response to moderate climatic conditions. The evidence presented in this study supports recognition of only one spe- cies in the L. alatum complex in the Southeast, with two varieties defined primarily by stature, leaf shape, and geographical range. The following nomenclatural notes, partial synonymies, and distribution map for L. alatum are based on specimens from the herbaria listed in the acknowledgments. They are not meant to be complete with respect to the species as it occurs beyond the range of the southeastern flora, as in the Caribbean region. The 95 few synonyms not included represent later names and hence their absence is of no nomenclatural consequence with respect to this primarily floristic study. Lythrum alatum Pursh, Fl. Am. Sept. 1: 334. 1814. Type: “In Lower Georgia. Enslen. June. July. v.s. in Herb. Enslen,’’ (W!; possible isotype, W!) Fig. 2, a & b. The holotype of L. alatum Pursh at Vienna bears the epithet in the author’s own hand, as attested to on the holotype by Prof. Joseph Ewan, who has made a thorough study of the work of Frederick Pursh (Ewan, 1952). The holotype and possible isotype are typical of the northern and midwestern plants of this species with ovate, round, or cor- date-based leaves. The leaves of the holotype, described by Pursh as sub- cordate and opposite, are primarily alternate, though the lowermost leaves, which are often opposite in Lythrum, are lacking on the type. In size and shape the leaves fall within the variability displayed by the complex (see Fig. 1, area 5). The flower length of 5 mm also lies within the range of the complex. On the basis of morphology there is no doubt that the epithet alatum was correctly applied by Torrey & Gray and later authors and should not be reserved for an unknown rare endemic of coastal or southern Georgia as has been suggested (Shinners, 1953). The exact geographical location of the type locality, recorded by Pursh as “Lower Georgia,’ is unknown. According to Ewan (1955) ‘‘we can learn little from the specimens as to Enslen’s route in the South for there are no localities on the form labels.’ Any part of the state that could be con- sidered ‘“‘Lower Georgia,” interpreted by Shinners as coastal or southern Georgia, lies outside the present range of the cordate-leaved L. alatum. That the plant could actually have come from the northwestern corner of Georgia, within its present-day boundaries, is very remote. Early maps of Georgia (Lewis, 1805; Low, 1810) show only two major trails through only a part of the area, that part of Georgia being firmly in control of the Chero- kees until the 1830s (L. DeVorsey, Univ. Ga., Dept. of History, pers. comm., 1974). Ewan (pers. comm., 1974) suggests the possibility that the specimen was collected by John Lyon in the vicinity of Augusta, Georgia, in June 1803 and subsequently shared with Aloysius Enslen. Pursh apparently saw Enslen’s collections in Philadelphia (Fl. Am. Sept., p. 12). The Enslen her- barium subsequently was acquired by the Vienna Botanical Muscum in 1822. Two other collections of cordate-leaved L. alatum are known from outside the apparent present range. One is an 1898 collection of A. S. Hitchcock from Citrus Co., Fla (MO, Fig. 3). The other is the type collection of L. cordifolium Nieuwl., $.B. Buckley s.n., no date, ‘‘N.C. to Ga. & Fla.” (holo- type, US!) or more specifically ‘‘Florida’’ (isotypes, NY, GH!) Both the Hitchcock and Buckley collections have unusually small leaves for L. ala- tum, but the flowers are identical to those of L. alatum. The basal parts and lower leaves are lacking on all the specimens. Difference in leaf size alone is not a sufficient basis for considering L. cordifolium as a species apart from L. alatum in view of the variability of this character in the species 96 flowering branch of the type. she ee ni 6% Tyopus é a ey oa 98 and in the genus as a whole. Lyihrum cordifolium is regarded here as synonymous with the typical variety of L. alatum. The disjunct occurrence of these cordate-leaved specimens in Georgia and Florida would seem to indicate either that the range of this variety was once more extensive than it is at the present time, or that plants of this type occasionally are intro- duced into areas beyond the principal range. That the Hitchcock and Buck- ley specimens occur at the extremity of the species range may account for their unusually small leaves. It is of interest to note here that the range of a species very closely related to L. alatum, L. californicum S$. Wats., has been severely restricted in northern California in recent years by man’s activities and is now rare except for a few areas in southern California (Ornduff, pers. comm., 1974). Possibly L. alatum var. alatum has been elim- inated from the southeastern coastal plain for the same reasons. Lythrum dacotanum is also relegated to synonymy, the type represented by a robust, cordate-leaved specimen scarecly differing from the type of L. alatum except in its larger leaf size. 4a. LYTHRUM ALATUM Pursh var. ALATUM. L. cordifolium Nieuwl., Am. Midl. Nat. 3: 265-266. 1914. Type: N.C. to Ga. & Fla., S. B. Buckley s.n. (US no. 48371!; isotypes, GH!, NY), non L. cordifolium Sessé & Moc., 1888 lL. dacotanum Nieuwl., Am. Midl. Nat. 3: 266-267. 1914. Type: Sioux Falls, S.Dak., Aug 1892, J. J. Thornber s.n. (US no. 240979!; isotype, US no. 516395!) The nominal variety is distinguished from var. lanceolatum by its generally slender stems up to 8 dm tall and ovate to oblong leaves with subcordate to rounded leaf bases. (n—10) All prov. Ark, Ala, Ga, Miss, Tenn, Ky, WVa, Va." /All, except Tex/ Lythrum alatum var. alaftum is incorrectly recorded as occurring in Louisi- ana and Texas (Fernald, 1950) and north-central Texas and the Texas pan- handle (Correll and Johnston, 1970). These plants are, for the most part, L. californicum, which also has rounded leaf bases but linear-lanccolate, slightly fleshy, glaucous leaves, or L. alatum var. lanceolatum, or possibly hybrids or introgressants of the two. Probable hybrids of L. alatum var. alatum and L. californicum have been collected in northern Oklahoma and southwestern Kansas. Lythrum californicum and L. alatwm are obviously closely related as indicated by the similar morphology and apparent ability to interbreed. A study of L. californicum in western United States and Mex- ico is needed to determine if its specific rank is justified or if it should be included within the L. alatwm complex. _~ 4b. LYTHRUM ALATUM var. LANCEOLATUM (EIl.) T. & G. ex Rothrock in G. M. Wheeler, Rept, U.S. Geographical Surveys West of the 100th Me- ridian, VI-Botany: 120. 1879. L. lanceolatum Ell., Sketch. 1: 544-545. 1821. “An additional southern locality for L. alafwm var. alafuim has come to my attention, but T have not seen the specimen. S.B. Jones in Lythraceae for the Mississippi Bione Project area cites a single specimen from Desoto Co., which is the yunty in the state and lies along the southern line of distribution of the variety. Type: { e ° e 1 e e Ce e ' ° %e +. ee oe o * .s—e-4 ee oF eg 5 00 8 © ° ee e e The term “experimental taxonomy” was originally applied to studies that investigated the nature of species by means of reciprocal transplants (Clements and Hall, 1920); such studies were experimental in the strict sense. In the decades after its introduction, however, the term acquired a broader meaning (e.g., Clausen, Keck, and Hiesey, 1941) and thereafter became confused with “biosystematics’” (Camp and Gilly, 1943). 106 upon revisions by focusing on the clarification of relationships that are still not well understood. Difficult taxonomic situations that may lead to experi- mental studies often result from the evolutionary dynamics of hybridization and/or introgression, inbreeding, and apomixis. CONTRIBUTIONS OF REVISIONARY STUDIES TO PLANT SYSTEMATICS I believe revisions to be useful in plant systematics primarily in four ways: (1) as a source of classifications, (2) as an aid to identification, (3) as a source of biological data, and (4) as a stimulus for further study. Each of these uses will be discussed in turn. The most significant contribution of revisionary studies is as a source of hierarchies of classification. Classifications are fundamental to the growth and development of plant systematics by playing at least three important roles (in part from Warburton, 1967). First, from the classification one can infer ancestral evolutionary relationships (— phylogenies) among all the included taxa, which gives some idea of the patterns of evolution through long periods of time. Second, the classification allows the biological data presented in the revision (see discussion below) to be retrieved more easily because they are ordered in an hierarchical arrangement corresponding to the classification itself. Third, the classification allows for the prediction of unknown attributes of taxa included in the revision. For example, if cer- tain pharmacologically-active compounds are found in one species, it can be predicted that the most closely related species in the classification might have the same or similar compounds. Another important use of revisions is for identification purposes. Many types of people, such as wildlife specialists, herbarium curators, etc., desire to know the names of particular plants. If no modern flora exists for an area, or if the group in question is not adequately treated in an available flora, then workers will turn to the keys, distribution maps, and descriptions in the revisionary study for help. The information contained in a revision—in descriptions, distribution maps, and statements on phenology and ecology—represents biological data on what the plants are like, how they differ from each other, where they grow, and when they flower. Many types of systematists—from horticultur- ists to phytogeographers—as well as other biologists, seek this information. Phytogeographers are so dependent upon up-to-date revisionary studies that very few meaningful phytogeographie interpretations can be made for gen- era that are not well understood from a revisionary perspective (Axelrod and Raven, 1972; Thorne, 1972). Many additional data such as nomenclature, discussions of generic relationships, and taxonomic history are usually in- cluded in revisions, but this information is generally of lesser interest ex- cept to floristic and other revisionary workers. The revision also serves as a stimulus for further study. Because basic relationships are clearly outlined, other workers such as cytologists or anatomists may want to pursue an analysis of the relationships further. 107 The revision may prove to be very stimulating to an anatomist, for example, even though no specific anatomical problems are mentioned. The interest may develop simply from his having read the descriptions of the different species or having seen their distributional patterns. On the other hand, specific problems for further study may be emphasized deliberately in the paper. For example, a discussion of difficulties of classification brought on by suspected hybridization and introgression might well catch the interest of a cytogeneticist. REQUIREMENTS FOR GOOD REVISIONARY WORK Although admitting the value of revisionary studies, some botanists might not realize that significant differences exist in the quality of revisionary work produced. Considerable variations among finished revisions do prevail, however, even to the extent that some studies are so inadequate that they prove nearly worthless for those purposes I mentioned earlier. Asa Gray, almost a century ago (1875), put it this way (p. 353): ‘‘Easy as the work may seem, the number of botanists who are able to elaborate a genus and draw up fairly good botanical descriptions is wonderfully small. The thing is quite possible if mere literary compilation is intended; but something more than this is needed.”’ The general requirements for excellence in revisionary work apply basic- ally to most descriptive sciences and to a lesser degree even to experimental sciences. Many criteria might be formulated for evaluating the ability of a revisionary worker (and, therefore, also the quality of his published revi- sions), but I have selected six of what to me seem the most important: (1) precision and thoroughness in gathering of comparative data of all types; (2) ability to recognize discontinuities in sets of comparative data (= pattern recognition); (3) ability to relate observed discontinuities in sets of data to the various fixed ranks of the taxonomic hierarchy; (4) pre- cision and thoroughness of description of recognized taxa; (5) precision and thoroughness of documentation in literature, specimen citations, and nomen- clature; and (6) precision, thoroughness, and clarity of expression in the final written treatment. These six general criteria must be kept in mind when pursuing each of the three major aspects of revisionary studies: field, library, and herbarium investigations. But in addition to these general cri- teria, specific requirements for excellence in each of these activities also must be considered. Although the completion of ficld investigations is not absolutely essential for good revisionary studies, there is little doubt that all other factors being equal, the more field work a researcher is able to do, the better will be his resultant treatment. The value of field observations for the revisionary worker lies mainly in improving his ability to recognize discontinuities among local populations and to relate the resultant groups to categories in the taxonomic hierarchy. Refined classifications can be produced solely from herbarium material, as has been done by many workers in the past 108 (e.g., S. F. Blake, B. L. Robinson), but such studies tend to be myopic in their treatment of relationships. Library investigations are very important primarily because revisionary studies have a strong historical orientation. Not only must past aspects of nomenclature be understood fully, but also the Hterature must be searched extensively for taxonomic opinions or dispositions that may bear on the present attempts at classification. To be skilled in library work demands traits characteristic of professional botanical bibliographers: knowledge of languages, familiarity with literature on itineraries and biographies of col- lectors, understanding of abbreviations of older taxonomic literature, fa- miliarity with the International Code of Botanical Nomenclature, and—most importantly—patience. Herbarium investigations likewise are a very important part of the re- visionary effort. These center on the core of the revision proper, which is the development of coneepts of taxa and their subsequent ranking. It is at this point that the natural ability of the taxonomist must show through. He must have the talent to recognize discontinuities in data sets, or to put this same idea another way, he must have a ‘‘taxonomist’s eye.” The recogni- tion of discontinuities sometimes can be very challenging, especially if the group in question shows unusual patterns of variation. In these cases a strong background in evolutionary theory is most helpful, if not absolutely necessary, to correctly identify apomixis, hybridization, or ecotypic differ- entiation and to make the proper taxonomic dispositions. Furthermore, to be good in pattern recognition necessitates knowing something about the kinds of data that have been collected and whether or not they have been collected correctly and in sufficient quantity. A good understanding of the morphology of the group in question is therefore essential, as is the full understanding of other types of data, such as anatomy, cytology, or chem- istry, that may be available for use. INNOVATIONS IN METHODS OF REVISIONARY STUDY As with any aspect of plant systematics, innovations in approaches to re- visionary studies must occur periodically or the studies themselves. will never approximate their full utility. The basic formats and methodologies for revisionary work have been around since the time of Linnaeus.’ From that beginning the approaches have remained largely unchanged except for the use of additional types of data on which to base the classifications. Be- cause over 2 centuries have passed since the appearance of the early re- visionary literature, it is appropriate at this time to consider if changes might be beneficial. Unlike most experimental systematic studies, the format for presentation of data in revisionary studies is very rigid. The reason for this rigidity de- = 7 Although taxonomic studies with “revision” in the title did not appear until the late 19th century, the earlier world floras of Linnaeus and DeCandolle are clearly revisionary in format and perspective. 109 rives from the objectives or uses of the revision, for which ease of data re- trieval is paramount. The descriptions, keys, lists of synonyms, citations of representative specimens, and distribution maps are effective ways to insure accessibility of information. Innovations in methods of revisionary study should therefore be concerned primarily with new ways to develop and process data necessary to fill the revisionary structure, rather than with developing a new structure altogether. Some space-saving changes in structure may be desirable, however, such as the substitution of data ma- trices for keys and descriptions (Solbrig, pers. comm.) Teamwork as a means of speeding up completion of revisionary studies has been emphasized by Leenhouts (1968) and by Blair and Turner (1972), and I agree with their suggestions. In the past, some joint ventures have been very successful (e.g., B. L. Robinson and J. M. Greenman; M. Mathias and L. Constance), and thee should serve to encourage others to follow suit in the future. It is realized that productive joint ventures rest with com- patible personalities that cannot be forced, but perhaps more overtures can be made toward colleagues with whom collaboration might succeed. The gathering together of library materials is now a much easier task than ever before. This does not mean that interpretation of data is any easier but simply that the relevant publications can be located more quickly through numerous indexes and obtained much faster through copy machines or on microform. Reader-printer machines now available allow for hard- copy to be obtained cheaply and rapidly from microform publications. The gathering together of herbarium specimens also is much easier than ever before. With the publication of the recent edition of the Index Her- bariorum (Holmgren and Keuken, 1974) the existence of collections from which loans may be made is now known to all workers, even those in isolat- ed geographic regions. Furthermore, the availability of Inter Documentation Company (Zug, Switzerland) microfiche editions of herbaria from which loans are not permitted (e.g., Linnacan and DeCandolle herbaria) has made examination of these collections almost routine. my opinion, one of the most significant innovations in methods of re- visionary study will come with the use of machine-assisted operations. Because many of the procedures used in revisionary work deal with data manipulation, data-processing machines are ideally suited for aiding this work. The following paragraphs explain briefly how these machines might be useful in some of the individual procedures The preparation of lists of representative specimens during the course of a revisionary study is a laborious, time-consuming, and error-prone task. To speed up this aspect of the work, in our laboratory we have developed a computer program (Meacham and Stuessy, 1974) that allows for specimens to be cited automatically once the data are on machine-readable cards. The details of this program will be reported elsewhere, but I should mention that the procedures for coding of data are so simple that any taxonomist can use the program with almost no prior knowledge of computers. An ad- 110 ditional step that we have not done, but that has been done already by some workers (Soper, 1964; G6mez-Pompa and Nevling, 1973; Adams, 1974), is plotting specimen localities by computer on a base map. nother time-saving procedure is use of a flexowriter, which essentially is a typewriter that generates a punched paper (or magnetic) tape at the same time the first manuscript draft is produced. This machine is very useful for the typing of descriptions and other textual portions that are not likely to be highly modified after the initial draft. At the time of final manu- seript production, the paper tape is used to produce an edited typescript with no detailed proofreading being necessary. One more computer procedure that will help make revisionary work more efficient is the preparation of keys directly from descriptions. It remains to be seen whether computers can produce as good a key as a taxonomist can, but it is very clear that preliminary keys can be generated that might be further modified by the worker with a net saving of time. Although this is a complex manipulation for computers, and although we are still a long way from having truly serviceable algorithms, preliminary studies (Hall, 1970; Pankhurst, 1971, 1974; Morse, 1971; Pettigrew and Watson, 1973; Dall- witz, 1974) suggest that some of the difficulties will be overcome soon. This procedure need not involve additional laborious coding of descriptive data cither, because once the data from the descriptions are on paper or mag- netic tape, as with use of the flexowriter described above, they could be fed with proper conversion into the computer without additional manual coding. RELATIONSHIP OF REVISIONARY STUDIES TO THIE ECOLOGICAL CRISIS With the growing realization that the world’s biota is diminishing at an alarming rate primarily as a result of man’s modification of the environ- ment, a new term, the ‘‘ecological crisis,’’? has been coined to dramatize this loss of organic diversity. Implicit in this concept is that organisms higher up in the food webs, such as man himself, eventually will be harmed by the loss of diversity at lower trophic levels. Many scientists have recognized this problem (e.g., Iltis, 1967; Fosberg, 1972; Holdren and Ehrlich, 1974), and many books have been written (e.g., Ehrlich, 1968; Johnson, 1970; Matthews, Smith and Goldberg, 1971) that speak to the dilemma and its resolution, at least in part. Because the present decline in diversity of the world’s flora is likely to continue in years ahead, the plant systematist is faced with the respon- sibility of deciding which types of studies now appear to be most useful or productive, not only in a restricted sense for systematic botany, but also for mankind in general. The plant systematist is in the unique position of possessing considerable knowledge about the world’s flora, and therefore presumably also in a good position to evaluate the kinds of studies that are important to stress at this time. Because of personal biases, different plant 111 systematists might give different values to the various kinds of studies. We do, it is true, need more experimental studies to learn as much as pos- sible about the evolutionary process before the key intermediate taxa be- come extinct. Likewise, we also need an increased effort in floristic studies so that at least we will have some idea of what plants grow in particular regions before the flora is decimated. Although respecting these different viewpoints, I believe that the revision- ary study is the type of investigation most needed at this time. Reasons for this position lie principally with my view of the revision as being central to all facets of systematic botany. Although floristic treatments sketch re- lationships among the included taxa, the depth of understanding in this type of study is necessarily limited and is not best suited for helping us under- stand phylogenetic and phytogeographic relationships of the groups con- cerned, nor for understanding mechanisms of speciation that have prevailed. Moreover, the floristic study may not clarify fully the nomenclatural prob- Iems, so that the names for the taxa themselves may not even be correct.‘ Iixxperimental studies on small groups of species are also needed at this time, but unless we have an accurate understanding of the basic relation- ships of most plant species, it will be impossible to extrapolate from these few in-depth studies on the process of evolution to the probable patterns of evolution in other groups of taxa. Therefore, to help document the world’s flora before it vanishes, I agree with Turner (1971) that plant systematists should give emphasis to revisionary investigations and, in particular, to en- couraging students to pursue these endeavors. Because of the above assessment of the importance of revisionary studies in systematic botany, especially at this point in time, I believe that we should give attention to the support of field endeavors that bear on these efforts. Two types of collecting programs will give maximum return in this direction: (1) those completed by the revisionary worker himself in any re- gion of the globe and (2) those completed by the floristic worker in very obscure regions of the earth as yet untouched markedly by European man (e.g., Amazon basin of South America) or in areas in danger of immediate and irreversible destruction (e.g., Mexican highlands). The revisionary worker is in the best position to make critical collections and observations that will allow the most significant data to be obtained from his plant ma- terials in the future. The floristic worker in obscure or threatened regions will serve the valuable function of collecting at least some materials that can be worked up at a later date as part of a revisionary investigation. I believe that both types of field programs are urgently needed. CONCLUSION The previous discussions have attempted to show that revisionary studies *T would agree that detailed and critical floras (of which few exist) avoid most of these limitations, but I contend that such works are basically revisionary in character because of their depth of coverage. 112 play a very important role in plant systematics. I have tried to indicate that being a productive and excellent revisionary worker involves many chal- lenges, including acquiring knowledge in several diverse areas—from_ bo- tanical bibliography to evolutionary theory. Such challenges should be point- ed out to graduate students so that a larger number of them will become stimulated to pursue these endeavors. With the world’s flora disappearing at an increasing rate, we need as many people as possible studying basic plant relationships before the plants themselves are completely extirpated. The hope is that we can obtain enough detailed information about the flora to give us a foundation for future studies leading to at least a partial un- derstanding of phylogeny and evolutionary processes in the plant kingdom. ACKNOWLEDGMENTS Thanks are due W. L. Phillips, EH. D. Rudolph, O. T. Solbrig, and B. L. Turner for offering helpful criticisms during the development of the manu- script. Partial support for the preparation and publication costs of this paper came from NSF Grant Number GB-37678. REFERENCES ADAMS, R. P. 1974. Computer graphic plotting and mapping of data in’ systematics. , 258 33 : AXELROD, D. I. and P. H. RAVEN. 1972. Evolutionary Fah cei! viewed from plate tectonic theory, pp. 218-236. Im J. A. Behnke (ed.), Challenging biological problems: directions toward their solution. Oxford University Press, New York. BLAIR, W. PF. ins - = TURNER. 1972. The integrative approach to biological classifica- tion, pp. 193 J. A. Behnke (ed.), Challenging biologic. al problems: directions toward their ton Oxford University Press, New York. Cc ae Rd H. and C. L. GILLY. 1943. The structure and origin of species. Brittonia 4: cr rene - J.. D. D. KECK, and a Fig HHESEY. 1941, Experimental taxonomy. Carnegie Inst. Washington Year Book No. 160-170. CLEMENTS, F. FE. and H. M. 1 HALL. 1920. Experimental taxonomy. Carnegie Inst. Wash- ington Year Book No. 18: 334-335 DALLWITZ, M. J. 1974. A flexible eaanaees program for generating identification keys. Syst. Zool. 23: 50-57. EHRLICH, P. R. 126s The ele bomb. asa Books, New York. FOSBE oo F. R. 19 The valu stematic the environmental crisis. Taxon 21: 631- eon -POMPA, A. and L. I. NEVLING, JR. 1973. The use of electronic data processing iecods in the Flora of Veracruz program. Contr. Gray Herb. 203: 49-64. GRAY, A. fg Bentham, on the recent progress and present state of systematic botany. Amer. J. Se a. oS A, . on A computer- based: system for forming identification keys. Taxon 19: HOL mre N, J. DP. os . ER REIOE: 1974. Human population and the global environ- me HOLMGREN, DP. oe pe W. “KEUKE ae boas The herbaria of the world. Index Her- sariorum. Part 1. Ed. 6. Regnum Veg : 1-397 ae TH. H. 1967. To the taxonomist ie ecologist whose fight is the preservation of ature. Bioscience 17: 886-890. JOHNSON, C. E. 1970. Feo-crisis. John Wiley & Sons, New York. UST, T. 1954. Generic synopses and their role in modern seem research, Taxon 3: 201-202. 113 LEENHOUTS, P. W. 1968. A guide to the practice of herbarium taxonomy. Regnum Veg. 58: 1-60. ene W. H., F. E. SMITH, and E. D. GOLDBERG (eds). 1971. Man’s impact estrial and oceanic ecosystems. The MIT Press, MEACHAM. C. and T. F. ene 1974. Procedural ingnvatane 2» 3 ‘=| oe 5 a: oie) oO > a s in revisionary studies: ao ee citations of representative specimens. Amer. J. Bot. 61: 47. ae, E. 1971. Specimen identification and key construction with time-sharing com- puters, . - eee PANKHURST. R. J. 1971. Botanical keys generated by computer. Watsonia 8: 357-368. 1974, Automate identification in systematics. Taxon 23: 45-51. PE T TIGREW, C. J. and L. WATSON, 1973. On the identification of sterile Acacias and Bae feasibility of pon rene an automatic key-generating system. Austral. J. Bot. 21: ROBINSON, B. Hs . The need of monographic activity in American botanical taxonomy. a SOPER, J. H. ce Mapping the distribution of plants by machine. Canad. J. Bot. 42: 1100 THORNE, R. Fe 1972. a disjunctions in the geographic ranges of seed plants. Quart. -4] URNER, B. L. 1971. Training of systematists for the seventies. Taxon 20: 3- WARBURTON, F. E. 1967. The purposes of classifications. Syst. Zool. 16 1- {2 24 NOTES ADDITIONS AND CORRECTIONS TO THE FLORA OF TEXAS.—Hibiscus esculentus L. (Malvaceae). Milam Co.: river bottom, 1 mile IX of Cameron. 30 Oct 1933, Simon E. Wolff 4890 (SMU). This specimen is an escape of a particularly coarse strain of the widely cultivated Southern vegetable, okra. Hydrilla verticillata (L. f.) Caspary (Hydrocharitaceae). Hays Co.: back- waters of San Marcos River, Southwest Texas State University campus, San Marcos. 27 Apr 1975. Jerry M. Flook 104 (SMU). This nuisance introduction has become so abundant in the San Marcos River that it is now being har- vested commercially and marketed statewide as an aquarium plant. Lythrum salicaria L. (Lythraceaec). Hardin Co.: almost in water of stream along roadside leading to Sour Lake, W of Hwy. 69. 25 June 1971. Peggy A. Amerson & Geraldine FE. Watson 540 (SMU). (See Sida 6(2): 100.) Rudbeckia subtomentosa Pursh (Compositae) has been included in the Texas flora in Britton and Brown’s Illustrated Flora (1897, 1913; Gleason 1952, New York Botanical Garden) and in Cory’s Catalogue of the Flora of the State of Texas (Tex. Agric. Exp. Sta. Bull. No. 550, 1937). It has, however, been omitted in the more recent major treatments of the Texas flora, L.e., Gould’s Texas Plants—A Checklist and Ecological Summary (Tex. Agric. Exp. Sta. MP-585, 1962; MP-585-Revised, 1989) and Correll and Johnston's Manual of the Vascular Plants of Texas (Tex. Res. Found., Renner, Tex., 1970). On the basis of the following collection R. subtomentosa should be reinstated in the Texas flora: Hardin Co., ditch bank along Ariola Rd. across from 100-1 Gun Club; pine savannah wetlands. 