WASH V-7 39 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY. New Series.—LXX. I. craw in the Boraginaceae.—II. . A Synopsis of the American native and immigrant Borages of the Subfamily Boraginoideae . . 2. A tentative Classification of the South Anierican Coldenias . II. Taxonomic Records concerning i has Spermatophytes.. 1. Parkinsonia and Cercidium . ‘ . New or otherwise noteworthy Plan mG eo oe . A neglected Paper by Jean Louis Bicandier . eS > 5 : On the Validity of Molina’s scientific Names. . . - 90 By I. M. JoHnston. PUBLISHED BY } THE GRAY HERBARIUM OF HARVARD UNIVERSITY e CAMBRIDGE, 38, MASS., U.S. A 3 = = * ° MissouR! BOTANICA _ : ARDEN hI BRARX I. STUDIES IN THE BORAGINACEAE,—II. 1. A SYNOPSIS OF THE AMERICAN NATIVE AND IMMIGRANT BoRAGES OF THE SUBFAMILY BoRAGINOIDEAE. Tue following summary of the American Boraginoideae is the result, first, of a critical examination and study of the generic lines within the subfamily, and, second, of a careful study of the bibliography of the group and a serious attempt to identify as many as possible of the poorly understood or unidentified species. The data accumulated have been fashioned into a synopsis containing new keys to both genera and species, and such precise information as to specific distribution as available specimens and reliable records will permit. For the ter- ritory south of Mexico the occurrence of introduced species has been given by countries, but to the north the occurrence has been given by states. An attempt has been made to distinguish between reports based upon published records, and those founded upon specimens personally examined, all records of the latter sort being indicated by the exclamation sign. No attempt has been made to treat the genus Cryptantha which is being reserved for a subsequent paper. Such genera as Amsinckia or Plagiobothrys which have had recent revision are not treated in detail; instead, merely the reference to the latest published review has been cited. The study of the South American members of the subfamily has been practicable only through the courtesy of Dr. W. R. Maxon and Dr. N. L. Britton who have loaned for my study, in conjunction with the material in the Gray Herbarium (G), that from South America contained in the United States National Herbarium (US) and the New York Botanical Garden (NY). Invaluable for my study of Pectocarya was the Southern Californian material of that genus con- tained in the Baker Herbarium of Pomona College (P) and loaned me by Dr. P. A. Munz. . Key To Re aes strongly convex or plug-like, ieaving a es upon the age or low-convex gynobase. Anchus Stamens appendaged dorsally, oassy crowded about thestyle; te, s Oe cried within tha dood a Sec ese 1. Borago. 3} rant 2 J Stamens ens unappendaged tubular — "asic faucal appendages or these represented by Seek 9 a oe RE Con ee Lei sw 2. Nonea. us spots Corolla with ec sg aa faucal appendages formed by the in- 4 JOHNSTON oS pple tubular; throat conspicuously developed anulate-dilate d; lobes short, erect or with og 2 maeeed: faucal piety sade lanceolate, acute, margins ee tt 3. Symphytum. Corolla funnel- or salver-form; throat ill-defined, ab- ruptly expanding or absent; lobes usually knits, spreaii ing or divergent; Aneel spear deltoid tusish, usually hair Corolla-tube distinetly Sent near 1 igh, limb subir- Ws es ee ne 4. Lycopsis. Corolla-tube straight, limb perfectly regular, not eee hs oot 1 eens fhe Kae ae 5. Anchusa. Attachment of nutlet without. an annular a Ssh or somewhat a avi Stigmas geminate or style bifid. Eoeperssiic Co tat hg more or less irregular, oblique; stamens unequal...... 6. Echium. more: E peenine or only very obscurely irregular, stamens Calyx Mew iediion’, merely pa git hisee ta grade triangular, t half length of tube, usually connivent over frui nits one anete ormally Alclad “this persistent and falling endlostd i in calyx; pubescence in part o uncinate hairs Inflorescence nake VRS OIG eo oe et es 7. Moriizia. Inflorescence leafy-brac $60, OPER, 55 es 8. 1 secant Calyx cut to near base; lobes ih or lanceolate, much onger tha ai ube; four nutlets normally developed; these falling separately and not enclosed in the oi ch is persistent; pabestence of straight Corolla hg large, 2.5-8 em. long, lobes acute; stamens ong, reaching at least to corolla-sinuses ane fre equen ntly Mnteh POVONE. ie yey ec: + Macromeria. Corolla of small or r medium size, less than Had long; stamens very short, included cons not ap- proachin prcliasaieione Filaments ligulate; anthers Benpily hirsute dorsally; corolla-lobes short, rounded, erect...... 10. Lasiarrhenum. Filaments filiform; anthers glabrous. Corolla: cute or ac Nakane erect; style long-exserted, protruded as the buds eaark anth rs sagitta 7 ea Te app aot eee a, Se Ue 1. Onosmodium. Corolla-lobes rounded or obtuse, | ascending ‘or rted, never rotruded cai flower is sexe pened; anthers oblon, y tuber- nel ecicnes. leaves ocr rate... .12. Lithospermum.. eee attached suprabasally and obliquely a pyramidal gynobase or attached to a yes USA OPP tam, STUDIES IN THE BORAGINACEAE _ 5 pies > i seal es — eee, capitate or disk-shaped, occa- nally « ema ivaricate or distinctly ge tleonep anterior ee bogs hp a right angle or a very broad acute angle with the floral axis. Cynoglosseae. i h glochidiate filiform or subulate appendages, eho eo and then merely by a wrinkle in the ' Nutlets lenticular, very depressed; cauline leaves on slender petioles at least half as long as the sya cordate blades; plants weak and trailing... oe Mimophytum, whic oblan melabe blades; plants erect............ 15. Cynoglossum. Dorsal surface of nutlets naked or occasionally covered ceous margin that is entire or serrate “A with uncin- margins nocragad armed with uncinate hairs 16. Pectocarya. iting i various, usually flex xuous, ~ much longer pears nutlets; corolla conspicuous, subrotate — or rotate-tubular; nutlets with cupulate margins hich ee uncinate hairs or teeth............ 17. Omphalodes. Nutlets — ar base or middle, more or less erect and parallel, aintien face paralleling ras axis or ormie: a very narrow acute angle neg —— calyces mgr eiatoreniat irregular r, ree lobes nearly distinct, the other united, Poo sing the fruit and Reonsaiie, orm a with 7-9 long glochidiate processes ; ovule es 2 bel se: Petgea se gas 18. Harpagonella. it tically so, not armed with comute “glochidiate | provemes: ovules usually 4. ritra Constaisuas convolute in the pene herbs with a ebracteate racemes an oth narrowly and ud. utlets 2, each commonly age Sag ee sagittate with auricles usua ; rather rank herbs with Prin deeply eta sessile ee SE as Seca ages la ea Dee eae 20. Cerinthe. Nutlets usually 4, these normally 1-celled: anthers not sagittate nor appendaged; leaves not cor- Fruiting calyx strongly accrescent, Pans veiny. irregularly toothed an d lobed, plie te....... 21. Asperugo. ¢ oderate at all no cuously At toothed or lobed, never plicate. Cotyledons 2-lobed; corollas unappendaged, range OF SUUOW. oe oo ose ee 22. Amsinckia. JOHNSTON Sale sien unlobed; corollas almost always aged, white or blue or very tel ria ‘yellow: ish. Shrub 9-1 | tl 1] ch ay for. pearly ‘ta whole fe eng th, margin lacerate, dorsum with eisehinliate appendag a Pe Pe PYG | oo ade eae 23. Selkirkaa. Herbaceous oe ada suffrutescent plants, less than 9 hi Nutlets with a Y dagniie medial ventral groove formed by the non-fusion of the pericarpial walls. .............. 24. fet Liceeals 25 Nutlets with the pericarpial walls fused east above the middle and com- urned rim or flange, almost always without glochidiate append- ages. Corolla white, hd esi! short _ mde llow, tube eeded by o we aig just ex shoadia calyx; sae et usually with a Salad dorsal eae ‘ile, aR shorter than wileie ee 26. Plagiobothrys. Caoclla trues throat cylindrical i fun “hing 2 oe pes nutlets usu soaty lacking a a mtiaL esa ‘gilt style ally greatly e cit Og nutlets. ae Mertensia. Gia of ues encircle tur: rim or flange which i ‘é pean toothed or ail commonly with a = hairs or glochidiate ap- Rte te flat; nutlets definitely tet- rahedral, "attached basally (at through a suprabasal substipitate oon aa ep Soe tbls oleh cays 28. Trigonotis. Gynobase idal or subulate; nutlets not at - tetrahedral, attached vent Nutlets eaualing the e gyno- = ar lacking a definite la; style dp barged nutlets; racteate annuals............. Lappula. Nutlets twice surpassing the stout gynobase STUDIES IN THE BORAGINACEAE 7 dp calyces erect or ascend- ing; low densely oe silky-pubescent plants with mall firm leaves........ 30. Eritrichium. ne ealyces reflexed; rank glabrous or sparsely OB. cau a 31. Hackelia. 1. Borago [Tourn.] L. Ba bag Bag Bs 4 ees Gen. Pl. 67 (1754). Borago officinalis L. Sp. Pl. 137 (1753).—Native of the Mediter- ranean region. Introduced in N. S., Me.!, N. H.!, Mass.!, Conn., N. Y., Penn., D. C., Va., Ont., Tenn., Ill.!, O., Mich., Wis., N. DD: Ore., B. C., Vera Cruz!, Jalisco, Nicaragua, Ecuador, Bolivial, Chile, Uruguay, Argentina. 2. Nonea Medik. Phil. Bot. i. 31 (1789). Key To Sprcizs. Nutlets i sare na attached on side, longer than tall, with the Fei coarse annular basal rim papillose-dentate ee ee aa tS ees Se ey 1. N. vesicaria N utlets a ety attached at or near the base, oti et and n sil coher tat fatlets brown, nearly eortial, Gay slightly oblique, sub- terete, very obscurely rugose, greatest breadth ca. 24 height; plant commonly Heap only near base; dorelie A VOW oer oe Be ee 2. N. lutea. Nutlets 9 hhteeiiny steels magpie y padi oblique, d, conspicuously eatest breadth nearly equalling he: on ely braictied: above; corolla rose-colored at least upon opening..................... 3. N. rosea. 1. Nonea vesicaria (L.) Reichenb. Fl. Germ. Excur. i. 338 (1831). Lycopsis vesicaria L. Sp. Pl. 138 (1753). LEchioides nigricans Desf. Fl. Atl. i. 163 (1798-1800). N. nigricans DC. FI. Fr. iii. 626 (1805).— Native of the western Mediterranean region. Reported, Mem. N. Y. Bot. Gard. v. 525 (1915), as adventive in the vicinity of New York. 2. N. lutea (Desr.) DC. Fl. Fr. iii. 626 (1805); Bornmiiller, Bull. Herb. Boiss. ser. 2, vii. 780 (1907). Lycopsis lutea Desr. in Lam, Dict. iii. mae (1791). L. ciliata Willd. Sp. Pl. i. 780 (1797). N. ciliata L. setosa Lehm. Asperif. ii. 269 (1818). N. setosa R. & S. Syst. iv. 154 (1819).—Native of southeastern Europe and adjacent Asia. Occasionally introduced in impure seed.’ Known from Penn.!, Ky.!, and N. Y.! 3. N. rosea (Marschall) Link, Enum. i. 167 (1821). Anchusa rosea Marschall, Fl. Taur.-Cauc. i. 125 (1808).—Native of southeastern Europe. Collected in a waste place at Bangor, Me., in 1902 by O. W. Knight. 8 ; JOHNSTON 3. Symphytum [Tourn.] L. Sp. Pl. 136 (1753): Gen. Pl. 66 (1754). Key To SPHCIES. Stems and cp ma covered with short bas Mew thornlet- or prickle-like hairs; tip of corolla-lobeserect.............. 1. S.asperum. Stems tna Setaceascnts glabrate or som cin mes hirsute, but Peogsee e not recurved and prickle-like; tips of corolla- obe a conspicuously decurrent; root branched, fusiform, out tuberous thickenings; nutlets lustrous, smooth, ‘ almost lacking the aie! thic oko ned base; base entir re. at S. officinale. tuberous ickeniies:: nutiete: dull, finely “aa ema with an annulate toothed base..................+.- 3. S. tuberosum. 1. Symphytum asperum Lepechin, Nov. Act. Acad. Petrop. xiv. 444, t. 7 (1805); Bucknall, Jour. Linn. Soc. xli. 510 (1913); Macbr. Rhodora xviii. 23 (1916). S. asperrimum Donn, Bot. Mag. xxiv. t. 929 (1806).—Native of the Caucasus region of southeastern Europe and adjacent Asia. Introduced in P. E. I.!, Que.!, N. S.!, N. B.!, Me.!, Vt.!, Conn.!, Mass.!, N. Y.!, Md., Mich.!, Wash.!, B. C. 2. S. officinale L. Sp. Pl. 136 (1753).—Native of temperate Europe. Introduced in Newf.!, Que., N. S.!, Me.!, N. H.!, Vt.!, Mass.!, R. [.1, Conn.!, N. Y.1, Md, Del, Do-C., Wi Va,; ¥a.), N. C.1, ee Ky., O., Ont., Wis., Ind., Mich., Mina. Wis., Mo. ., Mont., 3. 8. tuberosum L. Sp. Pi. 136 (1753).—Native = thiddic and southern Europe. Known in America only from Connecticut where reported from Southington! and Guilford. 4, Lycopsis L. Sp. Pl. 138 (1753); Gen. Pl. 68 (1754). Lycopsis arvensis L. Sp. Pl. 139 (1753).—Native of temperate Europe. Introduced in Que.!, N. B.!, N. S., Me.!, Mass.!, R. L.!,. N. Y.!, Penn.!, Va., D. C., O., Ont., Minn., Colo., Calif.!, Chile. 5. Anchusa L. Sp. Pl. 133 (1753); Gen. Pl. 64 (1754). Key To Spectzs. caper roma oblongish, nearly twice as tall as thick; calyx ar base, with nb . long, lo cong pp ‘linear- yx- 63 Nutlets horizontal, somewhat = less tall than long; calyx cut to near middle or below mm. long, lobes deltoid 4 ; corolla Sus tab Sita broad, appendages entirely pubescent. STUDIES IN THE BORAGINACEAE 9 Nutlets 1.5 mm. high, 2.2 mm. long; calyx-lobes select, evidently shorter than tube, tube frequently bladdery inflated in advanced fruiting; bracts linear to narrowly lanceolate; cauline leaves 3-8 or exc eptionally 13 mm. — inflorescence re a cylindrical panicle; a S25. Om Ione od coe t eee . A. capensis. N utlets 2 ory high, ‘3 m st long; calyx-lobes lanceolate < 6-12 ae ‘long re Pn rr RE meme Po Ae ren rae 3. A. officinalis 1. Anchusa azurea Mill. Gard. Dict. ed. 8, n. 9 (1768); Schinz & Thell. Vierteljahrs. Naturf. Ges. Ziirich liii. 557 (1908). A. ttalica Retz. Obs. Bot. i. 12 (1779.) A. paniculata Aiton, Hort. Kew i. 177 (1789). Buglossum elatum Moench, Meth. 418 (1794).—Native of the Mediterranean region. Material from Granby, Conn., and Salem, Ore., has been examined. It has been reported, Rhodora xxv. 28 (1923), from East Boston. 2. A. capensis Thunb. Prodr. Pl. Cap. 34 (1794).—Native of South Africa. Occasionally cultivated; occurring as an escape at Salem, Ore. 3. A. officinalis L. Sp. Pl. 133 (1753). A. angustifolia L. 1. ¢.; ee & Thell. Viertljahrs. Naturf. Ges. Ziirich liii. 557 (1908). A. officinalis, var. angustifolia DC. Prodr. x. 48 (1846); Macbr. Rho- dora xviii. 51 (1916). A. arvalis Reichenb. Fl. Germ. Excur. 343 (1831).—Native of temperate Europe and of Asia Minor. Introduced in Me.!, Mass.!, R. I.!, Conn.!, N. Y., N. J.!, O., Ut., Colo., Argentina. Attempts have been made to distinguish the var. angustifolia among the immigrant plants of A. officinalis. The variety, however, seems too unimportant and vague to warrant its recognition among the introduced material. “A. leptophylla Roem. & Schult.” is reported as a waif in the vicinity of New York by Taylor, Mem. N. Y. Bot. Gard. v. 525 (1915). The record may be based on a form of A. of- ficinalis L. since Boissier, Fl. Orient. iv. 154 (1875), suggests that Roemer & Schultes’s species may be only a narrow-leaved phase of A. officinalis. 6. Echium [Tourn.] L. Sp. Pl. 139 (1753); Gen. Pl. 68 (1754). Key To SPECIES. Basal leaves ovate or broadly ~— a nervation mani- fest; upper leaves narrowed from a broad cordate sub-' am plexie: aul spore — sy a uit texture, — or bat soeieny Sie. ee et ae — Basal leaves ir Alene secondary nervation obscure; eaves not cordate nor subamplexicaul ; co: ga of sadieoe Geis firm usually pubescent as sa ges 10 JOHNSTON Stamens very unequal, a pair conspicuously exserted and a shorter one subincluded; plants with an obscured central xis, the long branches rebranched................ 2. FE. Coincyanum. Stamens in two subequal exserted pairs; plants branched at base, central axis definite, with its laterals short and almost invariably unbranch Inflorescence a close narrow elongate thyrse 3-7 cm. thick; : a I oe & 1 J] I a RG g 1 4 e erbage with inconspicuous or no pustulate bases............. 3. E. vulgare. Inflorescence a very loose broad thyrse 7-30 cm. thick; erbage usually with very conspicuous pustulate bases to the very coarse spreading hairs............. da. H.». pustulatum. ® 1. Echium plantagineum L. Mant. Alt. 202 (1771): Coincy, Jour. de Bot. xiv. 328 (1900); Lacaita, Jour. Linn. Soc. xliv. 420 (1919). E. violaceum of authors, not L. (1767); Coiney, |. c.; Lacaita, 1. €. 423. E. bonariensis Poir. Eneye. viii. 674 (1808).—Native of the Medi- terranean region and the Near East. At an early date introduced into South America where now widely distributed over the states of Buenos Aires and Entrerios in Argentina, over all of Uruguay, and over parts of extreme southern Brazil. In North America reported merely as adventive in the vicinity of New York City. 2. E. Coincyanum Lacaita, Jour. Linn. Soc. xliv. 374 (1819). E: australe of Coincy, Jour. de Bot. xiv. 326 (1900), and Rouy' & Fouc- Fl. Fr. x. 309 (1908).—Native of the Spanish peninsula. Only two American collections have been seen, one made at Farmington, Me., in 1882 by Keyes, and the other made near Cambridge, Mass., in 1913 by Fernald & Long. The latter collection has the corollas merely setose and may not be correctly determined. 3. E. vulgare L. Sp. Pl. 139 (1753); Lacaita, Jour. Linn. Soc. xliv. 426 (1919).—Native of Europe. Long ago introduced into eastern United States and now a serious pest in the New England and Middle Atlantic states. Known from Que! N. 8.1, N. B.!, Ky., Tenn., Wis., Mich., Iowa, Mo.!, Neb., S. D., Kan., La., Texas, N. M. Reported from Chile by Reiche, Fl. Chile v. 240 (1910). 3a. E. vulgare, var. pustulatum (Sibth. & Sm.) Coiney, Jour. de Bot. xiv. 301 and 303 (1900). E. pustulatum Sibth. & Sm. Fl. Graec. 11. 68, t. 180 (1813); Lacaita, Jour. Linn. Soc. xliv. 386 (1919).—Native of the Mediterranean region. Adventive in America and known only from N. J.!, Penn.!, D. C.?, and W. Va.! Old mutilated plants of the typical phase of E. vulgare frequently take on the loosely branched habit of this variety, but should not be confused with it STUDIES IN THE BORAGINACEAE ll 7. Moritzia DC. in Meisn. Genera i. 280; ii. 188 (1840). Meratia A. DC. Prodr. x. 104 (1846). Kry To SpPEcIEs. See repel tyle 5-7 mm. long, greatly surpassing fruit as piste. corolla-tube ca. 5 mm. long, clearly exceed- Cale corolla-lobes Oblong-dettotd: corolla-a appendages ila — inflorescence shaggy-pubescent; calyces “ ca. er a Sntliguntes semicircular to short-ovate; nee opedicts represented merely by circular bearded areas, not ligulate nor protruded; inflorescence a sen gose; lowermost calyces with pedicels...................2.] 2. M. Lindenii. 1. Moritzia ciliata (Cham.) DC. in Meisn. Genera ii. 188 (1840): Fresen. in Mart. Fl. Brasil. viii. pt. 1, 63, t. 10, fig. 13 (1857); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3, 121, fig. 471 and 47m (1893). Anchusa ciliata Cham. Linnaea iv. 440 (1829). M: ciliata, var. hirsuta Fresen. |. c.—Braziu: Porto Alegre, Reineck (G); Prov. Rio Grande do Sul, 1897, Reineck & Czermak 1 (US); without locality, Sellow 672 (US, 1soryPE of A. ciliata Cham.).—The species is known positively Ses from extreme southern Brazil. oc ae oe nii (A. DC.) Benth. acc. Giirke in E. & P. Nat. Pflan- zenf. iv. ae 3, 121 (1897). Meratia Lindenii A. DC. Prodr. x. 104 (1846).—CoLomBIA: wet mossy paramo, Paramo de Cruz Verde, near Bogota, alt. 3300-3500 m., Pennell 2052 (US, NY, G); moist grassy paramo, alt. 3400-3700 m., Cerro Tatama, Dept. of Caldas, Pennell 10576 (G); forest below Parnine del Quindio, alt. 3500-3800 m., Dept. of Caldas, Pennell & Hazen 10078 (G). VENEZUELA: Paves de la Sal, Mérida, 3400 m., Jahn 515, 592 (US).—The type of this very distinct species is given as having come from Caracas, Venezuela. Although Giirke attributes the accepted combination to Bentham, the latter appears to have done no more than indicate in the Genera Plantarum, ii. 859 (1876), that Meratia was synonymous with Moritzia. 8. Thaumatocaryon Baill. Bull. Mens. Soc. Linn. Paris 839 (1890). — TO ineeeay. a eS | . Chatine leaves ovate or ay Seago ‘ee pecker Contes leaves lance-oblong to ovate-oblong, base rounded. 