31 July 1974. Geraldine Watson 1837 (SMU). In his list of additions and corrections to the Manual of the Vascular Plants of Texas D. S. Correll (Amer. Mdl. Naturalist 88: 494. 1972) reports Sabatia brachiata L. (Gentianaceae) as new to the Texas flora. Inclusion of that species was based on a collection from Hardin Co., Peggy A. Amerson & Geraldine FE. Watson 161. However, examination of this specimen in the S.M.U. Herbarium has revealed it to be Sabatia campanulata (L.) Torr., itself a new report for Texas. A second specimen from Hardin Co., Amerson & Watson 637, is also in the S.M.U. Herbarium. Sabatia campanulata is a plant of damp acid savannahs and boggy habitats and ranges through the southeastern states west to Arkansas and southeastern Texas. Vicia hirsuta (L.) S. F. Gray (Leguminosae). Shelby Co.: near Center. Simon E. Wolff, s.n., 11 May 1944 (SMU). This native of Europe is now found naturalized over most of the castern half of the continental U.sS.—Jerry M. Flook, Herbarium, Southern Methodist University, Dallas, Texas 75275, SCHIZACHYRIUM STOLONIFERUM NASH VAR. WOLFEI DeSELM, VAR. NOV. (GRAMINEAE).'—Named for John Nicholas Wolfe, American SIDA 6(2): 114. 1975. 115 botanist 1910-1974. —Haec planta ut in specie descripta praeter spiculam sessilem breviorem (5.1 mm.), aristam etiam saepe deficientem (ea pars brevis si praesens), atque spiculam pedicellarem (plerumque 1.8 mm.) et suam Eng (plerumque 0.8 mm.) breviores. Holotype, Florida, Osceola Co ur miles northwest of Loughman, Ray, Lakela, Patman 10494, 14 Rae 1960 (USF). The holotype appears to be of the same species as described by G. V. Nash, in Small, Flora of Southeastern United States, 19038. Exceptional characters are shorter (5.1 mm.) sessile spikelet length (aver- aging 6.3 mm. in the species,) awn absent to 7 mm. long, averaging 4 mm. (always present and averaging 6 mm. in the species), pedicellate spikelet length averaging 1.8 mm. (2.6 mm. in the species), its awn 0-1.5 mm., aver- aging 0.6 mm. (1.1 mm. in the species). DISTRIBUTION: sandhills, coastal strand and pine flatwoods, Central Flerida: Charlotte, Collier, Highlands, Manatee, Osceola, and Pinellas Counties. This small spikeleted, often non- awned variety of S. stoloniferum Nash was recognized during a study of Andropogon and related genera for the projected ‘““Manual of the Vascular Flora of the Southeastern United States.’’ Other collections of S. stoloniferum var. Wolfei in DU, UNC, UF, USF have been examined and annotated.—H. R. DeSelm, Botany Deen and Graduate Program in Ecology, Uni- versity of Tennessee, Knoxville, Tennessee 37916. HEMIGRAPHIS REPTANS (ACANTHACEAE), A GREENHOUSE WEED IN LOUISIANA.—The occurrence of Hemigraphis reptans T. Anders. as a greenhouse weed new to Florida was reported by Beckner in 1967 (Sida 3: 118-119). Shortly thereafter I noted the same species growing under the benches in the biology greenhouse at the University of Southwestern Louisi- ana at Lafayette. A visit to the other USL greenhouse (at the Horticulture Center) and to the commercial ‘‘Orchid Gardens’’ greenhouse in Lafayette revealed this rather drab Acanthacea growing as a weed there too. The Hemigraphis was still an innocuous weed in the biology greenhouse when I left USL (May 1973). The species can certainly be expected in areas other than Florida and Louisiana in the southeastern United States. The following collections will document the presence of H. reptans in Louisiana: LOUISI ANA. Lafayette Parish: weed in USL biology greenhouse, corolla white with a pinkish tinge, leaves purplish below, 7 Jun 1969, Thieret 31815 (SMU), 1 Apr 1973, Thieret 48400 USNK).—John W. Thieret, Department of Biological Sciences, Northern Kentucky State College, Highland Heights, Kentucky 1076. BRAZORIA (LABIATAE) IN MEXICO—Brazoria Engelm. & Gray, genus of four species, is closely allied to Physostegia. It has traditionally ' Department of Botany, Contribution N.S. No. 446. SIDA 6(2): 115. 1975, 116 been considered endemic primarily to Texas, with only one extra-Texas report in literature (a collection of B. scutellarioides from Murray Co., Ok- lahoma [L. H. Shinners, 1953, Field & Lab. 21: 154]). The idea of Texas- Oklahoma endemicity of Brazoria must now be modified in light of a record of B. scutellarioides from Coahuila, Mexico: ‘'15-285 km north of Saltillo,”’ elevation 1,200 ft, in sandy caliche soil, 26 May 1966, Robert Pearce 2129 (SMU). The habitat of the collection is consistent with the observed affinity of B. seutellarioides for well-drained rocky calcareous soils. This report extends the known range of B. scutellarioides 100-135 miles southward.— Jerry M. Flook, Herbarium, Southern Methodist University, Dallas, Texas are FATOUA VILLOSA (MORACEAE): ADDITIONAL NOTES ON DISTRI- BUTION IN THE SOUTHEASTERN UNITED STATES.—Thieret (Sida 1: 245) first reported Fatoua villosa (Thunb.) Nakai in the United States based on a Louisiana collection. DuQuesnay (Sida 5: 285) added a Pasco County, Florida station in 1974. Since these reports I have noted that specimens have appeared in our herbarium from additional stations in the southeastern states, We now have the opportunity to follow and study the spread of a new weed. The following citations, based on specimens deposited at NCU, indicate that this species is apparently being distributed by the transport of greenhouse plants and nursery stock. In some areas of the southeast, it is apparent that, once introduced, this species persists, ALABAMA: Mobile Co. Weed in flower bed, Mobile, 10 Jul 1972, Lelong 6638. FLORIDA: Leon Co. Weed in and around nursery beds, Capital Nursery, N. of Tallahassee, 20 Oct 1970, Godfrey 70116. GEORGIA: Lowndes Co. Weed of flower beds and shrub borders at Castle Park Shopping Center in Valdosta, 20 Jun 1970, Faircloth 6736. LOUISIANA: Concordia Parish. Haphazard Plantation off La. 566, See. 6, TSN, R8E, 11 Aug 1971, Thomas et al. 24566. Ouachita Parish. Weed in flower pots in greenhouses, Roach’s Nur- sery, west of West Monroe, 20 Dec 1969, Thomas & Scarborough 17467. MISSISSIPPI: Forest Co. Greenhouse, Fisher’s Flowers, 7th Street, Hatties- burg. 20 Mar 1972, Rogers 7902. NORTH CAROLINA: Mecklenburg Co. Growing in old pots of Tradescantia outside greenhouse at UNC-C, Charlotte, 15 Oct 1973, Matthews, s.n.— J. R. Massey, University of North Carolina, Chapel Hill, NC 27514. NEW AND VERIFIED ADDITIONS TO THE NEBRASKA VASCULAR FLORA.—Continued field and herbarium work has produced a number of new species for the state and has verified previous collections. All specimens cited are at the Nebraska State Herbarium (NEB). SIDA 6(2): 116. 1975, 117 Elymus junceus Fisch. Thurston Co.: 8.2 miles N. of Macy junction on Hwy 73, plant uncommon along roadside ditch on upland, 19 May 1973, Churchill 690. This introduced Eurasian grass is listed in only a few recent North American manuals. Festuca arundinacea Schreb. Lancaster Co.: north side of Branched Oak Lake NW. of Lincoln, edge of marsh shore in disturbed ground, 11 August 1973, Churchill 2355. Introduced from Europe, this grass was probably planted but apparently spreading at sites, Carex bushii Mack. Jefferson Co.: 5 miles S. of Fairbury on Hwy 15 & just E., seepage area on sandy grass-covered slope of north-facing bluff, 13 June 1974, Churchiil 3538A. This extends the range north from eastern Kansas. Carex frankiti Kunth. Gage Co.: 1.5 miles S. of Rockford, at Rockford State Lake area, edge of intermittent pool, 13 June 1974, Churchill 3569. Richardson . quarter of $26, R18E, TIN, rich wet alluvial soil by shallow pond, 9 Tuy 1974, Shildneck C-6679. An earlier collection of this species is Richard- son Co.: River read 1.5 miles S. of Indian Cave, (3 miles E. & 2 miles N. of Barada) low swampy ground, 27 June 1940, Reynolds 1553. This extends the range north from eastern Kansas. Carex trichocarpa Muhl. Burt Co.: 2.9 miles E. of Oakland on Hwy 32, along moist ditch between railroad tracks and highway, 18 June 1972, Churchill 158. This is an extension west from Iowa. Scirpus georgianus Harp. (Scirpus atrovirens Willd. var. georgianus (Harp.) Fern.). Jefferson Co.: 5 miles S. of Fairbury on Hwy 15 & just E., on lower slope of north-facing bluff in wet sandy soil, 13 June 1974, Churchill 3503, .9 miles S. of Endicot, disturbed sandy wet ditch, 13 June 1974, Chur- chill 3563. Saline Co.: 3.8 miles N. of Wilber on Hwy 103, small creek marsh, 13 June 1974, Churchill 3527. This extends the range north from eastern Kansas. Phlox paniculata L. Nance Co.: just S. of Fullerton on Hwy 14, along Loup River floodplain, in sandy soil at edge of young Salix-Cornus woods, 8 August 1974, Churchill & Kaul 4579. This cultivated plant is the first record collection for the state in the wild. Aster subulatus Michx. var. ligulatus Shinners. Lancaster Co.: just W. of Lincoln at Oak Lake, very common along saline shore, with Leptochloa & Eriochloa, 3 September 1973, Churchill 2652; 7 October 1974, Churchill 4579. This aster extends the range north from central Kansas. I am grateful to the following individuals for their verification or identifi- cation of specimens e Kolstad on Carex, Ronald McGregor on Aster, Alfred Schuyler on ene and David Sutherland on Elymus & Reduce Steven P. Churchill, University of Nebraska State Museum, Lincoln, Ne- braska 68508. SIDA 6(2): 117. 1975. REVIEWS FLOWERING PLANTS: EVOLUTION ABOVE THE SPECIES LEVEL. G. Ledyard Stebbins. XVIII + 399 pp., 59 figs. Belknap Press of Harvard Univ. Press, Cambridge, Mass. 1974. $18.50. This is a very exciting—and busy—time for anyone teaching advanced angiosperm morphology, phylogeny, or phylosystematics. Within the span of a few months we have seen the publication of this book by Stebbins, the 2nd Edition of Foster & Gifford (Freeman & Co.), a new look at distributions by Raven & Axelrod (Ann. Mo. Bot. Gard. 61:539-673. 1974), and a new treat- ment of the old Ranales by Thorne (Aliso 8:147-209. 1974). This new literature is like ‘“‘new bottles for old wine,’ with apologies to the late Sir Julian Huxley. Professor Stebbins has given us a fantastic book. It is full of ideas, new answers for old nagging questions, as well as new questions, and it is de- lightfully readable. Only a person with his vast knowledge and understanding of plants both living and fossil, and his understanding of genetics, develop- mental biology, and ecology could achieve this degree of conceptualism. As he says, it is based upon fifty years of observation, experimentation, and reading hundreds of books and journal articles. The book is divided into two parts. The first deals with ‘‘Factors that De- termine Evolutionary Trends.” In seven chapters he describes the basic evolutionary processes and argues for the concept of genetical uniformitarian- ism and the view that the same basic evolutionary processes give rise to all taxonomic levels from the races within a species to the higher categories. Concomitantly he argues against the necessity to conjure up a mechanism such as directed mutation pressure to explain those features of higher cate- gories that have no apparent adaptive value, and for the necessity to consid- er the importance of internal selection for a genotype governing a harmoni- ous course of development. Explanations are given for the maintenance of morphological similarities or constancy of characteristics of higher cate- gories in face of random fluctuations in the environment interacting with ran- dom mutations. Certain adaptations are discussed at length. The second part, also in seven chapters, concerns hypotheses about the origin of angiosperms and phylogenetic trends, and the adaptations of various character states. His diagram of relationships of orders, subclasses and classes shows affini- ties, degree of specialization, relative size of groups, and the amount of divergence within each group. The last chapter suggests kinds of research that are needed in the future. Each chapter is provided with an excellent and concise summary of the salient arguments and hypotheses. Undoubtedly the textbooks of the future will include such items as Steb- bins’ ‘‘corystosperm theory of angiosperm origin,” “‘internal selection theo- SIDA 6(2): 118. 1975. 119 ry,” and “principle of genetical uniformitarianism.’’ His diagram of evo- lutionary relationships of angiosperm orders will be added to those by Bessey, Hutchinson, Cronquist, Takhtajan, etc. We will add to our everyday vocabu- lary such terms as ‘‘evolutionary cradles,’’ ‘‘evolutionary museums,” ‘‘evo- lutionary canalization’’ and others learned here. This is a monumental work which will have a tremendous and lasting impact on our field. It is indispen- sable to botanists and evolutionary biologists. There are very few mistakes. My only criticism concerns the index. The bock will obviously be used as a reference to Stebbins’ views regarding evolutionary trends and adaptive significance of various features. Yet, the index is lacking in a number of references to such topics. For example, where is the discussion on the evolution of stipules? It could be in chapters 3, 9, 10, or 11 according to the Table of Contents. To help overcome this deficiency, a number of missing entries are given below. Some of these re flect a different choice of entry rather than a complete omission. Androecium, 220, 287 Aquatic acaplaton, 18 Chemical substances, 47 Cotyledons, number, (add) 21 Embryo sac; variation, 27 Epigenesis, (add) 23 Flowers; structure, 213 Fusions, (add) 18, 285, 295 Leaf, compound, 5, 18; deciduous, 253 Leaf arrangement, 212; shape, 212: venation, 212 Ovary position, 18, 301-305 Ovule; trends, 305 Parasitic adaptation, 19 Perianth, differentiation, (add) 220; merosity, 18, 20 Phyllode origin of monocot leaf, 330 Pollen grains, (add) 227 Roct adaptations, 42 Saprophytic adaptation, 19 Stamens; diadelphous, 5; number, 5 18 Svnimietty. floral, 18, 286 Tepals, 220, 284 James W. Hardin, Department of Botany, North Carolina State University, Raleigh, N.C. 27607 SIDA 6(2): 119. 1975. 120 LIST OF FLOWERING PLANTS IN HAWAII. Harold St. John. 519 pp. Mem- oir Number 1, Pacific Tropical Garden, Kauai, Hawaii, 1973. List of Flowering Plants in Hawaii sold for $22.50, but all remaining unsold copies were damaged in a disastrous flood and now sell for $5 to $15 depend- ing on their condition. This work is a vade mecum for the professional bot- anist and advanced student interested in the plant taxonomy of the Hawaiian Islands. The body of the book devotes pages 9 to 13 to Gymnospermae; 14 to 132 to Monocotyledones; 1383 to 369 to Dicotyledones; 369 to 374 to ‘““New Names or Combinations’’; 375 to 378 to an addendum; and finally an index ending with page 519. “The aim of this publication is to present a list of the flowering, or higher, plants known to be in the Hawaiian flora. For each is given the scientific and common names of the plant, genus, species, and infraspecific taxon. The name of the author of the scientific name is given in full or in abbreviation, and the date of publication is added. If the plant is restricted to one or more of the Hawaiian Islands, hence a native to that region, its scientific name is printed in bold face . .. and the islands where it occurs are listed. If it is native to the islands, but also to other regions, it is printed in bold face and is marked indig. ... If it is an introduced weed, it is printed in Roman type ... If the plant is described or mentioned in any of the four basic books on Hawaiian botany, those by Hillebrand, Rock, Degener, and Neal, a page ref- erence to it is given. Since the date of publication of each scientific name is given, it would have been helpful also to have given the full reference to its place of publication. Although these references were verified, this detail is deemed beyond the scope of the present summary.”’ To be sure, full citation of species would have added to the cost and bulk of the volume and would have been quite unnecessary as such information, except for dates, is readily available in the Index Kewensis. For the review- ers, however, full citation of trinomials ignored by the Index would have enhanced still more the value of the List by saving the reader the drudgery of scouring a library for such obscure references. Casual perusal of the index should alert the reader to check the orthogra- phy of names before using them. Of almost 8,000 approved in the List, well over a hundred go contrary to the International Code of Botanical Nomen- clature. To the reviewers the author’s statement that endemic species and infra- specific taxa number 2,668 is patently absurd; nor are they at all in agree- ment that “The endemic, indigenous, and adventive plants in the flora have been well collected and are now quite well known.”’ In fact it has long been their conviction that the flora of the Hawaiian Islands in Captain Cook’s time in 1778 did not consist of a mere 2,668 taxa but of 20,000 or more likely SIDA 6(2): 120. 1975. 121 30,000! Diligent monographic work on historical specimens collected since David Nelson’s botanizing during Cook’s voyage, and diligent collecting and studying of the presently surviving flora, should enable us to know about half the elements that were living 200 years ago. An inkling of the reviewers’ assertion of the number of taxa is shown, for example, by St. John’s treat- ment of the genus Cyrtandra (Gesneriaceae), beginning on page 308. Note the tabulation for the major islands of the Hawaiian Archipelago: ISLAND NUMBER OF TAXA SQUARE MILES SUMMIT IN FEET Oahu 128 604 4,045 Maui 29 728 10,025 Hawaii 23 4,030 13,792 Kauai 22 555 5,170 Molokai 13 260 4,970 Lanai 4 141 3,370 Cyrtandra taxa are partial to wet jungles, and these peter out above the inversion layer where the terrain becomes increasingly dry. This is at about 7,000 feet elevation. Hawaii and Maui, with high mountains, nevertheless have vast rainforests. Can it be true that they harbor but 23 and 29 Cyrtandra respectively? Though Kauai has about 50 square miles less area than Oahu, it has a somewhat greater elevation. This greater range in resulting tem- perature might well increase speciation. Kauai, according to the author, has 22 taxa to Oahu’s 128! In fact, while Oahu with its 604 square miles has 128; the other five islands with a combined total of 5,814 square miles have only 91. The explanation for such discrepancies is not botanical but human. Oahu has been the center of human activity for 200 years. It is the seat of the capital, Honolulu, where the Bishop Museum and the University are located. Most visiting botanists resided there and collected within easy walk- ing, riding, or driving distance of the city. Teachers, not excluding the author of the List, scoured Oahu with their students week-ends and holidays for its botanical riches. The ‘‘outside islands,’ in contrast, alawys have been neg- lected. What wealth of plants must still be growing there unknown to man! What applies to Cyrtandra, relatively unknown in the Archipelago excepting on Oahu, applies more or less to the remaining native genera. With this in mind, the reviewers appeal to the biological workers of the world to come to this Mecca to collect its neglected riches before ‘‘progress”’ destroys them. With the torch of knowledge feebly flickering during the last decade of questionable political ethics in Washington, federal funds for Hawaiian taxonomy have nigh dried up. Even the fabulous Marie C. Neal Herbarium is lying fallow in Honolulu for want of funds. As botanists can- not prevent the continuous slaughter of one endemic taxon after another, SIDA 6(2): 12 oy £97). they should at least attempt to collect, preserve, and record as much of the Hawaiian flora as is still extant so that future generations shall better under- stand what a splendid Paradise their forebears lost.—Otto & Isa Degener, Volcano, Hawaii. APPEAL Appeal for support for the INDEX HOLMENSIS project. The INDEX HOLMENSIS is an index of plant distribution maps with world-wide coverage. It is the only international bibliography on distribution of vascular plants in area and vegetation maps. We have so far published four volumes, i.e. volume I, covering vascular cryptogams and gymnosperms, volume IL containing Monocotyledoneae A-I, volume III Monocotyledoneae J-Z, and finally volume IV covering Dicotyle- doneae A-B, together more than 1000 pages. We intend to continue to publish one volume each year. The total number of distribution maps so far published is estimated to about 400,000, all of which will all finally be listed in the index or it’s supplement. Although the main work is at present done at the Swedish Museum of Natural History in Stockholm the indexing work is serv- ed by an international editorial board. Members of this board to some extent guarantee completeness of the file for their particular area. Still, the number of area and vegetation maps published annually is grow- ing rapidly owing to the increased importance that is felt for the geographic compound of plant taxa. Consequently, not only are there wide areas all over the world where the entire flora is mapped systematically, but mapping has become a common feature in monographs in different fields, as for instance economic botany, palaeo-botany, vegetational history, palynology, and last not least phytocoenology, In order to keep the file for the INDEX HOLMENSIS and it’s planned sup- plement volumes up-to-date we herewith ask our fellow botanists to inform us about their published maps and/or to send reprints of their publications. Needless to say, we shall also continue to supply all information on distribu- ticn maps so far not published in the INDEX HOLMENSIS to colleagues on request. All correspondence should be addressed to: Hans Tralau The Swedish Museum of Natural History S-104 05 STOCKHOLM 50 No — \o N ial SIDA 6(2): 122. . | : es SIDA tosctany VOLUME 6 NUMBER 3 MAY 1976 CONTENTS A biosystematic survey of Coreopsis in eastern United States and Canada. Edwin B. Smith. Varieties of Ipomoea trichocarpa (Convolvulaceae). Daniel F. Austin. Two species of the ‘'Filago germanica"’ group (Compositae— Inuleae) in the United States. Gerhard Wagenitz. Dryopteris YH Fig. 5. Coreopsis nuecensoides with achene (A) and insets of a bud of C nuecensoides (B), and a bud of C. nuecensis (C). 137 se Fig. 6. Distribution of Coreopsis nuecensis (circles) and C. nuecensoides (squares). Coreopsis nuecensoides may occur, also, in northern Mexico (Tamaulipas). ON . C, AURICULATA L., Sp. Pl. 908. 1753. Chrysanthemum hirsutum Plukenet, Phytographia f. 4, pl. 242. 1769. Anacis auriculata Schrank, Denkschr. K6énigl. Akad. Wiss. Munchen 9 (Math. Naturw.):7. 1817 Leachia trifoliata Cass., Dict. Sci. Nat. 25:389 (as to syn. C. auriculata L.). 1822. Chrysomelea auriculata Tausch, Hort. Canal. (15). 1823, Coreopsis auriculata var. diversifolia Ell., Bot. S.C. & Ga. 2:437, 1823. Coreopsis auriculata var. glabrata DC., Prodr. 5:571. 1836. Coreopsis diversifolia DC., Prodr. 5:571. 1836. (not C. diversifolia Jacq. nor C. diversifolia Hook.) Coreopsis auriculata var. B T. & G., Fl. N. Amer. 2:343. 1842. TYPE: United States: Virginia: John Clayton; Herb. Gronovii (LECTO- TYPE: D. Clayton Num. 290. BM!). 139 achene at upper left. oreopsis auriculata; Cc Fig. 7. 140 — | ae ee i > av) 4 "am er oo. © { ped . LA BA Fig. 8. Distribution of C. auriculata. Bleoming: Apr-May. Habitat: Wooded slopes near creeks or rivers, sandy hills. Chromosome number: n = 13. A species (Fig. 7; Fig. 8) rather similar to C. lanceolata but generally easily distinguished from that species by its few cauline nodes, vigorous basal offshoots, outer phyllary shape, and the incurved wing of its achenes. Attempted hybridizations of C. auriculata with C. grandiflora (var. grandt- flora, var. saxicola, and var. harveyana ‘“heterolepis’’ phase), C. nuecensis, C. nuecensoides, C. basalis, C. wrightii, and C. x delphinifolia all failed. In some of these crosses, normal-appearing but empty achenes were formed, suggesting that the hybridization succeeded but the hybrid embryo died at some early stage of development. Successful hybridization of C. auriculata with C. lanceolata was accomplished; see the latter species for details. 6. C. LANCEOLATA L., Sp. Pl. 908. 1753. Coreopsis crassifolia Ait., Hort. Kew. 3:253. 1789. (not C. crassifolia Sesse & Moc.). Coreopsoides lanceolata Moench, Meth. 594. 1794. Coreopsis lanceolata var. glabella Michx., Fl. Bor.-Am. 2:137. 1808. Coreopsis lanceolata var. villosa Michx., Fl. Bor.-Am. 2: 137. 1803. Type: United States: ‘‘Carolina’’ (Isotype P!). 141 Leachia lanceolata Tausch, Hort. Canal. (15). Coreopsis oblongifolia Nutt., J. Acad. Nat. oF cha 7:76. 1834. Coreopsis lanceolata var. suse aera DC., Prod. 5:570. 1 Coreopsis lanceolata var. crassifolia Ait.; Hes Nom. 1:219. “1840, Coreopsis lanceolata var. angustifolia T. & G., Fl. N. Amer. 2:344, 1842. Coreopsis heterogyna Fern., Rhodora 40:475, 1938. TYPE: None cited. I am here designating Martyn’s description and plate of Bidens caroliniana as the neotype (Martyn, Hist. Pl. Rar. 26 and plate. 1728. NEOTYPE !). Blooming: Apr-May south to Jun-Jul north. Habitat: Prairies, glades, sandy slopes, and roadsides. Chromosome number: n — 13 (+ 0-2 This species (Fig. 9; Fig. 10) exhibits considerable variability. Several intergrading phases have been recognized by previous authors at the variety or species level. The most common variation encountered in the field is the presence of a subglabrous and a pubescent phase often intermixed in the same population. Most medern treatments have recognized the pubes- cent phase as var. villosa Michx. I selected arbitrarily 70 herbarium speci- mens of C. lanceolata collected from Louisiana to northern Illinois and counted the number of hairs in a 6 mm circle from the middle of the basal leaves and median leaves. The pubescence of the first cauline internode (at least 1 cm long) was subjectively determined as densely hairy, medium hairy, or sparsely hairy to glabrous. Figure 11 presents the results of the study. It is clear that there is little correlation between pubescence of the lower leaves, median leaves, or stem of C. lanceolata. Individuals may have glabrous stems, hairy basal leaves, and glabrous upper leaves; or glabrous stems, glabrous basal leaves, and hairy upper leaves, etc. e ‘‘variety”’ villosa grades insensibly into the typical variety and is unworthy of recog- nition. In addition to the ‘“‘var. villosa’? phase, several other phases are worthy of mention: (a) Typical phase. Most individuals have a mixture of entire leaves and leaves with two auricles, the basal leaves commonly being entire or two- auricled and the median leaves two-auricled. (b) Entire-leaved phase. Some individuals have all of the leaves entire. This phase tends to be more frequent in the northern part of the range (Ontario, Michigan, Wisconsin, Indiana, Illinois) but is intermixed with the typical phase in that regio (c) Robust phase. Scattered through the range are individuals of unusually large size, commonly with five or six stem nodes and entire to 4(-+)-auricled leaves. An entire-leaved specimen of this phase was apparently the basis for Fernald’s C. heterogyna (Fernald, 1938); I agree completely with him that it was bold of him to propose the species. This phase may represent material of C. lanceolata that has been contaminated by hybridization with C. grandiflora. 142 Fig. 9. Coreopsis lanceolata; A = achene, B = several-auricied (pinnate) leaf appearing on some individuals. 143 yop NASA he TN ; tat \ yu gh ON FG Fig. 10. Distribution of C. lanceolata; also known from widely scattered localities (as an escape or waif?) further west, such as San Joaquin and Ventura counties in California, Larimer Co. in Colorado, and Culberson Co. in Texas. A somewhat peculiar dwarf form, with median leaves pinnately 5-lobed occurs in New Mexico (FE. R. Vasey, Aug. 1881, NY). (d) Dwarf hairy phase. On dry sandy roadsides of some southern states (Florida, Georgia, South Carolina, Louisiana) occurs a dwarf phase ca 2-3 dm tall with only two or three stem nodes and densely hairy spatulate leaves. This phase was the basis for C. crassifolia Ait. but overlaps the typical phase in all characters. Glabrous forms of the phase occur in the same areas; the typical phase is sometimes densely hairy and varies to as short as 2 dm with two stem nodes; it sometimes has spatulate leaves. 144 + on eee NO. HAIRS MEDIAN LEAF Le) o pee ea) eee eee lene ) 0 =) 10 15 20 25 30 35 40 45 50 55+ NO. HAIRS BASAL LEAF Fig. 11. Pubescence of leaves and stem in C. lanceolata. Open circles = sparsely hairy to glabrous stem; half-closed circles — medium- hairy stem; closed circles = densely hairy stem. The open circle with the « at lower left — ideal ‘‘var. lanceolata’; the closed circle with the « at upper right — ideal ‘‘var. villosa’. Further explanation in text. All of the phases intergrade such that it does not appear useful to segre- gate any of them from the typical phase. Attempted hybridizations between C. lanceolata and C. intermedia, C. nuecensoides, C. basalis, C. rosea, C. gladiata, C. tinctoria var. tinctoria, 145 and C. palmata all failed. Large but empty achenes, or selfed progeny, were produced in some of the crosses. Hybridizations that succeeded included: C. lanceolata x C. auriculata. Only two F, hybrids were obtained. They averaged 59% pollen stainability and resembled C. auriculata but bore achenes with both an incurved cartilaginous margin and a spreading wing. No F, was obtained, but two BC, plants (C. lanceolata the repeating parent) averaged 57% pollen stainability. C. lanceolata x C. grandiflora var. grandiflora f. grandiflora. Eighteen F, hybrids were obtained. Thus far 17 have bloomed; these averaged 61% pollen stainability. They had few cauline nodes (3-6) and unusually large heads (first head in bloom 5.0-7.7 em wide). No F, has been produced so far. C. lanceolata x C. grandiflora var. harveyana f. harveyana. Eleven F, hybrids were obtained. Thus far ten have bloomed; these averaged 87% pollen stainability and would have passed morphologically for C. grandiflora ar. grandiflora (or rather broad-leaved var. harveyana). The high fertility of these hybrids suggests that C. lanceolata and C. grandiflora could hy- bridize and intergrade in the field, and this may be the origin of the Robust phase. No F,, has been produced so far. C. lanceolata & C. grandiflora var. harveyana “‘heterolepis’’ phase. Thir- teen F, hybrids were obtained; these averaged 11% pollen stainability and would have passed morphologically for C. grandiflora var. harveyana. The large difference in pollen stainability between this cross and the previous one was unexpected, since the ‘‘heterolepis’’ phase cannot be separated satisfactorily from var. harveyana on morphological grounds. There is a suggestion here that C. lanceolata or C. grandiflora (or both) may contain unrecognized diversity in chromosome structure, which would not be sur- prising in species of such wide geographic range. A few F, achenes of this cross have been obtained but not yet grown. C. lanceolata * C. grandiflora var. saxicola. Achenes of this cross have been obtained but not yet grown. C. lanceolata x C. pubescens var. pubescens. Five F, hybrids were ob- tained. Three have bloomed so far; these averaged 20% pollen stainability. 