2. T. tetraquetrum. 12 JOHNSTON Leaves alternate pec py ox tee the plant; nutlets (at least in no. muricate an Anthers penne asians exceeded by length of filaments; plants somewhat hispid, harsh to touch; lowermost leaves ane pape developed, = og em. long, mm. beoed: crete: 06.5 fan lone i ee 3. T. dasyanthum. Anthers and filaments about ‘sacally long; plants evenly erty smoothish; lowermost leaves not well developed, 3-7 cm. long, 8- 18 mm. broad; corolla ca. 4 mm. i a Sellowianum. 1. Thaumatocaryon Hilarii Baill. Bull. Mens. Soc. Linn. Paris 839 (1890). Antiphytum Bornmiillert Pilger in Fedde, Repert. iii. 24 (1906).. A. Bornmiilleri, var. asperior Pilger, 1. c. 25.—This species is known to me only through descriptions, and it is with much doubt that I accord it full recognition. According to Pilger, |. c., it differs from T. tetraquetrum in its broader leaves, thicker shorter inflorescence, and smaller calyx-lobes. The Saint-Hilaire and the Bornmiiller material comes from Rio Grande do Sul, Brazil. 2. T. tetraquetrum (Cham.), comb. nov. Anchusa tetraquetra Cham. Linnaea viii. 113 (1833). Antiphytum tetraquetrum DC. Prodr. x. 122 (1846).—BraziL: without locality, Sellow (IsoTyPE, US). Paracuay: Alto Parand4, Fiebrig 5885, 6131 (G, US); “Sierra de Maracaytt,” Hassler 5731 (G).—Al! precisely labeled material comes from eastern Paraguay and seems to be exactly matched by isotypic material. It is not improbable that Sellow’s original col- lection came from outside the present limits of Brazil. This species has been almost universally regarded as a member of Antiphytum, being so treated in my recent revision of that genus, Contr. Gray Herb. n. s. lxviii. 51 (1923). The species, however, is a rank-growin broad-leaved herb with a tubular calyx that permanently encloses the single nutlet matured. These developments are not known in the other species of Antiphytum, but they do occur in and characterize Moritzia and Thauwmatocaryon. It is hence with considerable con- fidence that I now transfer the species from Antiphytum. ee dasyanthum (Cham.), comb. nov. Anchusa dasyantha Cham. Linnaea iv. 437 (1829). Moritzia dasyantha Fresen. in Mart. FI. Brasil. viii. pt. 1, 63 (1857).—Braziu: without locality, Sellow 7964 (isorype, US).—A mature collection of this species is greatly needed, as the only available material is a duplicate of the original collection and like it lacks fruit. Until mature fruit is available the status of the species is somewhat problematic. It seems most nearly related to T. Sellowianum, but is a coarser plant with longer pubescence and larger corollas. STUDIES IN THE BORAGINACEAE 13 4. T. Sellowianum (Cham.), comb. nov. Anchusa Sellowiana Cham. Linnaea iii. 115 (1833). Moritzia Sellowiana Fresen. in Mart. Fl. Brasil. viii. pt. 1, 63, fig. 14 (1851).—Braziu: without locality, Sellow (IsotypE, US). 9. Macromeria Don, Edinb. New Philos. Jour. xiii. 239 (1832). Philonomia DC. in Steud. Nom. ed. 2, ii. 320 (1841). Kay To Spectns. Corolla with lobes reflexed or recurved at least during pollina- tion, evidentl irregular, neh one 4 une a mage lea e bu Filaments 3-6 em. long, about twice bongth of af corolinicbes, excessively veatinie and conspicuous; corolla very large, -8 cm. long; lower leaves obovate; anthers linear- eI BO RON i ee is a 1. M. exserta. Filaments ca. 1 cm. lon ng, about equalling or a trifle exceeding c — -lobes, ot _ gly and conspicuously exserted; coro medium ae ee . long; ae leaves lanceo- lat anthers Henight. eblone 2 2-2.5m POOR oe ic eke 2. M. hispida. Corolla sci lo erect or ascending and never reflexed or without any abaxial distortion. § Macromer Stems conspicuously hirsute; anthers linear 3-4 m - tached between the base and middle, usually erect...3. M. longiflora. Stems gpm short-strigose; anthers oblong or linear -oblong, 2. m. long, attached just below the middle, 9 Sag uppers surface of leaves merely scabrous, without hairs or NA the a BN dy Soe he oe wee: 4. M. discolor, seer pares of leaves not only scabrous but also strigose or somewhat p ate. — a Rey mm. long, appressed silky-villous outside; ip ety ORR ep ten ye ee ey es 5. M. Pringlet. ip eh "2. 5 mm. long, tomentose outside; Peruvian. 6. M. cinerascens. 1. Macromeria exserta Don, Edinb. New Philos. Jour. xiii. 239 (1832); Bot. Reg. xxxiii. t. 26 (1847). Echium longiflorum Sesse & Moc. Pl. N. Hisp. 20 (1887).—Mountainous portions of southwestern Mexico from Jalisco to Oaxaca. Most abundant on the Pacific slope of Oaxaca or at least most collected there. This species is readily distinguished by its very coarse habit, obovate lower leaves, and its very large showy somewhat irregular corollas with enormously ex- serted stamens. Its corollas are the largest in the family. The plants are usually abundantly shaggy-hirsute on the stems, but in M. exserta, var. imparata Macbr., Contr. Gray Herb. n. s. xlix. 22 or uy; the stems are closely short-strigose. . M. ida Mart. & Gal. Bull. Acad. Brux. xi. 339 (1844). a longiftora, var. hispida A. DC. Prodr. x. 68 (1846). Onosmodium 14 JOHNSTON longiflorum, var. hispidum Macbr. Contr. Gray Herb. n. s. xlix. 21 (1917). M. longiflora of A. DC. |. ¢., not of Don (1832).—Known only from Michoacan. 3. M. longifiora (Sesse & Moc.) Don, Edinb. New Philos. Jour. xiii. 239 (1832). Lithospermum longiflorum Sesse & Moc. Fl. Mex. 29 (1894). Onosmodium longiflorum Macbr. Contr. Gray Herb. n. s. xlix. 21 (1917). M. viridiflorg DC. Prodr. x. 68 (1846); Moc. & Sesse, Calg. Fl. Mex. t. 904 (1874). O. Thurbert Gray, Synop. Fl. N. Am. ii. pt. 1, 205 (1878). M. Thurberi Mack. Bull. Torr. Bot. Cl. xxxii. 496 (1905).—Middle and southern New Mexico and Arizona, southward into Chihuahua and Sonora. Macromeria viridiflora DC. is based upon a Sesse & Mocifio plate which evidently pictures the species which Gray called 0. Thurberi, the drawing and Gray’s type agreeing in possessing such distinctive characters as spreading pu- bescence on leaves and stems, leaves with numerous veins, very: pubescent corollas with erect greenish lobes, and stamens that just equal or barely surpass the corolla-tube. Macromeria longiflora Don, was based upon “ Lithospermum longiflorum, Sesse et Mocino, mss. in herb. Lamb.” The description of L. longiflorum by Mocifio & Sesse, Fl. Mex. 29, applies to the plant figured in their plates, and their species is hence a synonym oi M. viridiflora DC. Macromeria lungiflora Don was based upon material labeled by Mocifio & Sesse, but the exceedingly short description written by Don gives the leaves as “smoothish” whereas in M. Thurberi and in the Mocifio & Sesse plates they are notably rough. There is, hence, a possibility that Don’s plant is different from the one described and figured by Mocifio & Sesse. In their flora the latter authors give the type locality of L. longiflorum as “ad oppidum S. Rosae prope Guanaxuatum.” This station could scarcely be within the present state of Guanajuato since the plant they illustrate and describe belongs to a northern species not known south of the southern boundary of Chihuahua. I suspect that the type was collected in southeastern Sonora where they are known to have collected within the range of the species treated. 4. M. discolor Benth. Pl. Hartw. 49 (1840). Onosmodium discolor Macbr. Contr. Gray Herb. n. s. xlix. 20 (1917).—Southern Jalisco southward into Oaxaca. Readily recognized by its glabrous, yet scabrous, upper leaf-surfaces which dry a rich chocolate brown. 5. M. Pringlei Greenm. Proc. Am. Acad. xxxiv. 570 (1899). Onos- modium Pringlei Macbr. Contr. Gray Herb. n. s. xlix. 20 (1917).— Known only from the Sierra Pachuca in Hidalgo. STUDIES IN THE BORAGINACEAE 15 5. M. cinerascens A. DC. Prodr. x. 69 (1846).—Known to me only from descriptions and perhaps improperly referred to Macromeria. The type came from Chachapoyas, Peru. 10. Lasiarrhenum, gen. nov. Calyx 5-partitus, segmentis linearibus. Corolla tubulosa, recta, fauce nuda; lobi 5, erecti, breves, rotundi vel obtusi. Stamina 5, conniventia, tubo affixa, inclusa, filamentis brevibus ligulatis; an- therae compresse fusiformes, dorso hirsutissimae, loculis subdistinctis, connectivo conspicue producto acuminato. Ovarii lobi 4, distincti, erecti, subglobosi, gynobasi planae impositi. Stylus filiformis, tarde non praecociter) exsertus; stigmatibus 2 cohaerentibus parvis. Nuculae 1-4, erectae, nitidae, albae, areola plana gynobasi planae affixae. Semina recta, cotyledones planae.—Herbae mexicanae erectae perennes rudes hirsutae et scabrae. Folia ‘alterna, venis primariis valde elevatis. Racemi terminalis rami scorpioidei bracteati primum densi deinde elongati dissitiflori. Flores pedicellati albi vel incarnati. (Name from Aéstoc, hairy, and &genv, male, with refer- ence to the hirsute anthers. enum strigosum (HBK.), comb. nov. Onosma strigcsum HBK. Nov. Gen. et Sp. iii. 93 (1818). Onosmodium strigosum Don, Gen. Syst. iv. 317 (1837). Onosma trinervium Lehm. Asperif. ii. 378 (1818); Icones i. 11, t. 9 (1821). Lithospermum longifolium Willd. in R. & S. Syst. iv. 742 (1819).—Pacific slope of southern Mexico from Jalisco to Oaxaca. This species has been usually taken as a member of the genus Onosmodium, but its reference there destroys the homogeneity of that genus and leaves the latter almost characterless, whereas its removal as a distinct genus allows Onosmodium and all its immediate relatives to be sharply and very naturally defined. As an iss to the above proposal, the genera Onosmodium and Mace might be submerged in Lithospermum. This latter wnat seems quite unwise, however, since it would bring under a single genus a great diversity of corolla-structures, thereby setting a new and much lower value upon them in the classification of the tribe Litho- spermeae and causing far reaching readjustments among the very numerous Old World members of the alliance. The genus Lasiarrhenum is most nearly allied to Onocmodicm, that genus being taken as defined by Mackenzie, Bull. Torr. Bot. Cl. xxxii. 495-6 (1905). Lasiarrhenum agrees with its relative in having sagittate anthers and cylindrical corollas with erect lobes, but differs 16 JOHNSTON in having rounded corolla-lobes, ligulate filaments, large hirsute anthers, and shortly and tardily protruded styles. Onosmodium is a very natural genus characterized by its tubular corolla with erect acute or acuminate lobes, short narrow filaments, narrowly sagittate glabrous anthers, and precociously long Sete style that is ex- serted before the corolla is even a third o Macbride, Contr. Gray Herb. n. s. aes 19-21 (1917), has recently amplified Onosmodium so as to include in it the majority of the species usually referred to Macromeria and restricted the latter genus to M. exserta. With this arrangement I am unable to agree, since it destroys the homogeneity of Onosmodium and obscures its characters, and since it separates generically such obviously related species as M. exserta and M. hismda because of fruit-characters of very dubious phylogenetic significance. Macromeria, as defined by De Candolle, Prodr. x. 68-9 (1846), is a very satisfactory and natural genus characterized by its exceptionally large usually trumpet- shaped occasionally irregular corollas with ascending or recurved ovate or oblongish usually acute lobes, very long filiform filaments that just exceed or very conspicuously surpass the throat of the corolla, oblong obtuse anthers, and tardily protruded style. The genus has the acute corolla-lobes of Onosmodium, but that is the only character which suggests a relation with the latter genus rather than with Lithospermum, the genus which seems to be its closest relation. Onosmodium differs from Lithospermum in its protruded stamens and large corollas with acutish lobes. Lasiarrhenum differs from Macro- meria in its smaller cylindrical corollas with erect rounded lobes, ligulate included filaments, and sagittate hirsute anthers. Lithospermum is a large and variable genus characterized by its ascending or divergent or recurved rounded or obtuse corolla-lobes, oblong included anthers, and style that is tardily and but little if at all exserted. It is also to be recognized by the frequent occurrence in its species of veinless leaves and obviously geminate styles, the species of Macromeria, Onosmodium, and Lasiarrhenum always having prominently veined leaves and obscurely geminate styles. In Lithospermum the corolla is extremely variable in size and form, though constant in the position and shape of its lobes. Faucal ap- pendages have various developments, and among the species of Lithosperms all stages can be found between conspicuous intruded ligulate structures and mere downy spots or naked fauces. The genus is readily separated from Lasiarrhenum by its non-erect. corolla-lobes, oblong glabrous anthers, and slender filiform filaments. Ra ne i Seat ee een eee Beene 7 : ae i yy STUDIES IN THE BORAGINACEAE 17 11. Onosmodium Michx. Fl. Bor. Am. i. 132 (1803). Osmodium Raf. Med. Repos. N. Y. v. 350 or 352 (1808); Am. Mo. Mag. ii. 176 (1818); Am. Mo. Mag. iv. 191 (1819). Purshia Spreng. Anleit. ed. 2, ii. 450 (1817). Onosmidium Walp. Ann. iii. 134 (1853). Key To SpEcizs. Corolla a clear light- or erence yell; lobes narrowly lanceolate- acuminate, 2-3 times as long a _ Re pas Sa O. virginianum. Corolla sordid - greenish-white or merely yellowish; lobes wide. Corolla lobes ages r less ac e. diaper Lyris well Sieorneers 8-15 mm. long; bracts obtusish; corolla-lobes pho 3 ae long as tube, S10 mm. WS ee ees ea eee 2. O. Helleri. igre pedicels not greatly elongated, long; acts lanceolate or ovate-laneeolate, pie oy peed __ lobes about half as long as rolla-lobes glabrate or pubescent. 5 a the middie: ghar = O. bejariense. Stems short appressed villous; seg bes pubescen over most of pera muriace; Mexican... 2. oes hcoc5 ne O. unicum. Corolla lobes merely a Stems smooth 2 Bai er usually shiny below...... 5. O. subsetosum. Stems roughly pubescent and dull throughout. Nutlets strongly — nstricted at the base; plant always coarsely spre g Ses Ser Eee ispidissimum. bse with Fae rounded or if ae constricted e plant senha My OS es i 7. O. molle. 1. Onosmodium virginianum (L.) A. DC. Prodr. x. 70 (1846). Iithospermum virginianum L. Sp. Pl. 132 (1753). 0. hispidum Michx. FI). Bor. Am. i. 133 (1803). Purshia hispida Lehm. Asperif. ii. 382 (1818). O. scabrum R. & S. Syst. iv. 57 (1819). P. scabra Nutt. Trans. Am. Philos. Soc. v. 189 (1837). (?) O. nigrum Raf. Atl. Jour. i. 148 (1832). O. virginianum, var. hirsutum Mack. Bull. Torr. Bot. - Cl. xxxii. 499 (1905)—Connecticut to Florida, and westward along the Gulf to Louisiana. The southern material is frequently spreading hirsute and has been called the variety hirsutum. 2. O. Helleri Small, Fl. Southeast. U. S. 1000 (1903); Mack. Bull. Torr. Bot. Cl. xxxii. 497 (1905); Schulz, Wild Fl. San Antonio 176 (1922).—Known only from Texas. I have seen no flowers of this species, the floral characters used in the key having been derived from Schulz’s, |. c., account of the species. 3. O. bejariense DC. Prodr. x. 70 (1846); Mack. Bull. Torr. Bot. Cl. xxxii. 505 (1905).—Known only from Texas 4. O. unicum Macbr. Contr. Gray Herb. n. s. xlix. 21 (1917).— Known to me only from the type collection made at Alvarez, San Luis 18 JOHNSTON Potosi. The plant is scarcely more than a pubescent geographic form of O. bejariense and perhaps does not merit specific rank. 5. O. subsetosum Mack. & Bush in Small, FI. Southeast. U.S. 1001 (1903); Mack. Bull. Torr. Bot. Cl. xxxii. 505 (1905).—Missouri and Arkansas. 6. O. hispidissimum Mack. Bull. Torr. Bot. Cl. xxxii. 500 (1905). O. hispidissimum, var. macrospermum Mack. |. c. 502.—New York to Maryland, and westward to Louisiana, Texas, and Nebraska. 7. O. molle Michx. Fl. Am. Bor. i. 133, t. 15 (1803); Mack. Bull. Torr. Bot. Cl. xxxii. 499 (1905). Lithospermum molle Muhl. Cat. 19 (1813). Purshia mollis Lehm. Asperif. ii. 383 (1818). O. cariolini- anum, var. molle Gray, Synop. Fl. N. A. ii, pt. 1, 206 (1878).—In its extreme form known only from the vicinity of Nashville, Tenn. 7a, O. molle, var. occidentale (Mack.), comb. nov. 0. occidentale Mack. Bull. Torr. Bot. Cl. xxxii. 502 (1905). O. occidentale, var. sylvestre Mack. 1. ce. 504.—Minnesota and Illinois, westward to Texas, New Mexico, Utah, and Saskatchewan. This is the common Onos- modium west of the Mississippi River. The plant is not accorded specific rank due to the material from Illinois which shows repeated tendencies to intergrade with 0. molle. The variety usually differs from the typical form of the species in its coarser stems, rough spreading pubescence, larger calyces (8-12 mm. long), and unpitted non-constricted nutlets, but these characters and their complements occur in various combinations in the collections from Illinois. 12. Lithospermum L. Sp. PI. 132 (1753); Gen. Pl. 64 (1754). Batschia Gmel. Syst. ii. 315 (1791). Cyphorima Raf. Jour. Phys. Ixxxix. 98 (1819); Am. Mo. Mag. iv. 191 and 357 (1819). Pentalophus A. DC. Prodr. x. 86 (1846). Key to Spszcies. Plant evidently annual. Nutlets dull, ma tuberculate; calyx not cut to base... .. 3 L. arvense. Nutlets glossy, not tuberculate; calyx divided. Leaves obtuse; fruiting pormesy , mm. long; corolla 4-5 mm. ong, Sh ca. 2 mm. Jong ——— sigs eee 2. L. matamorense. : ca. Ante Selon nutlets pocorn ...0. L. Pringlei. a definitely strigose emia - Middle casting leaves veinless, usually rounded or aac commonly above the middle, two surface weekly, contrasted, both more or less Paice silica & : ~\y we ay ys tee eoenlin. Yellow. 6... 4. L. lasiosiphon. STUDIES IN THE BORAGINACEAE 19 — eee Saab. nvideatty veined, acute, broadest at r below the middle, two surfaces strongly contrasted, face; middle cauline tac. 1-2 cm. broad... :.... 5. L. discolor. Plants herbaceous, a pir branched above; seeker rt nnd eng thick, veining not i mpressed o fac iddle cauline leaves 1. 5 -2.5 em ‘tena BL. epapeeacinier barons i, hare ‘inside or ee with a fiom iee ring at fe rho finitely w —-14 mm. nai limb 10-16 mm. broad.......... 7. L. Nelsonii. Tube 2-8 mm. long, limb 5-12 mm. broad. Cakes. ‘ovate or ate ipec not exceeding nutlets; soe ts low, decumbent or ascending, with a com- woody caudex; high altitu Pam Aig Peru. .8. L. peruvianum. Calylobes linear, och recapong the rane plants erect ending, lacking an evident compact. lage Doe Corolla yellow or cele. Nutlets e-t MUR tS ses Sts 10. L. mirabile Nutlets smooth or merely pitted rm, limb 9-18 road, lobes com- y fimbria ee tube slender, ante 15-33 mm. am 1-3 mm. thick; later whens cleisto- gamous,. abaulantts fructiferous, in fruit usually with recurved MOUND 8 ios oes a hg is oa 8 . 11. L. angustifolium. re | wae 1 b 2-7 mm. broad and yrronag never ae flowers never cleistogamous; fruiting pedicels (except in no. 32) linots erect. Basal leaves persistent and forming a rosette at anthesis. Plant with a short m J ous-thickened costal cauline thee ovate to obovate-oblong, 1-2'em. broad; corolla 3-4 mm. broad, lobes erect........... ewok 12. L. tuberosum. e pe Dip gas ign ia omiay's ete Sues Ga, vous 2 3. L. cobrense. Radical leaves obovate, — broad.. ore L. obovatum. Basal leaves lacking a t anthes' Corolla inco loess 4 mm. tok = less or with tube soe early not exceeding the length of yx Leaves herbaceous, veining eviden Leaves 20-45 mm. broad, ovate lanceolate; | ghOnely SPPRONNOE.. 2266. ok. ee. 15. L. latifolium. Leaves 6-15 | ape mm. broad, ee nutlets 20-28 mm. long: p loosely appressed or some vlst weenie 16. L. officinale. 20 - JOHNSTON Leaves firm, apparently titi lowers in a very leafy evlitidrical thryse, scattered along rather numerous short ane Cen eS 27. L. ruderale. re Flowers racemose or disposed along several orymbose bra sa 8; inflorescence not conspicuous] Fruiting calyx shinee, lobes about twice seme of mutleta sok oe oe ee: 17. L. calcicola. Soar nea or Eon encomsto; 5 a 9- 13 ONG Oo eee . L. revolutum. ey Ene . obtusifolium. —— an or sl tube aed exceed- calyx Leaves ‘Guinsipal ones). 5-3 ad, usuall conspicuous in the in w facie root aoe at all thickened; haan terete and not obviously keeled. Leaves with sian vei eaves glaucescent, gpa Re corolla unappen se vapi Vabeee ee een tae . L. californicum. Corolla at most only 1 em. io 3 ube cy. sg lobes divergent, leaves broad. 4 Stems Aric one hirsute, light-colored, usually branched; calyx — .21. L. Seleri. ome strigose, bro ah. lly Case inceenaeieee calyx Wore Preece es 22. L. medial :. Corolla 12-27 mm. fins leaves 4-25 mm. Oe 28. L: guatemalense. te eee wet eG es 1. Odidifeltinn. — oe — i eindrea, a fe ges por at danaely ‘aeeek hirsute stri Plant green; leaves ves scabrous above .25. L. Palmeri. Plant canescent; pane’ velutinous acai otter TT eee .26. L. viride. Ba an uae Sa Sri ia ae ge ii Si oe eh eee a ee rene ka pen ol ee eee as STUDIES IN THE BORAGINACEAE 21 Leaves appeals veinless Flowers in a very leafy > cgihedeioal ne along rather numerou sp ranches; corollas ae i nish or pale yellow, aly less than 1 en ong, more or less obscured by ins foliaceous lanceolate bracts........ 27. L. ruderale. Flowers simply racemose or distributed along several elongati ing _ ascending not noticeably obscured by the small or reduc ts. Corolla oyldetoat lacking a differ- entia hroat, lobes divergent, pr oe of limb not alias length tu Calyx-lobes twice length of nutlets. .28. L. calycosum. eae — about half longer than ilated throat and asc aiding. lobes or if doubtfully so then with the breadth of limb equalling or exceed- ing the length of tube. Later floral leaves reduced, scarcely if exceeding the calyx-lobes, somewhat simulating the latter. lobes and not Gentes the rar Young go silky-strigose; sea 8 aes Peat ealyx-lobes 5 . L. multiflorum. aes paste sof plant usually Na ier eoaag nutlets 3-4 mm. long; lobes 6-13 mm. long. Hodes green; corolla 10-23 desi Herbacs ‘Gaia: pass ss mm. broad, lobes Se: calyx aciently recurve fruit; Pacific drainage. "32. be californicum, 1, Lithospermum arvense L. Sp. Pl. 132 (1753).—Native of Europe and adjacent Asia and Africa. ‘Introduced in Me.!, N. H.!, Vt.!, Mass.!, R. 1.!, Conn.!, N. Y.!, N. J.!, Penn.!, Del.!, Md. BBG C., Va.!, W. Va, N. Ce S.. 4.1, Ga, Fis., Ont.!, ‘Mich.t, Chae Ind., kes Tenn., Miss., Ala., Wis., Minn., las. Mo.!, Ark., La., Nok. Kan.!, Mont.!, Utah, Calif., B. C. 2. L. matamorense A. DC. Prodr. x. 76 (1846). L. prostratum 99 JOHNSTON Buckley, Proce. Acad. Philad. 1861, — (1861).—Central Texas southward into Nuevo Leon and Tamau 3. L. Pringlei, sp. nov., annuum i dm. altum rigidiusculum ; caulibus paucis simplicibus vel laxe ascendenterque ramosis strigosis et hirsutis foliosis; foliis nervosis asperis hirsuto-strigosis saepe acutis infra pallidis, inferioribus ante anthesis torrentibus oblanceolatis, superioribus late lanceolatis brevissime petiolatis paullo reductis; - racemis simplicibus vel rare furcatis foliosissimis statu fructifero conspicue dissitifloris ad 1-2 dm. longis; corollis flavis 8-10 mm longis, tubo cylindrato 1.5-2 mm. crasso extus pubescenti calycem aequanti vel paullo superanti, lobis laxe ascendentibus vel (?) diver- gentibus ca. 1.5 mm. longis, faucibus appendices 5 orbiculares nuloso-pubescentes gerentibus; staminibus inclusis, antheris ob- longis longitudine filamenta aequantibus; pedicellis 4-5 mm. longis strictis; calycibus fructiferis lineari-lobatis 8-10 mm. longis; nuculis saepe 4 ca. 2.5-3 mm. longis teretibus.—Mexico: Pedigral, Valley of Mexico, Federal District, 2250 m. alt., Pringle 6416; rocky hills near Metepec Station, Hidalgo, 2500 m. alt., June 25, 1904, Pringle 12123 (rypr, Gray Herb.).—This very distinct species has been confused with L. distichum Ort., though not very closely related to it. In gross habit the new species much suggests L. latifolium Michx. and L. officinale L., but these species are perennial and have strong tap- roots. Lithospermum latifolium differs further by having broader opr and more closely appressed strigose pubescence. The European L. officinale has thicker leaves, smaller corollas, smaller somewhat de- ciduous calyx-lobes, ms — and fewer nutlets, L. lasiosiphon, s ., perenne laxe caulibus erecti Soot 3-6 dm. ee foliosis: foliis obtusis vel rotundatis herbaceis supra viridibus infra glaucescentibus, inferioribus spathulatis 2.5-7 em. longis 8-14 mm. latis laxe rosulatis paullo obscureque nervosis, superioribus oblanceolatis vel linearibus strictis numerosis gradatim reductis enerviis; racemis geminatis terminalibus strictis ad 15 em. longis dissitifloris, bracteis linearibus vel lanceo-linearibus saepe calycem superantibus ; pedicellis strictis longitudine calycem aequan- tibus vel vix brevioribus; calycibus fructiferis 5-7 mm. longis ad basin 5-partitis, lobis herbaceis linearibus vel lanceolatis; corollis infundibuliformibus 13-17. mm. longis dilute luteis extus breviter villosis intus sparse villosis lobis laxe ascendentibus limbus 8-10 mm. latis faucibus exappendiculatis; staminibus inclusis, filamentis quam ; nuculis lucidis ca. 2.5 mm. longis laevibus.— Maou at 2250 m. alt. in the Sierra Madre near Colonia Garcia, STUDIES IN THE BORAGINACEAE 23 Chihuahua, Aug. 10, 1899, Townsend & Barber 248 (rypE, Gray Herb.).—In inflorescence and in the shape of its flowers L. lasiosiphon strongly suggests L. multiflorum Torr., but differs in having persistent basal foliage, less pubescent non-acute leaves that are broadest above rather than below the middle, and corollas that are pubescent inside. The new species is probably most related to L. cobrense Greene, but differs from that species in its funnel-formed internally pubescent corollas, and larger and broader not abruptly reduced non-involute leaves. The species is a very distinct one. 5. L. discolor Mart. & Gal. Bull. Acad. Brux. xi. 337 (1844). L. chersinum Macbr. Contr. Gray Herb. n. s. xlix. 22 (1917). L. dis- color, var. candicans Kuntze, Rev. Gen. ii. 439 (1891).—Pacifie slope of southwestern Mexico from Oaxaca to Tepic. Due to its branching habit and very thick = bicolored leaves this is an exception- ally well marked species 6. L. hypoleucum, a: nov., perenne herbaceum ca. 7 dm. altum; caulibus erectis simplicibus vel apicem versus breviter ramosis dense retrorseque pubescentibus et hirsutis; foliis subsessilibus acutis obo- vato-oblongis vel oblongo-ovatis 5-7 cm. longis 2-2.5 em. latis con- spicue nervosis, supra viridibus scabris, subtus albidis vel canescenti- bus dense tomentosis; racemis terminalibus furcatis; calycibus frue- tiferis ca. 1 cm. longis, lobis linearibus, pedicellis crassis ca. 2 mm. longis bracteas saepe (?) aequantibus vel superantibus; corollis ca. 12 mm. longis (ut videtur albescentibus); tubo cylindrato 2.5-3 mm. crasso ) calycem 1-2 mm. superante intus medium versus villoso, lobis 2.5mm longis, faucibus ad basin tubi adfixis; filamentis brevibus; nuculis ignotis. —Mexico: Loma St. Maria near Morelia, Michoacan, 1850 m. alt. July 29, 1909, Arséne 2511 (typE, Gray Herb.).—A very well marked species proba- bly most nearly related to L. discolor Mart. & Gal. It agrees with its relative in having white or whitish corollas that are pubescent inside, and by having conspicuously bicolored leaves that are tomentose beneath. It differs in having simple coarser herbaceous stems, larger much less firm leaves, tubular rather than funnelform corollas, and longer calyx-lobes. The type was distributed as Onosmodium bejartense DC. 7. L. Nelsonii Greenm. Proc. Am. Acad. xl. 31 (1904).—Known only from the vicinity of Monterey, Nuevo Leon, and very distinct in its large white salverform corollas, though nevertheless an evident relative of L. distichum Ort. 8. L. peruvianum A. DC. Prodr. x. 77 (1846). L. andinum 24 JOHNSTON Krause, Engler Bot. Jahrb. xxxvii. 636 (1906).—Peruvian Andes at altitudes over 3000 m. The DeCandollean material came from Chachapoyas in northern Peru, and that of Krause from Tarma in central Peru. I have seen a collection from Puno in southern Peru. Lithospermum incanum R. & P., Fl. Peruv. ii. 4 (1799), not Forst. (1788), the type of which came from arid hil!s near Tarma, may be found synonymous with L. perurianum. It was described as an herbaceous annual with a fusiform fibrous tap-root, but may have been based on young plants. 9. L. distichum Ortega, Hort. Matr. Dec., i. 8 (1797). Bat- schia disticha Don, Gen. Syst. iv. 326 (1838). L. spathulatum Mart. & Gal. Bull. Acad. Brux. xi. 337 (1844). Anchusa mexicana Sesse & Moc. Pl. Nov. Hisp. 21 (1887). L. laevigatum Sesse & Moc. FI. Mex. 30 (1894). (?) L. discolor, var. subviride Kuntze, Rev. Gen. ti. 439 (1891).—Southern San Luis Potosi southward to Chiapas. Originally described from garden plants grown from seed said to have come from Cuba. The seeds probably came from Mexico, since only Mexican plants agree with the original description and with Jacquin’s plate, Frag. Bot. 42, t. 48, fig. 3 (1809), and since Dr. N. L Britton writes me that no native Lithospermum is known either from Cuba or from any other of the West Indian islands. 10. L. mirabile Small, Fl. Southeast. U. S. 999 and 1337 (1903); Schulz, Wild Fl. San Antonio 176 (1922)—Known only from the - vicinity of San Antonio, Texas. I doubtfully associate with Small’s species a plant in the Gray Herbarium collected by Palmer in Sept. 1880 at Soledad, Coahuila. The specimen mentioned has nutlets as coarsely roughened as those on isotypic material of L. mirabile, but has nutlets lacking the conspicuous angular margins of the authentic ~ _ material. The specimen is further atypical in having a funnel-form corolla about as long as the calyx, rather than one much exceeding the calyx. The plant from about San Antonio has a characteristic aspect, a fact which, coupled with its peculiar angled and roughened nutlets, seems to make it desirable for the present, at least, to consider the species as one of restricted range, although very close to and somewhat doubtfully distinct from L. angustifoliwm Michx. 11. L. angustifolium Michx. Fl. Bor. Am. i. 130 (1803). Cyphorima angustifolia Nieuw]. Am. Midland Nat. iii. 194 (1914). L. fimbriatum Nutt. in Fras. Cat. no. 11 (1813), nom. nud. Batschia longiflora Nutt. in Pursh, Fl. Sept. Am. i. 132 (1814); Nutt. Gen. i. 114 (1818). L. longiflorum Spreng. Syst. i. 544 (1825). Pentalophus longiflorus A. DC. Prodr. x. 86 (1846). C. longiflora Raf. in Hook. f. & Jacks. STUDIES IN THE BORAGINACEAE 25 Ind. Kew. i. 701 (1893). B. decumbens Nutt. Gen. i. 114 (1818). L. decumbens Torr. Ann. Lyceum N. Y. ii. 225 (1826). OC. decumbens Raf. in Hook. f. & Jacks. Ind. Kew. i. 701 (1893). L. inciswm Lehm. Asperif. ii. 303 (1818). L. lineartfolium Goldie, Edinb. Philos. Jour. vi. 322 (1822). C. linearifolia Lunell, Am. Midland Nat. iv. 514 (1916). L. mandanense Spreng. Syst. i. 544 (1825); Hook. Fl. Bor. Am. ii. 88, t. 166 (1838). P. mandanensis A. DC. Prodr. x. 87 (1846). C. mandanensis Lunell, Am. Midland Nat. iv. 515 (1916). L. brevi- florum Engelm. & Gray, Jour. Boston Soe: Nat. Hist. v. 252 (1845). L. asperum Nels. Bull. Torr. Bot. Cl. xxvi. 244 (1899). L. albicans Greene, Pittonia iv. 91 (1899). L. ciliolatum Greene, |. c. 92. L. oblongum Greene, |. c. 92.—Extreme southern Ontario to Illinois and Manitoba, thence to Coahuila, Chihuahua, eastern Arizona, Utah and eastern British Columbia. This is a very variable, but certainly a natural assemblage which intensive study will probably show divis- able into a number of good geographic varieties. Among normal plants there is considerable variation in size and form of corolla. The eastern material has very large corollas with conspicuously fringed lobes; material from Wyoming has very much smaller flowers; whereas material from Texas has small corollas with short entire lobes. Over most of its range the plant grows 2.5-4 dm. high, but in Wyoming it is commonly about 1 dm. tall. Since the appearance of the note by Bebb, Am. Nat. vii. 691 pedir and the acceptance of Bebb’s findings by Gray, Synop. Fl. N. Am pt. 1, 205 (1878), this species has been considered dimorphic, ‘ectidine a phase with very conspicuous salverform normal corollas, and one with very inconspicuous cleistogamous flowers. These two phases are so different in appearance that they were formerly treated not only as of different species, but even of separate genera. Indeed so different do the two forms appear, that only the repeated acceptance of Bebb’s findings and the complete lack of dissenting opinion during the last fifty years makes me at all content to associate the two forms under one specific name. I am somewhat troubled, however, at the apparent lack in the herbarium of forms transitional between the two phases. The two states differ in the following points. The normal form has simple erect stems terminated by a compact very floriferous inflorescence composed of short crowded branches bearing showy salverform apparently sterile corollas on erect pedicels. The cleisto- gamous phase is diffusely branched and bears its inconspicuous fertile flowers in very elongate racemes and its fruit on recurved pedicels. The type of L. linearifolium Goldie, grew “only on sandy beaches 26 JOHNSTON at the head of Lake Ontario.” Recent collectors do not appear to have collected L. angustifolium there, and so it may be that Goldie’s species is not properly referred to the older one of Michaux. How- ever, “nuculis impresso-punctatis” and “foliis linearibus” of the short original description agree better with L. angustifoliwm than with any of the species that are known from the general region. Several attempts have been made at giving this species special generic recognition. Justification for such steps have been found in the shape of the corolla, in its fimbriate lobes, and in its large faucal appendages. These, however, are all characters which are variable within the species and which would utterly fail to characterize a genus decisively. Halsted, Bot. Gaz. xiv. 202 (1889), has pointed out certain peculiarities of the pollen of this species which deserve further study, since they may furnish additional specific characters. 12. L. tuberosum Rugel in DC. Prodr. x. 76 (1846).—Tennessee to Georgia and southward to Alabama and Florida. This species can at once be recognized by its tuberous roots. Gray reported the species from Texas, but his record was based upon fruiting material of Onosmodium Hellert Small. 13. L. cobrense Greene, Bot. Gaz. vi. 157 (1881). (?) L. tubuli- florum Greene, Pittonia i. 155 (1888).—Southern Arizona and New Mexico southward to southern Durango. A species well characterized - by its rather large very yellow corollas, rosettes of large basal leaves, and abruptly reduced cauline foliage. 14. L. Macbr. Contr. Gray Herb. n. s. xlviii. 56 (1916). —Known only from the type collection made at Quebrada Honda, Durango. It is perhaps better treated as a variety of L. cobrense, since it differs merely in having broader, more definitely obovate basal leaves. 15. L. latifolium Michx. Fl. Bor. Am. i. 131 (1803). L. officinale, var. latifolium Lehm. Asperif. ii. 311 (1818). Cyphorima latifolia Raf. in DC. Prodr. x. 76 (1846). C. lutea Raf. Cat. 13 (1824). L. luteum House, Bull. N. Y. State Mus. 243-244, pg. 61 (1923). L. lutescens Coleman, Cat. Pl. Grand Rapids 29 (1874).—Western New York to southern Minnesota and southward to eastern Ten- nessee. The Kew Index, iii. 99 (1894), gives the binomial, L. lat:- folium, as having been proposed by Forskal as well as by Michaux. Forskal did not publish such a binomial at the citation given by the Kew Index, i. e. “Fl. Aegypt. Arab. 39,” sai there appears to be no evidence that he ever published such a 16. L. officinale L. ‘Sp. Pl. 132 (1753). he of Europe and the < STUDIES IN THE BORAGINACEAE 27 Mediterranean region. Introduced in Que.!, N. B., Me.!, N. H.!, Vt.!, Mass.!, R. I., Conn.!, N. Y.!, N. J.!, Ont., Mich.!, Wis., Ohio, Ind., Minn. 17. L. calcicola Robins. Proc. Am. Acad. xxvii. 182 (1892).—Nuevo Leon southward to Oaxaca. 17a. L. calcicola, var. Conzattii (Greenm.), comb. nov. L. Con- sattti Greenm. Bull. Field. Mus., Bot. ii. 339 (1912)—Known only from the type collection made in Oaxaca. Macbride, Contr. Gray Herb. n. s. xviii. 55 (1916), considers this to be “only a young state of L. calcicola,’ but because of its relatively large exserted corollas it appears to merit varietal rank at least. 18. L. revolutum Robins. Proc. Am. Acad. xxvii. 182 (1892). Onosmodium revolutum Macbr. Contr. Gray Herb. n. s. xlix. 21 (1917). —Known only from San Luis Potosi. Macbride, |. c., has referred this species to Onosmodium, considering the species most related in that genus yet with characters more or less intermediate between Onos- modium and Lithospermum. However, the species has a corolla with a funnelform throat and ascending lobes, a comparatively short and lobed style that is not exserted until the corolla is completely opened, oblong obtuse anthers, and unnerved leaves—all characters which unmistakably and definitely ally it with Lithospermum and exclude it from the habitally very different Onosmodium. 19. L. obtusifolium, sp. nov., perenne 1.5-3 dm. altum; caulibus paucis e radice longa et recte descendenti orientibus ascendentibus vel decumbentibus strigosis ramosis; foliis concoloribus enerviis strigosis sessilibus 2-3.5 cm. longis 3-9 mm. latis apicem versus latioribus oblongis vel linearibus inferioribus praecociter deciduis superioribus paullo reductis; floribus parvis inconspicuis 3-5 mm. longis; corolla calycem subaequali, lobis saepe laxe ascendentibus minus quam 1.4 mm. longis, tubo cylindrato intus glabro, faucibus appendiculatis; antheris oblongis longitudine filamenta aequantibus; pedicellis strictis 3-10 mm. longis; nuculis saepe 3 lucidis ca. 2.5 mm. longis.—Mextco: sandy places near San Miguelito, San Luis Potosi, 1876, Schaffner 728 (TYPE, Gray Herb.); Levios, 45 mi. E. of Saltillo, Coahuila, Palmer 900; Soledad, 25 mi. W. of Monclova, 1880, Palmer, - Unrrep States: Chiricahua Mine, Chiricahua Mts., Arizona, Blumer 1796.—The status of this species is somewhat dubious. The Mexican material comes from a small natural and is probably homogeneous, although Palmer 900 has very narrow somewhat involute leaves. The Arizonian material is without flowers. It is readily separable from all the species occurring in the United States, and matches sur- ~ 28 JOHNSTON prisingly Shaffner 728. Two of the cited specimens are mutilated, Shaffner 728 and Palmer 900 having had the main axis nipped off, apparently at an early age. These specimens may represent the cleistogamous phase of some species, but if so I have been unable to correlate it with any of the well known species. Since the plants form a definable group they are here described as representing an unnamed species. Part of the material was referred by Macbride to his L. strictum, var. calycosum. 20. L. strictum Lehm. Asperif. ii. 303 (1818). -Anchusa tuberosa HBK. Noy. Gen. et Sp. iii. 92 (1818). Heliotropium lithospermoides R. & S. Syst. iv. 737 (1819).