7. C. INTERMEDIA Sherff, Bot. Gaz 88:299, 1929. TYPE: United States: Texas: Wood Co., Sandy woods, Mineola. J. Rever- chon 2077, 12 Jun 1900 (HOLOTYPE B, destroyed; ISOTYPE MO). Blooming: late May-early Jul(Aug). Habitat: Sandy woods. Chromosome number: n = 13 (+ 0-1 B). I have not seen type material of this species but have examined material of the same collection number (J. Reverchon 2077) from NY. This is a rare species (Fig. 12) previously known only from the type locality and vicinity (Sherff, 1955). I have observed material from several other localities (Fig. 13), including two collections from northwestern Louisiana (Thieret 31204, Smith 1834; UARK). The species varies from nearly glabrous to rather 146 ry! at CAN ).. y Lt ea \\ oe ft A [ 7 DH A | Cn \ Fig. 12. Coreopsis intermedia; A = achene with entire wing, B = achene with irregularly fimbriate wing. densely short pubescent; the achene wings vary from entire to irregularly partially fimbrate. Attempted hybridization of C. intermedia with C. grandiflora var. harvey- ana (both the normal and the “heterolepis’’ phase), C. basalis, C. wrightii, C auriculata, C. lanceolata, C. rosea, C. linifolia (diploid race), C. nudata, C. gladiata, C. tinctoria var. tinctoria, C. major f. major, C. major f. oemlert (hexaploid race), C. tripteris and C. latifolia all failed. Suecessful hybridization included: C. intermedia « C. grandiflora var. grandiflora. Mature achenes were obtained but have not yet been grown. C. intermedia < C. grandiflora var. saxicola. Thirteen F, hybrids were obtained; these averaged 66% pollen stainability. One analyzed meiotically 147 Fig. 13. Distribution of C. intermedia; dots = semi-glabrous phase, squares = pubescent phase. exhibited » = 13 at diakinesis and regular distribution (13:13) at anaphase J. Thirty-three F, hybrids of this cross were grown. Thus far, 27 have flowered; these 27 averaged 52% pollen stainability. Some individuals pro- duced very little pollen. C. intermedia * C. pubescens var. debilis. Eleven F, hybrids were ob- tained; these averaged 39% pollen stainability. A few F, achenes have been obtained but not yet grown. C. intermedia * C. pubescens var. pubescens. Sixteen F, hybrids were obtained; 14 of these averaged 49% pollen stainability. In a reciprocal cross, four F, hybrids averaged 43% pollen stainability. The C. pubescens parent in this cross had a pair of B chromosomes (n = 13!! + 1!! B ). Three F, hybrids analyzed meiotically had n = 10!!, 1!) 1IV and nm = 11!!, 1!V, 1! at diakinesis, suggesting that C. intermedia and C. pubescens var. pubescens differ chromosomally by one reciprocal translocation and that the B chromo- some sometimes pairs with another of the normal pairs of chromosomes (i.e., is at least partly trisomic). Eighteen F, hybrids of this cross were obtained; only ten have bloomed. These ten averaged 19% pollen stain- ability; several produced very little pollen. Eight BC, hybrids (C. pubescens var. pubescens the repeating parent) were obtained; six of these averaged 59% pollen stainability. 148 C. intermedia « nuecensis. Four F, hybrids of this cross were obtained; three of these averaged 0.5% pollen stainability, One analyzed at diakinesis of meiosis had n = 1!, 2!!, 1!V, 1v. The heads produced by the F, hybrids had pubescent inner phyllaries and ligules with darker yellow bases, with some ligules basally flecked and some non-flecked. It is clear from the above hybridization data that C. intermedia is closely allied with C. grandiflora and C. pubescens. 8a. C. PUBESCENS EIl., Bot. S. C. & Ga. 2:44. 1823, var. PUBESCENS Coreopsis auriculata sensu Schkuhr, Handb. 3:156. 1803. (not C. auricu- lata L.) Coreopsis auriculata T. & G. vars. y and 6, Fl. N. Amer. 2:343-344. 1842. Coreopsis pubescens var. typica Sherff, Brittonia 6:341, 1948. TYPE: None cited; habitat given as ‘‘western districts of Georgia’. I am here designating a neotype: NEOTYPE: United States: Arkansas: Carroll Co., common along fencerow by hwy. 68, 13.8 mi. ENE of the Carroll- Madison co line, FE. B. Smith 1759, 25 Jul 1973 (US!). In designating a neo- type for this species, I am deliberately avoiding selection of a specimen from the ‘‘western districts of Georgia,’’ since this area is near the edge of the range of C. pubescens var. pubescens and contamination of the variety pubescens with C. grandiflora var. saxicola or C. pubescens var. robusta is not uncommon there. Little or no contamination of var. pubescens occurs in the Arkansas area. Blooming: Jun-Aug. Habitat: Alluvial banks, open woods, sandy fields. Chromosome number: n = 13 (+ 0-2 B). A species (Fig. 14, A-D; Fig. 15, dots) resembling C. auriculata but differ- ing in its much larger size with many stem nodes, narrow outer phyllaries, and winged achenes. Attempted hybridizations of C. pubescens var. pubescens with C. nuecen- soides, C. basalis, C. palmata, and C. tripteris all failed. Successful hybridizations included: C. pubescens var. pubescens < C. auriculata. Only three F, hybrids were obtained; only one has bloomed and it had 10% pollen stainability. C. pubescens var. pubescens x C. intermedia. Four F, hybrids are men- tioned under the latter species. C. pubescens var. pubescens x C. grandiflora var. saxicola. Four F, hy- brids of this cross averaged 27% pollen stainability. The hybrids looked much like C. grandiflora var. grandiflora f. pilosa. Nine BC, hybrids (C. pubescens var. pubescens the repeating parent) were obtained; eight of these averaged 62% pollen stainability. Three reciprocal BC, hybrids (C. grandiflora var. saxicola the repeating parent) were obtained; they aver- aged 55% pollen stainability. The parent material of both var. pubescens and var. saxicola in the above crosses was from Arkansas. In one reciprocal cross, C. grandiflora var. saxicola « C. pubescens var. pubescens, where the var. saxicola parent Fig. 14. Coreopsis pubescens var. pubescens (A-D) and var. robusta (E). A = lower stem; B = median stem; C = upper stem and flowers; D = achene; E = median leaf of var. robusta. 150 zat q y WO *, sen 00s) car we ners ™, om pois = ~ : . 4 ays 5 eon = Fraee, Far A FS fa soe 1s | 2 Ve a ae e be NZ Rs ge ey USS 6 A fer | |z xy} OY | mi 2 one - he LL weds ‘ ~/ ye as mwa] | |3 5 Pot eee eres | —, ‘ . dan AS Ue eem Te ec. sare * — | wee Pa - > \ ( ‘ / 5 / re as a . [ os i 4 eT wom] . : [ se GL J % : | we : Jiroree EP 7] aver vee” bets o NLA e. = ee bee oe = ‘ po 4 = per Tene SY NE es ee | ee ee ee . N VLGN_I, A pee wert [eine eS ee ae pee] i = Sa Teck bee Re pee eer . 5 gers FS [fees triangles bilis, sparsely = var. pubescens; Fig. 15. Distribution of C. pubescens. Dots white star in black circle hairy to essentially glabrous; black stars = var. debilis, apparently intergrading with var. pubescens and definitely hairy. robusta; 151 material was from Georgia, eight F, hybrids averaged 3% pollen stain- ability, suggesting that the var. saxicola material from Georgia differs chromosomally from the Arkansas material of var. saxicola C. pubescens var. pubescens x C. pubescens var. debilis. Thirteen F, hybrids of this cross were obtained; they averaged 90% pollen stainability, indicating very close relationship between these two varieties. All F, plants were pubescent. Twenty-four F, hybrids were obtained; 21 of these averaged 60% pollen stainability, with a range from 0 to 98%. Twenty-one of the 24 F, were pubescent and three were essentially glabrous. Ten BC, hybrids (var. debilis the repeating parent) were obtained; these averaged 91% pollen stainability. Eight of the ten were pubescent and two were essentially glabrous. The glabrous plants approached the morphology of C. pubescens var. robusta, C. pubescens var, pubescens * C. lanceolata. Five F, hybrids are men- tioned under the latter species. 8b. C. PUBESCENS Ell. var. ROBUSTA A. Gray ex Eames, Rhodora 18:239. 1916. Coreopsis testudinea Shuttl., nom. nud., in Sherff, Field Mus. Publ. Bot. 11:351. 1936 TYPE: United States: North Carolina: North Branch of New River, A. Gray and J. Carey s.n., Jul 1841 (HOLOTYPE GH!). Blooming: Jun-Aug. Habitat: Wooded slopes, glades, meadows, edge of cliffs. Chromosome number: unknown I collected only a few living plants of var. robusta (Fig. 14, E; Fig. 15, triangles), and all of these died in the greenhouse prior to my being able to determine their chromosome number or use them in attempted hybridi- zations. The var. robusta intergrades somewhat with var. pubescens in the mountainous districts of western North Carolina and northeastern Georgia ut can usually be distinguished with little difficulty. I believe it should be retained at the varietal level. 8c. C PUBESCENS Ell. var. debilis (Sherff) IX. B. Smith, comb. & stat. nov. Coreopsis debilis Sherff, Bot. Gaz. 89:366. 1930. Coreopsis corninsularis Sherff, Bot. Gaz. 94:597. 1933. Type: United States: Mississippi: Horn Island, S$. M. Tracy 8542, 25 May 1903 (HOLO- TYPE MO!; ISOTYPE F'!). TYPE: United States: Mississippi: Harrison Co., In dry, sandy, old fields along coast of Mississippi Sound, J. D. Smith 602, 15 Sept. 1885 (HOLOTYPE F!; ISOTYPE US!, 2 sheets). Blooming: late May-Aug(Sep). Habitat: Old sandy fields, rocky outcrops, slopes. Chromosome number: n = 13. Coreopsis debilis is a peculiar segregant of C. pubescens, as indicated by its interfertility with C. pubescens var. pubescens (see hybridizations under var. pubescens) and its intergradation with var. pubescens in the 152 northern Alabama-Mississippi-central Tennessee area. The extreme coastal form (Fig. 16; Fig. 15, white star on black circle) with its bushy habit, glabrous stem and blades, and two-auricled or weakly pinnate leaves ap- pears like an unusual form of C. grandiflora, Some individuals with entire to two-auricled leaves tend to mimic glabrous C. lanceolata; in the area of intergradation with var. pubescens, pubescent individuals appear (Fig. 15, stars) that seem to be rather peculiar material of C. lanceolata ‘“‘var. villosa’”’ (a variety which is itself merely a populational segregant). Gen- erally, it can be distinguished by its bushy habit, small heads, and inter- mediate number of cauline nodes. I believe it should be recognized at the varietal level. The rare “species” C. corninsularis Sherff (known only from Horn Island, Mississippi; Sherff, 1955) is scarcely different from the subentire-leaved forms of C. debilis. I have been unable to obtain living material of C. corninsularis but am convinced that it would be freely interfertile with C. debilis and should therefore be merged with it as C. pubescens var. debilis, Attempted hybridization of C. pubescens var. debilis with C. tinctoria var. tinctoria failed. Coreopsis pubescens var, debilis is highly interfertile with C. pubescens var. pubescens (see above) and shows low to intermediate interfertility with C. intermedia, C. grandiflora var. harveyana, and C. lanceolata. See the latter three species for details. 9a. C. GRANDIFLORA Hogg ex Sweet, Brit. Fl. Gard. 2: pl. 175. 1826, var. GRANDIFLORA forma GRANDIFLORA Coreopsis boykiniana Nutt., Trans. Amer. Philos. Soc, H. 7:38. 1841. Coreopsis heterophylla Nutt., ibid. (not C. heterophylla of others). Coreopsis grandiflora var. subintegrifolia T. & G., Fl. N. Amer. 2: 345. 1842 TYPE: United States: Sent by a Mr. Hogg from New York to England, represented by Robert Sweet’s Type Plate (loc. cit.)! Sherff (1936) specified that Sweet’s plate would serve as the type for C. grandiflora. I inquired for possible type material at BM, but none exists there. Blooming: late May-Jun. Habitat: Roadsides, prairies, sandy glades, railroads. Chromosome number: n = -+ 0-4 B) Coreopsis grandiflora (Fig. 17), a complex species of wide range (Fig. 18), has several somewhat intergrading varieties and a number of forms. Its complexity probably stems from local differentiation over its wide range plus its general promiscuity as indicated in artificial hybridizations. Attempted hybridizations of C. grandiflora var. grandiflora f. grandiflora with C. grandiflora var. longipes, C. nuecensoides, C. basalis, C. rosea, and C. major f. major all failed. But C. grandiflora var. grandiflora f. grandt- flora exhibited medium to high interfertility with C. grandiflora var. grandi- flora f. pilosa, C. grandiflora var. harveyana f. harveyana, and C. lanceo- 153 debilis; achene at lower right. Fig. 16. Coreopsis pubescens var. 154 lata: see these latter taxa for details. 9b. C. GRANDIFLORA Hogg ex Sweet var. GRANDIFLORA forma pilosa (Sherff) EX, B. Smith, comb. & stat. nov. Coreopsis grandiflora var. pilosa Sherff, Bot. Gaz. 89:368. 1930. TYPE: Cultivated at Agricultural Station, Island of Bermuda. S. Brown, N. L. Britton & P. Bisset 2011, 22 May-6 Jun 1914 (HOLOTYPE NY!). Blooming: May-Jul. Habitat: Roadsides, fields. Chromosome number: n This form, treated as a variety by Sherff (1936, 1955), was recently known in the wild state only from northern Georgia (Sherff, 1955). I have seen material of f. pilosa from Georgia, Alabama, Mississippi, eastern Texas, and eastern Oklahoma (Fig. 18, white star on black circle). The form tends to occur at the limits of the range of C. pubescens var, pubescens and probably has arisen through the hybridization of C. grandiflora var. grandi- flora (and, possibly var. harveyana) with that species. Artificial hybridiza- tions between C. pubescens var. pubescens and C. grandiflora var. saxicola have produced offspring that in some cases mimic the naturally occurring f. pilosa, The pubescent f. pilosa occurs intermixed with the glabrous f. grandiflora. I raised four plants in the greenhouse from achenes harvested from a f. pilosa plant in northern Georgia; two were glabrous and two were pubescent. I crossed one of the pubescent ones with one of the glabrous ones: of 12 offspring, five were pubescent and seven glabrous. I also crossed two of the pubescent plants; of 16 offspring, all were pubescent. These results indicate that f. pilosa differs from f. grandiflora by a single gene, with the glabrous condition dominant and the pubescent condition homo- zygous recessive. 9c. C. GRANDIFLORA Hogg ex Sweet var. SAXICOLA (Alex.) E. B. Smith, Brittonia 25:207. 1973. Coreopsis saxicola Alex., Torreya 31:161. 1932. Coreopsis saxicola Alex. var. duncanii Sherff, Bot. Leafl. 6:2. 1952. Type: United States: Georgia: Oglethorpe Co., in shallow soil at edge of thickly vegetated drain at Echol’s Mill SE of Point Peter, W. H. Duncan 3832, 4 Aug 1941 (HOLOTYPE GA!). TYPE: United States: Georgia: DeKalb Co., on thin soil over granite, Stone Mountain, F. W. Pennell 4029, 2 Aug 1912 (HOLOTYPE NY! Blooming: Jun (in Arkansas), Jul-Aug (in Georgia & Alabama). Habitaks Shallow sandy soil over sandstone or granite. Chromosome number: n = 18 (+ 0-4 B I have given detailed reasons elsewhere (Smith, 1973) for the reduction of C. saxicola to a variety of C. grandiflora and for the merging of var. duncanii with var. saxicola, Variety saxicola (Fig. 17, B; Fig. 18, open stars) is not homogeneous; the eastern phase (Georgia, Alabama) blooms later than the western phase (Arkansas), has somewhat wider leaf segments ae ui Ot Fig. 17. Coreopsis grandiflora var. harveyana (plant and achene A) and var. saxicola (achene B). 156 than the western phase, and apparently differs somewhat chromosomaliy (see hybridization, below). Attempted hybridizations of C. grandiflora var. saxicola with C. grandi- flora var. longipes, C. basalis, C. auriculata, C. nudata, C. major f. oemlert (hexaploid race), C. tripteris, and C. latifolia all failed. Successful hybridizations included: C. grandiflora var. saxicola (Arkansas phase) x C. grandiflora var. harveyana, Six F, hybrids of this cross averaged 96% pollen stainability. Kighteen IF, plants were highly fertile. In a repeat of this cross performed a year later, with the var. harveyana material represented by the “‘hetero- lepis”’ phase, five F, hybrids averaged 31% pollen stainability; in diakinesis of meiosis these F, hybrids exhibited mostly n = 8!, 1!'V, 1V!, indicating at least four structural differences in the parental chromosome complements. C. grandiflora var. saxicola (western phase; n = 13 + 2B) Xx C. grandi- flora var. savicola (eastern phase; n = 13). Twelve F, hybrids of this combination were raised and these averaged 379% pollen stainability; at diakinesis one of these exhibited n — 13 + 1 B. The eastern and western phases of var. saxicola appear to differ somewhat, but the lowered F, fertility is evidently not due to major structural differences in the chromo- some complements of the two. Several crosses were made with respect to inheritance of the B chromo- somes commonly present in Arkansas material of var. saxicola, The crosses made and the results were: n— 134+ 2B x n = 18. Eighteen F, hybrids of this cross were obtained. Chromosome counts were obtained for 15 plants. Two were n = 138 and aver- aged 95% pollen stainability; 13 were n = 13 + 1B and 12 of these aver- aged 899% pollen stainability. n— 134+ 3B x n = 18. Seventeen F, hybrids of this cross were ob- tained. Chromosome counts were obtained for 14 plants. Nine were n = 13 + 1B and eight of these averaged 94% pollen stainability; five were n = 13 + 2 B and averaged 83% pollen stainability. n— 13x« n = 13 + 4B. Ten F, hybrids of this cross were obtained. Chromosome counts were obtained for nine plants. Four were n = 13 2B and averaged 97% pollen stainability; one was n = 13 +4 3 B and had 16% pollen stainability; and four were n = 13 + 4 B and averaged 23% pollen stainability. The above crosses indicate that the B chromosomes in var. saxicola material from Arkansas behave normally on the maternal side but tend to increase (or at least be maintained) on the paternal side. The B chromo- somes have no noticeable effect on the morphology but, as in Haplopappus (Smith, 1968c), reduce pollen stainability significantly when three or more are present. Several meiotic cells seen in the plants of this series of crosses were autotetraploid, the observed frequency varying in individual plants from very low to as high as 50% in one individual of n = 13 + 4 B con- 157 an is Eph rn oO Se at at Ded a ico & ont Ai | ry Hq fi Ay 13 I ! H ! I = C. grandiflora var. = C. grandi- g. 18. Distribution of C. grandiflora: semicircles grandiflora f. grandiflora; white star in black circles = flora var. grandiflora f. pilosa; dots = C. grandiflora var, harvey- = C. grandiflora var. harveyana f. Fi ana f. harveyana; closed stars = = C. grandiflora var. saxicola; triangles = demareei; open stars C. grandiflora var. longipes. a 158 stitution (the autotetraploid cells at diakinesis were mostly n = 23!!, QIU, IV C. grandiflora var. saxicola (eastern phase) x C. pubescens var. pubes- cens. Eight F, hybrids of this combination averaged 3% pollen stainability and morphologically resembled C. grandiflora var. grandiflora f. pilosa. The reciprocal of this cross (with the western phase of var. saxicola), as well as BC, hyrids, was considerably more fertile. See C. pubescens var. pubescens hybridizations for details. C. grandflora var. saxicola hybridized successfully with C. intermedia; see the latter species for details. 9d. C. GRANDIFLORA Hogg ex Sweet var, HARVEYANA (A. Gray) Sherff, Bot. Gaz. 94:593. 1933, forma HARVEYANA Coreopsis harveyana A. Gray, Syn. Fl. N. Amer. [?:292. 1884. Coreopsis heterolepis Sherff, Bot. Gaz. 89:365. 1930. Type: United States: Arkansas: Cleburne Co., on dry sandy bluffs, Heber Springs, EF. J. Palmer 6962A, 30 Oct. 1914 (HOLOTYPE MO!; ISOTYPE MO!, two sheets). TYPE: United States: Arkansas: on cliffs near Fort Smith, IF’. L. Harvey s.n., Jun (HOLOTYPE GH!). Blooming: May-Jun (Sep-Oct in late season forms). Habitat: Sandy ridges, sandstone glades, rocky hillsides and prairies, roadsides, near railroad tracks. Chromcesome number: n = 138. This variety (Fig. 17; Fig 18, dots) is similar to var. grandiflora but has very narrow upper leaf segments and generally smaller heads. It reaches its greatest abundance in Arkansas. The ‘species’? C. heterolepis, named by Sherff on the basis of its “densely aggregated basal leaves’? and small heads, is merely a late season form of var. harveyana with clusters of leaves at old leaf bases (basically aerial rosettes) and with a few late heads in bloom, Several species of Coreopsis in this section and section Eublepharis commonly form such late season aerial rosettes. The ‘‘species” overlaps var. harveyana in all respects and cannot be maintained on any presently known morphological basis. However, there apparently are chromosomal differences between the typical var. harveyana and_ the “heterolepis” phase (see the C. grandiflora var. saxicola hybridizations and some of the crosses below). Localized chromosomal heterogeneity might be expected in a wide-ranging species such as C. grandiflora. Forma harveyana is essentially glabrous on the stem and the flat sur- faces of the blades. Attempted hybridizations of C. grandiflora var. harveyana f. harveyana with C. grandiflora var. longipes and C. tinctoria var. tinctoria failed. Suecessful hybridizations included: C. grandiflora var. harveyana f. harveyana « C. pubescens var. debilis. Ten F, hybrids of this combination averaged 35% pollen stainability; one of these analyzed meiotically exhibited n = 10!!, 2!!! at diakinesis. 159 C. grandiflora var. harveyana f. harveyana (typical phase) « C. grandi- flora var. harveyana f. harveyana (‘‘heterolepis’’ phase). Four F, hybrids of this cross averaged 20% stainability. At diakinesis, these exhibited n = g!!, 21V 4. a small fragment: at anaphase I, some cells showed a bridge and fragment, indicating heterozygosity for an inversion. C. grandiflora var. harveyana f. harveyana « C. grandiflora var. grandi- flora f. grandiflora. Twenty-six F, hybrids of this cross were raised. Thus far, only 16 have bloomed; these 16 averaged 74% pollen stainability. One of the F, analyzed meiotically exhibited n = 13 at diakinesis and 13:13 distribution at anaphase I. As the male parent, C. grandiflora var. harveyana f. harveyana was suc- cessfully hybridized with C. grandiflora var. saxicola and C. lanceolata; see these latter taxa for details. 9e. C. GRANDIFLORA Hogg ex Sweet var. HARVEYANA (A. Gray) Sherff forma DEMAREEI Sherff, Amer. J. Bot. 31:277. 1944 TYPE: United States: Arkansas: Bradley Co., on prairies, Warren, D. Demaree 19,262A, 7 Jun 1939 (HOLOTYPE F'!; ISOTYPE GA!). Blooming: May-Jun. Habitat: Sandy prairie. Chromosome number: n — 13. This form differs from f. harveyana in being rather densely short-pilose on the stem and flat surfaces of the leaves. Previously known only from the type locality (Sherff, 1955), it has been discovered by Marie P. Locke of Pine Bluff, Arkansas, growing also in Grant County in the parking lot of the Red Cockaded Woodpecker bird sanctuary, southeast of Sheridan (Fig. 18, c’osed stars). At the Grant County location both typical f. demareei and f. harveyana occur, as well as individuals with an intermediate degree of pubescence (leaf surfaces moderately short-pilose, stem mostly glabrous on internodes but rather densely short-pilose on and near the nodes. This is the only location known where apparent natural hybridization takes place between the two forms. Forma demareei and f. harveyana were hybridized; four F, plants were medium-pilose with 97% average pollen stainability. In the F,, 11 of 21 seedlings were albino; five of the 10 green seedlings survived to flowering. Only stem pubescence and pollen stainability were recorded on these: two were essentially glabrous (somewhat pubescent at the nodes), two were medium-pilose, and one was densely pilose; average pollen. stainability was 89%. Forma demareei appears to differ by a single incompletely domi- nant gene from f. harveyana. The albino F, seedlings may have been just a chance occurrence in this particular cross or may represent products of a balanced lethal system which would help prevent gene exchange between the two forms. 9f. C. GRANDIFLORA Hogg ex Sweet var. LONGIPES (Hook.) T. & G., Fl. N. Amer. 2:345. 1842 Coreopsis longipes Hook., Bot. Mag., pl. 3586. 1837. TYPE: United States: Texas: T. Drummond s.n. (HOLOTYPE BM!). 160 Type material of the variety is still extant and would take precedence over the plate by Hooker (indicated as the type by Sherff, 1936). Blooming: May. Habitat: Sandy readsides, sandy woods, near railroad tracks. Chromosome number: n = 39 (hexaploid). This variety is sometimes difficult to distinguish from var. grandiflora and var. harveyana but can usually be recognized by the combination of its broader leaf segments and larger achene aristae. It tends to be strongly stoloniferous and is hexaploid, while the other varicties are weakly stoloni- ferous and diploid. It is endemic to Texas (Fig. 18, triangles). So far, I have not been able to secure hybrids of var. longipes with any other taxon in Coreopsis. Attempted hybridizations of var. longipes with C. grandiflora var. grandiflora, C. grandiflora var. harveyana (both typical and ‘“heterolepis’’ phases), grandiflora var. saxicola, C. leavenworthii, C. intermedia, C. nudata, C. tinctoria var. tinctoria, and C. tripteris all failed. 1 II. Section SULPHIDIUM T. & G. ea A. Gray, Syn. . N. Amer. [1°:294. 1884. Sect. Eucoreopsis subsect. Silphidium T. & G., . Amer. 2:341. 1842. Subg. Silphidium F. Boynton in Small, Man. SE. | 1446. 1933. Sect. Latifoliae I. Boynton in Small, Man. SE Fl. 1446, 1933. Leiodon Shuttl.; Sherff, Field Mus. Nat. Hist., Bot. Ser. 11:416, as genus. 1936. £3 Herbaceous perennial; leaves opposite, simple, serrate; ray flowers four or five, lemon-yellow, sterile, the ligule elliptic-oblong and entire at the apex; disk flowers ca. 10-18, the corolla ca. 5.5-6 mm long, lemon-yellow, 5-toothed apically, with five anthers, the style tips deltoid in outline; palea linear, obtuse, glabrous, deciduous; achenes obcompressed, ca. 7 mm long, not winged, glabrous, bald at the apex. Section monospecific (C. latifolia). = 13 10. C. LATIFOLIA Michx., Fl. Bor.-Amer. 2:187. 1808. Leiodon latifolium (Michx.) Shuttl., nom. nud., in Sherff, Field Mus. Nat. Hist. Bot. Ser. 11:416. 1936. TYPE: United States: at higher altitude in the Carolina Mountains, A. Michaux s.n. ISOTYPE P! Blooming: Aug.-Sep. Habit: Rich, moist, shaded higher slopes. Chromo- some number: n = 18 Coreopsis latifolia (Fig. 19) is a rare, very distinctive, and primitive (Smith, 1975) species endemic to the mountains of the Carolinas and north- eastern Georgia (Fig. 20, stars). Its closest relatives are in section Electra of the genus and occur in Mexico. In general habit, especially in the serrate leaf margins, it appears to have had some ancestry in common with Bidens (as perhaps, does the entire section Electra; cf. Smith, 1975) I have not been able to obtain hybrids of C. latifolia with any other species of Coreopsis. Attempted hybridizations of C. latifolia with C. grandiflora var. saxicola, C. grandiflora var. harveyana “‘heterolepis” phase, C. nuecen- soides, C. basalis, C. major f. major, C. major f. oemleri (hexaploid race), = 161 Fig. 19. Coreopsis latifolia; achene at upper left. 162 mur kins BUTLER \ srk a v " OF | peuney, and Lome, i ADAIR =a picets i CHRISTIAN : © PUL ASR ‘dvnit Lestit ae ‘ 1000 * mu sste RUSSELL LOGAN BARREN * a cumpes SL warne HARLAN Scort can (AND mercy ¢ . i Ae JON . WASHING: es SIMPSON AON ROE —_ matey ( ROBERTSON SUMNER [MACON ‘ CLAY mANcOce SULLIVAN Mecatany scon CLAIBORNE anne UNCOLN | . 1 NION | ANSON LAUDEROALE €] MADISON JACKSON | a ahs = rc “s 2 CHESTERFIELD %s DLBERT ~ we s FAIRFIELD KERSHAW MARSHALL ae ; L* Ute NEWBERRY ig S () | T H. FLOREN SALUDA LUD Cexincron’s RICHLAND ( SUMTER CAA RKO_L INA CALHOUN ARLNOON way AIKEN BERKELEY BARNWELL LCEND COLLETON ° = FSCREVEN i as HARLESTON : e PS lannee Fig. 20. Distribution of C. latifolia (black stars) and C. pulchra (white star in black circle). The exact county in northeastern Georgia for C. latifolia is unknown. Coreopsis pulchra may also occur in north- western Georgia (Sherff, 1955). C. pulchra, C. tripteris, C. failed. In several of these crosses, good achene set occurred, but the seed- lings grown from the achenes were all normal C. latifolia. Coreopsis lati- folia is apparently self-fertile (or possibly apomictic) when stimulated by pollen from related species. When heads of plants of C. latifolia were bagged and left bagged in the greenhouse, no achene set occurred. linifolia (diploid race), and C. intermedia all Ill. Section PALMATAE F. Boynton in Small, Man. SE. FI. 1446. 1983. Anacis Schrank (in part: A. tripteris), Denkschr. Konigl. Akad. Wiss. Mtinchen 5(Math. Naturw.):5, as genus, 1817. Chrysostemma Less., Syn. Gen, Compos. 227, as genus. 1832. Sect. Gyrophyllum Nutt., Trans. Am. Philos. Soc. Il. 7:358. 1841. Subg. Anacis F. Boynton in Small, Man. SE. FI. 1446. 1933. Herbaceous perennials; leaves opposite, simple and entire or commonly palmately three-parted, the three leaflets entire to lobed or dissected; outer phyllaries adnate to the inner phyllaries and therefore arising about 1 or 2 mm above the base of the head; ray flowers about 6 to 8, lemon-yellow, the ligule elliptic-oblong and entire to slightly denticulate at the apex; disk flowers ea. 25-55, the corolla ca. 4.5-6 mm long, greenish-yellow to lemon- 163 yellow (or wine-purple apically), 5-toothed apically, with 5 anthers, the style tips deltoid to deltoid-caudate in outline; palea linear to filiform, dilated apically and obtuse or acute, glabrous, very tardily deciduous; achenes ob- compressed, ca. 3-7 mm long, narrowly winged, glabrous, bidentate (the teeth often lacerate or plurisetose) at the apex. x — 13 11. C. TRIPTERIS L., Sp. Pl. 908. 1753. Anacis tripteris Schrank, Denkschr. Konigl. Akad. Wiss. Munchen 5(Math Naturw.):7. 1817 Chrysostemma tripteris Less., Syn. Gen. Compos. 227. 1832. Coreopsis tripteris var. § T. & G., Fl. N. Amer. 2:341. 1842. Coreopsis tripteris var. smithii Sherff, Bot. Gaz. 88:301. 1929. Type: United States: Alabama: Low open woods near Montgomery, J. D. Smith s.n. 26 Aug 1885 (HOLOTYPE F'!). Coreopsis tripteris var. subrhomboidea Sherff, Bot. Gaz. 88:303. 1929. Type: United States: Texas: Bowie Co., sandy open woods bordering bog, near Texarkana, EF. J. Palmer 29421, 27 Oct 1925 (HOLOTYPE H!). Coreopsis tripteris var. intercedens Standley, Rhodora 32:34. 1930. Type: United States: Illinois: Edgewater, Chicago, F’. Gates 808, 23 Aug 1905 (HOLOTYPE F) Coreopsis tripteris var. deamii Standley, Rhodora 32:33. 1930. Type: United States: Illinois: Henderson Co., dry bluffs, H. N. Patter- son s.n., Aug 1871 (HOLOTYPE F, Acq. No. 17866). TYPE: None cited by Linnaeus. Habitat given as ‘‘Virginia’’. I am here designating the following neotype: United States: Virginia: Greensville Co., open woods NW of Taylor’s Millpond. M. L. Fernald & B. Long 11202, 29 Aug 1939 (NEOTYPE F'!). Blooming: Jul-Sep. Habitat: Open woods, creek banks, roadside, prairies, railroad right-of-way. Chromosome number: n = 13. Several varieties have been proposed within C. eer (Fig. 21; Fig. 22) on the basis of pubescence (vars. intercedens & deamit), entire leaves (var. smithii) versus the usual three-lobed leaves, and leaflet shape (var. subr- homboidea). In my opinion, all of these varieties lack merit. The pubescence of C. tripteris is variable, from essentially none to rather densely short- pubescent on the leaves and phyllaries. When observation is limited to local material, these variations may appear to merit recognition, but when viewed over the entire range of the species there is no clear-cut correlation of the pubescence types with geographical or ecological differences. Variously pubescent types often occur intermixed in the same local areas. It is pos- sible that these pubescence types might be validly recognized as formae, but I prefer to simply merge them with the typical type. The var. smithii, with ‘‘all or nearly all’’ leaves simple (Sherff, 1955), seems exceedingly weak to me since the ‘‘typical’’ variety bears simple leaves on the upper stem above the three-lobed median leaves. Depauperate 164 Fig. 21. Coreopsis tripteris; achene at upper center. 165 Mig. 22. Distribution of C. tripteris. 166 specimens often have leaves reduced in size as well as dissection; the var. smithii almost certainly represents merely a reduced growth form of the typical material. While the most common leaflet shape in C, tripteris is narrowly lanceolate, the leaflets vary from nearly linear to oblong, elliptic-oblong, or even rhombic-elliptical. The somewhat rhomboidal shape of the leaflets of ‘‘var. subrhomboidea”’ represents merely population variation within C. tripteris. The long-petiolate three-lobed leaves of this species clearly distinguish it from all other species of Coreopsis in eastern United States and Canada. Attempted hybridizations of C. tripteris with C. grandiflora var. grandi- flora, C. grandiflora var. longipes, C. grandiflora var. saxicola, C. pubescens var. pubescens, C. intermedia, C. nuecensoides, C. lanceolata, C. rosea, C. leavenworthii, C. palmata, C. major f. major, C. verticillata (tetraploid race), and C. xdelphinifolia all failed. In a few of these crosses, empty full-sized achenes were formed or a few selfed achenes. Successful hybridizations involved only: C. tripteris x C. major f. oemleri (n = 39). This hybridization was only very slightly successful: a single F, plant was obtained. Morphologically, the hybrid could have passed for C. major f. oemleri, although the leaves were short-petiolate (petioles ca. 2 mm long on median leaves; up to 6 mm long on lower leaves). The plant at pollen cell diakinesis exhibited n = ca. 15!, Bll, 3'll, 3IV but had pollen stainability of 899%. The plant has thus far produced no progeny. Because of the similarity of the F, hybrid to the C. major parent, it seems likely that natural hybrids between C. tripteris and C. major (n = 39), if they occur, would probably often pass undetected. C. tripteris « C. verticillata (n = 13). This hybridization produced some very interesting F, hybrids (Fig. 23). The F, hybrids mimicked exactly the type material of C. xdelphinifolia! See that species for more details. Eight F, hybrids were grown; five of these averaged 43% pollen stainability. Two of the F, analyzed meiotically were n = 13 at diakinesis and exhibited 13:13 distribution at anaphase I. Despite the moderate pollen stainability and apparently normal chromosome pairing in meiosis, attempts to produce an F,, and a BC, generation have thus far failed. 12. C. PALMATA Nutt., Gen. 2:180. 1818. Calliopsis palmata Spreng., Syst. 3:611. 1826. Coreopsis pauciflora Lehm.; Schlecht., Linnaea 10(Litt.-Ber.):76. 1836. (cited in Sherff, 1955). Coreopsis praecox Fresen.; Schlecht., Linnaea 13(Litt.-Ber.):93. 1839. (cited in Sherff, 1955). TYPE: None cited. Habitat given as ‘‘on the open plains of the Michigan Territory, Illinois, and Lower Louisiana.’’ I am here designating the follow- ing neotype: United States: Arkansas: Benton Co., mesic grass prairie about 1 mi. NW of Rogers, H. H. & G. S. Iltis 5549, 19 Jun 1955 (NEOTYPE MINN!; ISONEOTYPE UARK!). 167 ranean Se eae Fig. 23. Median leaves of C. tripteris (T), C. verticillata (V), and their artificial F, hybrid (F,). The leaves of the artificial F, hybrid look exactly like the leaves of the type material for C. x delphinifolia. Blooming: Jun-Jul. Habitat: Open woods, rocky ridges, prairies, railroad right-of-way. Chromosome number 2n = 26. This species (Fig. 24; Fig. 25) is very distinct in its palmately-lobed leaves. It grew poorly in the greenhouse and consequently was not used in as many attempted hybridizations as I would have liked. All hybridizations attempted with C. palmata as one of the parents failed, including C. palmata with C. major f. major, C. major f. oemleri (both n = 13 and n = 339), verticillata (n = 13), C. pulchra, C. tripteris, C. pubescens var. pubescens, C. lanceolata. 168 é yan Fig. 24. Coreopsis palmata; achene at upper center. 169 I quis zh ko, lo OF a t 08 lE es FETT 4 Fig. 25. Distribution of C. palmata., 170 13a. C. MAJOR Walt., Fl. Car. 214. 1788, forma MAJOR Coreopsis senifolia Michx., Fl. Bor.-Am, 2:188, 1803. TYPE: None cited. Sherff (1955) stated that ‘‘the remnants of Walter’s collection . . . are in London’’ at the British Museum. Apparently no type material of C. major f. major is preserved there, however, since Dr. R. Ross of that institution (pers. comm.) has indicated that none of the ma- terial in the Walter folio can be referred to C. major. I am here designating the following neotype: United States: Mississippi: Jackson Co., Ocean Springs, S. M. Tracy 4360, 10 Jun 1898 (NEOTYPE NY!;, ISONEOTYPE F!). Blooming: late May-Aug(Sep). Habitat: Open woods, rocky slopes, sandy roadsides. Chromosome number: n = 13. This form (Fig. 27, open stars) has previously been recognized as var. major. It commonly occurs intermixed in the same population with the nearly glabrous f. oemleri (n = 18 race). The two forms are completely interfertile, based on the results of artificial hybridizations mentioned below. Attempted hybridizations of C. major f. major with C. grandiflora var. grandiflora, C. grandiflora var. harveyana ‘“‘heterolepis”’ phase, C. inter- media, C. gladiata, C. linifolia (n = 18 race), C. xdelphinifolia, C. verti- cillata (n = 13 race), C. palmata, and C. tripteris all failed. The only successful cross completed was: C. major f. major x C. major f. oemleri (n = 13 race). Twelve F, hybrids were obtained. Four of these had sparsely pubescent leaves and essentially glabrous outer phyllaries and would have keyed to C. major f. oemleri; eight had sparsely to medium- pubescent leaves and rather densely pubescent outer phyllaries and would have keyed to f. major. Four of the F, hybrids averaged 71% pollen stain- ability. The cross has thus far not been carried to the F,. The results sug- gest that f. major and f. oemleri (n = 18) differ by a single probably incompletely dominant gene. 13b. C. MAJOR Walt. forma OEMLERI (EIl.) Sherff, Bot. Leafl. 6:4. 1952. ee senifolia Michx. var. rigida Nutt., Gen, 2:180. 1818. Coreopsis oemleri Ell., Bot. 8.C. & Ga, 2:435. 1828. Coreopsis wrayi Nutt., J. Acad. Nat. Sci. Philadelphia 7:76. 1834. Coreopsis stellata Nutt., J. Acad. Nat. Sci. Philadelphia 7:76, 1834. Coreopsis senifolia sensu Hook., Bot. Mag. pl. 3484. 1836 (not C. senifolia ichx 1803) Coreopsis rigida Nutt., Trans. Amer. Philos. Soc, II. 7:358. 1841. Coreopsis delphinifolia Lam. var. rigida T. & G., Fl. N. Amer. 2:342. 1842. Coreopsis senifolia Michx. var. stellata T. & G., Fl. N. Amer, 2:342. 1842. Coreopsis major Walt. var. oemleri Britton ex Small & Vail, Mem. Torrey Bot. Club 4:131. 1894. Coreopsis major Walt. var. rigida (Nutt.) F. Boynton in Small, Fl. SE. 171 U.S. 1276 & 1340. 1903. Coreopsis major Walt. var. stel 10:68 Coreopsis major Walt. var. stellata (Nutt.) B. L. Robinson forma oemleri (Ell.) Sherff, Bot. Leafl. 6:4. 1952. The name oemleri must be taken up as the forma designation (as con- trasted to stellata) since it has priority in the forma category. TYPE: United States: South Carolina: Richland Co., near the junction of the Broad and Saluda rivers, vicinity of Columbia, Oemler s.n. While Sherff (1936, 1955) mentioned the collector and type locality, he did not indicate the herbarium in which the type resides. I have thus far not located the type specimen. ~~ ata (Nutt.) B. L. Robinson, Rhodora Blooming: mid May-Aug(Sep). Habitat: Open woods, rocky slopes, sandy roadsides, Chromosome number: n = 13, 39 (two races). A great deal of variation exists in this form (Fig. 26; Fig. 27, dots), as indicated in part by the long list of synonymy. The leaves vary from entire (formerly recognized as var. stellata f. oemleri) to palmately divided into three leaflets that are flexible and spreading (formerly recognized as var. stellata f. stellata) or somewhat rigid and rather erect (formerly recognized as var. rigida). In addition, the leaflets vary from narrowly to broadly elliptical and the middle leaflet is sometimes three-lobed (Fig. 27, white star in black circle). Individuals with three-lobed middle leaflets are con- centrated in eastern Georgia and southern South Carolina, indicating prob- able introgression of the hexaploid phase of f. oemleri with C. x delphini- folia in this area. I am including individuals with narrowly linear leaflets (formerly recognized as var. linearis) in C. x delphinifolia. To complicate matters, there are two known chromosomal races in the form: n = 13 and n = 39. A very preliminary chromosome number survey suggests that populations west of the Allegheny Mountains (Tenn., W. Va., Ky., Miss., Ala., western bootheel of Fla., and western Ga.) are n = 13, while populations east and south of the Allegheny Mountains (Va., N.C., S.C., and northeastern Ga.) are n = 39. Many more chromosome counts would be necessary to establish these ranges for the chromosome races. The n = 13 race has pollen diameter of about 24-28(-30) Te the n = race has pollen diameter of about (28-)30-33 4. The two are so similar morphologically that they cannot be distinguished with certainty, without chromoscme counts. The n = 18 race is interfertile with f. major (see above) and apparently differs from it in a single gene. The n = 39 race is probably interfertile with and does morphologically grade into C. x delphinifolia. This entire complex has been called the “‘C. major complex” (Gilbert, 1940), with good justification. It evidently represents a polyploid pillar complex where three clearly distinct diploid species (C. major, C. tripteris, C. verticillata) have hybridized in a complex fashion, perhaps at several polyploid levels (see discussion under C. x delphinifolia) and have 172 reer: Eee Stee ESS ES eeneees 7 panies, eeeeas a Fig. 26. Coreopsis major f. oemleri; achene at upper center. 173 rae mm, ML fy is A ae; ee) pe aS aaa | ay ie : Z f i he oe Ae \ Se eae, a nw, es = ss ‘ P —y * mf oA SY - a : Ww Voa aaa oS Ore “tho ee ee eS 1 f 5 ee dots = C. major; = C. major 4, stars major. Distribution of C. major f. oemleri, with all leaflets entire; Fig. 27. white stars in black circles C. major f. oemleri, with the middle leaflets of some leaves 2- or 3-lobed. ? 174 also given rise to autopolyploid races (tetraploid, hexaploid), some of which are interfertile with the complex polyploid hybrid derivatives. This results in a mixing of the characteristics of the diploid species at the polyploid level such that a complex series of intergrading polyploid forms tend to morphologically interconnect the otherwise distinct diploids. Figure 28, drawn from field-collected material, indicates some of the intergradation exhibited by the C. major complex. Figure 29 indicates the intergradation that occurs in the complex in leaflet length and width, based on 128 speci- mens sampled from NCU. I have indicated above that the ‘‘varieties’? major and stellata (n — 13) are interfertile and should be reduced to formae. Form oemleri was based on material with all or nearly all leaves entire. But this entire-leaved material is rather clearly merely a growth phase of the usual three-leaflet state. I transplanted entire-leaved individuals into the greenhouse for use in hybridizations, and they produced three-leaflet leaves in the second year of growth. Form oemleri should hence be expanded to include both the entire-leaved and three-leaflet leaved phases. Variety rigida is represented by the n = 39 race, but cannot be adequately distinguished on the basis of rigidity of leaves nor the tendency for the leaflets to be held erect; there is complete intergradation with the typical phase in these characters, so that “var. rigida’ should be merged with f. oemleri. This presents a taxonomic dilemma: one could segregate these species on the basis of chromosome number and breeding behavior (in which case one would place the nm = 18 and n = 39 races of f. oemleri in separate species, even though they are not distinguishable morphologically) or one could segregate them on the basis of morphology (in which case one would place the probably interfertile n = 39 race of oemleri and C. x delphinifolia in separate species). In the first case one is at odds with the morphological species concept and in the latter case one is at odds with the biological species concept. I have chosen the latter course, largely in recognition of the fact that a taxonomic scheme must have utility to the non-specialist. In attempted hybridizations with the diploid (n = 138) race, C. major f. oemleri x C. gladiata and C. palmata failed. Successful hybridizations included: C. major f. oemleri (n = 138) & C. major f. major. This hybridization is ’ mentioned under forma major. C. major f. oemleri (n = 13) x C. pulchra. Five F, hybrids of this com- bination were obtained. They resembled ordinary f. oemleri, but had disk Fig. 28. Intergradation in leaf morphology in the C. major complex, drawn from field-collected material. A = C. major f. oemleri; major f. major; C = C. major f. oemleri, entire-leaved phase; D-E = material approaching C. x delphinifolia; F-J = various phases of C. x : “var linearis” 0: . : as : ; . renee 0 1.0 2.0 3.0 40 50 6.0 10 8.0 90 10.0 11.0 12.0 MIDDLE LEAFLET LENGTH (cm.) Fig. 29. Intergradation in leaflet length with width in the C. major complex. The ‘varieties’ major, stellata, rigida, and linearis are indicated on the basis of leaflet widths given by Sherff (1955). Open circles = leaves sparsely pubescent to glabrous; closed circles = leaves medium to densely pubescent. flowers apically reddish-brown. Three of these averaged 52% pollen stain- ability. One F, plant at diakinesis of pollen meiosis was n = 13. The F, plants failed to set fruit. In attempted hybridization of the hexaploid (n = 39) race, C. major f. oemleri « C. intermedia, C. latifolia, C. grandiflora var. sawzicola, C. grandiflora var. harveyana “‘heterolepis’? phase, C. pubescens var. pubes- cens, C. nuecensoides, C. basalis, C. wrightii, C. lanceolata, C. rosea, C. gladiata, C. falcata, C. leavenworthii, C. palmata, and C. tripteris all failed. The hexaploid race did hybridize with C. x delphinifolia; see that species for details. ’ 14. C. * DELPHINIFOLIA Lam., Encyc. 2:108. 1786. Ceratocephalus delphiniifolius Vaill. ex Ehret, Pl. et Pap. Rar. Depict. pl. 9, I. Coreopsis verticillata of Curtis, Bot. Mag. pl. 156. 1791. Coreopsis verticillata L. var. linearis Michx., Fl. Bor.-Am. 2:139. 1803. Coreopsis discolor Link., enum. Horti Berol. 2:353. 1822. Coreopsis major Walt. var. linearis Small, Bull, Torrey Bot. Club 22:48. 177 1895. Type: United States: Georgia: DeKalb Co., Little Stone Mt., J. Kk. Small s.n., 7 Jul 1893 (HOLOTYPE NY!; ISOTYPE NY!), Coreopsis delphiniijolia Lam. ex Ind. Lond. 2:294. 1930 (orthographic variant) Coreopsis delphinifolia Lam. var. chlooidea Sherff, Bot. Gaz. 94:596. 1933. Type: United States: Alabama: in pine woods, mountains of Alabama, S. B. Buckley s.n. (HOLOTYPE GH!, specimen on right side of sheet). Coreopsis delphinifolia Lam. var. linearis Sherff, Brittonia 6:341, 1948. Coreopsis delphinifolia Lam. var. delphinifolia forma concolor Sherff, Bot. Leafl. 6:5. 1952. Type: United States: Georgia: Tattnall Co., moist swale in sandhills 3 mi NW of Reidsville, A. Cronquist 5598, 3 Aug 1948 (HOLOTYPE GA!; TOPOTYPE, A. Cronquist 5345 GA!). TYPE: Founded on the illustration of Ceratocephalus delphiniifolius Vaill. by Ehret (loc. cit.) and supported by material cultivated in the King’s Garden at Paris (ISOTYPE P!). Blooming: late May-Jul(Sep). Habitat: Open rocky woods, sand_ hills, savannas, pine barrens, near sphagnum swamps. Chromosome number: n — 26, 39, ca. 52. As mentioned under C. tripteris, the type material of this species matches exactly the morphology of the C. tripteris ~« C. verticillata (n = 13) arti- ficial F, hybrid. The species occurs in South Carolina and Georgia (Fig. 30), with ‘typical’? material (leaflets dissected) being concentrated in eastern Georgia and southern South Carolina. I was puzzled, upon examining the type material of C. x delphinifolia, that none of the field-collected material I had observed matched the morph- ology of the type. The field-collected material commonly has less dissected leaves and wider leaf segments than the type material, and the terminal portions of the leaf segments tend to be acuminate rather than blunt as in the type material. After having made the C. tripteris « C. verticillata (n = 13) hybrid, and in view of the chromosome numbers and considerable morphological diversity of the field populations, I came to realize that the species was founded on a cultivated F, hybrid but that the field populations represent a complex of intergrading polyploids at the tetraploid, hexaploid and octoploid levels, derived from earlier hybridization between C. tripteris, C. verticillata, and (probably) C. major. The polyploid levels are charac- terized by narrow leaflets, commonly with the median leaflet at least deeply three-lobed and the lateral leaflets at least deeply two-lobed, producing a dissected leaf with the segments commonly 2-5 mm wide (Fig. 28, F-J). Unfortunately (for purposes of clarity in taxonomy) the species apparently is interfertile with and does intergrade somewhat with the hexaploid race of C. major f. oemleri. It seems proper to recognize the complex of forms with narrow leaflets that are commonly dissected as C. x delphinifolia. I do not believe it is useful to segregate var. linearis and var. chlooidea, both based on very narrow leaflets. Leaflet width is variable in the complex, o- SRA -? ny : A ee ¢ ex Ls 4 ee . | 3} 2 <)> Sfs Wh Ey 7 2 aR: a a oa v ( [ | Peele | . y F Be f ] \ G aes BoM Am ; ae . (re " a ye wale t Fig. 30. Distribution of C. x delphinifolia, white star in black circle = leaf- lets divided into several segments; black star = middle leaflets 2- or 3-lobed, lateral leaflets entire; dot = all leaflets entire. within populations, and to some degree within individuals. One specimen of C. x delphinifolia that I transplanted to the greenhouse and trimmed back (to increase the probability of surviving the transplantation) had produced leaflets 3-4 mm wide, the median one apically three-lobed in the field; leaflets on the regrowth in the greenhouse were all entire and as much as 9 mm wide! Form concolor is likewise of doubtful utility. The disk flowers of C. verticillata and C, major in this section are at first lemon-yellow (or green- ish-yellow) and become purplish-brown with age after anthesis. Those of C. tripteris darken similarly, or are purplish from the beginning. The disk flowers of C. xdelphinifolia are commonly (greenish-) lIemon-yellow to slightly tinged purplish-brown and darken irregularly with age. I can see little utility in recognizing a few individuals as a separate form because they happen to have entirely yellow disk flowers in the beginning. Attempted hybridizations of C. x delphinifolia (n = 39) with C. verticillata (n = 18), C. tripteris, C. major f. major, and C. duriculata all failed. One hybridization with C. major f. oemleri (n = 39) was successful. The C, major parent had entire leaflets ca. 9 mm wide; the C. xdelphinifolia parent had the middle leaflet three-lobed and the lateral leaflets two-lobed or entire, 179 with the leaf segments ca. 3 mm wide. The F, hybrids had middle leaflets three-lobed and lateral leaflets two-lobed or entire, with leaf segments 3 to 8.5 mm wide. Of nine hybrids, seven would have keyed to C. x delphini- folia and two would have keyed to C. major f. oemleri. The hybrids were vegetatively vigorous and had high pollen stainability (the exact percentage was not measured). Another successful hybridization with C. x delphinifolia was an intraspecific cross involving individuals that both had linear, entire leaflets (leaflets not dissected). Nineteen F, progeny of this cross all had narrow, entire leaflets. The octoploid material (n = 52) of C. x. Icones et descriptiones plantarum cultarum et colendarum addita colendi ratione. Vo ROBINSON, B. L. 1908, _ States. Rhodora 10: 64-68. SCHKUHR, C. 1803. Botanisches Handbuch der mehresten Theils. Vol. 3. MCHRANK, FP. L817. Anacis, cine neue Pflanzengattung. Denkschr. Kénigl. Akad. Wiss. Miinchen 5 (Math. Naturw. 54 ELTON S. 1968 hes notes on the vascular plants of the northeastern United SHELTON, L. S. 8. os flat species of Corcopsis on a granite monadnock. Bull. Vorrey Bot. Club 95: SHEREF, BE. E j dis tee 1929. cons or ae ise pote worthy Compositae. HI. Bot. Gaz. 88: 285-309. . 1930. New or otherwise noteworthy Compositae. IV. Bot. a 89: 362-373. 1933. New or otherwise noteworthy Compositae. VIII. Bot. Gaz. 94: 589-597, 1936, Revision of the genus Coreopsis. Field Mus. Nat. — Bot. Ser. 11: 277-475. . 1944. New or otherwise noteworthy American and Hawaiian Coreopsideas. Amer. J. Bot. 31: 277-281. 19 A name for the “alpha” variety or forma of miscellancous dicotyledonous plants. Ba uant 6: 332-342 1952 in) . Notes on Biles L. and Coreopsis L. (fam. Compositae) in the United States. Bot: se 6: 2- 19 gies pp. 4-40 i E. EL Sherff & FE. J. Alexander, Compositac- Helianthene Coreopiina North Amer. Flora II. Pe. 2. 962. New Coreopsidinae from Mexico and the United States. Brittonia 14: 172- 73. SMALL, J. - 1895, Studies in the botany of the southeastern United States. Bull. Bot. C lub 22: 43-48. : 1898. Studies in the botany of the southeastern United States. XIII. Bull. Vorrey Bot. Club 25: 134-151. Torrey a Flora of the southeastern United States. Pub. privately. . 1933. Manual os the southeastern flora. New York. SMALL, J. K., and A. M. VAIL. 1894. Report of the cae cane of southwestern Virginia during the season of 1892. Mem. Torrey Club 3-201. SMITH, FE. B. 1968a. In IOPB ea number reports. XPEX a xon 17: 573-577 1968b. ray see lex var. duncanii Sherf (Camposivne) In eon Proc. Arkansas Acad. 22 2-63 1968c. Stpermuiersy Gheoiihontnns in Haplopappus validus (Rydb.) Cory. Evolution 22: 748-750. 1969a. . IOPB chromosome number reports. XXII. Taxon 18: 433 2. 1969b. In TOPB chromosome number reports. XXIV. Taxon 18: 683-684. 1971. In IOPB chromosome number reports. XXXIV. Taxon 20: 785-797. 1972. Lobing of the disk flower corollas in North American Coreopsis (Com- positae). “Brittonia 24: 87-89 n IOPB Shieceeroni number reports. XXXIX. Taxon 22: 115-118. 1973b. A biosystematic study of Coreopsis saxicola (Compositae). Brittonia 25: 200- 208. 973¢. In IOPB chromosome number reports. XLI. Taxon 22: . 1974a. moro nuecensis (Compositae) and a related new species from south- ern ‘Texas. Brittonia 26: 161-171. +b. Wasi in ligule spotting in Coreopsis basalis. Proc. Arkansas Acad. 1975. The Se numbers of North American Coreopsis, with phyletic inkerphetations, Bot. Ga 136: 78-86. SMITH, E. B., and H. _ PARKER. 1971. A’ biosystematic — of Coreopsis tinctforia and ©. cardaminefolia | (Compositac). Brittonia 23: 161-170 215 SPRENGEL, K. P. J. 1826. Caroli Linnaei systema vegetabilium. Vol. 3. 1 varieties of Indiana plants. Rhodes 32: 32-34. ene Iowa State eee Press, Ames. rum rariorun a _ rum, 1859. ec eee pronouncing gazetteer of the world. Lippincott & Co., Philadelphia. , J. and A. GRAY. 1842. Flora of North America. TER, T. 1788. Flora caroliniana, secundum systema veuceabiGum, VARIETIES OF POMOEA TRICHOCARPA (CONVOLVULACEAE) DANIEL F. AUSTIN Department of Biological Sciences Florida Atlantic University Boca Raton, Fla. 33432 In the past ninety-five years several varieties in the Ipomoea trichocarpa alliance have been named. The first of these was proposed by Gray (1878) under the species I. trifida. Both Shinners (1953) and O’Donell (1960) have indicated the distinctness of I. trichocarpa and I. trifida; I will discuss their distinctness further in another paper (Austin, in preparation). During the same year that Shinners transferred Gray’s variety to I. trichocarpa, O’Donell (1953) described a new variety of this species from South America, Little has been written about the South American taxon since (O’Donell, 1959). During my studies of the genus Ipomoea (Abel & Austin, 1973; Austin, 1973, 1974, in preparation) I have examined specimens of all these named varieties; a resumé now seems in order. The type of I. trichocarpa originated in the Carolinas of the eastern United States (Dillenius, 1732), Linnaeus used the Dillenius specimen as the basis for his Convolvulus carolinus (Linnaeus, 1753: 154; see also Schmidt, 1965), the senior synonym for Ipomoea trichocarpa. As Gray (1878) and Shinners (1953; Correll & Johnston, 1970) have pointed out, there are two distinctive morphological variations of I. trichocarpa in Texas. The eastern morph is the tautonymic variety, while the western morph is var. torreyana. Shinners (1953) proposed that, while these two intergrade at their point of contact, they ought to be recognized as distinct taxonomic units. O’Donell (1953) questioned the distinctness of var. torreyana from var. trichocarpa. His major basis for this was the occurrence of glabrous forms known to him from Florida (Shinners, 1953). I have seen and collected glabrous forms in Florida; my student W. E. Abel has studied glabrous forms in South Carolina; and I have examined herbarium material of the glabrous forms from Louisiana. Thus far every glabrous “I. trichocarpa”’ we have seen from Florida and South Carolina has been a hybrid between I. trichocarpa and I. lacunosa (Abel & Austin, 1973). Glabrous forms are rare in Louisiana, but those seen usually occur with integrades of var. torreyana and var. trichocarpa. Hybrids have also been found between I. lacunosa and both I. trichocarpa varieties (Abel, in preparation). Shinners neglected to point out as supporting evidence that the two U.S. varieties occupy different ecological habitats in Texas. It is only where the Plains and Prairies meet the Deciduous Forest that the morphs intergrade. SIDA 6(3): 216—220. 1976. 217 In fact, var. torreyana is a population almost exclusively contained in the Plains and Prairies provinces of Texas (Fig. 1). Some Texas plants of this variety are known from east of the major line of integrades, but Shinners (Correll & Johnston, 1970) suggested that they were introduced. The var. trichocarpa is mostly contained within the boundaries of the Deciduous Forest vegetation. Again, those few collections of var. trichocarpa from west of this floristic province are most likely introductions by man (Fig. 1). Ipomoea trichocarpa is easily dispersed outside its home range in the United States as witnessed by casual plants from northwestern Mexico (Matuda, 1963), Cuba (Leon & Alain, 1957), Jamaica (Adams, 1970), and Guatemala, Honduras, and Colombia (Austin, in preparation). Shinners (1953) was of the opinion that the presence of glabrous forms in Florida represented a natural bicentric distribution of var. torreyana as is known in some other families. This hypothesis cannot be substantiated with the present data, nor can it be completely negated. It has been our observation that the glabrous forms outside central Texas occur in some association with a man-made disturbance, e.g. motel trash dumps, filling stations, roadside ditches. Our data suggest that var. torreyana did not exist east of Texas until man became the major dispersal agent for the species. The South American var. australis is known only from a phytogeographic association called the ‘‘Chaco.’’ The earliest collections were apparently made in the 1880’s; the ‘‘Chaco”’ has been an area of difficult access, and few botanists visited it until fairly recently. Morphologically this popula- tion is similar to var. trichocarpa, and its origin appears to have been an introduction and naturalization of that population. TAXONOMY IPOMOEA TRICHOCARPA EIll., Bot. S.C. & Ga. 1: 258. 1817. Type: based on Convolvulus folio hederaceo, etc. Dillenius collection (lectotype OXF). Convolvulus carolinus L., Sp. Pl. 154. 1753. Based on the Dillenius specimen at Oxford (Schmidt, 1965). Ipomoea carolina (L.) Pursh, Fl. Amer. Sept. 1: 145. 1814, not L. (1753) nom. illegit. I. commutata Roem. & Schult., Syst. Veg. 4: 228. 1819. New name for C. carolinus L. I. caroliniana Pursh in Small, Fl. Southeast. U.S. 963. 1903. New name for I. carolina (L.) Pursh. According to O’Donell (1953) the Dillenius specimen was at Oxford in the 1950’s. In 1974 the Curator of the Herbaria, Dr. F. White, was unable to locate the type specimen. The format of Linnaeus’ discussion of Convolvulus carolinus (1753: 154) indicates that he had the Dillenius specimen when the description was prepared, thus the plate should not be chosen as the lecto- Fig. 1. Distribution of varieties of Ipomoea trichocarpa in Texas. A. variety torreyana (Gray) Shinners; B. variety trichocarpa; C. intergrades. Symbols: open circle = var. torreyana; closed circle = intergrades; star = var. trichocarpa. Note that the major focus of intergrades follows a line about parallel with the border of the Deciduous Forest and the Plains and Prairie provinces. 219 type. The Dillenius collection was effectively chosen as lectotype by Schmidt (1965: 82). he type specimen and plate are not typical of var. trichocarpa and O’Donell (1953) suggested that the type could represent the same popula- tion as I. lacunosa f. purpurata Fernald. The type specimen of this form (Fernald & Long 7580 GH) I believe to be a backcross hybrid segregate of I. lacunosa and I. trichocarpa. The type of I. trichocarpa is also probably of hybrid origin. Due to introgressive hybridization between I. lacunosa and I. trichocarpa (Abel & Austin, 1973; Abel, in preparation), pure populations of I. trichocarpa appear to be rare if not non-existant. KEY TO VARIETIES a. Sepals glabrous. . oe ee ee). 62, Var, torreyana aa. Sepals pubescent or eT eae aie, b. Pedicels muricate; leaves 3-5-lobed, mostly 3-lobed with the base of the median lobe contracted . . 1. var. trichocarpa bb. Pedicels smooth, rarely somewhat cate: eaves entire to 3-lobed, rarely 5-lobed, the middle lobe not contracted at the base. 3. var. australis 1. I. trichocarpa var. trichocarpa A native of the southeastern United States, this variety is found in the Coastal Plain Province of the Eastern Deciduous Forest. Sepals are pubes- cent on the backs or at least ciliate. Those with only ciliate margins are mostly intergrades with var. torreyana. This variety ranges from eastern Texas to southern North Carolina, south into southern Florida. 2. I. trichocarpa var. torreyana (Gray) Shinners, Field & Lab. 21: 164, 1953. Basionym: I. trifida var. torreyana Gray, Syn. Fl. N. Amer. 2(1): 212. 18 Lectotype: Rio Grande, western Texas, anno 1848, Wright s.n. (GH). Syntypes: Sept. 1849, Lindheimer s.n. (GH, MO, NY); Bigelow, (pre- sumably GH, not found); Schott s.n. (NY). Synonym: I. trifida var. berlandiert Gray, Syn. Fl. N. Amer. 2(1): 212. 1878. Type: Bexar, Texas, Berlandier 546 (1931) (holotype GH; isotypes MO, NY) When he published var. berlandieri Gray suggested that it might be only a depauperate form. The type is but one of the apparently ecophenotypic variations of the species. The glabrous var. torreyana is normally found within the Plains and Prairies Provinces of Texas and adjacent Mexico (Tamulipas). Some of the locations in Mexico (Chihuahua, Mazatlan, Guaymas) are introductions. There are a few plants known from east of Texas in the United States. 220 3. I. trichocarpa var. australis O’Donell, Bol. Soc. Argentina Bot. 4: 260. 1953 Type: Lillo 12909 (holotype LIL, not seen; isotype NY). Additional specimens seen: Argentina: Peirano 9497 (NY); Schreiter 946 (NY). Paraguay: Morong 253 (NY). O’Donell’s variety is known only from the Gran Chaco region of Bolivia, Paraguay, and Argentina, From the specimens that I have seen, this variety also has larger trichomes on the calyx, and sepals which are more herbaceous than var. trichocarpa. ACKNOWLEDGMENTS This study was partly supported by a Grant from the Division of Spon- sored Research, Florida Atlantic University. Curators and staff members at A, FSU, GH, MO, NY, UNC, and US loaned material and helped in many ways during my visits. I would like to thank Dr. F. White (University of Oxford) for searching the Dillenius collection. REFERENCES ABEL, W. E. preparation. Introgressive hybridization between Ipomoea lacunosa and | pomoea ia M.S. thesis: Florida Atlantic University and D. F. AUSTIN. 