—Southérn Coahuila southward to the states of Mexico and Puebla; most frequently collected in San Luis Potosi. This is an exceptionately well marked species characterized by its thickened fusiform or bottle-shaped root, very narrow leaves, compressed sharply keeled nutlets, and very long simple incon- spicuously bracted unilateral racemes of elongate salverform corollas. In gross aspect the plant somewhat suggests Amsinckia Douglasiana 21. L. Seleri, sp. nov., perenne; caulibus ascendentibus stramineis simplicibus strigosis et hirsutis ca. 3 dm. altis; foliis lanceolatis acutis 2.5-4.5 em. longis 4-10 mm. latis breviter petiolatis conspicue nervosis, supra pustulatis, infra pallidioribus strigosis; racemis geminatis terminalibus hispidis; floribus luteis ca. 1 em. longis, corollae tubo eylindrato extus villoso intus glabro calycem paullo longiori, lobis laxe ascendentibus ca. 2 mm. longis, faucibus granulatis; antheris oblongis quam filamenta beapinetinea: calycibus fructiferis ca. 8 mm. longis hispidis, lobis linearibus quam nuculae duplo longioribus: nuculis ca. 2.6 mm. longis. Mexico: near Puebla, July 1888, Seler 847 (type, Gray Herb.).—This species appears to be most closely related to L. mediale Johnston, but differs from that in its spreading stiff pubescence, stramineous stems, and longer calyx-lobes. The type has been mutilated and has a number of ascending branches. It seems to have a weak perhaps only biennial root. 22. L. mediale, sp. nov., perenne 3-5 dm. altum canescens; caulibus erectis simplicibus vel apicem versus breviter ramosis compluribus a radice simplici recta orientibus foliosis strigosis; foliis strictis vel ascendentibus firmis oblongo-ovatis vel lanceolatis saepe concoloribus strigosis rare supra pustulatis (saepe cum nervis. paucis validisque) subsessilibus 3-6(-9) em. longis 4-13(-17) mm. latis superne paullo reductis; racemis terminalibus geminatis vel ternatis; floribus ochro- leucis ca. 1 cm. longis, corollae tubo cylindrato calyce saepe duplo STUDIES IN THE BORAGINACEAE 29 longiori extus villoso intus glabro; lobis laxe ascendentibus vel di- vergentibus 1-2 mm. longis, faucibus appendiculatis; antheris oblongis quam filamenta duplo longioribus; pedicellis 1-2 mm. longis strictis; calycibus fructiferis ca. 5 mm. longis, lobis linearibus quam nuculae vix longioribus; nuculis saepe 2-3 albis nitidis 2.5-3 mm. longis.— GUATEMALA: on cultivated ground, trachyte mountain Chi Lahuh K’ik, Dept. Quezaltenango, June 10, 1896, Seler 2893 (vyeR, Gray Herb.); San Lucas, Dept. La Antigua, Seler 2462; Santiago, Dept. Zacatepeque, Gomez 812. Corompta: Sibate, Dept. Cudinamarca, Pennell 2468 (NY, G); Zipacon, Dept. Cundinamarca, Rusby & Pennell 1217 (G); El Dintel, Bro. Ariste-Joseph A.525 (G); San Anton- io, Bro. Ariste-Joseph A.447 (G); Facatative, André 669 (NY); vicinity of Bogoté near Tequendama Falls, Dec. 1852, Holton (NY); without locality, Bro. Ariste-Joseph B-44 (US). VENEZUELA: between Mucurulea and Mucuchies, Dept. Los Andes, Jahn 801 (NY).—This very characteristic species is apparently restricted to central and northern South America. It has been mainly confused with the very different L. discolor Mart. & Gal., although its closest relative appears rather to be L. distichum Ort. The new species can be usually recog- nized by its canescence, firm leaves with few or no strong veins, short calyx-lobes that just exceed the nutlets, and very pale yellowish corol- las. Material from Guatemala and Colombia seems quite indistin- guishable, and is evidently referable to a single species. I have a suspicion that Kuntze’s L. discolor, var. subviride, Rev. Gen. ii. 439 (1891), was based on a plant of this species. L. guatemalense Donn. Sm. Bot. Gaz. xxvii. 436 (1899).—This was described from a collection made in the Dept. of Heuheutenango, in southeastern Guatemala. I have seen no authentic material of the species, but associate with it the left-hand plant on the sheet of Ghiesbreght 804 now preserved in the Gray Herbarium. I have strong suspicions that L. guatemalense will be found to be merely a narrow-leaved phase of L. oblongifolium Greenm. 24. L. oblongifolium Greenm. Proc. Am. Acad. xxxii. 300 (1897). (?) L. discolor, var. canescens Kuntze, Rev. Gen. ii. 439 (1891).—I know of only two collections, one, the type, collected in the Federal District of Mexico, the other collected by Ghiesbreght (no. 804 i in part) in the pine forests of southern Mexico. The latter is hirsute and I associate with it Kuntze’s varietal name. 25. L. Palmeri Wats. Proc. Am. Acad. xviii. 122 (1883).—Nuevo Leon and adjacent Coahuila. This differs from L. viride Greene, in pubescence, somewhat coarser habit, and isolated southern range, 30 JOHNSTON and might with good reasons be treated as a geographic variety of that species. 26. L. viride Greene, Bot. Gaz. vi. 158 (1881).—Southern New Mexico and Arizona. 27. L. ruderale Dougl. in Lehm. Pug. ii. 28 (1830). L. pilosum Nutt. Jour. Acad. Philad. vii. 43 (1834). Batschia pilosa Don, Gen. Syst. iv. 372 (1838). L. Torreyi Nutt. 1. c. 44. B. Torreyi Don, |. ¢. L. ruderale, var. Torreyi Macbr. Contr. Gray Herb. n. s, xlviii. 55 (1916). L. larum Greene, Pittonia iii. 263 (1898). L. lanceolatum Rydb. Mem. N. Y. Bot. Gard. i. 333 (1900). L. ruderale, var. lance- olatum Nels. Bot. Gaz. lii. 272 (1911). L. ruderale, var. macrospermum Macbr. 1. c.—Southern Alberta to western Colorado, westward to British Columbia and northern California. Although constant in its very characteristic leafy inflorescence, the species varies in size of nutlets and corollas, in pubescence, and in leaf-shape. The most striking and important of these variations is the var. macrospermum, the prevailing if not the only form of the species inseastern Oregon, and adjacent Idaho, Nevada and California. This variety has com- paratively large nutlets, 5-8 mm. long (the largest in the genus), which have a very prominent basal flange. 28. L. calycosum (Macbr.), comb. nov. L. strictum, var. calycosum Macbr. Contr. Gray Herb. n. s. xlviii. 56 (1916).—Southern San Luis Potosi, Guanajuato and Hidalgo. Although originally described as a variety of L. strictum this species is abundantly distinct, differing in its unthickened root, branched habit, broader leaves, and terete un- keeled nutlets. 29. L. multiflorum Torr. in Gray, Proc. Am. Acad. x. 51 (1874); Wats. Bot. King Exped. 238 (1871 ), nom. nud.—Southern Wyoming southward to Arizona and northern Chihuahua. 30. L. canescens (Michx.) Lehm. Asperif. ii. 305 (1818). Batschia canescens Michx. Fl. Bor. Am. i. 130, t. 14 (1803). Anchusa canescens Muhl. Cat. 19 (1813). A. virginiana L. Sp. Pl. 133 (1753); not L. virginianum L. 1. ¢. 132. L. sericeum Lehm. Asperif. ii. 306 (1818). B. sericea R. & S. Syst. iv. 743 (1819). B. conspicua R. Br. in Richards. Bot. App. to Frankl. Jour. 732 (1823).—Along the Ap- palachians from southern Pennsylvania south into Alabama and westward through Ohio to Kansas and Saskatchewan. 31. L. caroliniense (Walt.) MacMill. Metasp. Minn. Valley 438 (1892). Anonymos caroliniense Walt. Fl. Carolina 91 (1788). Batschia caroliniensis Gmel. Syst. i. 315 (1791). L. carolinianum _ Lam. Tab. Eneye. i. 397 (1791); Mack. Bull. Torr. Bot. Cl. xxxii. 501 STUDIES IN THE BORAGINACEAE 31 (1905). Onosmodium carolinianum A. DC. Prodr. x. 70 (1846). B. Gmelini Michx. Fl. Bor. Am. i. 130 (1803). L. Gmelini Hitche. Spring. Fl. Manhattan 30 (1894). Anchusa hirta Muhl. Cat. 19 (1813); nom. nud. L. hirtum Lehm. Asperif. ii. 304 (1818). L. _ strigosum Raf. New Fl. N. Am. pt. 4, 18 (1836). L. bejariense A. DC. Prodr. x. 79 (1846).—Western New York west to Montana, and south- ward through Nebraska to Texas, Tamaulipas and Louisiana. Also in South Carolina, Georgia, and Florida. 32. L. californicum Gray, Proc. Am. Acad. x. 51 (1875).—Northern California and adjacent Oregon. 13. Antiphytum DC. in Meisner, Genera i. 280; ii. 188 (1840). Amblynotopsis Macbr. Contr. Gray Herb. n. s. xlviii. 41 (1916).— A review of this genus. has been recently published, Contr. Gray Herb. n. s. Ixviii. 48-52 (1923). Recent study has shown, however, that the species there called Antiphytum tetraquetrum is properly excluded from the-genus and should be referred to Thawmatocaryon. 14. Mimophytum Greenm. Proc. Am. Acad. xli. 242 (1905). Mimophytum omphalodoides Greenm. Proc. Am. Acad. xli. 242 (1905).—This monotype is known only from the type collection taken from under a wet cliff in the barranca below Honey Station, 1500 m. alt., Puebla, Mexico, by Pringle in 1904. In gross aspect the plant remarkably simulates the Mexican species of Omphalodes, although in technical fruiting characters it seems quite distinct and much nearer Cynoglossum. 15. Cynoglossum [Tourn.] L. Sp. Pl. 134 (1753); Gen. PI. 65 (1754). ae Key TO SPECcIEs. b peer ee 1 hd a ll a > ee eee ae. 7 4 J Di... 1 J ¢. , wt : eae 1 4] n ,2-3mm long, seca? ares gested rea eee . C. Wallichii. ENG Lows ee eg ay oy oo cece ide as 2. C. officinale. N utlets 8-1C 8-1 omic. long, suborbicular in dorsal outline; aie weakly develo 1-2 mm. long, obscured and go Ones: POR oc. tie, C. Pringles. Nutlets not sunken in dorsally and lacking a definite Shi margin. Plant b biennial; European introductions. Corolla small, 4-5 mm. ne! nutlets small, 2.5—4 mm. long; leaves all acute, 4. C. zeylanicum. 32 JOHNSTON —s large, 6-8 mm. long; nutlets large, ca. 6 mm. long; of leaves usually obtusish, at least upper with : subamplexica Plant perennial native Ameri pec N metas mall, 2-3 mm. oe son se 2s rien by calyx- bes Fruiting pedicels 2-3 em. long, stiffly — clap ore sg calyx-lobes reflexed - ma ay" ~2. g, about e ee TUNG Ga C. paniculatum. Fruiting ear 0.5-1 em. long, teas oa PRES ing at matu ity. 3-5 mm. long, much Satis’ it BR ey oe Oe eo Pi ee 7. C. limense. Nutlets1 larger, 4-10 mm. long, if less than 5 mm. long then conspicuously exceeding the calyx-lobe Mature leaves glabrous on a. gustace, stems glabrate or sparingly hirsute or Leaves taper ing at base; nejacied are above; pedicels recurved in fruit; corolla- et bien me semilunate, included; Sout et aA ee CA TE 8. C. Trianaeum. Leaves a Sl sieuatad- to a slender petiole; stems glabrous above; fruiting pedicels ascending ‘ed gracefully outcurved; hag dages ob- protruded; ‘Nort h "America C. grande. eee Se ee eee el eae ® a Calyx-lobes 5-10mm. long in fruit, Hingat orlancédlate: aching aeg 35 mm. long, about twice length of fey mm. long, exceeded by <_ ae Aa style obscured by the nutlets, 1-2 Natiete 6.6-F-mm: lone. a a c. Ak apa Nutlets 3.5-5 mm. ‘long PES AEN eons) 8 arene eg eee 3. C. boreale. 1. Cynoglossum Wallichii Don, Gen. Syst. iv. 354 (1837). C. glochidatum Wall. Cat. 26 (1829), nom. nud.; Benth. in Royle, mn % 306 (1839).—Native of central Asia and dis. Adventive in 1902 in : garden at Hardwick, Mass . C, officinale L. Sp. Pl. 134 (1753). —Native of Eurasia, extending Met the western Mediterranean region eastward to central Asia. Introduced in Que.!, Me., N. H., Vt.!, Mass.!, R. I.!, Conn.|!, N. Y.!, N. J., Penn.!, Md., Del., Va.!, W. Va.!, N. C. Ont.!, Mich.!, Ohio!, Ind.!, Ill.!, Tenn., as. Ma, Wis., Iowa!, Mo. !, Ark., Neb., Kan.!, -, Mont., Wot Colo.!, Ut, N. M., Ore.! Pringlei G 3. Cc. reenm. Proc. Am. head, Ix. 30 (1904).—Mexico: Fultenango Canyon, State of pene Pringle 11350 (rypr); eo of Moralia, n, Arsene 5949; Guanajuato, 1893, Although closely related to C. officinale this species, a native of Jot vis beet Seca aie a Bee Pad oe) oe RE . C. creticum. SN PA te eee aN et STUDIES IN THE BORAGINACEAE 33 central Mexico, is clearly distinct in its larger and differently shaped nutlets and short style. Frequently only a single nutlet is matured from each flower. 4. C. zeylanicum (Vahl.) Thunb. in Lehm. Neue Schrift. Naturf. Gess. Halle ser. 3, ii. 20 (1817). Anchusa zeylanica Vahl. in Hornm. Enum. Hafn. 3 (1807), hyponym? C. furcatum Wall. in Roxb. FI. Ind. ii. 6 (1824).—Native of southeastern Asia. In the United States National Herbarium there is a specimen of this species collected in 1899 by Dr. O. Buchtien at Valdivia, Chile. The species is not other- wise known from America. 5. C. creticum Mill. Gard. Dict. ed. 8. (1768); Brand, Pflanzenr. iv. Fam. 252, i. 129 (1921). (C. pictum Soland, in Ait. Hort. Kew. i. 179 (1789). C. molle Phil. Linnaea xxix. 18 (1857).—Native of the Mediterranean region. Long ago naturalized in south central Chile whence it was described as a native species by Philippi. 6. C. paniculatum H. & A. Bot. Beech. 37 (1830). C. Azocarti Phil. Anal. Uniy. Chile Ixv. 62 (1884). C. paniculatum, var. Azocarti Reiche, Anal. Univ. Chile exxi. 248 (1908). C. paniculatum, forma Azocarti Brand, Pflanzenr. iv. Fam. 252, i. 137 (1921). C. paniculatum, forma Philippianum Brand, |. c.—South central Chile. Reiche, I. c., attributes the species to Peru. The plant is well marked by its ex- ceptionally elongate pedicels. 7. C. limense Willd. Sp. Pl. i. 762 (1798); Brand, Pflanzenr. iv. Fam. 252, i. 142 (1921). C. decurrens, var. limense A. DC. Prodr. x. 153 (1846). C. decurrens R. & P. Fl. Peruv. ii. 6 (1799). C. alatum Molina, Sagg. Chile ed. 2, 280 (1810)——Middle and southern Chile. The type of the species is reputed to have come from the vicinity of Lima, Peru. . ©. Trianasum Wedd. Chlor. And. ii. 90 (1859).—Mountains of Colombia and adjacent Ecuador. 9. C. grande Doigl. in Lehm. Pug. ii. 25 (1830). C. laeve es Synop. Fl. N. Am. ii. pt. 1, 188 (1878). C. grande, var. laeve Gray, ]. ec. suppl. 421 (1886) (?) C. Austinae Eastw. Bull. Torr. Bot. Cl. xxxii. 203 (1905).—From Monterey and Calaveras counties, California, wey to extreme southern British Columbia. 10. Proc. Calif. Acad. Sci. ser. 2, vi. 428, t..59 (1897) gure Nevada, California, from Calaveras to Tulare counties. 11. C. occidentale Gray, Proc. Am. Acad. x. 58 (1875)—From _ Plumas and Glen counties, California, northward to central Oregon. 12. C. virginianum L. Sp. Pl. 134 (1753). C. virginicum L. Syst. 34 JOHNSTON ed. 12, 146 (1767). C. amplexicaule Michx. Fl. Bor. Am. i. 132 (1803). C. lucidum Stokes, Bot. Mat. Med. i. 277 (1812). eget Connecticut southward to northern Florida, west to Louisiana and issouri. 13. C. boreale Fernald, Rhodora vii. 250 (1905)—New Brunswick and adjacent Quebec, southward to northern Connecticut, westward through New York and southern Ontario to northern Michigan; reappearing in southern British Columbia. All the vegetative char- acters of this species can be matched, after a short search, among indubitable material of C. virginianum. It usually differs, however, in its more slender habit, less stiff pubescence, more elongate slender pedicels, and smaller floral parts; and seems to be always distinct in its smaller nutlets and northern range. Brand, Pflanzenr. iv. Fam. 252, i. 131 (1921), reduces the species outright to C. virginianum, but surely it is deserving of varietal rank at the very least. 16. Pectocarya DC. in Meisner, Genera i. 279; ii. 188 (1840). Ktenospermum Lehm. Del. Sem. Hort. Hamburg 17 (1837), nom. nud. Gruvelia A. DC. Prodr. x. 119 (1846). Key To Species AND VARIETIES. Nutlet-margins lacerate or undulate or “coins usually _, ase ending or erect. § Eupect ocarya. t those forming the margin................ See Ae a: swale. Back of — without coarse subulate appendages; margin Plant erect, Ascondiaaly branched above; body of nutlets obovate ; calyx-lobes surpassing the nutlets.. .2. P. lateriflora. Plant prostrate or decumbent, much branched from the base >; body of nutlets mt linear or oblong, nite £ utlets Nutlets with margins pectnatly lacerate or dentate to e and in addition usually uncinate-bristly aed se ip AMO SE CT Bilis hoe Ese eee | 3. P. gracilis. Fruit tees throughout Fa ge a et cenronpliis. Bod ene ones very narrow got inco so seit .Var. genuina. very conspicuous... .var. platycarpa. Body of eggs ae acer or 0 oblor g-obova Bre ge iviana. ee eee ek, ee ee ee var Nutlets with margins entire or undulate alon sides, armed only at apex where densely uncinate-bristly. Nutlets similar, all margined; fruit not dimorphic, — similar in all parts of plant Spat ae Reg RE Er STUDIES IN THE BORAGINACEAE ye Nutlets dhiteilae fruit on the branches consisting of two eviden ntly margined and two unmargined nu tlets, — borne at base of plant with con- sl Seige con reflexed nutlets. ..var. heterocarpa. Nutlet-margin entire, actine uncinate-hairs or lacinae or teeth. N aes w es eer divergent, marginless, obviously angular and in outline; calyx-lobes strigose “— appressed ee ate hairs owards the ip; plant ve y, Ren branched mainly near base, sparsely prot ti Bee Nan UP . P. pusilla. ae divergent in pairs, some usually w 2 I urely angular, obovate or rhomboid-obovate in out- line e; calyx-lobes strigose and sparsely spreading hirsute’ e} plant stiffish, testa streaky branched, very strigose. . . wing cf setosa. Fruit with nutlets all wite-tharwined.. iiss hen Ovi Ss holoptera. Fruit with some of n — ts lacking a wing-margin. be Ee ree ee Cee eT ee ry een Oe var. aptera, Nutlets margined ea unmargined in the same pcan var. genui 1. Pectocarya anomala, sp. nov., erecta 7-15 cm. alta dichotoma strigosa supra laxe ramosa; foliis iritaribus 1-3 em. longis 0.8-1 mm. latis sparse hispidis inferioribus oppositis; pedicellis recurvatis quam bracteae foliaque multo brevioribus; sepalis lanceolatis quam nuculae paullo longioribus; nuculis obovato-oblongis ca. 0.7-0.8 mm. latis 1.5-1.7 mm. longis granulatis, dorso et marginibus cum appendicibus valdis uncinatis subulatis subteretibus uncinato-pubescentibus ca. 1-1.2 mm. longis munitis, marginibus inconspicuis——PERU: wenn pampa on the southern slope of Chachani Mountain near Arequip 2400 m. alt., Mr. & Mrs. F. E. Hinkley 41 (ryver, Gray Sty am Although in habit clearly a Pectocarya this species is quite anomalous in the genus in having the back of the nutlet, not unarmed and margined, but studded with coarse uncinate subulate uncinately pubescent appendages which are quite indistinguishable from those of the proper margin. The plant is erect-growing with a few loosely ascending branches from above the middle. In this habit it differs from P. gracilis and P. penicillata, and agrees with P. lateriflora. It is, however, a lower, less stiff, and less strictly branched plant than the latter. According to its collectors the plant is locally known as “estrella gateadora.” i P. lateriflora (Lam.) DC. Prodr. x. 120 (1846). Cynoglossum m Lam. Encyel. ii. 239 (1786); Planch. t. 92, fig. 2 (1791). Mattia lateriflorum Don, Gen. Syst. iv. 310 (1838). C. pilosum R..& S. Fl. Peruv. tL 6, t. 111b (1799). M. pilosa Don,|.c. Rindera pilosa R. & S. in DC. l. c. (2) Ktenospermum linifolium Lehm. Linnaea xii. Lit. 84 (1838); nom. nud—Brand, Pflanzenr. iv. Fam. 252, i. 95 (1921), has taken P. lateriflora as including more ieee 30%: JOHNSTON species, but as here taken it is in a much more limited sense and restricted to the Peruvian plants originally described. The species is known only from the western strip of Peru, and is well characterized by its erect habit of growth. 3. P. gracilis (R. & P.), comb. nov. Myosotis gracilis R. & P. FI. Peruv. ii. 5 (1799). Echinospermum gracile Lehm. Asperif. i. 129 (1818). Rochelia gracilis R. & S. Syst. iv. 111 (1819). Cynoglossum lineare R. & P. 1. c. 6. P. linearis DC. Prodr. x. 120 (1846). chilensis DC. 1. c. (?) P. chilensis, var. californica Torr. Pacif. R. R. Rep. iv. 124 (1857). 3a. P. gracilis, var. genuina.—This is the common species of Pectocarya in Chile and Argentina. It is uncommon in North America, apparently restricted there to Southern California and to the islands. off that coast. The following cited collections represent all the material of true P. gracilis which I have seen from North America.— San Diego, Brandegee 1636 (G, P); Granite, San Diego County, Spencer 115 (G); Palm Springs, Margaret Ferguson 41 (Wellesley College Herb.); Butte west of Lakeview, 1920, Johnston (P); Temescal Canyon, Munz & Harwood 3387 (G); foothills of San Bernardino Mts., 1896, Cummings (G); Claremont, Munz 2005 (P), Baker 4135 (P); Surprise Canyon, Panamint Mts., Coville & Funston 721 (G); Santa Cruz Island, 1887, Ford (G); Catalina Island, Grant 913 (G); Clemente Island, Munz 6690 (G, P); Guadalupe Island, Palmer 69a (G), Anthony 237 (G). 3b. P. gracilis, var. platycarpa Munz & Johnston, var. nov., as- cendens saepe robustior; margine nuculae grosse dentato ca. 1 mm. lato valde conspicuo.—Arizona: Tempe, 1892, Ganong & Blaschka (G); Wickenburg, W. W. Jones 88 (G); Tucson, Greene 1109 (G); mesas near Camp Lowell, April 16, 1881, Pringle in part (TyPE, Gray Herb.) ; without locality, 1884, Pringle (G). Uran: valley of Virgin near St. George, Parry 167 (G). Catirornia: Agua Caliente, Parish Bros. 122 (G); Chuckawalla Valley, Munz & Keck: 4804 in part (G); Providence Mts., Munz & Harwood 3532 (P), 3535 in part (P); east of Daggett, Munz & Harwood 3673 (P).—This well marked variety seems to replace the var. genwina in the deserts of California, Arizona, and Utah. With only North American material at hand the plant seems so extreme and positive as to merit specific rank. However, the South American material of the var. genuina frequently has rather broadly margined nutlets that show an approach to the condition characteristic of the var. platycarpa, and it hence seems best to treat ‘ the latter merely as a variety of the same species. STUDIES IN THE BORAGINACEAE 37 3c. P. gracilis, var. boliviana, var. nov., ramosa prostrata; nuculis obovatis vel oblongo-obovatis——Bouivia: Chignana, 3700 m. alt., Asplund 3897 (rypr, U. S. Nat. Herb.); Challapata, 3900 m. alt., Asplund 5895 (US); Teneral Campero, 4200 m. alt., Asplund 5894 (US); Atocha, 3700 m. alt., Asplund 5896 (US).—This appears to be the northern phase of P. gracilis in South America. The cited suite of specimens is uniform in the crucial characters of the variety, the obovate body of the nutlet, but as to margining varies from con- spicuously pectinate-dentate to almost naked. 3d. P. gracilis, var. dimorpha, var. nov., ramosa prostrata; fructis dimorphis, superioribus nuculas divaricatas compressas normales gerentibus, infimis nuculas reflexas crassas persistentes. gerentibus; nuculis oblongis.—CuiLe: Vallemar, Rose 19331 (US); Desert of Atacama, Morong 1282 (typE, Gray Herb.; tsorypr, N. Y. B Gard.).—This plant of northern Chile is evidently not typical of the var. genuina and because of its conspicuously dimorphic nutlets appears to deserve special varietal recognition. The normal nutlets have a very broad, erect, merely toothed margin. 4. P. penicillata (H. & A.) A. DC. Prodr. x. 120 (1846). Cyno- glossum penicillatum H. & A. Bot. Beech. 371 (1840). P. linearis, var. penicillata Jones, Proc. Calif. Acad. Sci. ser. 2, v. 709 (1895). P. miser Nels. Bot. Gaz. xxxvii. 278 (1904).—In its two forms this is the most common and widely distributed of the North American Pectocaryas. Brand, Pflanzenr. iv. Fam. 252, i. 95 (1921); takes up Cynoglossum Nuttallii Spreng., Syst. i. 566 (1825), as the earliest name of the present species. Sprengel’s name, however, is based upon C. pilosum of Nuttall, Gen. i. 114 (1818), which is, as Gray, Synop. Fl. N. Am. ii. pt. 1, 190 (1878), has indicated, a species of Lappula. _ 4a. P. penicillata, var. genuina.—Ranging from British Columbia, Idaho and southwestern Wyoming southward to northern Nevada and Lower California. In Arizona, southern Nevada, and the deserts of California it is replaced by the following variety. 4b. P. penicillata, var. heterocarpa, var. nov., deserticola; nuculis heteromorphis, duabus marginatis, duabus emarginatis gracilioribus _paullo reflexis—Arizona: Tempe, 1892, Ganong & Blaschka (G); Camp Grant, Palmer 182 (G); near Camp Lowell, 1881, Pringle in part (G); without locality, 1876, Palmer (G). Nevapa: Moapa, Coodding 2200 (G); St. Thomas, Tidestrom 8642 (G). CALIFORNIA: Coahuilla Valley, 1901, Hall (G); Surprise Canyon, Panamint Mts., Coville & Funston 721 in part (G); Providence Mts., Munz & Harwood . 38 JOHNSTON 3435 in part (P); near Daggett, Munz & Harwood 3673 in part (P); north of Randsburg, 1922, Pierce (P); Corn Springs, Chuckawalla Valley, 1922, Munz & Keck 4870 (tyPE, Pomona College Herb.); Chuckawalla Valley, Munz & Keck 4804 in part (G); Paloverde Valley, 1905, Walder (P); Palm Springs, Spencer 115 (P). The following collections appear intermediate between var. genuina and var. hetero- carpa. Hemet, Baker 4139 (P); near Cabazon, Munz, Street & Wil- liams 2387 (P); Santa Susana Mts., Brewer 210 (G); Oil City, Heller 7586 (G).—The nutlets of this variety are dissimilar, two of each fruit usually being unmargined, thicker, and somewhat reflexed. The fruit borne about the base of the plant is different from that borne well out on the branches and perhaps is produced by cleistogamous flowers. The nutlets of this fruit are usually consimilar, all are strongly reflexed, very thickened, and broadly and permanently affixed to the gynobase. Similar fruit is found in the Chilian P. gracilis, var. dimorpha. 5. P. pusilla (A. DC.) Gray, Proc. Am. Acad. xii. 81 (1876). Gruvelia pusilla A. DC. Prodr. x. 119 (1846); Gay, Fl. Chile iv. 482, t. 52, fig. 3 (1849). P. pusilla, var. flagillaris Brand, Pflanzenr. iv. Fam. 252, i. 96 (1921).—West of the high mountains from southern Washington to middle (Fresno County) California; also in Chilé where very rare. Brand, 1. ¢., considers the North American plant varietally distinct from the South American, arguing that it has more slender, elongate stems. Though the northern material may fre- quently attain greater slenderness and height than the southern, the two are certainly indistinguishable at times. 6. P. setosa Gray, Proc. Am. Acad. xii. 81 (1876). Gruvelia setosa Rydb. Bull. Torr. Bot. Cl. xl. 479 (1913).—This species is confined to western North America and breaks up into three geographic varieties as follows :— 6a. P. setosa, var. genuina.—Washington and Idaho inthe arid interior, southward into Utah and Arizona, and along the Sierras to . the Mohave and northern Colorado Deserts. This is the common and typical phase of the species having two of the four nutlets winged, _ and two wingless with the latter pair partially hidden by the former. 6b. P. setosa, var. aptera, var. nov., australis; nuculis vix alatis vel omnino apteris.—CALiForNIA; Warners Hot Spring, Eastwood 2620 _(G); dry canyon floor near Campo, 1903, Abrams 3571 (rypE, Gray Herb.).—The nutlets of this variety are quite similar to those of - The specimens referred here come from the western STUDIES IN THE BORAGINACEAE 39 side of the Colorado Desert and represent the only material of the species seen from the area. c. B. setosa, var. holoptera, var. nov., mohavensis; nuculis om- nibus lage aequaliterque alatis. —Cartronnta! White Mts. east of Laws, Heller 8187 (G); Granite Wells, Mohave Desert, 1922, Johnston 6489 (tyPE, Pomona College Herb.); near Mohav re, 1920, Johnston (P); Mt. Pinos, Hall 6423 (P).—This form, characterized by having all its nutlets broadly and equally winged, occurs along the northwest edge of the Mohave Desert. 17. Omphalodes Moench, Meth. 419 (1794). Kry To SPEcigs. Cauline leaves sessile, oblong-linear or lanceolate; corolla mene introdu ced pla nts. Saints Pe area seme waa: iiss Lure sae, O. linifolia. ag esti s plants, mainly Mexi Flowers in dakea ¢ racemes with only the iswerniaet if any TACtOALG, ANOUA os Sis po oe 2. O. aliena. Flowers all axillary; oped aie perennials. Corolla large, 10-14 road; leaves herbaceous, sire es 2 mm. ak sia Dar ty A pce Cunelseh ay Sarg ate ge eal kigy oo: acuminata. Corolla small, 5-7 mm. broad;1 firm, larg tk one only 10- Leaves with evident veri sharply acute or acuminate, rather sparsely strigose; pedicels shorter or but peas Sxpopding the Hirhete i. ee 4. O. cardiophylla. Leaves with Gbachbad. veining, broadly acute, velutinous with a dense short pubescence; pedicels twice ex- ceeding the subtending leaves... : 2.2 5. O. mexicana. 1. Omphalodes linifolia (L.) Moench, Meth. 419 (1794). Cyno- glossum linifolium L. Sp. Pl. 134 (1753) eg of southwestern Europe. Reported by Hooker, Fl. Bor. Am. ii. 86 (1838), from Labrador. Collected in 1919 on street rey in Salem, Oregon, by J. C. Nelson. 2. O. aliena Gray in Hemsl. Biol. Centr. Amer. Pacey ii. 377 (1882).— Texas: common, rocky hillsides, Sanderson, H. C. Hanson 381. Nuevo Leon: limestone hills near Monterey, 600 m. alt., Pringle 10205, 10206; Monterey, 1880, Palmer 893 (TYPE). 3. O. acuminata Robins. Proc. Am. Acad. xxvi. 170 (1891).— Nuevo Leon: Sierra Madre near Monterey, Pringle 2220 (rrPr); limestone ledges of Sierra Madre above Monterey, Pringle 10162. 4. O. cardiophylla Gray in Hemsley, Biol. Centr. Amer. Bot. ii. 377 (1882).—Pursia: Boca del Monte on shaded mountain slopes, E Pq Purpus 2498. CoaHvta Saltillo, Palmer 894 (tyr). 40 JOHNSTON 5. O. mexicana Wats. Proc. Am. Acad. xxv. 158 (1890) —NvuEvo Leon: fissures of dry lime rock, Sierra Madre near Monterey, Pringle 1878 (TYPE). 18. Harpagonella Gray, Proc. Am. Acad. xi. 88 (1876). Harpagonella Palmeri Gray, Proc. Am. Acad. xi. 88 (1876); Baill. Bull. Mens. Soc. Linn. Paris no. 102, 812 (1889); Hist. Pl. x. 351, 266-8 (1890); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 130, fig. 53 (1891)—An anomalous monotype known only from Pima and Cochise counties in southern Arizona, extreme southwestern California in western San Diego and southwestern Riverside counties, extreme northwestern Lower California, and on the Guadelupe and Santa Catalina Islands. 19. Myosotis [Dill.] L. Sp. Pl. 131 (1753); Gen. Pi. 63 (1754). Key To Sprciss. gic he oa leafy-bracted nearly throughout; corolla-tube at t a third longer than calyx; prostrate antarctic peren- aaah si bec ee 1 iano, icc maine ea. 1.8 mm plant Fyanig ts SDORE-DAIG. os es M. albiflora. mm. broad; calyx subsessile; Erm to n r base, lower flowers ag Sealer sank me blue, tube about equalling alexi style always shorter than ce MR iG ai cg core ght Pere Ren eee ee 10. M. micrantha. in yle Hastie much surpassing nutlets. . AL M. versicolor. 1. Myosotis albiflora Banks & Solander in Hook. f. Fl. Antarct. ii. 329 (1846). Eritrichium albiflorum Griseb. Abhandl. Ges. Wis Goett. vi. 131 (1857). Allocarya albiflora Greene, Erythea iii. 57 (1895).—Endemic to the vicinity of the Straits of Magellan. In the literature this plant has been repeatedly confused with the plant recently called Plagiobothrys Lechleri Johnston. My knowledge of this Myosotis has been largely derived from Skottsberg, Svenska Vet. Akad. Hand. lvi. pt. 5, 290, t. 23, fig. 8a, ¢ (1916), who besides dis- cussing it has given figures and a new description. M. antarctica Hook. f. Fl. Antarct. i. 57, t. 38 (1844). M. an- ‘Gerobion, fears pmectior Botts. Hvshala Vee Aled: hauled ot 5 291 (1916).—Extreme southern Patagonia; also in New Zealand and on the Stewart and Campbell Islands. The species has been confused with M. albiflora. Skottsberg, |. c., t. 23, fig. 8b, d, gives a few figures and a new description of the species. 3. M. scorpioides L. Sp. Pl. 131 (1753); Rendle & Brit. Jour. Bot. xlv. 440 (1907). M. scorpioides, var. palustris L. 1. c. M. palustris Lam. FI. Fr. ii. 283 (1778).—Native of Europe. Precquentiy cultiv- ated. Naturalized in Newf.!, Que.!, N. S.!, Me.!, N. H.!, Vt.!, Mass.!, R. I., Conn.!, N. Y.!, N. J., Penn., Md.!, D. C., Ga., Mich.!, Tenn.!, La., Ont.!, Calif.!, B. C., Hidalgo. 4. M. laxa Lehm. Asperif. i. 83 (1818). M. caespitosa, var. lara A. DC. Prodr. x. 105 (1846). M. palustris, var. lara Gray, Manual 42 JOHNSTON ed. 5, 365 (1867). M. lingulata Lehm. Asperif. i. 110 (1818). M. caespitosa Schultz, Fl. Starg. Suppl. i. 11 (1819). M. palustris, var. micrantha Lehm. in Hook. Fl. Bor. Am. ii. 81 (1838).—Newfoundland to Georgia, westward to Ontario and Tennessee; also on the Pacific Coast from California to British Columbia. : . M. azorica H. C. Wats. Bot. Mag. Ixx. t. 4122 (Nov. 1844). (?) M. maritima Hochst. in Seubert. Fl. Azor. 37 (1844).—AmeERICA: Mill Valley, California, 1913, Suksdorf 520 (G); Bogota, Colombia, 1876, Bayon (US); canyon of La Paz River, Bolivia, Shepard 174 (G); Valdivia, Chile, 1896, Buchtien (US).—Native of the Azores and Canary Islands. Cultivated in America and occasionally escaping. I have not examined any indubitable material of M. maritima, but Watson’s note, Lond. Jour. Bot. vi. 388 (1847), would seem to indicate that M. azorica and M. maritima merely represent ecological phases of one species. The preface of Seubert’s Flora Azorica is dated April 1844, and it hence seems probable that the name M. maritima has priority over M. azorica. However, due to my lack of precise know- ledge of M. maritima I am treating the present species under the well understood name given by Watson. In Petard & Proust’s flora of the Canaries, pg. 274, as well as in Bornmiiller’s list, Engler, Bot. * Jahrb. xxxiii. 466 (1903), the present species is treated as M. macro- ealycina Coss. Myosotis macrocalycina Coss., Batt. & Trab. Fl. Alg. i. 604 (1888-90), is a very closely related plant of northern Algeria, which has the pubescence and habit of M. azorica and appears to differ only in its slightly longer fruiting pedicels and larger fruiting calyces. It seems not at all improbable that M. azorica may have still earlier synonyms in M. latifolia Poir. Encye. Suppl. iv. 45 (1816), or M. oblongata Link in Buch, Phys. Beschr. Can. Isl. 144 (1825). The original descriptions of these latter species are so brief as to be indefinite and a study of the types must precede an attempt at pre- cisely settling their status. Both came from the Canary Islands. 6. M. alpestris Schmidt, Fl. Boehm. iii. 26 (1794); Schinz & Thell. Vierteljahrs. Naturf. Ges. Ziirich xviii. 469 (1923). M. sylvatica, var. alpestris Koch, Syn. Deutsch. Fl. 504 (1838). M. pyrenaica, var. alpestris Schinz & Thell. Vierteljahrs. Naturf. Ges. Ziirich liii. 558 (1908). M. olympica Boiss. Diagn. ser. 1, pt. 4, 50 (1844). M. rupicola Smith, Eng. Bot. xxxvi. t. 2559 (1814).—Aretic-alpine situations from Alaska and the lower Mackenzie River Valley, south- ward along the Rocky Mountains to Coloradc and Oregon (acc. Howell); also in the higher mountains of Europe. Many writers consider this plant to be only a phase of M. sylvatica, but in the STUDIES IN THE BORAGINACEAE 43 American and European material available it seems quite distinct in its very densely tufted, very scaly, perennial base, and arctic- alpine range. The garden plants current as M. alpestris are horti- cultural forms of M. sylvatica. . M. virginica (L.) BSP. Prel. Cat. N. Y. 37 (1888). Lycopsis virginica L. Sp. Pl. 139 (1753). M. verna Nutt. Gen. ii. add. (1818). M. macrosperma Engelm. Am. Jour. Sci. ser. 1, xlvi. 98 (1844). M. verna, var. macrosperma Chap. Fl. South. U. S. 333 (1860). M. vir- ginica, var. macrosperma Fernald, Rhodora x. 55 (1908). M. inflexa Engelm. |. e.—Maine to Florida, west to Ontario and Texas; reap- pearing in Idaho, Oregon, Washington, and British Columbia. In the South and West the plants tend to grow more rankly and to have the fruiting calyces over, rather than under 5:'mm. in length; this coarser plant is the var. macrosperma. ' §. M. sylvatica Hoffm. Deutsch. Fl. ed. 1, 61 (1791).—Native of Europe. This is the common forget-me-not of the garden of which there are a number of horticultural strains. It frequently escapes, as for instance in Newf.!, Que.!, N. H.!, Vt.!, Mass.!, Mich.!, and Calif.! 9. M. arvensis (L.) Hill, Veg. Syst. vii. 55 (1764); Schinz & Thell. Bull. Herb. Boiss. ser. 2, vii. 339 (1907). M. scorpioides, var. arvensis L. Sp. Pl. 131 (1753). M. intermedia Link in Schultz, Fl. Starg. suppl. i. 12 (1819).—Native of Europe. Introduced in Newf.!, Que.!, N. B.!, N. S.!, Me.!, Vt.!, Mass.!, R. I.!,; Conn.!, N. Y.!, Ohio!, Ore. |, B. C. 10. M. micrantha Pall. in Lehm. Neue Schr. Naturf. Ges. Halle iii. pt. 2, 24 (1817); Schinz & Thell. Bull. Herb. Boiss. ser. 2, vii. 339 (1907). M. arenaria Schrader in Schultz, Fl. Starg. suppl. i. 12 (1819). M. stricta Link in R. & S. Syst. iv. 104 (1819).—Native of Europe. Introduced in N. H.!, Mass.!, N. Y.!, N. J.!, Ohio!, Ind.!, Ont.! 11. M. versicolor (Pers.) Smith, Eng. Bot. xxxvi. sub. t. 2558 (1814); 1. e. vii. t. 480, fig. 1 (1798). M. arvensis, var. versicolor Pers. Synop. i. 156 (1805). Anchusa lutea Cav. Icones i. 50, t. 69, fig. 1 (1791), excl. syn. (?) M. lutea Pers. 1. c., not Lam. (1778); Schinz & Thell. Bull. Herb. Boiss. ser. 2, vii. 338 (1907); Rendle & Brit. Jour.’ Bot. xlv. 440 (1907).—Native of Europe. Introduced in Que.?, Mass.!, Del.!, Ore.!, B. C. 12. M. sp. indet.—There is a specimen in the Gray Herbarium col- lected July 2, 1912 by C. K. Dodge in partially shaded high ground on Mackinac Island, Port Huron, Mich., which in gross aspect sug- gests M. sylvatica, but has white corollas with concave limb 3-5 mm. broad, and leaves rather sharply acute. It also suggests M. virginica 44 JOHNSTON but differs in its large corollas, long spreading pedicels, and more loosely branched slender stems. More material of this form is desired and particularly information as to whether it be a native or escaped garden plant. The plant appears to be an undescribed spe- cies, but to describe it upon the data available seems entirely inad- visable. 20. Cerinthe [Tourn.] L. Sp. Pl. 136 (1753); Gen. Pl. 66 (1754). Cerinthe major L. Sp. Pl. 136 (1753).—Native of the Mediter- ranean region. Occasionally cultivated. Found escaping from a garden in Quebec by A. S. Pease in 1904. 21. Asperugo [Tourn.] L. Sp. Pl. 138 (1753); Gen, Pl. 67 (1754). Asperugo procumbens L. Sp. Pl. 138 (1753).—Native of Eurasia. An infrequent introduction. Known from Mass.!, N. Y., N. J.!, Penna.!, Del.!, D. C.!, Ohio, Minn., Alb., Utah!, Ore.!, Wash., B. C. 22. Amsinckia Lehm. Delect. Sem. Hort. Hamburg 7 (1831). Benthamia Lindl. Nat. Syst. 241 (1830).—A review of the Amsinc- kias north of Mexico has been published by Macbride, Contr. Gray Herb. n. s. xlix. 1-16 (1917). The genus consists of three very pro- nounced series. Of these Macbride has succeeded in giving a satis- fying elucidation of the two with lustrous and tessellate nutlets. The third and largest series, that with dull rugose nutlets, deserves further study though from its exceptional and baffling complexity there seems little hope that students can ever arrive at agreement regarding its treatment. The South American species of Amsinckia belong to the last section and need careful study. Superficially, however, the southern plants belong to a single species, the bibliography of which is as follows.—AMSINCKIA ANGUSTIFOLIA Lehm. Delect. Sem. Hort. Hamburg 7 (1831); Bettfr. Fl. Argent. iii. 177, t. 108 (1901). Ben- thamia angustifolia Druce, List Brit. Pl. 103 (1908). Lithospermum calycinum Moris, Mem. Acad. Torino xxxvii. 98, t. 22 (1833). chilense Colla, Mem. Acad. Torino xxxviii. 127 (1834). A. parviflora Bernh. Del. Sem. Hort. Erf. (1833). A. angustifolia, var. pseudo- lycopsioides Colla in Gay, Fl. Chile iv. 473 (1849). A. pseudolycop- sioides Speg. Anal. Soc. Cient. Argent. liii. 137 (1902). A. basi staminea Cesati, Atti Acc. Se. Nap. ser. 5, vii. 14 (1873). B. basi- staminea Druce, Rep. Exch. Cl. Brit. Isl. iv. 298 (1916). Eritrichium Mandonii Ball, Jour. Linn. Soc. Bot. xxii. 51 (1885). There is a species of doubtful status reported from Mexico. This STUDIES IN THE BORAGINACEAE 45 is A. mexicana Mart. & Gal. Bull. Acad. Brux. xi. 339 (1844), which is said to have come from near Morelia, Michoacan. I have seen no specimens of the species nor of any other Amsinckia coming from central or southern Mexico, nor is there, according to Mr. Paul C. Standley, any material of Amsinckia in the United States National Herbarium coming from that region. The plant is described as having corollas plicate in the throat which suggests that it may prove to be a Heliotropium. 