1973. Natural hybridization in Tiemuoey (Cc Ivul ) Amer. J. Bot. 60: 33-34. (abstract). ADAMS, C. D. 1970. Convoly aie In Flowering plants of Jamaica. pp. 601-614. Univ. W. Indies Press, Mona. D. F AUSTIN, D. FP. 1973. The sweet potato allies: a taxonomic review. Quart. J. Fla. Acad. Sci. 36 eee 1): 7. (abstract). 974. Convolvulaceae I Flora of Panama, Ann. Missouri Bot. Gard. (in press). : reparation. ao es d butte complex. CORRELL, D, - and M. JOHN 970. Convolvulaceae I Manual of the vas- cular plants of Texas. pp. ree 7. xas Research i Gk: Renner. DILLENIUS, J. J. 1732. Hortus pieqeags seu plantarum rariorum quas tn horto suo Elthami in Cantio coluit, ete. p. 100, 4. 84, f. 98. London. GRAY, A. 1878. Convolvulaceae In Synoptic flora of North America 2(1): 207-224 LEON, BRO. and BRO. ALAIN. 1957. Convolvulaceae 7” Flora of Cuba 4: 218-248. LINNAEUS, C. 1753. Species Plantarum. p. 154. Vindobonae MATUDA, E. 1963. El genero Ipomoea en i (1). An. Tasks Biol. Mex. 34: 85-145 O’DONELL, C. A. 1953. na nueva Convolvulacea Sudamericana. Bol. Soc. Argentina 63, 1959. Convolvulaceas Argentinas. Lilloa 29: 88-348. P I-V. . 1960. Notas sobre Convolvulaceas Americanas. Lillea 30: 39-69. ae ee H. ahs Der “Hortus Elthamensis” aus der ees Carl von Linnes. Fedde Q: 69-10 ae RSL; - 1953. Botanical notes. Field & Lab. 21: 164-165. TWO SPECIES OF THE ‘“FILAGO GERMANICA” GROUP (COMPOSITAE-INULEAE) IN THE UNITED STATES GERHARD WAGENITZ Systematisch—Geobotanisches Institut Universitdt Gottingen Untere Karspiile 2, D 3400 Gottingen, Federal Republic of Germany In the past there has been considerable discrepancy regarding specics- delimitation in the genus Filago especially as concerns the group centering around ‘‘Filago germanica”’ (a name which for nomenclatural reasons can- not be maintained and has to be replaced by Filago vulgaris Lam.— Wagenitz, 1965). Some authors have used the name ‘“‘Filago germanica”’ in a collective sense, but our monographic studies have confirmed the view of those authors, who have distinguished several species in this complex (Wagenitz, 1965, 1969). These species differ by inconspicuous but very con- stant characters. In North America “‘Filago germanica’’ has been known for a long time from the eastern part of the United States and has recently been reported also from Oregon and California. A closer inspection of these plants is necessary. 1. Eastern North America Clayton seems to have been the first American botanist who collected a Filago of this group in eastern North America, in Virginia. This find was published by Gronovius (1739). Although the name used (‘‘Gnaphalium mini- mum humile, Herba Impia dictum’’) may not be unequivocal, the fact is confirmed by a specimen from Virginia from the herbarium of Gronovius (in herb. Jacquin, Vienna). Later on the species is mentioned under the name of ‘‘Gnaphalium germanicum” or ‘‘Filago germanica”’ in most of the classical floras of this area (e.g. Pursh 1814, Barton 1818, Darlington 1837, Torrey 1843, Torrey & Gray 1843). Barton considered it to be a native species, but the opinion prevailed that it was introduced from Europe. This appears to be very probable, but as the spread of the plant apparently took place before the beginning of a floristic study of the continent no direct evidence is available. Uechtritz (1871: 190) seems to have been the first author who critically examined the material of ‘“‘Filago germanica”’ from North America after it had been established that there is more than one species of this group in Central Europe. He concluded that the American plants belong to ‘‘Filago canescens Jord.” (F. germanica s.str. of most later authors), now correctly called F. vulgaris Lam. This conclusion has SIDA 6 (3) > 221-223... 1976. 222 been fully confirmed by our studies. During the revision of material from numerous herbaria of Europe and several from the United States (GH, NY, POM, UC), 48 collections all belonging to Filago vulgaris Lam. have been seen from the following states: New York, New Jersey, Pennsylvania, Delaware, District of Columbia, Maryland, Virginia, West Virginia, North Carolina, South Carolina. According to Fernald (1950), the species is also known from Ohio and Georgia. It is remarkable that only 11 of these collec- tions are dated in this century, the latest from 1947. This may be due either to a diminishing activity of floristic botany in the last decades or to a decrease of the species on account of more intensive cultivation as noted in TIcurope. 2. Western North America The first record from this area was published by J.Th. Howell (1939) and referred to a plant collected by Mr. Lewis S. Rose in Roseburg, Douglas Co., Oregon. I have examined a fragment of this plant sent to me by the courtesy of Mr. Howell and can confirm that it is F. vulgaris Lam. This may have been a casual introduction from the Eastern states. Of more interest is the material collected in California since 1935 by Tracy and Lennon (see the short notes of Howell, 1942 and 1973 and Munz, 1968). All three collections so far known have been studied by me and they belong to Filago pyramidata L. (syn.: I. spathulata C. Presl), a mainly Mediter- ranean plant extending into the southern parts of Central Europe and the Near East (Wagenitz, 1970). It is not surprising that the Mediterranean representative of this group is introduced (and perhaps locally naturalized) in California as the high percentage of species with this type of distribution in the California weed-flora is very well known. In fact another Mediter- ranean Filago-species (belonging to another section) has been naturalized for a long time: F. gallica L. The distinction of the two species of the “Filago germanica’’ group is not difficult especially if the capitula are studied under the binocular. The main differences are: Filago pyramidata I’, vulgaris Leaves oblong-spathulate lanceolate (broadest near apex) (broadest near middle) Phyllaries in 5 distinet rows, not in 5 distinct rows, of involucre keeled on back rounded on back Median with short straight with spreading mucro phyllaries point Flowers in 5-6 hermaphrodite, 2-3 hermaphrodite, central part few female flowers numerous filiform female of involucre flowers Number of 20-30 (40) 8-16 -apitula in cluster 223 It seems appropriate to cite in full the three collections of Filago pyrami- data L. known from California: Mendocino County: near Hopland (along highway ten miles south), dry hill- side, 600 feet, 18.VIII.1935, J. P. Tracy 14741 (JEEPS) Mendocino County: between Hopland and Cloverdale, locally naturalized along highway on dry hillsides, 600 feet, 6.VI.1988, J. P. Tracy 15848 (DS, GH, JEPS, NY, UC) Marin County: dry open hills above Kirby Cove, Marin Headlands State Park, just west of Golden Gate Bridge, 28.V. & 3.VII. 1971, Elizabeth S. Lennon s.n. (CAS) Californian field-botanists should keep a close eye on this species, which may be expected to spread farther in the state. It is not a serious weed, but an interesting addition to the flora of California. REFERENCES BARTON, W.P.C. 1818. Compendium = Philadelphicae. Philadelphia. ieee ager W. 1837. Flora cestric Ed.2. West-Chester. FERNALD, M. L. 1950. Grek antal “ee botany. Fd.8. New York. one J. F. 1739. Flora Virginica. Pars 1. Lugduni Batavorum. HOWELL, ee 1939. Gifola adie (L.) Dumort. Leafl. Western Bot. 2: 192. ; 1942. Now in California. Leatl Western Bot. 3: 168 _ : By California record for Filugo vulgaris. iiede ofio 22: 151. MUNZ, P. A. 1968. me ne to a California flora. Berkeley and Los Angeles. PURSH, F. 1814. Flora Americae Septentrionalis. London, TORREY, J. 1843. Nn a of the state of New York. Vol. 1. Albany. (= Natural History of New York. Part 2. Botany.) a A. GRAY. 1843. A flora of North America. Vol. 2 Part 3. New York aid Londor UECHTRITZ. R.v. 1871. Flora Ungarns. Oesterr. Bot. Ztschr. 21: 185-191, 233-237, 26 ; WAGENITZ, G. 1965. Mi Syste smatik und Nomenklatur einiger Arten von Filago L. emend. earn subg Filago (‘Filago germanica’’—Gruppe). Willdenowia 4: 37-59. a 969, eae und Gliederung der Gattung Oh go L. s.1. (Compositae- Taulese), Wildenow §: 395-444 . Uber die Weibrcans einiger Filago-Arten. Feddes Repert. 81: 107-117. DRYOPTERIS XLEEDSH AND ITS WESTERNMOST STATION: WARREN H, WAGNER, JR. & W. CARL TAYLOR Department of Botany, The University of Michigan, Ann Arbor, 48104 Department of Botany, Southern Illinois University, Carbondale, 62901 Dryopteris . ‘. 1963. Studies in ppnere ie subgenus ae ie (DC.) A. Gray. If General morphological considerations in the taxonomy of the subgenus. Watsonia 4: 294-303. Sees 1966. A monographic study of Ranunculus subgenus Batrachinm (DC.) A. Gray. Mitt. Bot. Staatssamml. Miinchen 6: 47-237. DAMBOLT, J. 1974, Bug 232-317. In G. Hegt: ee Flora von Mittel- Europa, ed. 2. Bd. HII (Teil 3, Lief. 2/3, 4/5). Carl Hanser, Miinchen. DAVIS, K. C. 1900. Hae ind a ea of Nor he pene and segregated ‘ -50 DAVIS, P. H. 1960. ee fo a on of Turkey: IV. Ranunculaceae II. Notes Roy. Bot. Gard. Edinb. 23: 103- Een ay 1965, ce m Turkey, Vol. 1, Edinburgh Univ. Press, Edinbur DREW, W. B. 1936. The North American Represcatauee es of Ranunculus § Bairaohiuie Rhodora 38: 1-47. FASSETT, N. 1942. Mass collections: Ranunculus abortivus and its close allies. Amer. Midl. Nat. 27: 512-522. FERNALD, M. L. 1899. Two plant of the crowfoot family. Lage . ig 2: 1919. The variations of Ranunculus repens. Rho 21: . 1936. Contributions from the Gray Herbarium “i Harvard University—No. CXIIL. ae Memoranda on Ranunculus. Rhodora 38: 171-17 938. aes aaa of southeastern Virginia. Site ra 40: 364-424. ; i Last surv in the flora of Tidewater, Virginia. Rhodora 41: 465-504, 529-558, 564-575 erie on Ranunculus). . 1950, Gray’s manual of botany, 8th ed. American Book Co., New York FISHE R, F. J. F. 1965. The alpine Ranunculi of New Zealand. N.Z. Dept. Sci. Industrial Res. Bull. 165. Government Printer, Wellington. J. A. ROWLEY, and C. J. MARCHANT. 1973. T western snow- as Ranunculi of North America. C. R. Soc. Biogéo; aa J. J. B., BL. M. G. JONES, C. J. MARCHANT, J. McLEISH, and D. J. a ae . The eh ie of chromosome races of Ranunculus ficaria L. in the British a oe Bot. ace 36(144): 31-47. GLEASON, H. A. and A, CRONQUIST. 1963. Manual of vascular plants of northeastern United States i adjacent Canada. D. Van Nostrand Co., Princeton, N.J. GRAY, A. 1886. A revision of the North American Ranunculi. Proc. Amer. Acad. 21: 63-378 le gee of the ni 43. GREEN, P. S. and J. L. THOMAS. 1961. The bulbiferous Ranunculus ficaria. Rhodora 6 9 GREENE, FE. L. 1900. Some new or critical Ranunculi. Pittonia 4: 142-146, HARA, H. and S. KUROSAWA. 1956. eee notes on the Ranunculus acris group in Japan. Bot. ae ae 69: 345-3 HARPER, J. L. 1957. Biological flora of the bei Isles: aia acris L., Ranunculus repens L., and Runnels Sonn e L. J. Ecol. ‘ 42. HUTTON, E. 1971. Plants previously unknown in ue en Castanea 36: 166-167. MARCHANT, C. : and C. A. BRIGHTON. 1974. ales hg ae and triploid fre- ee in a complex ponileeon of Ranunculus ficaria L. Bot, (Lond.) 38(154): RADE ‘OND, A. FE. 1968. Ranunculaceae, p. 452-468. In A. E. Radford, H. E. Ahles, and Bell, anual of the vascular flora of the Carolinas. Univ. of North Carolina Press, San Hil oe Rh , J. W. HARDIN, and R. L. WILBUR. 1967. Contributor's guide ascular Nora - the southeastern United States. Dept. of Botany, Univ. of North ar SALISBURY, E. J. 1931. On the morphology and ecology of Ranunculus parviflorus L. 578 SARUKHAN and J. L. HARPE R. 1973. Studies on plant demography: Ranune ulus repens R. bulbosus L. and R. acris L. 1. Population flux and survivorship. J. 61: 16 283 SHINNERS, L. H. 1960. oe trilobus (Ranunculaceae) in southern Louisiana: new to the United States. Southw n Nat. 5(3): . 1962. metas is frachycarbs (Ranunculaceae) in south-central Louisiana: new to _ North Fa ca. Sida 1: 104-10 SMALL, J. K. Seal of the ee flora. Publ. by the author, New York. TAMURA, M. W907. Sa ecology and phylogeny ee a Ranunculaceae. VII. Sci. Rep. Osaka Univ. 21- TUTIN, T. G. oe Ramon subgen, oe p. 223-237. In T. G. Tutin et al. (eds.), Flora Europaea, Me . Cambridge Univ. Press, Londor WEATHERBY, C. A. 192 Oe variants of Rasunculus ceanaak Rhodora 31: 163-164. TAXONOMY OF THE GENUS POLYGALA SERIES DECURRENTES (POLYGALACEAE) ROBERT R. SMITH Department of Biology, Hartwick College, Oneonta, New York, 13820 DANIEL B. WARD Department of Botany, University of Florida, Gainesville, Florida, 32611 ABSTRACT A taxonomic revision of POLYGALA series Roeitensirarne is presented, based upon evidence from morphology, chromosome numbers, habitat, _ distribution. Seven. ences one with two ete are coca All a native to the s Selmer a United States. IKach species is illustrated with line drawings Chas ng habitat and floral morphology; distribution maps indicate sites of an populations. Selected specimens are cited, repre sentative of the 2,463 examined. P GALA BALDUINII var. carter (Small) Smith & Ward is proposed as a new combination, and POLYGALA smallii Smith & Ward is proposed in place of the illegitimate P. arenicola Small. The genus Polygala L. is widely distributed in the tropical and temperate regions of the world. It is absent only in New Zealand, Polynesia, and the arctic regions of North America and Asia. Approximately 475 species of Polygala have been described; of these, 127 species occur in North America, excluding the Caribbean, with about 50 in the United States. Series De- currentes, a natural grouping of seven species of the Southeastern Coastal Plain, known by Small (1933) as the genus Pilostawis Raf., is here revised. The numerous species in the genus Polygala exhibit a morphological diversity that is accommodated only by the use of hierarchical subgeneric classification. Three ranks may be usefully recognized below the level of genus and above that of species. Chodat (1893, 1896) divided the North American species into four sections, the largest of which is his Ortho- polygala, containing the type of the genus (P. vulgaris L.) and now under Art. 22 of the International Code known as sect. Polygala. This section is characterized by the presence on the abaxial petal (keel) of a lacerate or fimbriate crest, and by sepals that persist in fruit. Within sect. Polygala Chodat recognized three subsections, two of them found in North America, and weakly distinguished by the degree to which the cells of the capsule are margined or winged. Those species with unwinged capsules or with the two cells narrowly and equally margined were placed in subsect. Apterocarpae, again the largest taxon. Within this subsection Chodat de- scribed several series, among the most natural of these being ser. Decur- ‘This paper is Florida Agricultural Experiment Station Journal Series No. 5259, SIDA 6(4): 284—310. 1976. 285 rentes, in which the sepals are decurrent, forming narrow wings along the pedicels and onto the adjacent inflorescence axis. The first species of Polygala ser. Decurrentes to bear a legitimate name was P. lutea (1753), typified by two slender plants received by Linnaeus from Peter Kalm, his correspondent and agent in the New World. Kalm undoubtedly made his collection near Swedesboro (then Raccoon), New Jersey, his home in the spring of 1749, and close to the northern limit of this most far ranging member of the series. Other species were described by Thomas Walter (1788) from the vicinity of his rice and indigo plantation on the Santee River, Berkeley County, South Carolina (P. cymosa), by Andre Michaux (1803) from a moist meadow in the Carolinas (P. nana), by Thomas Nuttall (1818) from a specimen received in correspondence from near the mouth of the St. Mary’s River, northeastern Florida (P. balduinii), by Stephen Elliott (1822) from coastal South Carolina (P. ramosa), by A. W Chapman (1887) from Tampa Bay, western peninsular Florida (P. rugelii), and by J. K. Small (1905) from pinelands near Cutler, Dade County, Florida (P. smalliz). These species were not immediately recognized to constitute a natural grouping within the genus Polygala. Nuttall revealed some awareness of relationships by arranging three of the four species treated by him in one numerical sequence, although P. lutea was placed separately. Elliott was familiar with five species which he listed consecutively but otherwise with- out separation from other members of the genus. DeCandolle (1824) treated P. lutea and P. nana apart from the other species now in the Decurrentes, and Torrey and Gray (1838), also not completely aware of the naturalness of the group, did likewise. Rafinesque, again a resident of Philadelphia after his self-imposed exile to pioneer Kentucky, but still an outcast from the scientific community, first gave nomenclatural recognition to a grouping of the above species. His genus Pilostaxis (1838) contained only P. lutea, designated as the type, and P. nana (which he saw as two species), and was superficially or even erroneously characterized by capitate inflorescences and stamens in 4’s (rather than 6-8’s). Perhaps more for reasons of hauteur by the botanical establishment than from awareness of the inadequacies of his treatment, Rafinesque’s Pilostazis was disregarded for nearly one hundred years until disinterred by Small (1933). It remained for the versatile Swiss botanist Robert Chodat, then rising through the ranks of the University of Geneva and wholly isolated from living North American Polygala, to recognize (1893) that several of the species often placed together on general appearance, shared the character- istic of sepals that were decurrent on the pedicel. He placed the species with this property in his series Decurrentes. But the equally astute Blake (1924), an American with field experience, although continuing the close placement of these species, gave the group no formal designation. Recent workers with 286 the Polygalaceae in the southeastern United States have given little attention to subgeneric divisions within Polygala; Saulmon (197la, 1971b) followed Blake, and Miller (1971) was content to note that the Decurrentes of Chodat contained ‘‘clearly related species.” Recognition of infrageneric groups must always be subjective, and at times because of uncertain relationships elsewhere in a genus, one is tempted not to name even the more distinct aggregations. Since the time of Chodat, however, there has been general agreement as to the evolutionary homo- geneity of the Polygala species with decurrent sepals. Although the features which unite the group are not of a magnitude to confirm the judgment of Rafinesque and Small that these species constitute a recognizably distinct genus, their structural similarities, supported by their common range on the Southeastern Coastal Plain and uncontradicted by the constancy of their chromosome numbers, seems justification for formal recognition at a lower level. Designation with the rank of series, as Decurrentes Chodat, is an appropriate means of inserting this group at a suitable level in the genus Polygala. A total of 2,463 specimens from 12 herbaria was examined during this study. Lending institutions were: BUS, DUKE, F, FLAS, FSU, GA, MO, NCU, NSC, NY, PH, US. Distribution maps and specimen citations omit collections beyond those necessary to give indication of range. MORPHOLOGY HABIT. Plants of the series Decurrentes are all herbaceous, either erect or ascending, glabrous, and may appear succulent, particularly in the rosette stage. Their size varies from about 2.5-3 cm, a minimum height of P. smallii, to 11 dm, a maximum height of P. cymosa. The stems are round and smooth, or ridged. Most species may be either annual or biennial. Polygala ramosa appears to be consistently annual. Polygala lutea at least occasionally may be perennial, as was concluded by Holm (1929) following a three-year study. The biennial characteristic is readily revealed by the formation of a leafy rosette one year, then a stem with flowers the second year. No species ex- hibit a rhizome. Reproduction is solely by seed. Since the flowers are small and the in- florescences often densely crowded, the number of seeds produced is high. ROOTS. The root systems of the series vary from fibrous to tap-roots. Polygala cymosa has fibrous roots, while P. nana and P. smallii develop tap- roots. The rest of the species may have either, depending on the age of the plant and perhaps the influence of environment. The root system customarily has a wintergreen-like odor, a frequent con- dition throughout the genus Polygala. This fragrance gives these plants the common name of ‘“‘candy-weed’’ or ‘‘candy-root.’’ All species have a strong or fairly strong scent except P. cymosa, which has little of this character- istic odor. 287 LEAVES. The species in this series show two major types of leaves, basal and cauline. A basal rosette is found in all species; however in P. ramosa and P. balduinii the leaves are usually withered by anthesis. Except for P. cymosa, the leaf shape is of an obovate-spatulate type which gradates up- ward to an elliptic-obovate shape. Polygala cymosa has long, broad, linear- lanceolate leaves in the basal rosette that abruptly change to widely spaced cauline leaves which are narrow and much smaller. INFLORESCENCE. The inflorescences are of two recognizable types. One is the thick capitate raceme as found in P. rugelii, P. lutea, P. nana, and P. smallii. The other is a cymose panicle with racemose branches as in P. cymosa, P. ramosa, and P. balduinii. The branches may be dense as in P. balduinit var. balduinii or very loose as often occurs in P. balduinii var. carteri. BRACTS. Each flower is subtended by one small and two minute bracts. In the species that have cymose inflorescences, the bracts are persistent. In the thick head-like racemes the bracts are deciduous, except in P. nana where they are usually deciduous but sometimes persistent. FLOWERS. Flowers of the genus Polygala are zygomorphic and perfect. The calyx consists of a whorl of five free sepals. The two lateral sepals (wings) are large and petaloid and prominently project from the sides of the floral structure. These wings are orange, yellow, or yellowish in all species except P. balduinii in which they are white. The other three sepals are definitely sepaloid, the upper one being somewhat larger than the two lower ones. The corolla is more specialized than the calyx. In this series it consists of three petals, two upper and one lower. The two upper petals are connate to the lower one but not to each other. One of the upper petals overlaps the other, forming an involute tube. The lower petal (Keel) is distinctive, con- sisting of a blade which terminates in a hood-like flap (amella) that partially encloses the reproductive organs. The outer surface of the lower petal bears a fimbriate or lobed extension (crest) that in size, shape and number of lobes is variable and is rather characteristic for each species. The corolla is deciduous upon maturation of the capsule. The androecium consists of eight stamens in all species except in P. nana where there are usually six, although occasionally seven were observed. The general structure of the stamens is the same in all species. Each filament is adnate to the keel almost to the anther, the degree of adnation varying slightly with species. The anthers are all one-celled and when mature dehisce by a single oblique apical opening. The gynoecium is characterized by a deciduous style which bifurcates in a vertical plane, with the ventral lobe curving distally upward. Both stylar lobes thus appear dorsal to the axis of the style. The distal stigma is tufted with long once-forked hairs. The proximal stigma is indument-free and is closer to the stylar axis. The ovary is two-celled and superior. Both cells are 288 functional, each containing one ovule borne on the distal portion of the septum. FRUITS. All species have two-celled, wingless, thin-walled capsules. The locules are in the plane of the stylar lobes, each dehiscing by a single long loculicidal suture. SEEDS. One seed is produced in each cell, Seeds in this series are lightly pubescent with the exception of P. cymosa, which is glabrous. A blunt rostrum may be present at the base of the seed. The size and shape of both the seed body and the aril are usually distinct for each species. CYTOLOGICAL INVESTIGATION Considering the size of the genus Polygala, there have been very few publications concerning chromosome studies. Only Lewis and Davis (1962) have published a Polygala cytology in the New World. They have also summarized the earlier work of Old World authors, Lewis and Davis were impressed with the striking diversity of chromo- some numbers encountered. They suggested a basic number of x + 17 for the section of the genus in which ser. Decurrentes occurs. They specifically reported P. lutea, P. nana, P. ramosa, and P. rugelii as having gametic counts of 34. These figures are in conflict with all counts obtained in the present study. In all species of the series Decurrentes (excluding P. smallii, which was not counted) the gametic numbers here observed were 32. In the single species where a successful somatic count was obtained (P. lutea) the chromosomes numbered 64. Voucher specimens are recorded in Table 1. All chromosomes were small and without apparent differentiation, BLE I CHROMOSOME NUMBERS OF POLYGALA SER. DECURRENTES All counts were made from Florida materials. All vouchers are deposited in FLAS. Taxon 2n n Source P. cymosa 32 Clay Co. =? mi. n. of So Head Branch State Park, Smith & Myint 70. P. ramosa 32 Union Co., 3.8 mi. s. of ae Butler, Smith 94. P. balduinii var. balduinii 32 Lee Co., 7 mi. e. of Ft. Myers, Smith 380. var. carteri 32 Dade Co., in glade adjacent to Long Pine Key, Everglades National Park, Smith & Buchanan 108. P. rugelit 32 Pinellas Co., w. of 14th St., St. Petersburg, Smith 160. P. lutea 64 32 Alachua Co., 6 mi. n.e. of Gainesville, Smith (10 Apr. 1961). P. nana 32 Alachua Co., 6 mi. n.e. of Gainesville, Smith (10 Apr. 1 961). 289 GEOGRAPHICAL DISTRIBUTION AND ECOLOGY Representatives of the series Decurrentes are distributed along the Coastal Plain from New Jersey to Florida, westward to Texas, with P. balduinia var. carteri extending to Cuba. Their usual habitat is the moist to wet pine- barrens, pine-flatwoods, pine-palmetto associations, or open grassy savannas. Polygala nana and P. smallii also inhabit drier areas, the latter being found only in the dry sandy pine-palmetto areas of South Florida. Most of the species are able to withstand considerable competition, but P. nana and P. smallii are usually found in microhabitats of open areas void of other plants. The over-all distribution of series Decurrentes indicates the southeastern United States as a point of origin. The lack of a fossil record makes im- possible any certain knowledge of the age of the group, but the similar distribution of the various species strongly suggests an origin not too far in advance of the Pleistocene. The morphological diversity of the series is much greater than one might anticipate to develop in the relatively brief period since the Pleistocene. During th Pleistocene epoch much of the area now occupied by members of the series Decurrentes was at least temporarily submersed (James, 1961). But islands and archipelagos isolated from the continental mass would clearly provide conditions under which divergent evolution and, ulti- mately, speciation could take place. Such opportunities may well have been responsible for some of the morphological diversity found in the Decurrentes. Polygala lutea, P. nana, P. ramosa, P. cymosa and P. balduinii var. balduinii are relatively uniform throughout their ranges northward up the Coastal Plain and westward along the Gulf Coast. Their distribution either was not broken by Pleistocene flooding or, if broken, did not lead to mor- phological diversity. There is no reason to speculate that they have had any recent history more complicated than survival in the continental south- eastern United States during Pleistocene flooding, followed by migration into the newly exposed Coastal Plain and Gulf Coast Polygala rugelii, limited in range to the Florida peninsula and yet adapted to the same habitats as some of the above more widely ranging species, suggests a different history. Its distribution may perhaps best be explained by the assumption that its pre-Pleistocene range was restricted to the area of the Florida peninsula and that the species survived only on the Pleisto- cene islands from which it had since spread. The two remaining taxa suggest perhaps a third mode of origin. Polygala balduinii var. carteri and P. smallii show very close relationships to balduinii var. balduinii and P. nana respectively. Their morphological dif- ferences may result from Pleistocene isolation in a manner similar to that suggested for P. rugelii, but with the more widely distributed member of each pair surviving also on the mainland and since spreading into the 290 newly exposed peninsula. However, it seems at least equally probable that these two taxa have resulted from a post-Pleistocene isolation by adaptation to the special habitats of South Florida. SYSTEMATIC TREATMENT Series Decurrentes Chod., Mém. Soc. Phys, Genéve 31 (2)? :125, 1893. Pilos s Raf., New Fl. N. Am. 4: 1838. Type species: Polygala lutea L. (selected by Rafinesque, 1838) Annuals, biennials, occasionally perennials. Root system fibrous or tap- root. Stem erect to ascending, single or several, smooth or ridged. Basal leaves spatulate (gramineous in P. cymosa), usually thick. Cauline leaves alternate, entire. Stipules absent. Inflorescence a capitate raceme or a cymose panicle. Flowers yellow, orange, or white to greenish-white, Sepals 9, decurrent on pedicel and on inflorescence axis; lateral sepals (wings) large, petaloid. Petals 3; lower petal (keel) crested, connate to adjacent upper petals. Stamens 6-8, with filaments adnate to lower petal. Anthers single-loculed, opening by apical pore. Ovary 2-loculed; stigma 2-lobed with distal stigmatic surface tufted. Seed ellipsoidal to subspherical, short- pilose (glabrous only in P. cymosa); aril present, sometimes scarious or reduced to a minute scale, or absent. Key to Taxa of Series Decurrentes A. Inflorescence a cymose panicle of dense or elongated racemes, more or es flat-topped; lower leaves spatulate and withered or nearly so when ant flowers, or gramineous and persistin B Corolla and wings of calyx bright yellow; "pedicels greater than 1.0 mm long C. Basal leaves linear-lanceolate io linear, forming a large persistent rosette; stem more than 4 . Polygala cymosa. C. Basal leaves broadly oo usually withering once plant flowers; stem less than 4 dm olygala ramosa. B. Corolla yellowish es of calyx white to greenish- white: pedicels less than 1.2 mm | Polygala balduinii. D. Praselr ee 7 dense racemes; wings definitely white: seed less than 0.6 mm long; aril on seed usually 0.2 mm long, infrequently smaller. 3a balduinit. D. Inflorescence of more or less elongated racemes; one cream to a white e; seed more than 0.6 mm long; aril on seed a minute 3b. var. A. ie a solitary dense head-like raceme; lower leaves spatulate, persisting. E. Racemes bright yellow to orange; lobes of keel less than 1.5 mm long F. - monesctnice yellow (drying yellow); lobes of keel ae m 4, jala rugelii. F. ie eee orange (drying yellow); lobes of keel Fee oF ak nally ifurcated, less than 0.7 mm lon olygala aes E. Racemes apes -yellow, ee yellow, to green; lobes of keel greater than 1.5 mm long. G. Inflorescence lemon-yellow to greenish-yellow; wings elliptic, involute 291 at apex; seed 1.6 mm long or less. 6. Polygala nana. G. Inflorescence yellow-green to aren wings oblong-lanceolate, t involute at apex; seed 1.9-2.3 mm long 7. Polygala smallii. 