23. Selkirkia Hemsley, Bot. Challenger i. pt. 3, 47 (1884). Selkirkia Berteri (Colla) Hemsley, Bot. Challenger i. pt. 3, 48, t. 57 (1884); Brand, Pflanzenr. iv. Fam. 252, i. 16 and 163 (1921); Skottsberg, Nat. Hist. Juan Fernandez ii. 163 (1921). Cynoglossum Berteri Colla, Mem. Acad. Torino xxxviii. 132, t. 43 (1834).—A shrub endemic to Masatierra of the Juan Fernandez Islands. The genus is monotypic and very distinct, appearing to be most related to Hackelia. 24. Cryptantha Lehm. in F. & M. Ind. Sem. Hort. Petrop. ii. Krynitzkia F. & M. 1. ¢. vii. 52 (1841). Piptocalyx Torr. in Wats. Bot. King Exped. 240 (1871), not Oliver (1870); Johnston, Contr. Gray Herb. n. s. Ixviii. 55 (1923). Eremocarya Greene, Pittonia i. 58 (1887); Johnston, |. c.56. Creeneocharis Giirke & Harms in E. & P. Nat. Pflanzenf., Gesamtreg. 462 (1899). Wheelerella Grant, Bull. So. Calif. Acad. v. 28 (1906).—This is the largest and most difficult genus among the American members of the subfamily. It is re- stricted to America and has two definite centers, one in western United States and the other in middle and northern Chile. A critical descriptive monograph of the North American species is in preparation. The South American species, all different from those north of the Equator, are in distressing need of attention, but their satisfactory elucidation must await the attention of the student who can study them in the field and have access to Philippi's types at Santiago, Chile. Reiche, FI. Chile v. 217-236 (1910), has given most of the bibliography and a rough classification of the very numerous Chilian species. The treatment while difficult of use is nevertheless very valuable in showing the probable status and relationships of most of Philippi’s briefly described species. The following species from Peru appear never to have been properly named under Cryp- tantha. A 46 JOHNSTON Cryptantha limensis (A. DC.), comb. nov., based on Eritrichium limense A. DC., Prodr. x. 133 (1846), had its type collected at Lima, Peru. This has oblong obtuse leaves, leafy spikes, coarsely hirsute calyx-lobes, and inconspicuous tubular corollas. Cryptantha granulosa (R. & P.) Johnston is the only other member of the genus described from Peru, although material recently received from that country seems to contain several additional species, all undescribed. 25. Oreocarya Greene, Pittonia i. 57 (1887). This group has been recently revised by Macbride, Contr. Gray Herb. xlviii. 20-38 (1916). It may be strongly questioned whether it is generically distinct from Cryptantha, particularly so when the uth American members of the latter genus are considered. No one has yet pointed out characters for Oreocarya which are consistently diagnostic, nor has a rather superficial search on my part resulted in the discovery of any. The group is maintained here as a genus pend- ing detailed studies to be reported on later. 26. Plagiobothrys F. & M: Ind. Sem. Hort. Petrop. ii. 46 (1835). Echidiocarya Gray, Proc. Am. Acad. xi. 89 (1876). Sonnea Greene, Pittonia i. 22 (1887). Allocarya Greene, Pittonia i. 12 (1887).—For a synopsis of this genus see Johnston, Contr. Gray Herb. n. s. lxviii. 57-80 (1923). 27. Mertensia Roth, Cat. Bot. i. 34 (1797). Pneumaria Hill, Veg. Syst. vii. 40 (1764).—Macbride, Contr. Gray Herb. n. s. xlviii. 1-20 (1916), has recently revised the true Mertensias of western America. Bibliography and descriptions of the relatively few species left untreated by Macbride may be found by consulting Gray, Synop. Fl. N. Am. ii. pt. 1, 199-201 (1878), and Britton & Brown, Illust. FI. ed. 2, iii. 82-83 (1913). In the past the genus has been almost always referred to the tribe Lithospermeae, but the attachment of its nutlets and its simple stigma clearly point to a relationship among the Eritrichieae. Suggestive also of this latter relationship is the consistently boreal distribution, the Lithospermeae being a prevailingly austral group. 28. Trigonotis Stev. Bull. Soc. Nat. Moscou xxiv. 603 (1851). Trigonotis peduncularis (Trev.) Benth. Jour. Linn. Soc. xvij. 384 (1879). Myosotis peduncularis Trev. Mag. Ges. Nat. Fr. Berlin vii. 147, t. 2, fig. 6-9 (1816). Eritrichium pedunculare A. DC. Prodr. x. 128 (1846).—Native from extreme southeastern Europe through middle and northern China to Japan. Although apparently hereto- STUDIES IN THE BORAGINACEAE 47 fore unreported from America, this species was collected July 12, 1893 in the vicinity of Nanaimo, Victoria Island, British Columbia, by John Macoun (no. 694) and distributed as Myosotis arvensis. It is a roadside weed in Asia and probably introduced in America in ballast. 29. Lappula Moench, Meth. 416 (1794). Echinospermum Sw. in Lehm. Asperif. i. 113 (1818). Staurina Nutt. Jour. Acad. Philad. i. 182 (1848) Kry To SPECIES AND VARIETIES. Nutlets with marginal prickles in at least two r Fruit in situ ca. 3 mm. high; nutlets with oma granulations uniform; introduted plants. fi. . 6646 echo ccd oats L. echinata. Fruit in situ 4-5 mm. high; nutlets with dorsal cieaaiatae igh fobs middle and low towards sides; plants native av stern United States. Plant suffruticos, bushy-branched, green ; pubescence short, appre sparse; fruit in situ ca. 5 mm. high... 2. L. cenchrusoides. Plant nitirely” herbaceous, usually branched above and strict, ——, pabeoee rather long and — Tuit in e 4 wR es a ees 3. L. Fremontii. Nutlets pith mergin at lao de finitely i in a single r — distin psa or nearly so, not saiaaaiis upulate ua Style cineated. by atletes pais of nutlets ane ay oe wit th short obscure glochids. . ner ap saan 4 aes . brachystyla. well developed prickles....................... L. Redowskii. Tubercules on nutlets numerous, not in regular ae udinal rows; native of western America.....var. occidentalis. oe on — few, arranged in regular longitudi- ws; ballas cloaks native of Eurasia......... var. patula. Cee ae files Fone vats é forming a conspicuous maple ie structure on back of some or all the nutlets............ . L. texana. Nutlets gee opin pli one yal sabe with distinct prickles t with a cupulate m Plant eee branched above tall BS gi Sister ey EN var. genuina. Plant decumbent or Price Sa branched at base, te r. heterosperma. Nutlets homomorphous, all with confluent prickles ype Cc Ww spr Plant ched abov pi nce villous...... var. columbiana. Plant branched at base; ever eee eerie var. Cu margin deep, Ankevior parts of nutlets tubeless ee Fy var. homosperma. Anterior parts of nutlets smooth..................-- var. coronata. . Lappula echinata Gilib. F1. Lith. i. 25 (1781). Myosotis Lappula L. <. Pl. 131 (1753). Echinospermum Lappula Lehm. Asperif. i. 121 (1818). - Lappula Karst. Deutsch. fe 979 (1880-83).—Native of Eurasia. Introduced in Newf.!, VP Shi NN, BE. ST 48 JOHNSTON Me.!, N. H.!, Vt.!, Mass.!, R. I., Conn.!, N. Y.!, N. J.!, Penn., Ont.!, Ind.!, Mich.!, Wis.!, Tll.!, Minn.!, Mo.!, N. D.!, S. D.!, Neb.!, Kan., Tex., Alb.!, Calif., B. C.!, Alaska! 2. L. cenchrusoides Nels. Bull. Torr. Bot. Cl. xxvi. 243 (1899); Nels. & Macbr. Bot. Gaz. Ixi. 38 (1916).—A local species known only from southeastern Wyoming. Characterized by its habit and very large fruit. 3. L. Fremontii (Torr.) Greene, Pittonia iv. 96 (1899); Nels. & Macbr. Bot. Gaz. Ixi. 37 (1916). Echinospermum Fremontii Torr. Pacif. R. R. Rep. xii. pt. 2, 46 (1860). -L. erecta Nels. Bull. Torr. Bot. Cl. xxvii. 268 (1900).—Assiniboia to Wyoming, Utah, and Idaho. As here used the binomial covers the plant originally de- scribed, namely Suckley’s collection from the “Upper Missouri River.” The collection made by Fremont, which was originally cited and from which the species apparently got its name, is probably L. Redowskii, var. occidentalis inasmuch as the specimen is given as having come from the southern Sierra Nevada where only the latter species is known to occur. There are hence good reasons for con- sidering the Fremont collection as the type of L. Fremontii and for referring the species to synonymy under the variety of L. Redowskit. Indeed I should be inclined to do this were the status of the alternate name, L. erecta, an unclouded one. Nelson & Macbride, |. c., say “ L. erecta is clearly a synonym” of L. Fremontii as here accepted, yet in the original description of the former the nutlets are described as having “a single marginal series of about 10 aculeae,” whereas the plant concerned has numerous aculeae in several series. 4. L. brachystyla (Gray) Macbr. Contr. Gray Herb. n. s. xlviii. 40 (1916). Echinospermum brachycentrum, var. brachystylum Gray, Proc. Am. Acad. xxi. 413 (1886).—Known only from the type col- lected in 1885 at Spence’s Bridge, Thompson River, in southern British Columbia. The species is very distinct among the American representatives of the oe 5a. L. Redowskii (Horn m.) Greene, v ar. occidentalis (Wats.) Rydb. Contr. U. S. Nat. Herb, iii. 170 (1895); Nels. & Macbr. Bot. Gaz. Ixi. 38 (1916). LEchinospermum Redowskii, var. occidentale Wats. Bot. King Exped. 246, t. 23, fig. 9-10 (1871). L. occidentalis Greene, Pittonia iv. 97 (1899). Cynoglossum pilosum of Nutt. Gen. i. 114 (1818); not R. & P. (1799); T. & G. Pacif. R. R. Rep. ii. 171 (1855) and Nutt. Jour. Acad. Philad. i. 182 (1848). L. Redowskii, . var. ee MacMill. Metasp. Minn. Valley 441 (1892). L. pilosa Hitche. Spr. Fl. Manhattan 30 (1894). C. Nuttallii Spreng. Syst. i. ROOTS Te SEE 5 re Ty Ales we Onaga aa NEE ee eam STUDIES IN THE BORAGINACEAE 49 566 (1825). Pectocarya lateriflora, var. Nuttallii Braid, Pflanzenr. iv. Fam. 252, i. 95 (1921), name only. E. strictum Nees in App. Neuwied, Trav. 17 (1841), not Ledeb. (1829). E. Redowshkii, var. strictum Wats. 1. c. 247, name only. (?) E. pilosum Buckl. Proc. Acad. Philad. 1861, 462 (1861). (?) L. montana Greene, Pittonia iv. 96 (1899). L. calycosa Rydb. Bull. Torr. Bot. Cl. xxviii. 30 (1901). E. patagonicum Speg. Anal. Soc. Cient. Argentina liii. 79 (1902). L. patagonica Druce, Rep. Bot. Exch. Cl. Brit. Isl. iv. 630 (1917). L. leucotricha Rydb. Bull. Torr. Bot. Cl. xxxvi. 676 (1909).—Native of Asia and America. In the New World occurring in Patagonia and from Saskatchewan, North Dakota, Oklahoma, and New Mexico westward. Extending northward to Alaska. In United States introduced east of its natural range in Me.!, Mass.!, Mich.!, Wis.!, Minn.!, and Mo.! 5b. L. Redowskii, var. patula (Lehm.) Nels. & Macbr. Bot. Gaz. Ixi. 39 (1916). Echinospermum patulum Lehm. Asperif. i. 124 (1818). —Native of southeastern Europe and adjacent Asia. Occasional in America as a ballast plant, and known from Penn.!, and Ore.! I am not thoroughly satisfied that this introduced plant can always be ° separated from indigenous material by the characters given in the key. 6a. L. texana, var. genuina. L. terana (Scheele) Britt. Mem. Torr. Bot. Cl. v. 273 (1894); Nels. & Macbr. Bot. Gaz. Ixi. 40 (1916). Echinospermum texanum Scheele, Linnaea xxv. 260 (1852). E. Re- dowskii, var. cupulatum Gray, Bot. Calif. i. 530 (1876). L. cupulata Rydb. Bull. Torr. Bot. Cl. xxviii. 31 (1901). E. cupulatum K. Schum. in Just, Jahresb. xxix. pt. 1, 564 (1903). ZL. Redowskii, var. cupulatum Jones, Bull. Univ. Montana xv. 44 (1910). EF. scabrosum Buckl. Proc. Acad. Philad. 1861, 462 (1861).—This, the typical form of the species, is most common and apparently centers in Texas, and from there extends northward through Kansas to North Dakota, and southward into Coahuila. A discussion of the application of the name L. cupulatum will be found under L. texana, var. columbiana. In desperation I refer two puzzling collections to the present variety; these are Goodding 1010 from Modena, Utah, and Macbride & Payson 948 from the Grand Canyon, Arizona. Most of the fruits on these collections have heteromorphous nutlets and most of the latter have conspicuous margins which are high about the base of the nutlets and ‘very low or absent about the tip. Confusing is the fact that the later fruit produced by these plants appears to be practically homomorphous and quite like that produced in L. Redowskii, var. occidentalis. It becomes a question, therefore, as to what should be done with the 50 JOHNSTON collections and just how they affect the validity of L. texana. The status of that species having previously been none too satisfactory these collections have made me strongly consider reducing it to vari- etal rank under L. Redowskii. Since, however, the specimens may be abnormal or possibly of hybrid origin, pending the receipt of new data or specimens, I am for the present offering no solution of the problem and am not considering the matter in relation to the distinctness of L. Redowskii and L. texana. 6b. L. texana, var. heterosperma (Greene) Nels. & Macbr. Bot. Gaz. lxi. 41 (1916). L. heterosperma Greene Pittonia iv. 94 (1899). (?) L. desertorum Greene, l. c. 95. L. cucullata Nels. Bot. Gaz. xxxiv. 29 (1902)—Western and middle southern Colorado northward to western Wyoming. Readily distinguished from var. genuina by its low bushy habit and different range. 6c. L. texana, var. columbiana (Nels.), comb. nov. L. columbiana Nels. Bot. Gaz. xxxiv. 28 (1902). LL. infelix Greene, Pittonia iv. 235 (1901). (?) L. anoplocarpa Greene, Ottawa Nat. xvi. 39 (1902). L. ecupulata of Nels. & Macbr. Bot. Gaz. lxi. 39 (1916), and most authors.—Washington, northern Idaho, and northern Oregon. This plant has the habit of var. genuina, but has different nutlets and a widely different range. Most authors have maintained it as a distinct species, but it seems to intergrade with the var. homosperma through the var. foliosa. Nelson & Macbride, |. c., Piper, Contr. U. S. Nat. Herb. xi. 475 (1906), and Rydberg, Fl. Rocky Mts. 717 (1917), have treated the present plant as L. cupulata, typifying it by the specimen from Trinity Mountains, Nevada, cited by Watson, Bot. King Exped. 247 (1871), under “ Echinospermum Redowskii, var. strictum.” It is not clear why L. cupulatum has been so typified. The plant treated by Watson is what is here called L. terana, var. columbiana, but Wat- — son’s trinomial, E. Redowskii, var. strictum, is evidently based upon E. strictum Nees, Append. Neuwied, Trav. 17 (1841), which originally came from the upper Missouri River and appears to be L. Redowskvi, var. occidentalis. In the Botany of California, i. 530 (1876), Gray pro- posed a new trinomial, E. Redowskii, var. cupulatum, apparently be- cause Nees’s binomial was antedated by one of Ledebour’s. Under the new name Gray cited Watson’s trinomial and Nees’s binomial, clearly showing thereby that he considered them synonymous. Gray’s de- scription and the meaning of his varietal name, however, clearly apply to a plant different from that properly covered by the names he cites. When the variety was proposed Gray had six, or possibly se ven, specimens that he referred to it. Of this suite three, or possibly four STUDIES IN THE BORAGINACEAE 51 specimens, are L. texana, var. genuina, one is L. terana, var. hetero- sperma, one is L, texana, var. foliosa, and one is the plant here called L. texana, var. columbiana. It hence appears that E. Redowskii, var. cupulatum Gray is best treated, not as synonymous with E. strictum Nees (and consequently a synonym of L. Redowskii, var. occidentalis), but as synonymous with L. texana, var. genuina which includes the bulk of the original material cited by Gray and which is most ap- propriately covered by his varietal name. 6d. , var. foliosa (Nels.), comb. nov. L. desertorum, var. oliosa Nels. Bull. Torr. Bot. Cl. xxvii. 267 (1900). LL. foliosa Nels. Man. Rocky Mt. Bot. 413 (1909). L. cupulata, var. foliosa Nels. & Macbr. Bot. Gaz. Ixi. 40 (1916).—Plains of southern Idaho and ad- jacent Wyoming and Oregon. By L. texana, var. homosperma (Nels.) Nels. & Macbr. Bot. Gaz. . 41 (1916). L. heterosperma, var. homosperma Nels. Bot. Gaz. xxxiv. 29 (1902). (?) L. montana Greene, Pittonia iv. 96 (1899).— Alberta southward to Colorado. . L. texana, var. coronata (Greene) Nels. & Macbr. Bot. Gaz. Ixi. 41 (1916). JL. coronata Greene, Pittonia iv. 94 (1899).—Southern Arizona. 30. Eritrichium Schrad. in Gaud. FI. Helv. ii. 57 (1828). Eritrichum Schrad. Comm. Goett. iv. 186 (1820), hyponym. Key To Species AND VARIETIES Plant closely pion: pobre nce of leaves not Le Neg. be- yond apex as a conspicu ae osbe 7-13 mm : . broa — Nitlets bordered with tet teeth; plant green, wit arather poewenes: Fee age and Yukon Territory. ae 1. E. splendens. N utlets Sonledalt Wi ow unarmed rim; plant canescent, : a dense baton cence; Montana to Washi ington. ...2. E. Howardi. scoages ieiahé villous; pubescence on leaves = aia beyond pex as a conspicuous tuft; corolla 1-7 m Stern of nutlets with very capate teeth, the pere about Os equalling length of body of nutlet; arctic........ 3. E. aretioides. Margin of nutlets aren or with teeth enok: shorter than body of nutlet; montan Dorsal margin of nut itlet CUO ee 4. E. elongatum. margin of nutlet toothed: ./.2..6051502.5-.%. var. argenteum. 1. Eritrichium splendens Kearney in Wight, Bull. Torr. Bot. Cl. xxix. 410, fig. C (1902).—AxasKa: Old Man Creek, 1901, Mendenhall (type, US). Yuxon: long. 141° lat. 62° 5’, alt. 1800 m., Cazrnes 85868 (G).—A very distinct species which in habit suggests the species related to E. rupestris (Pall.) Bunge. It is known only from Alaska and adjacent Yukon. 52 JOHNSTON 2. E. Howardi (Gray) Rydb. Mem. N. Y. Bot. Gard. i. 327 (1900). Cynoglossum Howardi Gray, Synop. Fl. N. Am. ii. pt. 1, 188 (1878). Omphalodes Howardi Gray, |. c. 423 (1886).—The collections cited by Wight, Bull. Torr. Bot. Cl. xxix. 410 (1902), and by Rydberg, I. c., in addition to that preserved in the Gray Herbarium, shows the species to center on the eastern slope of the cordilleras in western Montana. It extends southward, according to Dr. Rydberg in lit., at least to the east slope of the Big Horn Mts. in northern Wyoming. The species is reported from the Cascade Mts. upon the basis of a collection by Tweedy, but its occurrence in western Washington seems incongruous with its indubitable range, and perhaps Tweedy’s specimen was attributed to the Cascades through mislabeling. 3. E. aretioides (Cham.) A. DC. Prodr. x. 125 (1846); Kurtz, Engler Bot. Jahrb. xix. 471 (1894). Myosotis aretioides Cham. Linnaea iv. 443 (1829). E. nanum, var. aretioides Herder, Act. Hort. Petrop. 1. 535 (1871-72). Omphalodes nana, var. aretioides Gray, Proc. Am. Acad. xx. 263 (1885). E. Chamissonis A. DC. |. ce. nanum, var. Chamissonis Herder, |. ec. O. nana, var. Chamissonis Gray, L e.—Like Macbride, Contr. Gray Herb. n. s. xlviii. 50 (1916), I am unable to maintain E. Chamissonis as distinct from FE. aretioides. The differences pointed out by Wight, Bull. Torr. Bot. Cl. xxix. 408 (1902), appear to be merely ecological. The species is known only from the vicinity of Bering Sea ; 4. E. elongatum (Rydb.) Wight, Bull. Torr. Bot. Cl. xxix. 408, D (1902). E. aretioides, var. elongatum Rydb. Mem. N. Y- Bot. Gard. i. 327 (1900). Oreocarya pulvinata Nels. Bot. Gaz. xl: 63 (1905).—Mountains from Montana to Idaho and eastern Oregon, southward to Utah and northern New Mexico. It is uncommon in the southern part of its range being largely replaced by the var. argenteum. As Macbride, Contr. Gray Herb. n. s. xlviii. 50 (1916), has pointed out, Eritrichium elongatum is very closely related to the ropean E. nanum (All.) Schrad. The only constant difference that I can detect is that the American plant is notably more canescent than its European congener. The differences in nutlets mentioned by Macbride, I fail to find. European plants vary greatly in the development of the dorsal rimming on the nutlets and seem to duplicate all the variations of this structure produced by the American relative. It can be said that E. nanum has a larger and less variable corolla than E. elongatum. On Pikes Peak, Colorado, the latter species is abundant on the alpine sward and shows a complete lack of STUDIES IN THE BORAGINACEAE 53 constancy in corolla-size, the corolla varying from 1-6 mm. in breadth without any particular size predominating. da. E. elongatum, var. argenteum (Wight), comb. nov. E. ar- genteum Wight, Bull. Torr. Bot. Cl. xxix. 411, fig. F (1902).—Most abundant in the mountains of Colorado, particularly in the southern part. It differs from the typical phase of the species only in having the margin of its nutlets jaggedly toothed. 31. Hackelia Opiz in Bercht. Fl. Boehm. ii. pt. 2, 146 (1839). The bibliography and status of this genus was recently reviewed by Johnston, Contr. Gray Herb. n. s. lxviii. 43-48 (1923). Descriptions and keys for most of the species may be found in the basic paper by Piper, Bull. Torr. Bot. Cl. xxix. 535-549 (1902). EXCLUDED AND UNIDENTIFIED NAMES. Anchusa depressa Sesse & Moc. Pl. N. Hisp. 21 (1887).—“ Habitat in Uruapam’’ [Michoacan, Mexico].= HELiorRoPIUM LIMBATUM Benth. _ANCHUSA INCANA Sesse & Moc. Fl. Mex. 30 (1894).—* Nascitur Quauhnahuacae agris” [(?) Cuernavaca, Morelos, Mexico]. Identity wholly obscure. Antiphytum mexicanum DC. Prodr. x. 121 (1846).—* In Mexico in Chilpansigi circuitibus” [Guerrero, Mexico]. = HELioTRoprum CAL- CICOLA Fernald. Batschia albiflora Raf. New Fl. N. Am. pt. 4, 19 (1836).—“ On the River Arkanzas on sand bars.” = Euploca convolvulacea Nutt. (1837), and should therefore bear the name Euploca albiflora (Raf.), comb. nov. CERINTHE LANCEOLATA Sesse & Moc. Pl. N. Hisp. 20 (1887).— “ Habitat Quahunahuacae” [(?) Cuernavaca, Morelos, Mexico]. i has a synonym in Heliotropium lancifolium Sesse & Moc. (1894), but both species are quite obscure. ERITRICHIUM PAMPEANUM Speg. Fl. Ventana 44 (1896).—Sierra de la Ventana, prov. of Buenos Aires, Argentina. Spegazzini, Rev. Fac. Agron. et Veternar. iii. 551 (1897), says regarding E. albiflorum (B. & S.) Griseb. (= Plagiobothrys Lechleri Johnston), “species habitu E. pampeano Speg. sat. similis atque valde affinis.” Eritrichium: pampeanum is described as having subtrigonous smooth shiny nutlets that are completely bordered by an entire coriaceous-membranous ridge or crest. I know of no Plagiobothrys with such nutlets and. ‘suspect that Spegazzini’s plant is a Cryptantha. 54 JOHNSTON Lithospermum aggregatum R. & P. Fl. Peruv. ii. 4 (1799).—* Habitat in collibus aridis Tarmae et Huanuci” [Peru]. = CoLDENIA PARONY- CHIOIDES Phil. Lithospermum angustifolium Sesse & Moc. Fl. Mex. 29 (1894); not Michx. (1803).—‘“ Habitat in Oppido Ario” [Michoacan, Mexico]. This species has an earlier synonym in L. ROSMARINIFOLIUM Sesse & Moe. (1887); not Tenore (1826), but neither species has been recog- nized. LiITHOSPERMUM CILIATUM Raf. New Fl. N. Am. pt. 4, 18 (1836).— “Found by Hart on the Red River and by Drummond in Texas.” Unrecognized but probably belonging in Lithospermum or Oreocarya. Lithospermum cuneifolium Pers. Synop. i. 158 (1805). Based upon L. rncanum R. & P. (1799). LITHOSPERMUM CUSPIDATUM Raf. New FI. N. Am. pt. 4, 18 (1836). —“In Alabama and Georgia.” The identity of this species is wholly obscure. Lithospermum dichotomum R. & P. FI. Peruv. ii. 5, t. 111¢ (1799).— “Habitat Peruviae arenosis, versus Lurin, prope castrum Pacha- camac.”’ = CoLDENIA DIcHOTOMA (R. & P.) Lehm. LITHOSPERMUM FLORIDANUM Raf. New FI. N. Am. pt. 4, 18 (1836). —“In Florida.” Seeds said to be pilose, and hence perhaps a - Heliotropium. LiTHOSPERMUM FLAVUM Sesse & Moc. Fl. Mex. 30 (1894).— “Habitat in montibus Zitdcudro inter et Malacatepec interjectis” {Michoacan, Mexico]. Wholly obscure. Lithospermum gracile Raf. New Fl. N. Am. pt. 4, 17 (1836).—“ Sent me from Alabama and found on Red River Arkansas and Texas.” = HELIOTROPIUM TENELLUM (Nutt.) Torr. LirHosPERMUM INCANUM R. & P. FI. Peruv. ii. 4 (1799) —“ Habitat in Peruviae collibus aridis versus Tarmae oppidum.” Probably a Lithospermum and perhaps L. peruvianum A. DC. LiITHOSPERMUM OBTUSIFLORUM Sesse & Moc. Fl. Mex. 29 (1894).— “Habitat in montibus Oppido de El Valle” [(?) state of Mexico]. Wholly =— : NIFOLIUM Sesse & Moc. Pl. Nov. Hisp. 20 (1887); not Pek (1826). —“ Habitat in Oppido Ario” [Michoacan, Mexico]. Unrecognized. Lithospermum um tenellum Nutt. Trans. Am. Philos. Soc: ser. 2, v. 188 (1837). —* In ee Paes cee ica th TENELLUM (Nutt.) Torr STUDIES IN THE BORAGINACEAE 55. Lycopsis AXILLARIS Raf. New FI. N. Am. pt. 4, 19 (1836).—“ Found in Florida.” Identity wholly obscure: Myosotis cymosa Nutt. in Hook. Kew Jour. Bot. iii. 294 (1851); nom. nud. Mrootted GRANDIFLORA HBK. Nov. Gen. et Sp. iii. 90, t. 199 (1818).—“ Crescit in Regno Quitensi, inter Tiscan et Alexiei ” Ecuador, Perhaps a Lithospermum. MyosorTis MEXICANA Sesse & Moe. Fl. Mex. 31 (1894).—“ Habitat in temperatis N. Hispan. montibus.” Probably Heliotropium cal- cicola Fernald. Myosotis sER1cEA Nutt. in Hook. Kew Jour. Bot. iii. 295 (1851), nom. subnud.—* On the hills of the Upper Platte River.” = Oreocarya sp. Myosotis URUGUAYENSIS Arechav. Anal. Mus. Nat. Montevid. ser. 2, i. 68, fig. 4 (1911); Fl. Urug. iv. 180, fig. 12 (1911).—Dept. of Tacuarembo, Uruguay. Unrecognized. ONOSMODIUM ERIOCAULON A. DC. Prodr. x. 70 (1846).—‘“In Mexico ad Cordilleram Guchilaqua.” Described from Berlandier 1020, material past flowering. It is probably a Lithospermum. PULMONARIA CANADENSIS Yong, Cat. 44 (1783), nom. subnud.— From the Virginias and Carolinas. Probably asynonym of Mertensia virginiana (L.) DC. PuLMONARIA ELLIPTICA Raf. New FI. N. Am. pt. 4, 17 (1836).— “In the Apalachian Mts. of Virginia and seeped Probably a synonym of Hackelia virginiana (L.) Johnsto SYMPHYTUM AMERICANUM Yong, Cat. 47 (1783), nom subnud.— From the Virginias and Carolinas. Wholly obscure Symphitum fruticosum Sesse & Moc. Fi. N. Hisp. 21 (1887); A. DC. Prodr. x. 121 (1846).—“ Habitat in frigidis Chilpanzingi montibus”” [Guerrero, Mexico]. = HeLiorroptum caLcicota Fernald. YMPHYTUM HIRSUTUM Raf. Med. FI. ii. 95 (1830)—“‘A native American species . , found west of the Mississippi, in prairies and glades, and cateoated a Bartram’s garden.”’ Probably S. officinale L. and not a native. SympHytumM Minus Yong, Cat. 47 (1 —_ nom. subnud.—From the Virginias and Carolinas. Wholly obscure 2. A TENTATIVE CLASSIFICATION OF THE SOUTH AMERICAN CoLDENIAS. __ The present paper has oo nai an _— at definitely al- described a by Ruiz & Pavon. ie a 56 JOHNSTON This problem has involved a study of the South American species of Coldenia. As these have never been treated comprehensively it is hoped that notes presented in the form of a tentative classification may be useful. Through the kindness of Dr. N. L. Britton and Dr. W. R. Maxon I have had the privilege of studying the South American material of the genus contained in the collections of the New York Botanical Garden and the United States National Herbarium, in addition to that of the Gray Herbarium. Mr. J. F. Macbride of the Field Museum of Natural History has also allowed me to study such material of the genus as he has collected in Peru. The material thus assembled is not so extensive nor so complete as desirable, more than half the collections coming from the Galapagos Islands, and the re- mainder containing much duplication of collections and localities. It has been consequently necessary to treat a number of species en- tirely upon the basis of published descriptions. In doing this I have derived much help from Reiche’s treatment of the Chilian species, FI. Chile v. 187-191 (1910), and have usually accepted his verdict on the status of Philippi’s rather numerous proposals. The genus Coldenia breaks up into four sections, Eucoldenia, Eddya, Sphaerocarya, and Tiquiliopsis. Of these only the last is unrepre- sented in South America. The sections of the genus may be distin- guished as follows:— Corcila appendaged within =>. 2.0.05 7 OO ee Tiquiliopsis. Corolla naked within sia ke pra oe at the base by a small circular areola, Or OF Wes PRN Se ke ’ Sphaerocarya. Netlets attached deren ge meee or less elongate: r small, in situ not closely crowded, ‘distinct, eriorly witha medial ventral keel nd Wie eos caw Eddya Nutlets portend when - situ closely juxt ae yeaidiony Fo age ibe the: fat faces 0 adjacent satin. gc age Oe ieiecc cress, Eucoldenia The section Eucoldenia is oe broken up into three distinct series. First, genuine Coldenia consisting of C. procumbens, a weedy annual herb of the Old World tropics, which has rugose veiny leaves, and lobed glandular-pubescent tuberculate acuminate fruit. Second, _ the series Stegnocarpus containing C. canescens and C. Gr reggit which is composed of suffruticose plants of the adjacent portions of United States and Mexico Possessing tomentose unveined leaves, and smooth STUDIES IN THE BORAGINACEAE 57 non-acuminate fruit. The genus Tiquilia Pers. was based entirely upon C. dichotoma, In the past it has been incorrectly used to cover species properly referred to Eddya and Tiquiliopsis. The nutlets of C. dichotoma and C. grandiflora are large and have flat closely appressed anterior faces, quite as in the other species of the section Eucoldenia, in fact were all four instead of merely two opposite nutlets developed, they would not be hemispherical but quite like those in C. canescens. The species of Coldenia on the Galapagos Islands, which was first taken as constituting a new genus and described as Galapagoa Darwini by the younger Hooker, is evidently a very close ally of the continental plant called C. paronychioides. Obviously of the same group of im- mediate relatives are the North American C. hispidissima, C. tomen- tosa, C. mexicana, C. Purpusti, and C. cuspidata. The seven species mentioned form so homogeneous a group that the synonymy of the genera Eddya and Galapagoa can be questioned by no one. The section to which I refer the seven species, the oldest name of which is Coldenia § Eddya Gray, is characterized by its rounded more or less ovoid not closely crowded dark roughened nutlets which are attached ventrally along a well-marked elongated ventral keel or groove. For the three species, C. litoralis, C. atacamensis, and C. parviflora, I am proposing a new section which may be called, Sphaerocarya (type, C. litoralis Phil.). These species have practically spherical nutlets which just touch one another and are basally attached by a small circular scar. The petioles in the species are usually without the villous-ciliations characteristic of the section Eddya. The species of the section Sphaerocarya center in northern Chile. The final section of Coldenia is Tiquiliopsis. The group is char- acterized by its appendaged corollas. The fruit in the three species referred to the section is widely variable and I am not at all satisfied that the division is a natural one. In C. Nuttallii the nutlets are elongate-ovoid with a ventral scar quite like that in the species of the section Eddya. The nutlets are nearly smooth, however, and the slender petioles are not villous-ciliate. In C. Palmeri (= C. brevi- calyx Wats.!) the nutlets are quite like those in the South American species of the section Sphaerocarya, in fact this species seems typical of the section in all but its appendaged corollas. Coldenia plicata (= C. Palmeri of ner, not Gray) has ovoid nutlets with a very short stipe-lil tt t unlike that in any other species of the genus. It perhaps finds its nearest relations in the section Sphaeroca ocarya. T have not attempted to place sectionally C. decumbens Hauman, a el l »;31pRARe 58 JOHNSTON since I know it only from description and am not confident that it is properly included in Coldenia. If actually a member of Coldenia it may — to the section Eddya or perhaps is worthy of special Chaveni L. Sp. FE-125 (1753); Gen: Pl: 61 (1754). Tiquilta Pers. Synop. i. 157 (1805). Monomesia Raf. FI. Tellur. iv. 87 (1836). Galapagoa Hook. f. Proce. Linn. Soc. i. 277 (1845); Trans. Linn. Soc. xx. 196 (1847). Stegnocarpus T. & G. Pacif. R. R. Rep. ii. pt. 2, 169 (1856). Ptilocalyx T. & G. 1. c. 170. Eddya T. & G. 1. ¢. 170. Lobophyllum F. Muell. in Hook. Kew Jour. Bot. ix. 21 (1857). Tiqui- liopsis Heller, Muhlenbergia ii. 239 (1906). Key To SpEcizs. Nutlets plano-convex, 1.5-2.5 mm. long, conspicuously mottled. Corolla 3-4 mm. broad; stamens pe: os esa exserted. ...1. C. dichotoma. Corolla 9 mm. broad; stamens included................ . C. grandiflora. Nutlets — to narrowly ovoid, aot at all plano-convex, concolorous. Nutlets 1-2. mm. long, smooth, ovoid; plant annual; Argen- at aide" Neral Pere ALIRT A Cena ik lr gamma 3. C. decumbens. Nutlets 0.5-1 mm. long, ovoid pebbled; plants perennial, . becoming gin ec ee base; Peru and Chile mainly. Nutlets narrowly greid with a narrow longitudinal ventral scar; petioles rather short, usually conspicuously Fruiting calyx ovoid, 1-2 mm. long, obscured by and simulating the f oliage, base of lobes closely investing back of nutlets; Galapagos Islands............... 4. C. Darwint. Fruiting calyx cccgno 2. 5-3. 5 mm. goa evident, ; C. paronychioides. i aac daigpond a or ee so, with a small crea b basal gape slender, usually not strikingly villous- Leet binds aban. < Sap with 2-3 pairs of veins, about es, margins straight; plant regularly CM OS a eee ia sea te es 6. C. litoralis. bend iene cables obtuse, with 4-5 pairs of veins, vidently longer than petioles, margins usually t irregularly branched. Co ‘8-10 mm. long, twice fonath of calyx..... a: “i Spam pt Corolla ca. 5 mm. long, about equalling calyx..... C. parviflora. 1. Co_pENntIA pDicHoTtoma (R. & P.) Lehm. Asperif. i. 9 (1818). Lithospermum dichotoma R. & P. Fi. Peruy. ii . 5, t. Mle (1799). ace. Steud. Nomencl. ed. 1, 212 (1821). C. Dombeyana Juss. acc. Buek, Index iii. 108 (al oss bee Pied from Peru. The type material ] arin, ahant 9 ‘ south ~~ ssssiwnw STUDIES IN THE BORAGINACEAE 59 of Lima, and near the fort of Pachacamac. I have studied an ill- labeled specimen preserved in the Gray Herbarium, and material gathered by Macbride (no. 5946) near Lima. From their detailed agreement with the original description and plate, the two collections are evidently typical C. dichotoma. These plants mature two single- seeded nutlets the backs of which are strongly convex, finely tessellate- granulate, and more or less ashy and strongly mottled with brown. The anterior face of the nutlet is flat with a small rim, and is orbicular or orbicular-ovate in outline. Four ovules are produced but only two opposite ones mature in each fruit. The peculiar hemispherical nutlets are very characteristic, but may not always be developed since it is likely that the plants, occasionally at least, produce more than two mature nutlets in a fruit. The exserted stamens are sufficient to identify the species in any case. The calyx-lobes are divided to the ase. 2. C. GRANDIFLORA Phil. Cat. Pl. Itin. Tarapaca 55 (1891); Reiche, Fl. Chile v. 188 (1910).—I know this species only from descriptions. It appears to have been known to Reiche and Philippi only from the type-collection made at Médanos de Pica in the province of Tarapaca, Chile. It is evidently a close ally of C. dichotoma, but is clearly different in its corolla. 3. C. DECUMBENS Hauman, Apuntes Hist. Nat. Buenos Aires i. 55 (1909); Anal. Soc. Cient. Argentina Ixxxvi. 301 (1918).—The type of this species came from 2350 m. altitude in the valley of the Rio Tupungal near the confluence of the latter with the Rio Mendoza in northwestern Mendoza, Argentina. I have seen no material of the species. Its nutlets, described as “ lisses, brillants, ovoides, 4 extrémité supérieure assez pointue,”’ are apparently quite different from those of any other South American species of the genus. The species is said to be a slender herbaceous annual, though perhaps it may be perennial as are the other South American species, and has like some of them been described from year-old material. This latter seems improbable, however, since Hauman mentions collections made on widely differ- ent dates. 4. C. Darwin (Hook. f.) Gray, Proc. Am. Acad. v. 341 (1862). Galapagoa Darwini Hook. f. Trans. Linn. Soc. xx. 196 (1847); Anderss. Stockholm Acad. Handl. 1853, 86, t. 16, fig. 1 (1854). G. fusca Hook. f. ]. e. 197; Anderss. I. c. 87, t. 16, fig. 2. C. fusca Gray, |. c—Endemic to the Galapagos Archipelago. I have studied a series of twenty- nine different collections from the Galapagos Islands and am forced to the conclusion that the archipelago has but one variable species of 60 JOHNSTON Coldenia. The island plants vary considerably in compactness, size, and pubescence of leaves, but these differences seem clearly responses to different habitats. Galapagoa Darwini and G. fusca appear to be merely trivial inconstant pubescence-forms which are not restricted to separate islands, but occur together on most of the latter. I find not the slightest tendency for the reoccurrence of the particular com- bination of characters present in Hooker’s original specimens. 5. C. PARONYCHIOIDES Phil. Cat. Itin. Tarapaca 55 (1891); Reiche, FI. Chile v. 190 (1910). Lithospermum aggregatum R. & P. FI. Peruv. li. 4 (1799). C. aggregata Rusby, Deser. New Sp. So. Amer. 106 (1920); not based upon L. aggregatum R. & P. (?) C. tenuis Phil. 1. ¢.; Reiche, |. c.—Prru: Payta, Safford 9 (US), Ball (NY, G), Williams 2913 (NY, type of C. aggregata Rusby); Yonga, Wilkes Exped. (US, NY, G). Bottvia: La Paz, Rusby 1430 (NY).—The species appears to range from northern Peru to northern Chile. Lithospermum ag- gregatum R. & P. is the oldest name applied to the species, but it is not taken up since the combination under Coldenia would have an homonym in Coldenia aggregata Rusby. Coldenia aggregata Rusby was based upon material from Payta in northwestern Peru and was named without any reference to the species, founded upon material from near Tarma and Huanuco in west central Peru, which was much earlier described as Lithospermum aggregatum by Ruiz & Pavon. Coldenia aggregata Rusby, being a synonym, L. aggregatum R. & P. might according to the International Rules of Nomenclature be combined under Coldenia and the resulting combination be taken as the proper name for the species; however, since this procedure would give the same species synonymous homonyms it is thought that the possibilities of confusion arising from this awkward situation justifies the acceptance of Philippi’s specific name. The original Philippian collection came from Pampa del Tamarugal, province of Tarapacd, Chile. 6. C. Lrrorais Phil. Fl. Atacamensis 37 (1860); Reiche, Fl. Chile v. 189 (1910). C. mitis Phil. Anal. Univ. Chile xc. 229 (1895). C. virens Phil. 1. c.—Cutre: deserts of Atacama, Morong.1257 (US, NY, G); Caldera, Ball (NY, G).—Apparently endemic to the province of Atacama, Chile. The type of C. litoralis came from the coastal sands at Caldera, Chile. I am following Reiche, 1. ¢., in reducing Philippi’s later species. 7. C. ATACAMENSIS Phil. Fl. Atacamensis 37 (1860); Reiche, FI. _ Chile v. 189 (1910)—Known to me only from description. Ac- TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 61 - cording to Reiche, |. ¢., it occurs in the interior of the province of Antofagasta, Chile. 8. C. PARVIFLORA Phil. Cat. Pl. Itin. Tarapaca 55 (1891); Reiche, Fl. Chile v. 190 (1910). .C. elongata Rusby, Deser. New Sp. So. Amer. 106 (1920).—Yura, “ Peru,”’ Williams 2562 (NY, type of C. elongata). Arequipa, “ Bolivia,” Williams 2521 (NY). Southern slopes of Cha- chani Mountain near Arequipa, Hinkley 11 (G).