1. POLYGALA CYMOSA Walter, Fl. Car, 179. 1788. TYPE: UNITED STATES: SOUTH CAROLINA: Berkeley County, T. Walter (HOLOTYPE: not seen and presumably no longer extant; the bulk of the Walter herbarium is lost, the small collection in the British Museum (Blake, 1915) appearing to be but a fragmentary representa- tion given John Fraser (1789)). Fig. 1. Polygala corymbosa Michx., Fl. Bor. Am. 2:54. 1803. Polygala graminifolia Poir. in Lam., ‘Ency cl. 5:500. 1804. Polygala attenuata Nutt., Gen. 2:90. 1818. Polygala acutifolia Torr. & Gray, Fl. N. Am. 1:128. 1838. Polygala cymosa Walt. var. graminifolia (Poir.) Torr. & Gray, FI. N. Am. 1:670. 1840. Pilostaxis cymosa (Walt.) Small, Man. S. E. Fl. 744. 1933. Biennial, erect, 45-110 cm tall, usually without branching below the in- florescence, from well-developed fibrous roots. Basal leaves 3.5-14 cm long, 0.2-0.6 cm broad, linear-lanceolate to linear, thin and flat, forming a large, persistent rosette; cauline leaves abruptly smaller, gradating in size to the bracts of the inflorescence. Inflorescence a single terminal cymose panicle, with numerous (to 150) racemose branches. Bracts 1-1.5 mm long, persistent. Pedicels 1-2 mm long. Flowers vivid yellow, turning from pale yellow to dark green upon drying. Upper sepaloid sepal 1.5-1.8 mm long, ovate, sparsely ciliolate; two lower sepaloid sepals 1.1-1.2 mm long, lance-ovate; petaloid sepals (wings) 2-4 mm long, 1.4-1.6 mm broad, oval to oval-oblong, 3-nerved, the apical portion involute. Petals 2.5-2.8 mm long, connate only at base; crest with 2-3 entire or bifurcating lobes. Seed 0.7-0.9 mm long, 0.6-0.7 mm broad, glabrous, rugose, with a small rostrum; aril minute, unlobed. Anthesis: April through August. Habitat and distribution: Moist to wet acid soil of open boggy communi- ties and marshlands, often with the rosette beneath 1-10 cm water; penin- sular Florida from Lake Okeechobee, northward on the Coastal Plain to southern Delaware and westward to Mississippi. Polygala cymosa, with its gramineous leaves, glabrous seeds, and scarcely odorous roots, is clearly a peripheral member of the series Decurrentes. It cannot, however, be seen to have closer relationships elsewhere in the genus. Representative specimens: UNITED STATES: sie ee a Baldwin Co.: Wet boggy meadow in pinelands, 1 mi. n. of Stapleton, Webster & Wilbur 3524 (NSC, eee Mob ile Co.: Wet ground, Spring Hill, Graves 1016 (MO, US). DE LAWARE: Sussex Co.: Ditches and iceadoars. Ellendale, Ce s.#. (NY). FLORIDA: Duval Co.: Pine barren swamps, near Jacksonville, Curtiss 511° (F, GA, MO, PH). Highlands Co.: In shallow water, border of I one pre n.e. of Sebring, a 143 (FLAS). Holmes Co.: Grassy pine ear: 2.5 f Bonifay, Ford 3553 AS). Martin Co.: Wet roadside ditch on Fla. 76, Smith S Buchanen 124 AS). Osce- a Co.: Moist open pine area, 3.5 mi. e. of St. Cloud. Siiids § Buchanan 95 AS). Santa Rosa Co.: In grassland, odes of cypress swamp, 7 mi. s. of B Bagdad, Ford eer (FLAS). 292 x el, ae a y SU ee Fs i wf Mt mY, We Qe oO Figure 1. | esnioee A, habit & 1/5; Be 3 ane at tip of peduncle X 2%; C, single flower with one wing bent forward to expos ), crest on keel (lower petal) X ae ‘, pistil & 10; F, seed & AG sista eed on 70 [FLAS], Goldhead State Park, as 293 GEORGIA: Appling Co.: Dry bed of pool area in drainage way, 7.5 mi. es.e. of Baxley, Duncan 11035 (GA, MO, NSC). Early Co.: Low undrained pinelands, 10 mi. s. of Blakely j w, 0.5 mi fF Mi - n : — GA). Sumter Co.: Shallow pine-barren pond, w. of Leslie, Harper 462 oo MISSIS- SIPPI: Hancock Co.: Low pinelands, Necaise, nd nf s. a. (US). Jackson Co 1 Springs, Seymour 91-8-31 (F, FSU, MO, NCU). Newton Co.: Tracy 5177 (MO). Nona CARO- LINA: Bladen Co.: In borrow pit, in standin r, 2.5 mi. e. of Elizabethtown, Fox & Godfrey 2645 (NSC). — ian 7 le aca pond, w. of Newport, Whitford s. on. (NSC). Dare Co.: Bog, of Manteo, alas 4639 (NCU). Wake Co.: Raleigh, Hyams 682 (NSC). ona Co Berkeley Co.: ae See pool, 5 mi. n. of Wanda, pe 5340 ee Fasper Co.: ek 2.1 f Tillman, Bell 3857 FLAS). Ce ; ow nnah, s. of Britton Neck, Bell 7871 (FSU, NCU). Sumter ie ioe sav mee on ae (NY, PH). 2. POLYGALA RAMOSA Elliott, Bot. S. Car. and Ga. 2:186. 1822. TYPE: UNITED STATES: SOUTH CAROLINA: ? Charleston County, “ponds in the flat pine barrens,’ S. Elliott (HOLOTYPE: Stephen Elliott Herbarium, The oe Museum, Charleston, S. C., not seen; as photo NY!, US!). Fig. Po aie cymosa sensu Poir. in Lam., Encycl. 5:500. 1804; non P. cymosa Walt. Polygala corymbosa Nutt., Gen. 2:89. 1818; non P. ci Mae Michx. Polygala corymbosa sensu Torr. & Gray, Fl. N. Am. 1:128. 1838; non P. corymbosa Michx. Polygala baldwinti var. chlorogena Torr. & Gray, Fl. N. Am. 1:129. 1838. Pilostaxis ramosa (Ell.) Small, Man. S. E. FI. 774. 1933. Annual, erect, 10-38 cm tall, stems single or several, from fibrous roots. Basal leaves 3-7 cm long, 0.4-0.8 cm broad, spatulate, narrowing into petiole- like base, usually withering once plant flowers; cauline leaves spatulate- elliptic to linear. Inflorescence a cymose panicle, with few to many (to 110) racemose branches. Bracts 1.4-1.7 mm long, usually persistent. Pedicels 1.3- 2.3 mm long. Flowers bright yellow, turning bright green or less frequently yellow-brown upon drying. Upper sepaloid sepal 1.2-1.6 mm long, wide lance- ovate; two lower sepaloid sepals 1.2-3 mm long, narrow lance-ovate; wings 2.5-3.5 mm long, 1.0-1.4 mm broad, obovate or elliptic-obovate, 3-nerved, involute at apex. Upper petals 2.5 mm long, % connate to lower petal (keel); keel 2.2 mm long; crest of simple or bifurcating lobes. Seed 0.6-0.7 mm long, 0.3-0.4 mm broad, with fine pubescence; aril minute to 34 length of seed, 2- lobed. Anthesis: April through October. Habitat and distribution: Moist to wet, acid, moderately leached soil of low, open pine flatwoods and grassy roadside ditches; peninsular Florida, north on the Coastal Plain to Delaware, west to eastern Texas. Polygala ramosa is closely related to P. balduinii. This similarity is ob- scured by the striking contrast between the yellow-flowered (green when dry) loosely branched inflorescence of P. ramosa and the white or near- white, more compact inflorescence of P. balduinii. Occasionally individuals eccur that suggest hybridization; these have the flower color of P. ramosa, ) 2? NY 7 | \ \ ) 7 ec Nr aera ee ja ta : \ $ yh pee ee ee eA e =, « —e Nos a e =. ’ ee | . \e a wae se ee e yet ents Ww, \ x, [yz jo Figure 2. Polygala ramosa. A, habit & %; B, single raceme C, crest on keel (lower petal) & 12; D, pistil & 8; E, seed & 15; F, deenbution. jee 94 [FLAS], Butler, Fla.) Lake 295 the compact inflorescences of P. balduinit, and pedicel lengths (1-2 mm) almost exactly intermediate. Such individuals have been called P. balduinti var. chlorgena Torr. & Gray (1838) Representative specimens: UNITED STATES: ALABAMA: Baldwin Co.: Low bar Magnolia Springs, Schallert 758 (DUKE). Mobile Co.: In swamp, ae Hill, Be. 214 (MO, NY, US Wa ers o.: Low Regie 4 mi. n. of Fruitdale, Saree Ss. w. (GA, NSC). DELA- WARE: t Co. n, Can n, (PH). Sussex Co.: Moist open sandy depression, e of Ea Pomel ee (MO). FLORIDA: Alachua Co.: Low ground near Orange Heights, Kea (FLAS). Dade Co.: Low pineland back of mat City, Small © Wilson so nm. (F, NY), pees Co.: Dune area near Coast Guard station on Santa Rosa Island, lord 4484 (FLAS). Lee Co.: Moist sandy pin mere Pine eer Moldenke 942a (NY). Osceola Co.: Shallowly flooded depression in cypress wooded area, Holopaw (DUKE, FLAS, MO, NSC, NCU, NY, PH). Wakulla Co.: Sandy peat of ditch, 6 mi. n. of ; E NY). GE St. Marks Lighthouse, Redfearn & Kral 2475 (DUKE, FSU, F : Baker Co.: Low open grassy meadow, 11 mi. n. of Newton, Thorne 5032 (F, MO, NY). Charlton Co.: St. Mary’s River swamp, below Trader’s Hill, Small s. 1 , NY). Chatham Co.: Pine bar- ren nea a on Lake, Eyles 6236 (GA). Laurens Co.: Moist sandy woodland, ni ublin, Pyron & McVaugh 748 (GA). Richmond Co.: Pineland bog, ta, Cuthbert son OUISIANA: Calcasieu Parish: Vicinity of Lake Charles, Allison 77 (US) SE S). St. Tammany Parish: Open pineland, 1 mi. n. of Abita Springs, Pennell 4148 Y, PH). Vernon Parish: Fairly wet grassy area, 2 mi. w. of Faas Webster : Wilbur 3242 (NSC, US). MISSISSIPPI: Covington Co.: Open pineland, 1.5 s.e. of Mt. Olive, Webster 3334 : Abundant on black i ee "ot hillside oe 1 mi. n.w. of Ala- bama line, Kral 1298 (FSU). Pearl ae Co.: Low pinelands, Heavens Sargent s, a. (US) NORTH eae Reg o.: Moist ae es and along seepage areas in large open savannah, 1.5 mi. n.e. na, Wiha 4204 (FSU, GA, NSC). Dare Co.: Along margin of swamp, n. of co a 1101 (NSC). Jones Co.: In savannah, 20 mi. s. of Kinston, Wilbur 4093 (FSU, GA, NSC). eee Co.: Salisbury, Porter & Heller s. n. (NSC). saci AROLINA: Clarendon Co.: Grass-sedge bog or savannah, 3 mi. s.e. of Manning, Zz = 5 & oO ° - oq oO oO °o Ables 3 Bel 15747 (NCU). Horry Co.: oe savannah n Myrtle Beach, Correll 5277 .: Boggy swale, 5 mi. s. of Colu mee pore & Tryon 1265 (NY, US). TEXAS: Houston Co.: Sandy open ae yao. aang ee CoG, NY, US). Jasper Co.: Jasper, Fisher 32123 (F). Jefferson Co.: umont, (MO). VIR- GINIA: Sussex Co.: Dry argillaceous field, n. of a Fernald ae 6263 (MO, US). 3. POLYGALA BALDUINII Nuttall, Gen. 2:90. 1818. TYPE: UNITED STATES: FLORIDA: Nassau County, W. Baldw (HOLOTYPE: Thomas ae Herbarium, Academy of Natural ek Philadelphia, not seen). . 3, Annual or biennial, stems 10-65 cm tall, rarely branched below inflores- cence, single to several from thick fibrous or indefinite tap-root base. Basal leaves 0.5-2.5 cm long, 0.3-1.2 cm broad, obovate to spatulate, obtuse or rounded, crowded into irregular rosette, withered or nearly so when plant flowers; cauline leaves graduating from broad-eliptic to linear. Inflorescence a cymose panicle of numerous (to 40) racemose branches. Bracts 1.5-2.2 mm long, persistent. Pedicels 0.6-1.2 mm long. Flowers white or cream to greenish-white, drying from white to brownish-green. Upper sepaloid sepal 1.8-2.4 mm long, lance-ovate; two lower sepaloid sepals 1.6-2.2 mm long, lanceolate or linear-lanceolate; wings 2.8-4.8 mm long, 0.9-1.7 mm broad, 3-nerved, narrowing into a cusp-like apex. Petals 1.7-2.8 mm long, yellowish, connate at base only; crest with large simple lobes at lateral edges and 296 HUA Ha ba xt a? 2A 8 a: f Sha a i # aN i dL Ba ges . a ue Soe Ya 8 ’ po& 4 pyde | 5 iz \ \ ae \ \ SS) BN Figure 3. Poly ie balduinii. A—D, var. balduinii: A, saa x 4; B, flower (adaxial view) showing crest on lower petal X 8; C, pistil & 10; D, seed & 15 — 122 ee lig Martin Co, Fla.). E, var. carferi: habit & (Smith 103 [FLAS], Everglades Nat. Park, Fla.). F, distribution: var. balduinit (dots), var. carferi (asters—S. Fla. and Cuba) 297 simple or bifurcating lobes in center. Seed 0.5-0.7 mm long, 0.4-0.5 mm broad, moderately pilose, with a small rostrum; aril scarious, of 2 appressed oval lobes, to 0.2 mm long, or a minute scale. When Nuttall first proposed this name he Latinized the epithet as ‘“‘Bald- uini’’; this was purposeful as indicated by its repetition in his discussion. Torrey & Gray (1838) improperly modified the spelling to ‘“‘Baldwinii’”? and have largely been followed by later authors. Article 73 of the International Code, however, compels the use of ‘‘balduinii,’’ reflecting corrections both in typography and orthography. Small (1905) described from South Florida a new species of Polygala which he considered related to P. balduinii. He named it P. carteri in honor of his fellow collector, J. J. Carter. He first distinguished it from P. balduinii by the presence of elongate racemes and cuspidate wings, and later (1933) by an additional series of partially overlapping characters including flower color, size of capsule, and length of seed. A few of Small’s criteria at first reading appear quite distinct, but they are not so apparent when applied to specimens from southern Florida. Blake (1924) recognized P. carteri without equivocation, but Miller (1971) thought it ‘possibly conspecific’? with P. balduinii, and Long & Lakela (1971) without comment reduced it to syn- onymy under that name. Even cursory examination of Polygala in the field in South Florida reveals that in that area plants of P. balduinii s. 1. are more variable than elsewhere in the range of the species. The flower color varies from white to cream to greenish-white; the racemose branches of the cymose panicle vary markedly in their degree of elongation; and a number of less conspicuous floral char- acters extend beyond the range of variability observed northward. Yet, as implied by the criteria cited by Small and by Blake, many individuals are not clearly assignable to either of the two species. Wing size appears to be among the more discriminating characters. The wings of the South Florida plants are more uniform in size (3-4 mm long, 1.5-1.7 mm bread) with a short apical cusp (0.3-0.5 mm long), while P. balduinit elsewhere in its range is more varied (2.8-4.8 mm long, 0.9-1.4 mm broad) with a longer cusp (0.6-0.8 mm long). Seed size definitely overlaps, but P. ‘“carteri” tends to have larger seeds, approaching 1 mm in length. The aril, which is characteristically present on the seed elsewhere in the range of P. balduinii, is often absent in southern Florida. A degree of additional objectiveness was provided by the use of a hybrid index, as popularized by Anderson (1949), for six characters: length of bract, wing width, length of upper sepal, length of cusp-like attenuation, length of seed, and presence or absence of aril. Equal weight was given each character (scaled from 0 to 4), with lowest values assigned to P. ‘“‘carteri’”’ and the highest to P. balduinii s. s. Data were obtained from 80 plants from 33 Florida counties. These have been plotted on a map of 298 \ @ r NO MINIMUM VALUES < 1 NO MINIMUM VALUES < 7 me NO MINIMUM VALUES < 4 Figures 4-5. Objective representation of variation within 80 plants of Polygala balduinii g J I a from 33 Florida counties, involving six characters. Fig. 4, a map of Florida showing low and lines indicate northern limits of low histogram values, restricting var. carferi morphotypes to southern end of the state, with var. balduinii occurring throughout. Figure 5, histogram, the abscissa being the possible totals obtainable by summation of values for the six measured characters, the ordinate being the number of individuals. high histogram values (low values = var. carferi, high values = var. balduinii); dashed 299 Florida (Fig. 4) and charted on a standard histogram (Fig. 5). Measurements of the six utilized characters, when converted to histogram values and plotted on a map, show a striking restriction of plants with low values (characteristic of P. ‘‘carteri’’) to the southern counties of Florida, with a perceptible increase in these minimum values as one moves north- ward in the state. Only within Dade County, the type locality for P. carteri, do values of 2 occur (with 0 the theoretical extreme for “‘pure’’ P. “‘carteri’’). Adjacent Collier and Monroe counties have plants with values as low as 4, while it is necessary to travel almost to the center of the peninsula to obtain values in excess of 9. In contrast, plants with high histogram values (24 being the theoretical extreme for P. balduinii s. s.) occur throughout the state, and presumably throughout the southeastern Coastal Plain, reflecting the widespread distribution of individuals typical of the species. The histogram (Fig. 5) incorporates values for all individuals measured. Its lack of bimodality supports observations in the field, that many individuals are intermediate in morphology. The definite regional tendency for South Florida plants to vary beyond the norms for the species does not seem sufficient for specific recognition. It is here proposed to give recognition to the differing individuals in South Florida only at the varietal leve 3a. POLYGALA BALDUINII var. BALDUINII Polygala polycephala Balddw. ex Nutt., Gen. 2:90. 1818; as manuscript nam Pol ygala “Baldwinii’”’ sensu Torr. & Gray, Fl. N. Am, 1:128. 1838. Pilostaxis “‘Baldwinii’’? (Nutt.) Small, Man. S. E. Fl. 774. 1933. Racemes generally short and dense, to 3 cm long, 0.6-1 cm thick; bracts usually over 2 mm long. Flowers white, drying white to brownish. Upper sepaloid sepal usually over 2 mm long; wings 2.8-4.8 mm long, 0.9-1.4 mm broad, narrowing into lengthened cusp-like apex 0.6-0.8 mm long. Seed usually less than 0.6 mm long; aril 0.2 mm long, occasionally much smaller. Anthesis: January through November. Habitat and distribution: Moist to wet open pine flatwoods and moist grassy roadsides, often intermittently flooded; frequently sympatric with P. ramosa; peninsular Florida, northward to central Georgia and west (locally) to southern Mississippi and eastern Texas. In gross appearance, Polygala balduinii var. balduinti is generally more robust than var. carteri, having taller stems and a larger and more con- spicuous inflorescence. Specimens from a population in Lee County, Florida, were noted to have seeds lacking the characteristic pilose pubescence [W. E. Liggett 1979a, s.e. of Ft. Myers (FLAS) ] Representative specimens: NITED STATES: ae Alachua Co.: In slash-longleaf a aa oods, between ea and Newn Lake, Ward 1281 (FLAS). Broward = pinelands, Ft. Lauderdale, Small & W Vilson 1629 (NY). Collier Co.: Very mo ass salmetto along U.S. 41, Smith 372 (FLAS). Dade oe Humbugus Prairie, "Suvall, Meee a) Small 300 6580 (NY, US). Duval ae Moist pine barrens, near Jacksonville, Curtiss 504° (F, NY, PE I). ‘he ambia Co.: Dune area near Coast Guard station on Santa Rosa Island, Ford Franklin .. Low pine barrens, Apalachicola, Saurman s. a. (MO PH). Indian River Co.: Pinelands near Fellsmere, Small 8883 (NY). Lake Co.: Swamp, age of pee Nash 1179 (PF, MO, NY, PH, US). Levy Co.: Peaty roadside ditch, 4 mi. e. of Cedar Key, Kral 6989 (NCU, LY). Polk Co.: Open grass-sedge meadow, 9 mi. s.w. ot Risin Wilbur & Webster 2638 (NSC). Volusia Co.: Low. flatwoods, Lake Helen, (MO ) f — aK So —~ i WS : » US). Su Moist shady pine-barre 2 s.w. of Leslie, Harper 05 (PF, MO, NY, US). Neer ies kson Co.: Horn “Island, has of Biloxi, rdcy (fF, MO, Y). TEXAS: Jefferson Co.: Smith s. (PH) 3b. POLYGALA BALDUINII var. earteri (Small) Smith & Ward, stat. nov. Polygata celles Small, N.Y. Bot. Gard. Bull. 3:426. 1905. T : TED STATES: FLORIDA: Dade County, ‘‘pinelands between Cutler a Black Point,” J. kK. Small & J. J. Carter 813 (HOLOTYPE: NY!). Pilostaxis carteri (Small) Small, Man. S. FE. Fl, 774. 1933. Racemes elongated and usually loose, 0.4-6 em long, 0.4-1.3 em thick; bracts usually less than 2 mm long. Flowers greenish-white to greenish-cream, dry- ing as such with brownish tinge. Upper sepaloid sepal usually less than 2 mm long; wings 3-4 mm long, 1.5-1.7 mm broad, with apex short-tipped, 0.3- 0.5 mm long. Seed usually 0.6-0.7 mm long; aril a minute scale, or absent. Anthesis: March through August; December. Habitat and distribution: Wet to seasonally very dry marl prairies, oc- casionally on calcareous waste areas or moist grassy roadsides; endemic to South Florida (Collier, Dade, and Monroe counties, including the Florida Keys), locally in western Cuba. Polygala balduinii var. carteri is a less striking plant than var. balduinii; the plants are on average smaller, and the inflorescences are much more green and less conspicuous. Its northern limit is set more as a matter of arbitrary convenience than of cbjective judgment, for individuals with par- tial manifestation of the characteristics of the more extreme South Florida populations may be found, particularly in coastal areas (Brevard County), appreciably further north (Fig. 4). Representative specimens: CUBA: Pinar del Rio: In wee places, Remates, Cienaga La Tumba, Ekman 1 1 ; UNITED STATES: FLORIDA: Collier Co.: In grassy glades, Copeland, Brass 15403 (US); Deep Lake, Sculf s. in. (FLAS). Dade Co.: an it-over sie 6 mt w. of Royal Palm Park, rea Cc. An teat Deam 60930 (DUKE); In mmoles pineland on limesto 5S omi. w. of Royal Palm Ham- mock, O'Neill 7591 (PF); Low pinelands, s. of Cutler, Small & Carter 3065 (NY)3 Ever glades near Cocoanut Grove, Syrall & Wilson s. a. (DUKE, FLAS); Glade, Key oe ee Tatnall 842 (PH). Monroe Co.: Prairie at abandoned lumber camp, Pinecrest, Ari (FLAS); Margin of pond along road to Watson Hammock, Big Pine Key, ei 149.84 (FLAS). 4. POLYGALA RUGELII Shuttleworth ex Chapman. Bot. Gaz. 3:4. 1878. TYPE: UNITED STATES: FLORIDA: Hillsborough County, ‘‘Tampa,’ IF. Rugel s. n. (HOLOTYPE: not seen, presumably among the Alvan Wentworth Chapman collections of the Gray Herbarium; Rugel in 1845 301 collected near Tampa Bay, sending his materials to Robert Shuttleworth in Berne, Switzerland, who marketed large numbers of them to institu- tions and aiiduals in North America). Fig. 6. Polygala reynoldsiae m., Fl. S. U.S. 1883. Pilostaxis rugelii (Shuttlew. ex Chapm.) ae Man. S. E. Fl. 774. 1933. Annual or biennial, erect or ascending, 20-78 cm tall, with long branches from main axis, arising from definite tap-root. Basal leaves 3-6 cm long, 0.5-1.5 em broad, spatulate, obtuse to rounded, succulent, drying rugose; cauline leaves gradating from oblanceolate to lanceolate. Inflorescence a head-like raceme, ovoid to thick cylindric, usually 2.5-3 cm long, 1.8-2.5 cm thick, occasionally elongated. Bracts 3.5-5 mm long, usually deciduous. Pedicels 2-2.5 mm long. Flowers bright yellow, drying pale yellow to dark ereenish-brown. Sepaloid sepals 1.5 mm long, the upper deltoid-subulate, cuspidate; wings 4.5-8 mm long, 2-4 mm broad, oblong to oblong-elliptic, 7- nerved, the apex shcort-acuminate to apiculate. Petals 5.2-5.5 mm long; crest with bifurcating lobes 1 mm long. Seed 1.2-1.6 mm long, moderately pubes- cent; rostrum 0.2 mm long, rounded; aril membranous, to length of seed or longer. Anthesis: April through November. Habitat and distribution: Low moist open pinelands and well-drained palmetto-pine associations; endemic to peninsular Florida. Polygala rugelii, although less common than P. lutea with which it is often found, when in flower is a conspicuous member of the Florida flora. Oc- casional populations [cf. M. C. Reynolds s. n., St. Johns Co. (NY)] exhibit an elongation of the raceme. This anomaly was observed by Chapman (1883) who introduced the name P. reynoldsiae (in honor of the collector, Mary Reynolds, St. Augustine, Florida) for what he thought to be a new species. Such an elongation has also been observed in P. lutea Representative specimens: UNITED STATES: FLORIDA: Alachua Co.: Openings in ce ag repens—L vonts lucida brush, 2.6 mi. w. of Island Grove, Ward 1940 (PLAS, FSU). evard Co.: Flatwoods, w. of Sharpes, Rdoads s. nm. (ELAS). Collier Co.: oom pine He 0.3 mi. s. of North Naples city limit, Ward 1805 (FLAS). Dade Co.: Low pinelands, back of Lemon City, Small & Wilson 1967 (NY). Flagler Co.: Low pinelands 1 near Bunnell, Wise s. 2. (FLAS larde o.: Pasture, e. of fo Springs, Cooley, Wood & Wilson 6092 (FS NY). Her- nando Co.: Flatwoods pond margin, Weeki Wachee Springs, Kral 7003 (FLA S). Indian River Co. ry sandy area on Fellsmere road Smith & Myint 704 (ELAS) Lafayette Co ee 3 mi. s. of Mayo, West & Arnold s. in. (FL! rke Co.: Wet soil of ditch, near road stor Park, Adanrs 606 (FSU). Lee Co.: Pineland, vicinity o f Myers, Standley 177° (NY, US). Levy Co.: Pi Godfrey & Lindsey 56995 (FSU). Martin Co.: Bank of wet ditch, 0.4 mi. s. of St. Luc Co. line, Smith oo 126 (FLAS). Orange Co.: Open pine-palmetto 3.4 nw. of Brevard Co. line, Smith & oe 141 (FLAS). Pasco Co.: Sandy wie flat. woods, w. of Zephyrh ills, Hood 3501 (FLAS). Pinellas Co.: Moist open pine-palmetto area, St. Petersburg, Smith 160 (FLAS). St. Lucie Co.: Open grassy field, 6.3 mi. s. of White City, Hansen 928 (FLAS). Sarasota Co.: Sandy peat of depression 7k pine-saw palmetto flats, 3 mi. s. of Venice, Kral 7481 (F LAS, NCU) 302 3. POLYGALA LUTEA Linnaeus, Sp. Pl. 705. 1753. TYPE: UNITED STATES: NEW JERSEY: ? Gloucester County, P. Kalm s.n. (HOLOTYPE: 882.39 as 1945], Linnaean Herbarium, ndon, not seen; as microfiche!). Fig. Polygala lutea L. var. elatior Michx., FI. 8 Am. 2:54. 1803. Polygala lutea L. var. pauciflora Raf., New FI. N. Am. 4:90. 1838. sede pseudosenega Bertol., Mem. "Accad. Bologna 5:400. 1854: with ex- eption of phrase, ‘‘Flores rosei.’ Pitostazis | lutea (L.) Small, Man. S. E. Fl. 744. 1933. Biennial or perennial, ascending and spreading or erect, 6-42 cm tall, with variations from a solitary unbranched axis to many stems branched or un- branched from a thick base; tap-root well developed in larger specimens, more fibrous in smaller plants. Basal leaves 1.5-5.8 cm long, 0.7-2 cm broad, spatulate to obovate, succulent with obscure venation, drying rugose, in an irregular rosette; cauline leaves obvate to oblanceolate. Inflorescence a head-like cylindric raceme, 8-33 mm long, 8-20 mm wide, rarely elongated (to 4 cm long). Bracts 2.5-3 mm long, deciduous. Pedicels 1.5-2.7 mm long. Flowers vivid orange to yellow-orange, usually turning a dull, pale yellow on drying. Sepaloid sepals 1.2-1.4 mm long, the upper ovate-acuminate, minutely ciliolate; wings 5.0-7.0 mm long, 2.7-3.6 mm broad, slightly oblique- elliptic, acuminate to short pointed, 5- to 6-nerved, partially involute at apex and slightly narrowed at base, minutely ciliolate. Petals 3.5-4.2 mm long, 2/3 connate; crest 0.5-0.7 mm long, of single or slightly bifurcated lobes. Seed 1.4-1.6 mm long, pilose; rostrum 0.3 mm long Anthesis: February through November. Habitat and distribution: Typically acid soil with high water table, com- mon in pine flatwoods, open grassy areas, or pine barrens; peninsular Florida north of Ft. Meyers, northward on the Coastal Plain to New Jersey and Long Island, New York, westward to eastern Louisiana. Polygala lutea is both the most widespread and the most abundant mem- ber of the series Decurrentes. Although the epithet was doubtless suggested by the color of the dried flowers, the living plants with their conspicuous orange inflorescences are a characteristic component of the acid flatwoods Coastal Plain flora. Hardin (1961) reported a color variation of lemon yellow for a single plant from Brunswick County, North Carolina. This species is closely related to P. nana, P. smallii, and P. rugelii, and specimens are frequently encountered bearing one of these names. It is most definitely separated by the very small, largely non-bifurcating lobes of the crest, the otherwise distinctive orange flowers fading with maturity or with drying. Intermediates have not been encountered with P. nana or P. smallii. A single possible hybrid with P. rugelii revealed over 65% of the pollen grains as apparently non-functional. Representative specimens: a 6. Polygala oes A, habit . 14; B, flov est on keel (lower petal) x ; D, pistil K 8; eed X 15; I Seen nn : TEL AS], Trenton, Fla.) as RSkeKe Kes ash Ne WARS Aes O48, WA EJ At Was Figure 7. Polygala lufea. A, habit XX } 3, flower wich upper oe se Pia showing : Pr . I 1 od ob C, | we. ium and style, with hidden ovar y Sn. y shadowin rolla with upper petals w ne ae showing androecium adnate to eal Gower a x 5; D, flower, with prominent. peti eae sepals and inrolled 3-petaled corolla xX 3; E, crest on 2; Fy seed & 15; G, distribution. (Drawn from fresh ae Alachua Co., Fla.) 305 Lae STATES: ALABAMA: Baldwin Co.: ae barrens, 5 mi. n. of Minctte, LeClair s. on. (NCU). Escambia Co.: Moist ground near spring, Escambia Exp. Forest, oan 235 Cala dae ene Co.: In swamps, Spring Hill, "Bush 221 (MO, NY). 4 Co.: Moist area on steep slopes on side of high ridge, 2 mi. w. of Phenix City, Duncan 9641 (GA). DE LAWARE: ‘Ker ent Co.: In swamps near Harrington, Gleason s. nu. (DUKE Sussex Co.: Wet thick near Georgetown, Briffon 58 (NY). FLORIDA: Collier Co.: Sand flat near ee Scull s. n. (FLAS). Columbia Co.: nee roadside, 6.6 mi. s.e. 0 Suwannee River, Smith . Myint 287 (FLAS). Flagler Co.: Flatwoods n. of oe West & Arnold s. n. LAS). Liberty Co.: Shallow pond in pineland, Bristol, West & Arnold s.n. (FLAS). Pens Co.: Bayhead in longleaf pine, n. of Fe. Drum, Wes? s. 1. AS). Putnam Co.: Low ground, Johnson, ee 2134 (NY). Santa Rosa Co.: Grassy pine flatwoods, 2 mi. s.e. of Milton, Ford 4073 (FLAS). Union Co.: Flatwoods, 2 mi. s.w. of Raiford, Murrill s. 1. (FLAS). GEORGIA: Brooks Co.: Moist savannah, 0.5 mi. n.e. of Barney, Wilbur oe (GA, NSC). Bryan Co.: Sandy humus in thin pine woods, 7.8 . sc. of Richmond Hill, Moore & Lawrence 748 (FLAS, GA). Decatur Co.: Wet san nde i headwaters swamp of Willacoochee Creek, near Faceville, Thorne, Muenscher & Smith 3023 (GA). Laurens Co.: Pine woods, s.w. of Cadwell, Duncan 5000 (GA). Richmond Co.: Sandy low barrens, Augusta, Cuthbert s. nu. (NY). LOUISIANA: Sc. Tammany Parish: ae ae woods, 5 mi. e. of Slidell, Harris gs. mn. (FSU). MARYLAND: o wt Sa ( Snow Hill, Vandegrilt 7253 (NY). MISSISSIPPI: Choctaw Co.: French Camp, Clute 58 Oe Forrest a ae oe lane & Anderson 1514 (MO). Hancock Co.: Wet ine barrens, Ba Demaree 31930 (NCU). Jackson Co.: Ocean Springs, Pollard 1168 (MO, NY, US). as Ww JERSEY: ee Pe Swampy ground, Lindenwood, Mere- dith s. in. (MO). Cape May Co.: Low ground, Cape May, Parker s. a. (NY) Ocean Co.: ae swamp, 10 mi w. ey Barnegat, Gleason 75 en 172 (NY). NEW YORK: Suffolk In bog, headwaters of Browns Creck, Parehoene, Muenscher & Curtis 6224 (PH, NY, c, NORTH CAROLINA: Brunswick Co.: Ditch bank, margin of shrub bog, 5 mi. n. of Southport, Godfrey 49168 (NSC). aes Co.: Savannah-like meadow near Smith- field, Garner s. nn. (DUKE). Lincoln Co.: Boggy sandy margins and adjacent woods, e. of U.S. 321, Bell s. mu. (NCU). Pamlico Co.: Pine —— 1.5 mi. s. of Reelsboro, SS 35935 (NCU). Sampson Co.: Flat pine-oak woods, 1.1 mi. w.s.w. of Newton Grove, Ables & Huacsloop 29946 (NCU). SOUTH CAROLINA: Beaufort Co.: Pine savan- nah, e. side of S. C. 170, Bell 3826 (NCU). Florence Co.: Low ditch by pine savannah, Mc Heller 912 (F, MO). James City Co.: Sandy soil in sphagnum-magnolia swamp, w. 0 Williamsburg, Grimes 3854 (NY). Nansemond Co.: Mossy depressions in sandy pine barrens, nal — Landing, Smith & Hodaion 964 (DUKE, F, FLAS, PSU, GA, MO, NSC, NC US). Prince Geors xe Co.: Sandy pineland, s.w. of New Bohemia, Wherry @& Pe anell 14423 ie MO, PH). j 6. POLYGALA NANA (Michaux) DeCandolle, Prodr. 1:328. 1824. Fig. 8. Polygala lutea L. var nana Michx., Fl. Bor. Am, 2:54. 1803. TYPE: oe STATES: ? SOUTH CAROLINA: ‘“‘in partis udis Caro- linae,’? A. Michaux (HOLOTYPE: in Herbarium Michaux, Museum Na- tional d’ Histoire Naturelle, Paris, ay seen). ee nana hx.) Raf., New N. Am. 4:89. 1838; combination indic Iso under Polygala, but ie made. Tne hyemalis Raf., N . 4:89. 1938; combination indi- ated also under Polygala, but not made Polygala nana (Michx.) D var. humillima Chod., Mem. Soc. Phys. Geneve 31 (2)? :200. 1893. Annual or biennial, erect, 3-13 cm tall, stems unbranched, single to several from a definite tap-root, variable in height within same plant, often with one 305 short inflorescence in center of basal rosette. Basal leaves 1.1-5.5 cm long, 0.4-2 cm broad, spatulate, rounded, occasionally apiculate, narrowed at base, 3- to 5-nerved, succulent, drying rugose; cauline leaves very few if present, oblanceolate. Inflorescence a head-like cylindric raceme, 1-3.8 em long, 1-1.7 cm thick, extending above the basal leaves. Bracts 4.5-6.5 mm long, usually deciduous, sometimes persistent. Pedicels less than 1 mm long. Flowers lemon yellow to greenish-yellow, drying to green or yellowish-green. Upper sepaloid sepal 0.4 mm long, subulate-lanceolate; lower sepaloid sepals 3- Thine y an } Ais an it \ : \ M ue 3 AEN ( a, jf | \ /) ay FD ES [|i \ mc ‘igure 8. Polygala nana. A, habit X Ys; B, hae as sepals flattened & 7; C, flower with petaloid sepal turned down, exposing petals X ), crest on keel ae pet a «K 6; FE, pistil, with pollen acon to oui stigma X i pee X= 15; G, distribution. (Drawn from fresh material, Alachua Co., Fla.) 307 5.86 mm long, lanceolate; wings 5.5-7.5 mm long, 1.8-2.8 mm broad, elliptic, 3- to 5-nerved, long-acuminate to cuspidate, involute at apex, sparsely cilio- late. Two upper petals 3.5-4 mm long; keel 3.5-5 mm long; crest with bifur- cating lobes 2 mm long. Seed 0.8-1.6 mm long, pilose; rostrum thick, 0.2-0.5 mm long; aril 4% to full length of seed. Anthesis: February through June; October. Habitat and distribution: Sandy well-leached soil, in dry palmetto-pine flat- woods, longleaf pine—turkey oak or scrub oak communities, or in moist grassy areas; peninsular Florida north of Ft. Myers, northward inland to western North Carolina, westward to eastern Texas. Polygala nana apparently cannot stand much competition. It is often in small sandy clearings where the rosettes may spread. When found on em- bankments it is usually observed on the drier upper slopes, often with P. lutea occurring on the moister soil below. In South Florida (Lee and Collier counties) this species approaches the habit of P. smallii, with the inflores- cences scarcely exceeding the leaves [cf. P. C. Standley 12751 (US)], but even in these populations the distinctions enumerated under P. smallii re- main valid. Representative specimens: UNITED STATES: ALABAMA: Baldwin Co.: Gateswood, oo 8683 (F, MO, NY US). Covington Co.: Upland pine Sceey between Lockhart and Wing, Godfrey & Harrison 55403 (FSU). Houston Co.: Open grassy area at ie nn : mi. s.s.e. of eo Hardin & a 14867 (GA). Mobile Co.: Pine flatwoods swamp on Dauphin Island, Lund 298 (FSU). FLORIDA: nee Co.: Moist ditchbank along Fla. 26, in slash-longleaf pine flatwoods between Gainesville and Newnan’s Lake, Ward 1277 (FLAS). Duval Co.: Dry la he near Jacksonville, Curtiss 518 (F, NY). Franklin Co.: In dry soil along U.S. 98, mi. e. of Gulf-Franklin Co. line, Ward & Smith 2612 (FLAS). Lee Co.: Among Been . - Fe. Myers, Standley 51 (PF, US). Martin Co.: On bank of wet ditch, 0.4 mi. s. St. Lucie Co. line, along U.S. 1, Smith © ae 125 FLAS). Santa Rosa oe es sand, lowland, at Yellow River, Rd. 87, of Mi Hood ee (FLAS). GEORGIA: Brooks Co.: Dry pinewoods, 10 mi. w. as Quitman, Pyron & McVaugh 2177 (GA). Camden Co.: Low pineland, 5 mi. n. of St. Mary, Blanton 6334 (US). Dade Co.: Sandy soil of abandoned oS summit - Lookout Mt. near Little River, McVaugh 9025 (GA, MO). Seminole Co. nelands ealy’s Plantation, southern tip of county, Thorne & Muenscher 2428 (GA). oo o. Low pineland, 1.5 mi. s. of Tennile, Duncan 17514 (GA). LOUISIANA: Calcasieu Parish: Moist sandy loam in longleaf a. DeQuincy, Pennell 10239 (NY). Natchitoches Parish: Sandy barrens, Chopin, oe 339 (PH, US). Quachita Parish: On moist sandy clay of oak-pine woods, 3 mi. s. ca 8758 (FSU). St. Taininony Parish: In pine woods, Mandeville, Langlois SOM, iF). is IPPI: Forrest Co.: Hattiesburg, Woodson @& Anderson 1569 (MO). Harris Co.: Low ground in See Long Beach, Joor s. #2. (MO). Jackson Co.: e A a i US). o = = 5 I Ocean Springs, Skeban 59 (FP, 1 Iktibbeha Co.: Starkville, Tracy s. a. ( Pike Co.: McComb, Earle s. 1. (NY ; MO). NORTH CAROLINA: Buncombe Co.: Ash- ville, Seymour 91-8-19-31 (DU KE). SOUTH CAROLINA: ee on Co.: Lumbered sandhill with seepage area containing pocasin plants, 0.4 mi. s. of Boiling Springs Church, Duke ee 2230 (NCU). TEXAS: Hardin Con Moist open pinewoods, Silsbee, eee 9565 ne ), US). Newton Co.: Newton, Fisher 3473 (F). 7. POLYGALA smallii Smith & Ward, nom, nov. Fig. 9. se la arenicola Small, N. Y. Bot. Gard. Bull. 3:426. 1905, nom. illeg., arenicola Giirke in Baum, Kunene Sambesi Expedition 273. 1903. TYPE: UNITED STATES: FLORIDA: Dade County, ‘‘pine- 308 lands between Cocoanut Grove and Cutler,” J. K. Small & J. J. Carter 1276 (HOLOTYPE: Pilostaxis arenicola (Small) Small, Man. S. E. Fl. 773. 1983. Biennial, erect, 2.0-7.5 em tall, with 1 (usually) to 4 or more short un- branched or branched stems from well-developed tap-root. Basal leaves 1-4.2 cm long, 0.2-1.4 em broad, oblanceolate to linear-oblanceolate, rarely near- spatulate, often absent, or in irregular, crowded rosette; cauline leaves simi- lar. Inflorescence a head-like, cylindric raceme, 0.4-3 cm long, 0.5-1.8 cm thick, usually surpassed by the basal leaves. Bracts to 3.5 mm long, decidu- ous. Pedicels less than 1 mm long. Flowers green or greenish-yellow, drying the same color. Sepaloid sepals 0.5 mm long, the upper lanceolate or lance- ovate, the lower linear-lanceolate; wings 5-8 mm long, 2-3 mm broad, oblong- lanceolate, 3- to 5-nerved, long-acuminate, remotely ciliolate. Upper petals 4-4.5 mm long; keel 4 mm long; crest with bifurcating lobes to 3 mm long. Seed 1.9-2.3 mm long, pilose; rostrum large, rounded, 0.6-0.7 long; aril % to almost length of seed. Anthesis: March through May. i B me Figure 9. Polygala ag A, habit x 4; B, SeAae sepal, flatte * 7; Cy crest on keel mie petal) & 6; D, pistil & 12; E, seed & 15; : dieebntion. 7 755 [FLAS], Dade Co., Fla.) 309 Habitat and distribution: Sandy or calcareous rocky areas of the open grassy pinelands; endemic to South Florida (Broward and Dade counties). Polygala smallii shares with P. nana the apparent inability to withstand competition and is characteristically found in small sandy openings in the dense herbaceous ground cover. It is clearly related to P. nana and has been relegated to that species by Blake (1924), Miller (1971), and Long & Lakela (1971). On close examination a number of consistent differences ap- pear. The leaves of P. smallit are more lanceolate and much narrower, the wings are more lanceolate, the flowers are much greener at anthesis, and the length of the seeds is significantly greater. The most closely adjacent populations of the two species are separated by approximately 32 miles. Small overlooked Gurke’s detailed description in 1903 of a Polygala from southeastern Africa, under the name P. arenicola, rendering his later use of the epithet homonymous. Since Small’s discovery of this species merits continued recognition, and since no conflict is created by its application to this most diminutive member of the series, it is here given the replacement name Polygala smallit. Representative specimens: UNITED STATES: FLORIDA: Broward Co.: In pinelands, Fort Lauderdale, Small Carter 1011 - Y). Da - fo ae Darcie: Bitton 149 (F, NY); Sandy pina on corner of 26th Rd. ve. Miami, Eyles 8172 (NSC); Miami, Garbe (FLAS, NY); Dry roeky. Soil, ane ‘Hood 72100 (FLAS); Miami, Porter s. mn. ap I dry pen palmetto-pine fetwood: along Coral Reef Drive, South Miami, Smith & Myint A yi Coconut Grove, Peattie s. nm. (F); Pinelands about Arch Greek Prairie, Small, Mosier & Small 6772 (NY, US); Pineland, Cutler, Spizk 106 (FLAS); Woodbury s. n. (BUS). w ACKNOWLEDGEMENTS The present paper is modified from a thesis submitted by R. R. Smith in partial fulfillment of the requirements for the degree of Master of Science at the University of Florida, June 1962. He is most grateful to Mr. George R. Cooley for financial support during one season’s field study and to the curators and staff of those herbaria from which specimens were borrowed. REFERENCES ANDERSON, es 1949, Introgressive Hybridization, New York: John Wiley & Sons. 109 p eae S. F. 1915. Some neglected names in Walter’s Flora Caroliniana. Rhodora 17: 29-137. = 24. Polygalaceae. In: North American Flora 25: 305-370. CHODAT, R 1893. Monographia Polygalacearum. Mém. Soc. Phys. Hist. Nat. Genéve 31(2): 120-140, 197-204. 196, Polygala. Im: A. Engler & K. Prantl, Die Natirlichen Pflanzenfamilien 3(4): 330- CA oon ‘a P. 1824. Polygaleae. In: Prodromus Systematis Naturalis Regni Vege- aes 1: 321-33 FRASER, J. 1789. - short history of the Agrostis sae aon London. 8 pp. HARDIN, J. W. 1961. Color variation in Polygala Infea. Castanea 26: 1 HOLM, T. 1929. Morphology of North American species Mf Polygala, Bot. Gas 88: 167-185. JAMES, C. W. 1961. Endemism in Flosida, Brittonia 13: 225-244 310 LEWIS, W. H. Bore 1962. Cytological observations of Polygala in eastern North America, Khoon 7" 102-113. x, R. & O. LAKELA. i 71. A Flora of Tropical Florida. Coral Gables: Univ. of LONG, | Miami oe 962 as [ Polygala, 525-531.] _ I o> G. 1971. The Polygalaceae of the Southeastern United States. J. Arnold Arbor. PED eg sree A revision of the Polygalaceae of me Southeastern mae States fesclsing Fria, ae D, Dissertation, West Virginia Univ., Morgantown. 187 ete. The genus Polygala in the Southeastern Cited States. Proc. West Va. ad Sc 9-13. Ac ap has E, S. 194s, A Catalogue of the Linnaean Herbarium. London: Taylor & Francis, Ltd. SMALL ‘ K. 1933. Manual ‘of the Southeastern Flora. New York. 1554 pp. [Péilostavis, eon a J. & A. GRAY. 1838. A Flora of North America. 1:126-129. A NEW SPECIES OF PERITYLE (ASTERACEAE) FROM COAHUILA, MEXICO A. MICHAEL POWELL Department of Biology, Sul Ross State University, Alpine, Texas, 79830 PERITYLE carmenensis Powell, sp. nov. Plantae perennes; caules 10-15 cm longi glabri vel minute sparseque pubescentes. Folia conferta opposita vel alterna 1-1.7 cm longa (petiolis inclusis), 0.5-0.9 cm lata glabra tenuia tripartita subcruciformia segmentis lobatis; petioli 5-8 mm longi. Capitula solitaria ca. 22-flora pedunculis tenuibus brevibus. Involucrum 4-5 mm altum anguste campanulatum, phyllariis 12-14. Flores radiati ca. 5 pistillati fertiles; tubus 2-2.5 mm longus glandulari-pubescens, ligula flava ca. 4 mm longa ca. 2 mm lata oblonga apice 3-dentata; flores disci corolla flava vel ad maturi- tatum purpurea 3.5-4.5 mm longa minute glandulosa fauce late cylindracea 1.8-2.2 mm longa, Rami styli ca. 2 mm longi tenues gradatim angustati apice acuti distaliter pubescentes adaxialiter planati. Achaenia compressa 2.8-3.3 mm longa anguste oblanceclata vel oblonga margine callosa utrinque brevi- pubescentes; pappus nullus. Chromosomatum numerus, n = 17 II TYPE: MEXICO: Coahuila. Maderas del Carmen, 102° 36’ 30’? W, 0.5 mi N of El Dos on trail to El Tres, 2325 m, along stream, on overhanging cliff of igneous rock (rhyolite?) in thin soil, associated with Pinus strobiformis, Pseudotsuga, Abies, Cornus, Physocarpus, Polypodium, 7 Aug 1974, A. Adamcewicz & T. Wendt 525 (Holotype, TEX; Isotypes, MWXU, and others to be distributed). Known only from the type collection. Perityle carmenensis is related to P. castillonii (sect. Laphamia; Powell, 1973) from which it is easily distinguished by radiate heads, leaf shape, and leaf glabrosity. The new yellow-rayed taxon belongs with an otherwise dis- coid group of three species with allopatric distributions in southeastern Arizona, west Texas, and southeastern Chihuahua. The related-species group also includes P. dissecta and P. lemmoni. Significantly, one atypical collec- tion of P. lemmoni from Arizona (Powell, 1973) does have radiate heads, and this along with P. carmenensis from Coahuila suggests that the related- species group in question might have originated from a widespread radiate ancestor. The specific epithet of P. carmenensis is taken after the Maderas del Carmen range of northern Coahuila to which it is probably restricted in distribution. The existence of the new species was brought to my attention by Tom Wendt who made the original collection and recognized it as prob- SIDA 6(4): 311—312. 1976. 312 ably undescribed. I am grateful to M. C. Johnston for providing the Latin translation. REFERENCES POWELL, A. M. 1973. Taxonomy of Perifyle section Laphamia (Compositae-Helenieac- Peritylinae). Sida 5: 61-128. A NEW SPECIES OF EUPHORBIA (SECTION CHAMAESYCE) FROM THE BAHAMAS STEVEN R. HILL Department of Biology, Texas A&M University College Station, Texas, 77840 In the recent upsurge of interest in the Bahama Flora, many new plant records have been reported (e.g. Gillis, Howard and Proctor, 1973; Correll, 1974; Hill, 1974), but few, if any, new species have come to light. Instead, as a result of more careful study, many of the old species of Britton and Millspaugh (1920) have been merged into more widespread species of the southeastern United States and the West Indies. In the present case, I wish to record the presence in the Bahamas of a previously undescribed species. Despite the many collections from the islands to date, there are occasional sites that have not yet been botanized, and which may harbor narrow endemics. The environment of the new species is a rocky limestone coast, windswept and harsh. Under such conditions, a colonizer must quickly adapt in order to survive. This adaptation is reflected in a distinct new morphology and, presumably, also in a new genetic makeup. The plant described has tiny succulent leaves with revolute margins that enable it to survive in its hot, dry, sometimes salty environment. The tough stipules have become elaborated to sheathe the leaf base, and a woolly indument within the sheath further reduces water loss from the tender point of juncture of stem and leaf and protects tender buds. The branches are lax and flexuous, and the plant is rather appressed to the rocks, allowing its survival in the frequent gale winds in these islands. To survive a hurricane a plant must either be very flexible or very strong. Many woody plants in the Bahamas have very strong trunks, but have narrow flexuous branches more likely to twist than to snap. Despite the apparently recent geological origin of the Bahamas, certain groups of colonizers have diverged sufficiently to be considered new species. The present case is such an example. EUPHORBIA longinsulicola S. R. Hill, sp. nov., sectionis Chamaesycis; suf- frutex perennis laxus 0.5-1 m altus, caule infime 6-9 mm crasso; rami glabri alterni, vaginis stipularibus persistentibus intra lanatis; laminae foliorum succulentae oppositae obovatae 3-7 mm longae 1.5-2 mm latae glabrae mar- ginibus valde revolutis, apicibus rotundatis vel retusis, basibus breve angustatis; petioli 0.5-1 mm longi; cyathia olivacea obpyramidalia extus glabra 1.2-1.5 mm longa, pedunculis 1.0-1.2 mm longis, intra lanata; glandes 4 deltoideae 0.8-1 mm longae 0.3-0.5 mm latae exappendiculatae; flores SIDA 6(4): 313—316. 1976. 314 staminati plerumque 8, antheris 2 vel 3; flores pistillati brevi-exserti, glabri, stylis tribus integris; fructus maturi non vidi. TYPE: BAHAMAS: Long Island, exposed seacoast at the end of the Buck- leys Road, Deadman’s Cay. Rooted in crevices of rock among wind-stunted shrubs. 1 July, 1974. Steven R. Hill 2383 (Holotype, NY; Isotypes FTG, US, VT). ADDITIONAL SPECIMEN EXAMINED: BAHAMAS: Long Island, exposed seacoast at Turtle Cove, between Clarence Town and Deadman’s Cay, 31 May, 1972, Steven R. Hill 832 (FTG, NY). Figure 1. The new species is a perennial subshrub 0.5-1 meter tall, rather woody with flexuous lax branches. The branches are 6-9 mm thick at the base, alternate, glabrous, with persistent stipular sheaths that are woolly within giving the stem a jointed appearance. The leaves are opposite, obovate, rounded or usually retuse at the tip, 3-7 mm long and 1.5-2 mm wide. The bases of the glabrous and fleshy leaves are shortly attenuate, and the edges of the blade are strongly revolute, the edges being fused to the nerve be- neath. The cyathia are solitary on branchlets, glabrous outside, in the axils of the leaves at the branchlet tips, obpyramidal, 1.2-1.5 mm long (measured from peduncle top to rim), olive-green with 4 dark brown (almost black) glands, without appendages: the deltoid glands are 0.8-1 mm wide (meas- ured along the rim). The peduncle is 1.0-1.2 mm long. All cyathia counted had 8 staminate flowers. The pistillate flowers are short exserted, glabrous, and fewer in number than the staminate. The stamens bear two or three anthers. The third anther is, when present, either in reduced, non-functional form or apparently complete and functional. The plant was in full flower in early July, in contrast to the May specimen which was nearly sterile. Euphorbia longinsulicola was found growing in an area of wind-swept limestone bedrock and stunted sea grape shrubs at a site previously un- collected. In more protected areas nearby the more common Euphorbia mesembrianthemifolia Jacq. grew. Field cbservations suggest that the species is restricted to a narrow strip of land exposed to wind and ocean influence along the coast from Turtle Cove to Deadman’s Cay, a distance of about 15 km. The obovate, fleshy leaves and sprawling, shrubby habit seem to sepa- rate the taxon easily from the species of Euphorbia sect. Chamaesyce pre- viously known from the Bahamas. It seems to be most clearly allied to another Bahamian endemic, Euphorbia vaginulata Griseb., but the latter species, which I examined in the field while on Inagua, has leaves that are rather brown and cinereous when fresh and which are significantly longer and narrower than those of the new species, and it has stipular sheaths that are prominently ciliate. In addition, I. vaginulata seems to prefer open sand or somewhat rocky areas, and has sparse, thin, erect branches. It is also not known to occur on Long Island. The species epithet indicates that it is an inhabitant of Long Island, the 315 Bahamas. I would like to thank Dr. Derek Burch for his help in examining the plant and verifying its uniqueness (personal communication), and also Dr. Dono- qd‘. ee Figure 1. Euphorbia longinsulicola 8. R. Hill a. stem axis and branchlet with cyathium. b-e portions of staminate flowers: b-c. typical staminate flower, top and lateral view; d. staminate flower with 2 normal and one abortive anther; e. staminate flower with three functional anthers. f. cya- thium, top view. g. pistillate flower. h. cross-section of leaf, freehand, show- ing revolute-fused margin. All drawn from the holotype. 316 van Correll who has been very encouraging in this work. I would also like to thank Dr. Paul A. Fryxell for his encouragement and comments on the manuscript. REFERENCES BRITTON, Sa L. & . PF. MILLSPAUGH. 1920. The Bahama Flora. Hafner Publishing Company, Inc. New Yor CORRE LL. D. ‘S. 1974, Flora of the Bahama Islands—new additions. Fairchild Trop. Gard. Bull. 29: 11-12; 15. GILLIS, W. T., R. A. HOWARD, & : R. een ates Additions to the Bahama flora since Britton and Millspaugh—I. Rhodora 75; -42 HILL, S. R. 1974. Range extensions = new records a aoa flora. Rhodora 76: 471-477. NEW GYPSOPHILIC SPECIES OF PSEUDOCLAPPIA AND SARTWELLIA (ASTERACEAE) FROM WEST TEXAS AND EASTERN CHIHUAHUA A. MICHAEL POWELL and B. L. TURNER Department of Biology, Sul Ross State University, Alpine, Texas 79730 and Department of Botany, The University of Texas, Austin, 78712 The species described here provide additional examples of the remarkable effect of gypsum-edaphic endemism on plant speciation and they add to the list of new taxa discovered during the course of our study of gypsophilic vegetation of the Chihuahuan Desert region (Powell, 1972, etc.; Turner, 1972; etc.) PSEUDOCLAPPIA watsonii Powell & Turner, sp. nov. Fig. 1. Frutices (25-) 30-75 cm alti intricate ramosi ramis filo metalico similibus validis in parte superiore pedunculiformibus. Folia alterna carnosa_ sub- teretia 1.2-1.7 cm longa glabra superiora reducta bracteiformia. Capitula radiata solitaria pedunculis erectis ca. 4 cm longis; involucra late obconica 1.1-1.4 em alta; bracteae inaequales 3-4-seriatae lineari-lanceolatae glabrae margine scariosae apice attenuatae (vel acutae); flores radiati 6-7(-11), ligulis flavis 1.0-1.8 em longis, oblongo-ellipticis vel linearibus apice 1-2- incisuratis vel irregulariter fissis; disci flores ca. 25 corollis flavis 8.0-8.5 mm longis tubis ca. 2 mm longis faucibus cylindricis basem versus leviter angustatis 5.5-6.0 mm longis, lobis ca. 0.5 mm longis erectis vel leviter in- curvatis; rami styli ca. 2 mm longi fere glabri lineares planati apicem versus leviter angustati apice truncate; antherae 3.5-4.0 mm_ longae. Achaenia matura nigra ca. 4 mm longa oblonga basem versus leviter angustata utrin- que dense pubescentia pilis adpressis ca. 0.5 mm longis in transectione rhom- bea vel fere quadrata; setae pappi plures inaequales 8-10 mm longae sub- teretes vel complanatae stramineae antrorsus barbellatae. Chromosomatum numerus, n= ca. 19 TYPE: TEXAS. Hudspeth Co., Clay hills and arroyos, ca. 1 mi E. of Tommy’s Town, 24 Aug 1974, A. M. Powell 2792 (Holotype, SRSC; Isotypes, MO, TEX, US). Additional specimen examined: TEXAS. Presidio Co., 53.8 road mi N of Candelaria, on slopes of a small hill along the primitive road to Chispa; Sierra Vieja foothills; with Acacia, Agave, Yucca, Larrea, Condalia, Erion- euron and Xylorhiza; rocky clay soil; only three plants observed; 21 May 1971, T. Watson 630. 'Field work supported in part by NSF Grants GB-37674 and 29576X. SIDA 6(4): 317—320. 1976. or ie a ~ a ue -_ 2 SPT TEL ~ 4 92); habit and achene. owell 27 Figure 1. Pseudoclappia watsonii (P 319 Pseudoclappia watsonii is only the second species of the genus to be de- scribed, the other one being P. arenaria, a halophytic-gypsophile of south- eastern New Mexico, north-central Mexico, and adjacent Texas. The new species is similar to P. arenaria except that most floral features and the heads are larger in P. watsonii. The two species are readily distinguished, however, by habit and habitat differences. Pseudoclappia watsonii is a strong, intricately branched, low shrub which apparently is restricted to dry, gypseous-clay soils; P. arenaria, while also a low shrub, evidently does not attain the size and branching habit of the other species and it occurs in more mesic saline-gypsum habitats such as along playa lakes, etc. The new taxon was first collected by Dr. Tom Watson, former student of both authors, after whom we are pleased to name the species. SARTWELLIA gypsophila Powell & Turner, sp. nov. Suffrutices perennes glabri 35-80 cm alti erecti vel caulibus effusis. Folia opposita anguste linearia 7-12 em longa 1.0-2.5 mm lata sessilia inferiora connato-vaginata, Capitulescentiae corymbiformes multiflorae apice planae. Capitula radiata 12-15 flora; involucra turbinata, bracteae 5 ellipticae acutae apice minute ciliato-laceratae; flores radiati saepe 5, ligulis flavis ovalibus vel oblongis vel ellipticis 1.5-1.8 mm longis 1.0-1.5 mm latis; flores disci corollis 2.0-2.5 mm longis tubis fere indistinctis, faucibus superius abrupte expansis lobis ca. 1 mm longis acutis maturis reflexis, styli rami ca. 0.5 mm Iongi vel breviores apice truncati penicillati. Achenia nigra 1.0-1.5 mm longa minute pubescentia; pappus coroniformis squamellis pluribus pro parte coalescentibus ca. 0.3 mm longis. Chromosomatum numerus, n = 18. TYPE: MEXICO: Chihuahua. Jurassic gypsum near the lake road and ca. o mi NW of Lake Granero, ca. 16 mi from Morri6n, 14 July 1973, A. M. Powell 2536 (Holotype, SRSC; Isotype, TEX). Known only from the type locality. Sartwellia gypsophila was discovered at a locality where several other novel gypsophilic taxa have been found in the past few years, including members of Argemone, Selinocarpus, Nerisyrenia, and Gaillardia. The discovery of a new Sartwellia is Surprising since the genus, containing only three species, was recently treated by the junior author (Turner, 1971); however, the gypsum area at the type locality was unexplored by botanists at that time which leads us to believe that yet other gypsophiles within this genus might remain undescribed from the area concerned. Judging from pappus structure, leaf morphology, head size, and floral characters, Sartwellia gypsophila seems most closely related to the more northerly S. flaveriae of southern New Mexico and adjacent Texas although it also shares characters with the more southerly S. puberula. Sartwellia gypsophila is readily distinguished by its tall, suffruticose habit and selected floral characters. Other species of Sartwellia are annuals or short-lived perennial herbs, but all are restricted to gypsum substrates (Turner, 1971). 320 REFERENCES POWELL, A. M 1972. A new species of Argemone (Papaveraceae) from Mexico. Southwest. Nat. 17: 106. TURNER, L. 1971. Taxonomy of Sartwellia (Compositae—Helenieae), Sida 4: 265-273. 72. Two new species of Isocoma (Compositae—Astereae) from north central ») B. . 19 Mexico. Sida 5: 23- ADDITIONS TO THE BAHAMA FLORA STEVEN R. HILL Department of Biology, Texas A&M University, College Station, Texas, 77840 The following listing of further additions to the Bahama Flora is intended to supplement those of Correll (1974), Gillis (1975), Gillis, Howard, and Proctor (1973), and Hill (1974). The present study was undertaken as a pro- ject sponsored by the Fairchild Tropical Garden where I was employed as an assistant to Dr. Donovan Correll during the Spring and Summer of 1975. Dr. Correll’s collections cited here with his permission were made by sup- port from a Grant (No. DEB74-02135-A02) to him from the National Science Foundation. In the listing, plant families are arranged in the order followed within the original Bahama Flora of Britton and Millspaugh (1920) for ease in com- parison. Within each family, species are arranged alphabetically. Unless otherwise specified, the determinations were made by Dr. Correll or myself. A full set of the cited specimens is housed at the Fairchild Tropical Garden Herbarium, Miami, Florida with duplicates sent to the New York Botanical Garden, Bronx, New York, the Field Museum Herbarium, Chicago, Iinois, and the Missouri Botanical Garden Herbarium, St. Louis, Missouri. They will serve as voucher specimens for the forthcoming revision of the Bahama Flora being undertaken at the Fairchild Tropical Garden. I would especially like to thank Dr. Donovan S. Correll and Dr. John Popenoe for their support towards this project. POACEAE BRACHIARIA SUBQUADRIPARA (Trin.) Hitchc.: Andros, weedy-brushy area along edge of cultivated field near San Andros airport, January 29, 1974, D. S. Correll & R. K. Godfrey 41312; Berry Islands, in clumps on edge of coppice along road on Anderson Cay, Great Harbour Cay, October 16, 1974, D. S. & H. B. Correll 43676; North Eleuthera, in sandy soil of open coppice about Preachers Cave, January 12, 1974, D. S. Correll 41176; San Salvador, in moist Bg about 1 mile south of Finger Lakes Station, November 21, 1974, D. S. Correll 43872. ERAGROSTIS bees (L.) Beauv.: North Andros, in open field at Owens Town, February 2, 1974, D. S. Correll & R. K. Godfrey 41519. NEYRAUDIA REYNAUDIANA (Kunth) Keng: South Bimini, in large clumps in depression on edge of coppice just north of airstrip, 4 m tall, April 18, 1974, D. S. Correll 42155. As far as we know this is the first report of the occurrence of this Asian species in the West Indies. PANICUM COMMUTATUM Schult. (P. joorii Vasey): Grand Bahama, on SIDA 6(4): 321—327. 1976. 322 rich slope in partial shade in dense hammocky coppice in Fortune Hills area of Freeport, November 4, 1973, D. S. Correll 40496. This is an extension of this species from Cuba and Hispaniola. PANICUM PORTORICENSE Desv. ex Hamilt.: North Eleuthera, in moist soil on edge of depression in palm-broadleaf savanna near intersection of road to Harbour Island ferry with Ridley Head Road, January 11, 1974, D. S. Correll 41162. PASPALUM ARUNDINACEUM Poir.: Great Inagua, in large clumps 1.5 m tall, in depression, Horse Pond, northeast of Matthew Town, July 31, 1975, D. S. Correll 45819. PASPALUM DILATATUM Poir.: Andros, in fields at Owens Town, Feb- ruary 2, 1974, D. S. Correll & R. K. Godfrey 41520; Grand Bahama, grassy soil bordering driveway at Jansel Courts, Freeport, August 17, 1974, D. S. Correll & Robert Kral 42967. PASPALUM DISTACHYON Poit. ex Trin.: Great Abaco, forming mats in drying marl sink on northwest edge of Marsh Harbour, July 11, 1974, D. S. Correll 42800. This is an extension of this species from Cuba and Hispaniola. PASPALUM FILJIFORME Sw.: Little Exuma, in large clumps in rocky area about Williams Town, June 27, 1974, D. S. & H. B. Correll 42485. This is an extension of this species from Cuba and Hispaniola. PASPALUM PLEOSTACHYUM Doell: Great Inagua, in open sandy soil along roadway between Matthew Town and Maroon Hill, August 19, 1975, D. S. Correll 46032; Long Island, roadside at the upper margin of the town salina, north side near clinic, Clarence Town, June 19, 1974, S. R. Hill 2200. PASPALUM ROCANUM Leén in Britt.: Great Inagua, in disturbed saline flats about Morton Bahamas Limited salt pans, August 1, 1975, D. S. Correll 45864. ~~ CYPERACEAE CYPERUS SURINAMENSIS Rottb.: Andros, in field at Owens Town, Feb- ruary 2, 1974, D. S. Correll and R. kK. Godfrey 41518; Grand Bahama, north end of Freeport airport, August 18, 1974, D. S. Correll & Robert Kral 43014. DICHROMENA FLORIDENSIS Britt.: Andros, on pitted limestone in cut- over pineland just north of Love Hill, January 29, 1974, D. S. Correll & R. K. Godfrey 41518; Long Island, dooryard in moist soil, near Baptist Church, Deadman’s Cay, July 1, 1974, S. R. Hill 2392; Grand Bahama, in and around freshwater ponds in pineland area in northcentral portion of island, about 2 miles from coast, November 6, 1973, D. S. Correll 40568. RHYNCHOSPORA ELLIOTTII A. Dietr.: Grand Bahama, in water of fresh- water ponds about 2 miles from coast, northcentral section of island, No- vember 6, 1973, D. S. Correll 40565, RHYNCHOSPORA INTERMIXTA C. Wr.: Grand Bahama, in water of pond in pineland, about 2 miles east of Freetown, November 6, 1973, D. S. Correll 40585. 323 BROMELIACEAE CATOPSIS FLORIBUNDA (Brongn.) Sw.: Andros, uncommon local epi- phyte, hardwood coppice along ridge 3.4 miles northwest of Love Hill, July 11, 1975, S. R. Hill 3377. COMMELINACEAE SETCREASEA PURPUREA B. K. Boom: Grand Bahama, in restricted area about 1 mile east of Freeport airport, along Queen’s Highway, August 20, 1974, D. S. Correll & Robert Kral 43059. ORCHIDACEAE ONCIDIUM FLORIDANUM Ames: Gillis (1975) states that he has been unable to locate specimens of this taxon in the Bahamas. This record may serve to verify its presence. Andros, hardwood coppice along ridge, 3.4 miles northwest of Love Hill, July 11, 1975, S. R. Hill 3374. This species is often confused with O. sphacelatum Lindl. CHENOPODIACEAE CHENOPODIUM ALBUM L.: Andros, disturbed ground along farm road, 2 miles south of San Andros airport, July 13, 1975, S. R. Hill 3445. NYMPHAEACEAE NYMPHAEA ELEGANS Hook.: Great Abaco, in water of ponds in open pinelands along Forest Drive, about 1.5 miles northwest of Marsh Harbour, March 18, 1975, D. S. Correll & F. G. Meyer 44603, BRASSICACEAE CARDAMINE HIRSUTA L.: Grand Bahama, in rich soil in foundation plantings about Freeport airport, May 25, 1975, D. S. Correll & John Popenoe 45446. LEGUMINOSAE ADENANTHERA PAVONINA L.: New Providence, off JFK Read to air- port, near Lake Cunningham, March 23, 1974, D. S. & H. B. Correll & John Popenoe 41926. This native of tropical Asia has become naturalized in various parts of America. AESCHYNOMENE AMERICANA L. var. GLANDULOSA (Poir.) Rudd: North Andros, in field at Owens Town, February 2, 1974, D. S. Correll & R. K. Godfrey 41514. This species represents a genus new to the Bahamas. CROTALARIA FALCATA Vahl ex DC.: Grand Bahama, in open fields about Freeport industrial area on way to harbor, plants 1.5 m tall, August 18, 1974, D. S. Correll & Robert Kral 42972; Great Abaco, in field along Great Abaco Highway, about 15 miles south of Marsh Harbour airport, July 7, 1974, D. S. Correll & John Popenoe 42687. These represent a new West Indian occurrence for this Old World species. 324 DESMODIUM CILIARE (Willd.) DC.: New Providence, in pineland near Mermaid’s Pool, March 26, 1974, D. S. & H. B. Correll 42028. MELILOTUS INDICUS (1.) All.: Eleuthera, in moist soil around lakes in dump area across from entrance to Cotton Bay Club, between Rock Sound and Green Castle, May 14, 1975, D. S. Correll & S. R. Hill 45067. SESBANIA VESICARIA (Jacq.) Ell.: Andros, disturbed ground along farm road, 2 miles south of San Andros airport, July 18, 1975, S. R. Hill 3457. EUPHORBIACEAE ALEURITES MOLUCCANA (L.) Willd.: Eleuthera, large trees in coppice about an ancient dump, Half Sound area between Tarpum Bay and Rock Sound, May 18, 1975, D. S. Correll & S. R. Hill 45270. This native East Indian tree has become naturalized in the West Indies. EUPHORBIA OPHTHALMICA Pers.: North Eleuthera, in sandy soil in open coppice near Preacher’s Cave, January 12, 1974, D. S. Correll 41192. Adams, (1972, p. 431) includes the Bahamas in the area of distribution of this species. This collection, however, is the first we have seen of this species from the archipelago. I choose to accept Chamaesyce as a_ section of Euphorbia, PHYLLANTHUS TENELLUS Roxb.: Grand Bahama, weed in border plantings at Jansel Courts, Freeport, August 17, 1974, D. S. Correll & Robert Kral 42956. MALVACEAE SIDA RHOMBIFOLIA L.: Andros, common in disturbed ground along farm road, 2 miles south of San Andros Airport, July 13, 1975, S. R. Hill 3453. LYTHRACEAE AMMANNIA COCCINEA Rottb.: Exuma, Hummingbird Cay, in low swale in dry meadow, January 19, 1971, N. H. Nickerson & J. Dalton 3330. Both common species of Ammannia in the Southeastern United States are now seen to occur also in the Bahamas. The second, A. latifolia L., is seen to be pres- ent in both the petaliferous and apetalous state, the former formerly being referred to the taxon A. teres Raf., now a synonym of A. latifolia L. (Gra- ham, 1975). ARALIACEAE BRASSAIA ACTINOPHYLLA F. Muell.: Andros, an established escape, vicinity of main road and AUTEC road, Andros Town, July 12, 1975, S. R. Hill 3413. An abundance of young plants was found in the vicinity of a large, abundantly fruiting tree. LOGANIACEAE CYNOCTONUM SESSILIFOLIUM (Walt.) J. F. Gmel.: North Andros, on moist flat in opening in pineyard, Twin Lakes Region, January 29, 1974, D. S. 325 Correll 41295. Although Mrs. Northrop (1902, p. 57) reported this species as from Red Bays, Andros, Britton and Millspaugh disallowed it by referring her collection to C. mitreola (L.) Britt CONVOLVULACEAE IPOMOEA FISTULOSA Mart. ex Choisy: Great Abaco, in disturbed soil among rubbish at edge of coppice-covered marsh in pineland area along Forest Drive, about 2 miles north of Marsh Harbour, July 11, 1974, D. S. Correll 42809-A; Little Exuma, on rocky open hillside at Ferry, December 13, 1973, D. S. Correll 40961; a suffruticose Ipomoea, escaped from cultiva- tion, and now naturalized. VERBENACEAE VERBENA BONARIENSIS L.: Andros, disturbed ground along farm road, 2 miles south of San Andros Airport, July 13, 1975, S. R. Hill 3435. This col- lection represents a genus new to the Bahamas. SOLANACEAE LYCOPERSICON ESCULENTUM Mill. var. CERASIFORME (Dun.) A. Gray: New Providence, in waste area on edge of coppice, Yamacraw Beach area, March 23, 1974, D. S. & H. B. Correll & John Popenoe 41914. LENTIBULARIACEAE UTRICULARIA FIBROSA Walt.: North Andros, in fresh water of swamp near Behring Point, January 31, 1974, D. S. Correll & R. Kk. Godfrey 41453. ULTRICULARIA PURPUREA Walt.: Grand Bahama, in water of canal just west of Pelican Lake, May 24, 1975, D. S. Correll & John Popenoe 45422; Great Abaco, floating on small pond on edge of marsh along Marsh Harbour- Treasure Cay road a little north of Marsh Harbour, July 5, 1975, D. S. Correll & John Popenoe 42634; North Andros, about 1 mile from Queen’s Highway going west from Love Hill, in shallow water of swamp in pineland, February 1, 1974, D. S. Correll & R. K. Godfrey 41475; same location as preceding, July 11, 1975, S. R. Hill 3365. ACANTHACEAE BARLERIA CRISTATA L.: oo established around Andros Town air- port, May 24, 1975, S. R. Hill 2 DYSCHORISTE ren (Michx.) Ktze. var ANGUSTA (Gray) R. W. Long: Grand Bahama, several plants scattered over open pine savannah along Queen’s Highway going east to North Riding Point, August 19, 1974, D. S. Correll & Robert Kral 43030. JACOBINIA SPICIGERA (Schlecht.) Bailey: New Providence, in coppice behind beach near Delaport Point, March 24, 1974, D. S. & H. B. Correll 41968. This native of Mexico is now well established in a coppice on New Providence. 326 GOODENIACEAE SCAEVOLA TACCADA (Gaertn.) Roxb.: Bimini, escaping along beach area on southeast coast of Cat Cay, July 12, 1975, D. S. & H. B. Correll 45693; Abaco, between airstrip and water on Walker’s Cay, June 23, 1975, D. S. Correll 45480. ASTERACEAE BORRICHIA FRUTESCENS (L.) DC.: Grand Bahama, scattered over mud flats in marsh on edge of Pelican Lake, August 16, 1974, D. S. Correll & Robert Kral 42902. This species is maintained as distinct from B. arbores- cens (L.) DC. based especially on its reflexed, spine-tipped phyllaries. CALYPTOCARPUS VIALIS Less.: Great Abaco, in lawns and vacant lots along waterfront at Marsh Harbour, July 8, 1974, D. S. Correll & John Pope- noe 42729. This is an extension of this tropical American plant from Florida and Cuba. CIRSIUM VITTATUM Small: Great Abaco, in open rocky soil, Treasure Cay, July 6, 1974, D. S. Correll & John Popenoe 42666; Great Abaco, in marshy area 1 mile north of Bail Sugar Mill, south of Marsh Harbour, July 7, 1974, D. S. Correll & John Popenoe 42725. Cirsium horridulum Michx. (C. smallii Britt.) and the above Floridian plant are maintained separately in Dr. Correll’s forthcoming Flora of the Bahamas. HELIANTHUS ARGOPHYLLUS T. & G.: Eleuthera, on open slopes around dump, Tan Bay, about 2 miles south of Palmetto Point, May 17, 1975, D. S. Correll & S. R. Hill 45222. This Texas endemic has also been seen under cultivation in George Town, Exuma. TITHONIA DIVERSIFOLIA (Hemsl.) A. Gray: pre in coppice between Small Hope Bay and Hard Bargain, January 29, 1974, S. Correll & R. K. Godfrey 41261. This and the following species eee a genus new to the Bahamas. It is a native of Mexico and Central America that has become naturalized, TITHONIA TUBAEFORMIS (Jacq.) Cass.: New Providence, forming colony among bushes and grasses along south side of runway at Nassau Air- port, November 25, 1975, D. S. Correll 46302. Determined by Dr. B. L. Turner. POLY PODIACEKAE NEPHROLEPIS RIVULARIS (Vahl) Mett. ex Krug.: Crooked Island, cov- ering north facing cliffs at base of hill in open area just northeast of Land- rail Point, February 19, 1975, D. S. Correll 44404 REFERENCES ADAMS, C. D. 1972. Flowering plants of Jamaica. University of the West Indies, Mona, Jamaica, 848 pp. BRITTON, N. L. & C.F. MILLSPAUGH. 1920. The Bahama Flora. Hafner Publishing Company, Inc. New York. CORRELL, D. S. 1974, Flora of the Bahama Islands—new additions. Fairchild Trop. Gard. Bull. 29: 11-12; 15. 327 GILLIS, W. T. 1975. Additions and corrections to the Bahama Flora—lIl. Sida 6: 52-62. rILLIS, W. T., Ro. A. HOW? ,& G. R. PROCTOR. 1973. Additions to the Bahama and Millspaugh—TIf. Rhodora 75: 411-425. 75. Taxonomy of th flora since Britton GRAHAM, S. A. I¢ Sida 6: 80-10 HILL, S. R. 1 A. 19 e Lythraceae in the southeastern United States. 3 974. Range extensions and new records for the Bahama flora. Rhodora 76: 47 77. NORTHROP, A. R. 1902. Flora of New Providence and Andros. Mem. | 12: 1-98 Torr. Bot. Club POLLEN MORPHOLOGY OF DALEA SECTION THEODORA (LEGUMINOSAE ne WM. F. MAHLER Herbarium, Southern Methodist University Dallas, Texas, 75275 ABSTRACT. The pollen mor phology of Dalea section Theodora in the south- western United States and Mexico has been sama Four of the five species in section Theodora nen a unique pollen type commonly referred to as “dogbone’’ with Dalea mollis possessing the eee ue prolate pollen type of the genus. A radial development of the pollen shape is evident from the prolate pollen type in the center of the geographic range of the section Theodora to the ‘‘dogbone” pollen type radi ating outward to the north, cast, and southeast. The pollen grains of D. mollissima represent the intermediate form between the prolate and the ‘“‘dogbone™ ec. After a preliminary study of the species of Daica within the United States possessing an unusual pollen shape commonly referred to as “dogbone”’ pollen grains (Mahler, 1970), Barneby (pers. comm.) suggested that the pollen grains of section Theodora be examined. The section Theodora, as circumscribed by Barneby (pers. cemm.), includes the following taxa: Dalea mollis, D. mollissima, D. neonexicana var. neomexicanda, var. longipila, var. megaladenia, D, simulatriv, and D. verna, MATERIALS AND METHODS. Poilen samples were taken from herbarium specimens and permanent acetolyzed pollen slides were prepared using the acetolysis method (Erdtman, 1960). Anthers were acetolyzed and the screen- ing process was eliminated as the length of the ‘“‘dogbone” grains, some over 140 um long, did not pass through the 50 um mesh screen except when oriented perpendicular to the screen. The pollen grains were mounted in glycerin on glass slides and sealed with paraffin. He rbarium specimens and the acetclyzed pollen slides are on deposit at their respective herbaria with duplicate slides at SMU (Table 1). The author is indebted to Rupert C. Barneby for access to his studies, additional materials, and suggestions ‘egarding the manuscript. Gratitude is also extended to Blanche W. Meeson for the preparation of the pollen and photographs (MSM-2 Mini-SEM) and to Linda W. Laury for preparation of the plates ~ RESULTS. The pollen grains of section Theodora are tricolpate, prolate to cylindrical (intermediate) to cylindrical with enlarged polar regions (dog- scriptions of Dalea sect. Theodora, D. verna, D. simulatrix, neomenicdnda var 1 at: megaladenia, and ce technical combination of D. meomexicana var. eon will be pub- lished in a revision of genus Dalea by R. C. Barneby, currently in manuscript (Mem. N.Y Bot. Gard. 7). SIDA 6(4): 328—331. 1976. TABLE 1. Pollen and collection data of Dualea section Theodora. Ave i sti ” Le ie Pe Taxon Collector & No. or Date ollectio Dalea mollis *Nelson 11135a (M) 47.6 “Orcutt 1889 (MO) 49.1] *Oreutt 1890 (MO) 51.0 *Parker, et al 7796 (MO) 52.3 aa et al 4131 (M 53.1 Pinkava, et al 103382 (SMU) 53.3 Wiegand, et al 3593 (MO) 54.1 Mahler 6046 (SM 55.1 Mahler 6074 (SMU) 56.4 Mahler 5501b (SMU 58.1 *Wiegand, et al 3590 (MO) 58.9 Palmer 67 (10020) (MO 59.3 Mahler 6045 (SMU 59,5 arker 7976 (SMU 59.9 Mahler 6030 (SMU) 66.3 a 194 (SMU) 67.1 Dalea mollissima Brenckle 51136 (SMU) 85.2 *Maguire, et al 4856 (MO) 89.8 Isely, nl 7228 (ISC) 91.5 Heller 15962 (MO) 103.2 Hitchcock 12362 (MO) 104.0 Rose 38341 (MO) 106.3 Kennedy, et al 1906 (MO) 107.7 Goodding 2237 (MO) 109.7 Munz 12605 (MO) 112.5 Spahr 26489 (SMU) 116.1 Jones 1897 (MQ) 118.5 Clover 8243 (SMU) 136.4 Dalea neomexicana rar. nNeomeNnicana Mahler 6095 (SMU) 97.8 Beasley, Finzel 865 (NY) 113.5 Warnock, Hinckley BG 208 (SMU) 116.9 *Tsely 8386 (ISC) 120.1 Correll 13734 (SMU) 123.6 Grimes 317b (SMU) 123.6 7 re, Steyermark 3270 (MO) L 25:7 rimes 317a (SMU) 129.4 ae 26 » (MO) 130.0 Thieret 30981 (SMU) 142.9 var. longipila Mahler 5681 (SMU) 142.8 var. mega nid Ripley, Barneby 14207 (NY) 141.8 Dalcea verna Mahler 5746 (SMU) 125.4 * Previously reported by Mahler (1970). bone-shaped) and 47-142 um in polar length (Fig. 1). The average pollen grain length per species is plotted in Figure 1, using horizontal bars, The average pollen grain polar length is relative as the number of collections per taxon differs because of the availability of suitable flowering material: > D. neomexicana var. neo- = Dalea mollis—56 um; D. mollissima—106 mexicana—122 um, var. longipila—143 um, var. megaladenia—142 um; D. verna—125 um (Table 1). Dalea simulatrix possesses long ‘‘dogbone’’ pollen grains that are easily observed with a dissecting microscope but measure- 330 ments of acetolyzed pollen are unavailable (ig. 2). The locations of the collections used in this study are plotted in Figure 1 with Dalea mollis possessing the prolate and shortest grains. This taxon is correctly delimited in the preliminary study (Mahler, 1970). However, Dalea mollissima extends northward into southern Nevada according to Barneby (pers. comm.) with D. neomexicana var. neomexicana extending westward only as far as Cochise County, Arizona. DISCUSSION. The tricolpate pollen grains of section Theodora are prolate (Dalea mollis), cylindrical (D, mollissima in part), or cylindrical with en- larged polar regions (dogbone-shaped: D. mollissima, D. neomexicana, D. simulatriz, and D. verna). The pollen of Dalea mollissima ranges in pollen length from an average per sample of 85 to 136 um and is thus, intermediate in both shape and length between D. mollis and the other taxa of section Theodora. The development of the pollen shape appears to be radial from the prolate pollen type (assumed to be the primitive type) in the center ee] 30 um Figure 1. pecmpnoge of collections ot Dalea section Theodora. The average pollen polar length per species or variety is eae by the horizontal bars (um). 1. Dalea mollis (circles); 2. D. m isin (squares); 3a. D. neomenxicana var. neonmtexicana aa 3b. var. pail, (eBanele points down), 3c. var. megaladenia (tria an points up); Verna (diamonds) 331 of the geographic range with the other taxa possessing the ‘‘dogbone’’ type at the periphery of the range of section Theodora. In summary, Dalea section Theodora may be considered as the ‘‘dogbone”’ pollen grain group that includes D. mollis with the typical prolate pollen type characteristic of the genus. igure 2. Pollen grains of Dalea simulatrix. A. 79 note tricolpate condition and smooth surface; OX, note shape and furrow; B. 2000x, 700X% REFERENCES ERDTMAN, G. 1960. The acetolysis method. A revised description. Svensk Bot. Tidskr. 54 561-564. MAHLER, WM. F. 1970. Pollen morphology of the Dalea = mollis-neomexicana complex (Leguminosae) in the United States. Southw. Naturalist 15; 15 - 4 i 7 Bee SID CONTRIBUTIONS TO BOTANY Index to Volume 6 Names of contributing authors are in capital letters. New scientific names are in beldface. Synonyms are in italics. Page numbers in boldface are the main entry for the taxon involved. Page numbers in italics are of those on which an illustration appears. Generic names in article titles are indexed to the first page of the article only. Abies 311; concolor 48 Acacia 317 Acanthaceae 115, 325 Acer saccharum 225 Adenanthera pavonina 323 Adiantum pedatum 226 Aeschynomene glandulosa 323 Agalinus maritima 59, var. grandi- flora 59; spiciformis 59 Agave 317 americana var. koehnei 85; latifolia 80, 84, 85, 86, 324; teres 80, 85, 86, 324, var. exauriculata 86 Amomodon attenuatus 226; rostra- tus 226 Anacis 130; auriculata 138; tripteris 162, 163 Andropogon 114; pertusus 53 Apium leptophyllum 58 Apocynaceae 58 Araliaceae 324 Arctostaphylos uva-ursi 48, 49 Arenaria patula 252 Argemone 319 Argythamnia candicans 54; lucay- ana 934 Asarum candense 226 Asimina triloba 225 Aster 60, 117; subulatus var. ligula- tus 117 Asteraceae 311, 317, 326 Atragene americana 37 AUSTIN, DANIEL F. 216 Ayenia insulicola 54; tenuicaulis 54 Barleria cristata 325 BARNEBY, R. C. 7 Batrachium 280; flaccidwm 281; tri- chophyllum 281 Bidens caroliniana 141; serrulata 133; verticillata 179 Boraginaceae 55, 235 Borrichia arborescens 326; frutes- cens 326 Bothriochloa 53; pertusa 53 Brachiaria subquadripara 321 3rassaia actinophylla 324 Brassicaceae 323 3razoria 115; scutellarioides 116 Bromeliaceae 323 Calliopsis 201; basalis 133; bicolor 206; cardaminefolia 207; drum- mondii 133; nudata 187; palmata 166; rosea 184; tinctoria 206, var. atropurpurea 207 Calyptocarpus vialis 326 Cardamine hirsuta 323 Cardiolepis nigra 70; rubra 70 Carex 117; bushii 117; trichocarpa 117 Carpinus caroliniana 225 Carya 225 334 Casearia 57; guidonia 57 Cassia 7: bauhiniodem 11, 13: bauhinioides 11, 12, 13, 14, var. pilosior 10; crotalarioides 8, 9, var. vogeliana 7, 8; demissa var. diffusa 9, var. radicans 7, 9, 10; durangensis 10, 11, var. duran- gensis 10, 11, var. iselyi 7, 11; mensicola 7, 11, 12, 13; mexicana 8; monozyx 7, 14, 16, 17, 18; nana 7, 8; parralensis 7, 18; pilosior 7, 10; pringlei 14; pumilio 8, 14; ripleyana 7, 8, 13, 15; roemerana 7, 13; roemeranam 11; unijuga 14, 16, 18; vogeliana 8, 9, 10 wislizeni 17, var. villosa 7, 16, 17; var. Catopsis floribunda 323 Celtis occidentalis 225 Centaureum cyanus 248 Centella 58; asiatica 58; erecta 58 Ceratocephalus delphiniifolius 176, Cerbera 58; thevetia 58 Cercis canadensis 68, 225 Chamaecrista goldmanti 7, 18 Chenopodiaceae 323 Chenopodium album 323 Chrysanthemum hirsutum 138 Chrysomelea 130; auriculata 138 Chrysostemma 162; tripteris 168 CHURCHILL, STEVEN P. 117 Cirsium horridulum 326; smallu 326; vittatum 326 Clematis 33; addisonii 34, 35, 36, 41, 45, 46; albicoma 34, 35, 40, 42, 48, var. coactilis 40; bahamica 56; baldwinii 34, 35, 40, var. latius- cula 40; catesbyana 34, 37, 38, 39; coactilis 34, 35, 40, 42, 438; crispa 34, 36, 41, 45, var. walteri 45; dioica 56; dioscoreifolia 34, 36, var. robusta 36; glaucophylla 34, 36, 41, 45, 46; ligusticifolia 37, 38, 39; maximowicziana 34, 36; micrantha 37; occidentalis 34, 35, 37, var. occidentalis 37; ochro- leuca 38, 35, 37, 40, 41, 42, 43; paniculata 34, 36; pitcheri 34, 36, 41, 44: reticulata 34, 36, 41, 44; versicolor 34, 36, 41, 44, 45; verticillaris 34, 37, var. cacu- minis 37; viorna 34, 36, 45, var. flaccida 45; virginiana 34, 36, ‘ var. cacuminis 37; viticaulis 34, 35, 40, 42, 48 Coccothrinax argentata 56; argen- tea 56 Coldenia 235; atacamensis 236; can- escens 236, var. pulchella 236; conspicua 236; cuspidata 236; elongata 236; ferreyrae 236; his- pidissima var. latior 236; litoralis 236; mexicana 236; nesiotica 236; nuttallii 236; palmeri 236; par- onychioides 236; plicata 237; pro- cumbens 235; purpustt 237; simu- lans 237 Commelinaceae 323 Compositae 31, 55, 60, 114, 123, 221 Condalia 317 Convolvulaceae 55, 216, 325 Convolvulus carolinus 216, 217 Coreopsis 123; angustata 204; an- gustifolia 195; auriculata 126, 133, 138, 139, 140, 145, 146, 156, 178, var. B 138, var. 6 148, var. y 148, var. diversifolia 138, var. glab- 156, 160, 176, 184, 187, 192, 198, wrightiit 130, 1383; boykintana 152; callosa 187; cardaminifolia 207, var. angustiloba 207, var. lineari- loba 207; corninsularis 151, 152; coronata 136: crassifolia 140, 143; debilis 151, 152; X delphinifolia 128, 140, 167, 170, 171, 175, 176, 177, 178, 179, 181; delphinifolia 177, var. chlooidea 177, var. del- phinifolia f. concolor 177, 178, ar. linearis 177, var. rigida 170; dichotoma 195; discolor 176; di- versifolia 133, 138; drummondii 130, 138, var. wrightii 180; fal- veyana 127, 145, 152, 157, 158, 159, var. longipes 127, 152, 156, 157, 158, 159, 160, 166, 187, 204, var. subintegrifolia 152, harvey- ana 158; helianthoides 195, 196, 198; heterogyna 141; heterolepis 158; heterophylla 152; integrifolia 129, 196, 201, 202, 203; intermedia 192, 198, 201, 207; lanceolata 126, 140, 141, 142, 143, 144, 145, 146 2 > 151, 152, 159, 166, 167, 176, 187, 207, 212, var. angustifolia 141, var. crassifolia 141, var. glabella 339 140, var. succisaefolia 141, var. villosa 140, 141; latifolia 127, 146 2 199, 204, 207; longifolia 195, 196, var. godfreyi 195, 196; longipes 159; major 171, 173, 174, 176, ane, 9 var. rigida 170, 171, 174, 176, var. , 133, 176, 184, 185, 186, 192, 198, 201, 207, 210, f. leucantha 184: saxi- 336 cola 154, var. duncanii 154; saxi- coloidea 187, 192, 198; senifolia 170, var. rigida 170, var. stellata 170; similis 210; stellata 170; stenophylla 207; tenuifolia 179; testudinea 151; tinctoria 199, 204, var. atkinsoniana 133, 207, var. atropurpurea 207, var. imminuta 207, var. similis 130, 138, 184, 206, 210, 211, var. tinctoria 130, 133, 209, 210, f. atropurpurea 207; tripteris 128, 146, 156, 160, 162, 163, 164, 165, 166, 167, 170, 171 176, 177, 178, 181, 184, var. 163, var. deamii 163, var. inter- cedens 163, var. smithii 163, 166, var. subrhomboidea 168, 166; verticillata 128, 166, 167, 170, 171, 716, 177, 178, 179, 181, 182, 183, 199, var. linearis 176, var. tenui- folia 179; wrayt 170; wrightii 125, 130, 131, 132, 1383, 140, 176, 184, 198, 204, 207, 210 Coreopsoides 130; lanceolata 140 Cornus 117, 311; florida 225 Corylus americana 225 Cranichis muscosa 53 Crotalaria falcata 323 CROVELLO, THEODORE: J. 242 Cuphea 81, 100; asper 101; cartha- genensis 101; glutinosa 80, 101, 102; hyssopoides 101; 102; procumbens 102; viscosissi- ma 101, 102 Cuscuta globulosus 55 Cuscutaceae 54 Cynoctonum mitreola 325; sessill- folium 324 Cyperaceae 53, 322 Cyperus surinamensis 322 Cyrtandra 121 petiolata 04: umbellata Cyrtopodium punctatum 53 Dalea 328: mollis 328, 329, 330, 331; mollissima 328, 329, 330; neo- mexicana var. longipila 328, 329, 330, var. megaladenia 328, 329, 330, var. neomexicana 328, 329, 330; simulatrix 329, 331; verna 329, 330 Dalechampia scandens 53 Daucus 249 Decodon 81, 82; verticillatus 88, var. laevigatus 82 DEGENER, OTTO & ISA 122 Delphinastrum carolinianum 258; exaltum 255; tricorne 258; vires- cens 260 Delphinium 243; ajacis 247; ala- bamieum 246, 247, 250, 251, 252 258, 258, 263; ambiguum 243, 245, 246, 247, 249, 263; azureum 258; carolinianum 244, 245, 246, 247, 253, 258, 259, 262, 263; elatum 252: exaltatum 243, 244, 246, 247, 252, 256, 263; tricorne 2438, 246, , 253, 254, 263; tridactylum 255; urceolatum 252; virescens 245, 246, 247, 260, 261, 2638, var. virescens 262; vimineum 244 Deschampsia caespitosa 49 DESELM, H. R. 115 Desmodium ciliare 324, tortuosum Dichromena floridensis 322 Didiplis 80, 86; diandra 86, 87 Digitaria adscendens 56; ciliaris 56; sanguinalis 56 Diplosastera 203; tinctoria 206 Draba 19 Dryropteris 224; celsa 225, 226, 228, 230, 232; clintoniana 230; cristata 224, 225, 228: goldiana 226, 230, 232, 233; intermedia 228; X leedsii 224, 226, 228, 230, 231, 232, 233; ludoviciana 233: marginalis 225, 226, 228, 230, 232, 233, XK neo- wherryi 226, 232; spinulosa 225, 228; separabilis 225, 230; wherryi 230 Dyschoriste oblongifolia 325 Earleocassia roemerana 7; vogeli- ana 8 Eddya gossypina 236; hispidissima 236 Elymus 117; junceus 117 Eragrostis pilosa 321 Erigeron bellioides 55 Eriochloa 117 FEricneuron 317 Euonymus americana 225 Euphorbia 313, 324; longinsulicola 313, 314, 315; mesembrianthemi- folia 314; ophthalmica 324; vagi- nulata 314 Euphorbiaceae 53, 57, 324 Fatoua villosa 116 Festuca 117; arundinacea 117 Filago gallica 222; germanica 221; pyramidata 222, 223; spathulata 222; vulgaris 221, 222 Fimbristylis annua 53 Flacourtiaceae 57 FLOOK, JERRY M. 114, 116 Foeniculum 57; foeniculum 57; vul- gare 57 Fraxinus americana 225 FRYXELL, PAUL A. 1 Gaillardia 319 Galapagoa darwinii 236; fuscua 236 Galium 59; aparine 59; bermudense 59; hispidulum 59; pilosum 59 Gentianaceae 114 Gesneriaceae 121 GILLIS, WILLIAM T. 52 Gnaphalium germanicum 221; hel- leri 30, var. helleri 30, 31, 32, var. micradenium 32; polycephalum 337 30, var. Bp 30, var. helleri 30 Goodeniaceae 326 GRAHAM, SHIRLEY A. 80 Gramineae 53, 56, 114 Haloragidaceae 59 Haplopappus 156 HARDIN, JAMES W. 119 Hedwigia ciliata 226 Hedysarum tortuosum 57 Helianthus argophyllus 326 Helosciadium ammi 58 Hemigraphis 115; reptans 115 HESS, WILLIAM J. 48 Hibiscus esculentus 114 HILL, STEVEN R. 313, 321 Hydastylus 49: elmeri 49 Hydrangea arborescens 225 Hydrilla verticillata 114 Hydrocharitaceae 114 Hydrocotyle erecta 58 Ipomoea 325; carolina 217; carolini- ana 217; commutata 217; fistulosa 325; lacunosa 216, f. purpurata 219; nil 55; trichocarpa 216, 217, 218, var. australis 220, var. ber- landieri 219; var. torreyana 216, 217, 218, 219, var. trichocarpa 216, 217, 218, - 220; trifida 216 IRWIN, 5 Jacobina spicigera 325 JOHNSTON, LAVERNE A. 67 Juniperus 75; communis var. de- Labiatae 115 Lagerstroemia 81, 82; indica 82 Larrea 317 Lauraceae 19 Leachia 130; lanceolata 141; trifoli- ata 138 Leavenworthia 252 Leguminosae 56, 114, 323 Leiodon 160; latifolium 160 338 Lentibulariaceae 55, 325 Leptochloa 117; uninervia 53 Lindera benzoin 225 Loganiaceae 324 Ludwigia palustris 83 Lycopersicon esculentum var. cer- asiforme 325 Lythraceae 50, 114, 324 Lythrum 80; acinifolium 88; alatum 80, 87, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, var. alatum 98, 99, var. lanceolatum 80, 88, 89, 90, 91, 98, 99, 100; borysthenicum 86; californicum 89, 98, 100; cor- difolium 90, 91, 95, 98; curtissii 87, 89; dacotanum 90, 91, 93, 98; flagelare 87, 88; lanceolatum 80, 90, 91, 93, 94, 98, 100; lineare 87, 88, 89, i100; salicaria 87, 100, 114: thesioides 86; thymifoha 86: tri- bracteatum 86 Macradena lutescens 53 MAHLER, WM. F. 30, 328 Malachra alceifolia 54; capitata 54; urens 54 Malvaceae 54, 114, 324 Malvaviscus arboreus var. cuben- Sis 54 MASSEY, J. R. 116 Mastichodendron foetidissimum 58 Melilotus indicus 324 Melochia tomentosa var. frutescens D4, var. tomentosa 54 Mcraceae 116 Nephrolepis exalta 52; hirsutula 52: multiflora 52; rivularis 326 Nerisyrenia 319 Neyraudia reynaudiana 321 Nymphaea elegans 323 Nymphaeaceae 323 Oncidium floridanum 53; luridum 93; Sphacelatum 323 Onosmodium molle 252 Orchidaceae 53, 323 we) Palmae 56 Palmerocassia 14; painteri 16; vil- losa 16 Panicum caerulescens 56; ciliare 56; coerulescens 56; commuta- tum 322; joorii 322; portoricense 322; sanguinale 56 Parsonsia 101; balsamona 101; ly- throides 101 Paspalum arundinaceum 322; dila- tatum 322; distachyon 322; fili- forme 322; pleostachyum — 322; reeanum 322 PEARMAN, RONALD W. 19 Pedicularis angustifolia 49, 50; angustissima 49, 50 Peplis 86; diandra 86 Perityle 311i; ecarmensis 311; castil- loni 311 Persea 19; borbonia 19, 21, 22, 23; humilis 19, 20, 21; littoralis 19, 21, 22, 23; palustris 19, 20, 21 Petalostemon gattingeri 252 Phlox paniculata 117 Phyllanthus amarus 57; niruri 57; tenellus 324 Physocarpus 311 Physostegia 115 Picea engelmannii 48, 50; pungens 48 Pilostavis 285, 290; arenicola 308; baldwinti 299; carteri 300: cymosa 291; hyemalis 305; lutea 302; nana 305; ramosa 293; rugelii 301 Pimpinella leptophylla 58 Pinus engelmannii 48; flexilis 48; ponderosa 48; strobiformis 311 Poaceae 321 Polygala 284; acutifolia 291; areni- cola 307, 309; attenuata 291; bald- uinil 285, 287, 288, 290, 293, 295, 296, 297, 298, 299, var. balduinii 287, 288, 289, 290, 296, 298, 299, 300, var. carieri 284, 287, 288, 289, 290, 296, 298, 299, 300, var. chlorogena 293, 295: baldwinii 299; carteri 297, 299; corymbosa 291, 293; cymosa 285, 286, 287, 288, 289, 290, 291, 292, 293, var. graminifolia 291; — graminifolia 291; lutea 285, 286, 287, 288, 289, 290, 301, 302, 304, 307, var. elatior 302, var. nana 305, var. pauci- flora 302; nana 285, 286, 287, 288, 289, 290, 291, 302, 305, 306, 307, 309, var. humillima 305; poly- cephala 299; pseudosenega 302: 290, 293, 294, 298, reynoldsiae 301: rugelii 285, 287, 288, 289, 290, 300, 301, 302, 303; smallii 284, 285, 286, 289, 291, 302, 307, 308, 309: vul- garis 284 Polygalaceae 284 Polymnia canadensis 226 Polypodiaceae 52, 326 Polypodium 311 Polyrrhiza lindenii 53 Polystichium acrostichoides 226, 228 Ponthieva racemosa 53 Populus tremuloides 48 POWELL, MICHAEL A. 311, 317 PROCTOR, GEORGE R. 52 Proserpinaca palustris 57; platy- carpa 957 Prunus serotina 68 Pseudoclappia 317; arenaria 319: watsonii 317, 318, 319 Pseudotsuga 311; menziesii 48 Psoralea subacaulis 252 Ptelea trifoliata 68 Quercus gambelii 48; muehlien- bergil 225; rubra 225 Ranunculaceae 33, 56, 266 Ranunculus 263; abortivus 266, 268, 270, 271, var. indivisus 270, 271 acris 267, 269, 274; allegheniensis 267, 268, 271; ambigens 267, 270 277, 278, 279, 280; apricus 275; aquatilis var. capillaceus 281; arvensis 269, 272, var. tubercula- tus 272; bonariensis 280; bulbosus 267, 269, 272, 274, 275, ssp. ad- scendens 274, ssp. bulbosus 274, var. dissectus 274, var. valdepu- bens 274; carolinianus 267, 270, 276, 277; cymbalaria 270, 280; delphinifolius 273; fascicularis 269, 275, 276, var. apricus 275: ficaria 267, 270, 281; flabellaris 267, 269, 273; harveyi 267, 268, 271, var. pilosus 271; hederaceus 270, 280; hispidus 267, 269, 275, 276, var. eurylobus 275, var. falsus 275, var. marilandicus 275; hydro- charoides 49; laxicaulis 267, 270, 279, 280; lindheimeri 277, 279; longirostris 267, 270, 281; macounii 269, 273; marginatus 267, 269, 272, var. trachycarpus 272; micranthus 267, 268, 271, var. cymbalistes 271, var. delitescens 271, var. micranthus 271: mis- sissippiensis 279, 280; muricatus 268, 272; oblongifolius 279: ob- tusiusculus 279, 280; palmatus 277; parviflorus 268, 272; parvulus 272; pensylvanicus 267, 269, 273; platensis 267, 269, 271, 272, 273; pusillus 267, 270, 277, 278, 279, var. angustifolius 279, var. pusillus 279; recurvatus 269, 273, var. adpressipilis 273; repens 267, 269, 273, 274, var. pleniflorus 274; sardosus 268, 269, 272; scleratus 269, 273; septentrionalis 267, 270, 275, 276, var. caricetorum 275, var. pterocarpus 277; subcordatus 267, 270, 278, 280; tener 277, 279, var. angustifolius 277: texensis 279, 280; trachycarpus 272: tri- 340 chophyllus 270, 281; trilobus 268, 272 Rhamnus 67; fasciculata 67, 72, 77, lanceolata 67, 68, 69, 70, 76, 77, 78, 79, var. glabrata 71; parvifol- ius 70: serrata 67, 68, 69, 72, 76, 77, 78, 79, var. guatemalensis 68, 69, 74, 79, var. serrata 68, 69, 72; serrulata 72; shortit 70; smithii 67, 68, 71, 76, 77, 78, 79, ssp. fasciculata 67, 72, 77, var. mulleri 72. 77, ssp. typica 67; standleyana 68, 69, 76 Rhynchospora_ elliottii 322; inter- mixta 322 RICHARDSON, ALFRED 235 Rotala 81, 82; indica 80, 88, 103; ramosior 83, var. interior 83 Rubiaceae 09 Rudbeckia subtomentosa 114 Sabatia brachiata 114; campanu- lata 114 Sageretia lanceolata 70 Salix 48, 117; fragilis 66 Sambucus neomexicana 48 Sapotaceae 58 Sarcomphalus shortianus 70 Sartwellia 314: flaveriae 319; gyp- scphila 319; puberula 319 Scaevola taccada 326 Schizachyrium stoloniferum = 115, var. wolfei 114, 115 Schoenolirion croceum 252 Scirpus 117; atrovirens var. geor- gianus 117; georgianus 117 Scrophulariaceae 59 Selinocarpus 31! Sesbana vesicaria 324 Setcreasea purpurea 323 Sida 1: anomala 1, 3; ciliaris 3, var. involucrata 54; collina 4, 95; corymbosa 4; costata 4; fulva 1, meridiana 1, 2, 3; rhombifolia 4, 5. 324: sidarum 1; woronowii 4 Sisyrinchium demissum 49; elmeri 49: longipes 49 Smilax 225 SMITH, EDWIN B. 123 SMITH, ROBERT R. 284 Solanaceae 58, 325 Solanum aculeatissimum 58; ameri- canum 58 59, var. nodiflorum 58, 59: capsicoides 58; ciliatum 58; nigrum 58, 59 Sophora tomentosa 56, ssp. baha- mensis 56, 57 Spermacoce confusa 59; tenuior 959 Spilanthes iodiscaea 55 Spiranthes cranichoides 53; elata 53 Sterculiaceae 54 STUCKEY, RONALD L. 24 STUESSY, TOD F. 104 Syntherisma sanguinalis 56 Tamaricaceae 57 Tamarix canariensis 57; gallica 97 TAYLOR, CARL W. 224 Tetramicra canaliculata 53 Tharpia 8 Thevetia 58; peruviana 58 THIERET, JOHN W. 115 Thrinax microcarpa 56; morrisil 56 Tiquilia 285; atacamensis 236, 240, brevifolia var. plicata 237; can- escens 236, var. pulchella 236; conspicua 236; cuspidata 236; darwinii 236: elongata 236, 238, 239, 240: ferreyrae 236; fusca 236; eossypina 236, 237, 238; hispidissi- ma 236: hunteri 238, 239, 240; latior 236; litoralis 236, 238, 239, 240: mexicana 236; nesiotica 236; nuttallii 236; palmeri 236; par- onychioides 236; plicata 237; pur- pusii 237; simulans 237: taenensis 238, 239, 240; tuberculata 237; turneri 237, 238 Tithonia diversifolia 326; tubae- formis 326 Tournefortia stenophylla 55; volu- bilis 55 Toxicodendron radicans 225 Tribulus terrestris 54 Triphora gentianoides 53 Tropidia polystachya 53 TURNER, B. L Umbelliferae 57 Utricularia 55; fibrosa 325; foliosa 5d; gibba 55; purpurea 325 Verbena bonariensis 325 Verbenaceae 325 Vernonia 19, 6 arbuscula 60; bahamensis 60; blodgettii 60; cin- Barney L Lipscomb Herbarium Southern Methodist University Dallas, Texas 75275 341 erea 60; insularis 60; obcordata 60 Veronica americana 49 Vicia hirsuta 114 Viorna addisonii 45; baldwinii 40; beadlei 41; crispa 41; flaccida 45; gattingeri 45; glaucopylla 45; obliqua 41, 45; ochroleuca 37; pitcheri 41; reticulata 41; sub- reticulata 41; viorna Zuelania 57; guidonia 57 Zygophyllaceae 54 Oo