—Apparently ranging from southern Peru and adjacent Bolivia, southward to the province of Tarapacd, Chile. The species seems to differ from C. atacamensis in its small corollas and more northern range. EXcLUDED SPECIES. Coldenia glabra Phil. Anal. Univ. Chile xe. 230 (1895). According to Reiche, Fl. Chile v. 191 (1910), this is a synonym of FRANKINIA . GLABRATA Phil. Coldenia phaenocarpa Phil., Cat. Pl. Itin. Tarapaca 55 (1891). Reiche, |. ¢., has shown this to be a Heliotropium, H. PHAENOCARPA (Phil.) Reiche. II. TAXONOMIC RECORDS CONCERNING - AMERICAN SPERMATOPHYTES. 1. PARKINSONIA AND CERCIDIUM. Both Cercidium and Parkinsonia appear to be valid. genera, but their acceptance can be justified only after a slight change in their traditional limits and the stressing of characters not heretofore emphasized by those who have maintained the genera. All authors have sanctioned the recognition of Parkinsonia, and most recent students have insisted on the validity of Cerctdium. The writers on the flora of Argentina, disregarding the obviously albuminous seeds produced by their “Brea,” have persisted in treating their species of Cercidium as a Caesalpinia, but there are good morphological char- acters and overwhelming usage arguing against this treatment. Karsten, Fl. Colomb. ii. 25, t. 113 (1862), proposed a segregate genus Rhetinophloem, but later, Engler’s Jahrb. viii. 346 (1887), reduced it to a subgenus of Cercidium. In 1876, Watson, Proc. Am. Acad. xi. 135 (1876), concluded that the characters used to separate Cercidium and Parkinsonia failed when applied to the North American species and proceeded to merge the two genera. Watson’s opinion prevailed until 1889 when Sargent, Gard. & Forest ii. 388 (1889), reopened the 62 JOHNSTON matter and reestablished Cercidium. This decision was based on his finding that “the valvate imbrication of the calyx-lobes is constant in Cercidium and that the thickened glandular claw of its petals does not appear in our species of Parkinsonia,”’ and further and more important that “the fruit which in Parkinsonia is linear, rounded and torose,” is in Cercidium “linear-oblong compressed, and in one of our species slightly contracted between the seeds.” Since 1889 the acceptance of Cercidiwm has been almost universal. A study of the material in the Gray Herb. (G), and some very in- teresting collections from the United States National Herbarium (US) which were loaned me through thei of Dr. J. N. Rose, has shown that the sepals are not always simply valvate in Cercidium, but rather that there seems to be intraspecific variation between valvate and induplicate-valvate conditions. In fact these variations were used y Karsten as characters of his subgenera, Eucercidium and Rhe- tinophloem. I am unable to see even specific differences in this character since both conditions frequently occur on the same specimen. In Parkinsonia, as currently taken, the calyx-lobes are not always imbricate, since P. microphylla has valvate calyx-lobes quite indis- tinguishable from those in indubitable Cercidiums. It is only in Parlinsonia aculeata (the type species of Parkinsonia) and P. africana that very definitely imbricate aestivation occurs, and then in such an extreme form as not to be even approached by the conditions in the _ species of Cercidium or the other species of Parkinsonia. A study of the claw of the petal has revealed no difference between the two genera. The characters of generic import currently ascribed to the fruit of Cercidium and Parkinsonia fail miserably when they are applied to all the known species of the genera. Parkinsonia aculeata and P. africana are very closely related and are obviously congeneric, but the latter has the legumes somewhat compressed with thickened almost straight margins, so that they appear quite different from the terete torose pods of P. aculeata. The fruits of P. africana and C. teranum are very similar as to margin, width, and thickness, in fact the only conspicuous difference is in the length. Practically every species of Cercidium frequently has the margin of its legumes more or less sinuate or even contracted between the seeds. Sargent, Man. Trees N. Am. 585 (1922), finds generic difference in the relations of the spine to the leaves, giving the leaf-rachis as spin- escent in Parkinsonia and not so in Cercidium. In fact the leaf-rachis _is spinescent only in P. aculeata and P. africana. In the plant called TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 63 P. microphylla there are no spines, whereas in its near relative, Cercidium sonorae, the spines are clearly below the leaf-rachises and exactly as in other indubitable Cercidiums. Past authors have accepted Parkinsonia microphylla Torr., as a mem- ber of Parkinsonia, being largely influenced in so doing, no doubt, by the extreme similarity in its fruit with that of P. aculeata. However, P. microphylla differs from P. aculeata and P. africana in having short corymbose rather than very elongate racemes, short filiform terete rather than extremely long phyllodial leaf-rachises, and calyx-lobes that are valvate or a trifle induplicate-valvate rather than extremely imbricate in the bud. It is significant that these characters are pos- sessed by all the species of Cercidium, and that P. microphylla in habit resembles the species of Cercidium much more than it does the two species of genuine Parkinsonia. In brief, Parkinsonia microphylla is clearly a species of Cercidium and is consequently transferred to that genus. Parkinsonia thereby reduced to homogeneity consists only of P. aculeata and P. africana. Parkinsonia and Cercidium, then, become two well marked and very natural genera eminently worthy of recogni- tion, as may be appreciated by a study of the following key :— Inflorescence an elongate raceme, 8-18 te long; rachis of pinnae llodial, 1-6 dm. long, apparently borne on the spines; cal Sate very vane arigiog ae ere Parkinsonia. Inflorescence 5 short cor 1-3 ¢ ong; rachis of pinnae cm. long, ithe ee phyllodial, clearly borne below tie spines; calyx-lobes valvate or induplicate-valvate in bud . é Cercidium. Parkinson [Plum.] L. Sp. Pl. 375 (1753); Gen. Pl. 177 (1754). _ Keay To SpEctrgs. Rachis of pinnae winged, 1.5-3 mm. wide, 15-60cm. sgt oar ca. 25-jugate; legumes to orose, — margin very strongly uadubeee but scarcely thickenéd..... ......-+.. ss a,- + 1. P. aculeata. his of pinnae not winged, 1 mm. im ide, 5-15 5-15 sist long; P pnae 15-jugate or less so; legumes Ww i margins ns slightly adhoate and prikectty Gictened 2. P. africana. 1. PARKINSONIA ACULEATA L. Sp. Pl. 375 (1753). P. spinosa - HBK. Nov. Gen. et Sp. vi. 335 (1823). P. Thornberi Jones, Contr. W. Bot. xii. 12 (1908).—Native of America, but widely paps, over the warmer parts of the world. For notes regarding distri- bution see the account by Sargent, Silva N. Am. iii. 87, t. "81 | Aisa The name “ Parkinsonia spinos nosa HBK” is evidently a lapsus meant for P. aculeata L. The specimen cited by Kunth is apparently not 64 JOHNSTON P. aculeata, however, but is the collection which finally became the type of Cercidium spinosum Tul. 2. P. arricana Sond. Linnaea xxiii. 38 (1850); Engler, Veg. der Erde ix. Band 3, Heft 1, 501, fig. 267 (1915).—Known only from South Africa.—ArFrica: on a stony plain, Damaraland, Marloth 1216; without locality, Zeyher 557. Cercipium Tul. Arch. Mus. Paris iv. 133 (1844). Rhetinophloeum Karsten, Fl. Colomb. ii. 25, t. 113 (1862) Key To SpEctzs. Pinnae 1-3-jugat sta’ € on ps ala ey pinnae 1- or ee eas PES Bans EG ek EONS A fe a or Bg ae are . C, texanum. Bees pee Uae Gog ten ronan Wales antares . C. macrum. 3. C. floridum. wig . C. peninsulare. Pinnae 4-15-jugate Leaves muse simply pinnate, primary rachises un- saghtege pinnules minute, ca. 1-1.5 mm. broad, 1-2.5 e;N Oth Ameria... ck. he ‘microphyllum. Pict “with Rage Be spines: le a 7. C. andicola. ong. Plant unarmed; cso orbicular or ovate; fruit we -lard ican Leaves ooneisads pubescent, cinereous Leaves glabrate or —— glaucous or ellow-green. aan s 4-7 cm ong, 4 mm. thick, margin strongly te; ; Sono OF Sa ee a ee 6. C. sonorae. Lae ce 3-5 cm. long, i mm. thick, margin not at all undulate; Argentina... ..2.%............ .:. 9. C. australe. 1. CeRcIDIUM TEXANUM Gray, PI. Wright. i. 58 (1852). Parkin- sonia texana Wats. Proc. Am. Acad. xi. 136 (1876).—Southwestern Texas. Texas: Uralde, 90 mi. northwest of San Antonio, 1879, Palmer 270; Eagle bd _ Palmer 271; Rio Grande, 1848, Wright; without locality, Pope; western Texas, 1849, Wright 149; without locality, 1852 (Wright); « New Mexico,” ret, Wright 1 16. 2G. macrum, sp. nov glab Ped spinis 0.7 mm. longis; fol opacis sparse inconspicueque Goss unijugis, rachibus primariis 4-8 mm. longis; pinnis trijugis, rachillis 7-15 mm. longis; pinnulis oblongis vel oblon ovatis emarginatis vel rotundatis sessilibus 46 mm. longis 2-3.5 mm. latis, basi acutis TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 65 saepe paullo obliquis, costis conspicuis; floribus in racemas pauciflores axillares subcorymbosas congestis; ovario glabro; leguminibus valde ‘compressis 2-3 mm. crassis 8-11 mm. latis 3-6 cm. longis, margine rectis vel paullo undulatis; seminibus brunneis oblongis ca. 7 mm longis 4 mm. latis—C. floridum of authors. C. floridum of Sargent, Silva N. Am. iii. 83-4, t. 129 (1892), as to plant described and pictured. Hoopesia arborea Buckley, Proc. Acad. Philad. 1861, 453 (1861), in part; 1870, 137 (1870); Gray, Proc. Acad. Philad. 1862, 163 (1862).— Texas: Rio Hondo, Cameron Co., Chandler 7034; Fordyce; 1905, Tracy 9065; hills, La Salle Co., 1881, Buckley. Tamauuipas: San Fernando to Ji imeney, 1902, N. Aobh 6608; vicinity of Victoria, 1907, Palmer 125. Nuevo Leon: mesas near Monterey, 1889, Pringle 2537 (rypE, Gray Herb.); near Pesquerea Grande, northwest of Monterey, 1847, Gregg; without locality, Gregg; Nuevo Leon, Berlandier 3142 (707).—This species is the well known one of Texas and northeastern Mexico cur- rent as Cercidium floridum Benth. That species, however, is based upon a flowering specimen made by Coulter who did not collect within the range of C. macrum. Coulter did, on the other hand, collect within the range of the common “Palo Verde” of Arizona. It therefore seems evident that Bentham’s name cannot properly be applied to the plant here called C. macrum, since it was impossible for Coulter to have collected it, and since it seems more than probable that Bentham’s name should be applied to the closely related species of Arizona which Coulter could scarcely have missed collecting. The Gregg and Wislizenus collections cited under Gray’s original de- scription of “Cercidium floridum Benth.” represent C. macrum, those of Emory and Fremont are true C. floridum. From their source, Corpus Christi, it seems probable that the fragments of Cercidiwm in the complex type of Hoopesia arborea belong to C. macrum rather than to C. teranum. Hoopesia arborea was based, according to Gray, Proc. Philad. 1862, 163 (1862), upon material representing three distinct leguminous genera (Cercidium, Acacia, and Pithecollobium). No particular one of the three elements predominates in Buckley's description nor in his material. I am not definitely associating Hoopesia arborea with any species, but am dropping it from con- sideration as a nomen confusum. 3. C. rLormpuM Benth. in Gray, PI. Wright. i. 58 (1852). Parkin- sonia florida Wats. Proc. Am. Acad. xi. 135 (1876). P. Torreyanu Wats. l. c. C. Torreyanum Sarg. Gard. & Forest ii. 388 (1889); Silva N. Am. iii. 85, t. 80 (1892)—Southern Arizona to the Colorado Desert of California, southward into adjacent Sonora and Lower 66 JOHNSTON California.—ARrIzona: near Fort Yuma, 1880, Lemmon 39, 40; rocky hills at Camp Grant, 1867, Palmer 62; Gila [River], Sutton Hayes 209; valley of the Gila, Emory; Gila [River], 1852, Thurber; near the Colorado [River], Bigelow; without locality, 1867, Palmer. Catt- FORNIA: Palm Springs, alt. 150-200 m., Parish 4115; in desert sand, Palm Spring, alt. 120 m., Spencer 755; Palm Springs, Eastwood 3004. —For the present use of the name “Cercidiwn floridum Benth.” see discussion under C’. macrum. 4. C. PENINSULARE Rose, Contr. U. S. Nat. Herb. viii. 301 (1905); Goldman, Contr. U. S. Nat. Herb. xvi. 336, t. 114 (1916).—Southern Lower California—Lower CALIFORNIA: La Paz, 1890, Palmer 112; San José del Cabo, 1897, Anthony 363.—This differs from C. floridum in scarcely more than pubescence. In range, however, it is widely separated from that latter species. 5. ©. microphyllum (Torr.) Rose & Johnston, comb. nov. Parkin- sonia microphylla Torr. Bot. Mex. Bound. 59 (1859); Pacif. R. R. Rep. iv. 82 (1857); Sargent, Silva iii. 91, t. 132 (1892).—Southern Arizona, Sonora, and middle and northern Lower California.— ARIZONA: poor rocky hills, Camp Grant, 1867, Palmer 63; Gila [River], Sutton Hayes 217; Williams River, Bigelow; desert near Fort Yuma, 1880, Lemmon 41; Maricopa, 1881, Parry; Lowell, 1884, W. F. Parish 49; without locality, 1871, Lieut. Wheeler. 6. C. sonorae Rose & Johnston, sp. nov., arboreum; ramulis breviter strigosis canescentibus; spinis 0-9 mm. longis; foliis sparse pubes- centibus 1—2-jugis, rachibus primariis 0-5 mm. longis apice subulato _caduco 1-2 mm. longo; pinnis 2-4-jugis 12-32 mm. longis; pinnulis subsessilibus 1.5-2 mm. latis 2-3 mm. longis oblongis basi oblique rotundatis, apice obtusis mucronatis costa conspicua; inflorescentiis axillaribus corym cemosis paucifloris; dense breviterque pubes- centibus; leguminibus 4-7 em. longis 4 mm. crassis 7-8 mm. latis’ margine valde undulatis; seminibus oblongis 1 em. longis 4 mm. latis. —Sonora: Guadeloupe, Feb., Parry 321a (G); Torres, 1903, Coville 1664 (US); Guaymas, 1922, Oreutt 1196 (US); dry hills in the vicinity of Guaymas, 1910, Rose, Standley & Russell 12586 (typr, U. S. Nat. Herb.).—A very well marked species apparently nearest to C. micro- phyllum, but differing in its thorny stems, larger, petioled and fre- quently several-jugate leaves, more compressed less torose legumes, and much smaller seeds. The species is known only from Sonora. 7. C. ANDICOLA Griseb. Abh. Ges. Wiss. Goett. xxiv. 114 (1879). Caesalpinia praecox, var. andicola Hoss. Bol. Acad. Cérdoba’ xxvi. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 67 145 (1921).—Northern Argentina and southern Bolivia.—Bourv1a: Toldos near Bermejo, 1800 m. alt., 1903, Fiebrig 2193 (G). 8. C. prancox (R. & P.) Harms in Engler Jahrb. xlii. 91 (1908). Sappania praecox R. & P. Fl. Peruv. t. 376, ined. Caesalpinia praecox H. & A. Bot. Miscl. iii. 208 (1833), as to description and accepted name. Cercidium spinosum Tul. Arch. Mus. Hist. Nat. Paris iv. 134 (1845). Rhetinophloem viride Karsten, Fl. Colomb. ii. 25, t. 113 (1862). Cercidiwm viride Karsten in Engler, Jahrb. viii. 346 (1887). Cercidium plurifoliolatum Micheli, Mém. Soc. Phys. et Hist. Nat. Genéve xxxiv. 269, t. 18 (1903). Cercidium Goldmani Rose, Contr. U. S. Nat. Herb. viii. 301 (1905). Cercidium unijugum Rose, |. e.-—Extreme middle-western Peru to northern Venezuela, the Leeward Islands (Curagao fide Urban, Margarita), Santo Domingo (fide Urban), and southern and western Mexico.—Ecvuapor: between Loja and Portovelo, Rose, Pachano & Rose 23330 (US). VENEZUELA: El Valle, Margarita Island, alt. 200 m., 1903, J. R. Johnston 31 (G). La Vela de Coro, 1917, Curran & Haman 423, 468, 499 (G); Cumana, 1917, Curran & Haman 1251 (G). Mexico: Santa Lucia, Oaxaca, 1908, Purpus 3187 (G); hills, Dominguillo, Oaxaca, alt. 900 m., 1895, L. C. Smith 437 (G); between Teotitlan and San Antonio, Oaxaca, alt. 900 m., 1907, Conzatti 2119 (G); San Luis, Guerrero, alt. 100 m., 1899, Langlassé 933 (G); Orizaba, Vera Cruz, Bottert 994 (G); Maria Madre Island, Tepic, 1897, Maltby 62 (US); vicinity of Fuerte, Sinaloa, 1910, Rose, Standley & Russell 13516 (US); San José del Cabo, Lower Calif., 1911, Rose 14466 (US); Guaymas, Sonora, 1897, Maltby 186 (US); dry cactus plain near Navojoa, Sonora, 1910, Rose, Standley & Russell 13176 (US); La Tinajo, Sonora, 1890, Hartman 241 (G).—I have searched in vain for characters of sufficient constancy to justify the treating of the Mexican material as representing a valid species. It can be generally said that the South American material has smaller flowers and darker branches than the Mexican, but these characters are erratic and can scarcely be used for specific dis- tinction. The type of Ruiz & Pavon’s species came from middle western Peru and is distinct from the plant of the Argentine which authors, following Hooker & Arnott, have confused with it. A dis- cussion of this subject will be found under the next species. 9. C. australe, sp. nov., arboreum vel arbusculum 1-8 m. altum spinosum; ramulis glabratis vel sparsissime strigoso-pilosis striatis pallide viridibus; spinis 8-15 mm. longis rectis; foliis bipinnatis 1- jugis glabratis vel sparsissime adpresseque pilosis’ flavo-viridibus, rachibus primariis 5-10 mm. longis; pinnis 5~7-jugis rachibus 1-2 68 JOHNSTON cm. longis; pinnulis oblongis 1-2 mm. latis 2-5 mm. longis obtusis basi subobliquis; petiolulis ca. 0.2 mm. longis; floribus breviter corymboso- racemosis; petalis luteis immaculatis maxime ca. 1 cm. longis; legum- inibus membranaceis valde compressis acutis reticulatim venosis glabratis 3-5 cm. longis 8-10 mm. latis 1-1.5 mm. crassis margine rectis.—Caesalpinia praecox of H. & A., Bot. Miscl. iii. 208 (1833); as to plant of Gillies, not as to R. & P. name or plate. Cercidium praecox of Manganara, Anal. Soc. Cient. Argent. Ixxxvii. 163 (1919), and other recent authors. Cercidium andicola, var. petiolata Chod. & Hass. Bull. Herb. Boiss. n. s. iv. 828 (1904).—Central Argentina and adjacent Paraguay.—ARGENTINA: Andalgala, Prov. Catamarca, Jorgensen 1119 (G); vicinity of General Roca, Rio Negro, alt. 250- 360 m., 1915, W. Fischer 20 (ryPE, Gray Herb.); San J uan, Mendoza, Jameson (G).—Although this tree of the Argentine has long borne the name Caesalpinia praecox or Cercidium praecox, it seems clear that it has been improperly so called. Caesalpinia praecox was published by Hooker & Arnott in 1833 after they had concluded that a plant collected by Gillies in Mendoza was specifically identical with Ruiz & Pavon’s manuscript species, Sappania praecoz, and its unpublished plate. Gillies’s specimen represents the Argentine plant that has ed as C. praecox and which is here named, C. australe. Since Ruiz & Pavon did not collect in Argentina or at all close to the range of C. australe, and described in their flora only material collected by themselves or Dombey in either Chile or Peru, it must be clear that Sappania praecor R. & P. cannot be the plant that Hooker & Arnott thought it to be. Since there is a Cercidium which occurs in middle-western Peru where Ruiz & Pavon did much collecting, it certainly seems much more natural to assume this to be the Cercidium collected and illustrated by them. This plant I doubtfully associate with the Venezuelan species, current as Cercidium spinosum Tul. Hooker & Arnott’s description of “ Caesalpinia (Sappania) praecor agrees with the diagnosis as well as does the Argentinian plant, I am and am consequently giving a new name to the “ Brea”’ of the Ar- gentine. The plant from Mendoza differs from the common form of C. australe in having its ovary velvety-pubescent. It may be worthy TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 69 2. NEW oR OTHERWISE NOTEWORTHY PLANTs. Persea Hartmanii, sp. nov., glabrata arborea 6-9 m. alta; ramis juventate fulvo-tomentane: foliis anguste ellipticis vel lanceolatis glabratis 6-11 cm. longis 15-36 mm. latis basi cuneatis apice saepe acutis subtus pallidioribus; petiolis 1-2 em. longis vetustis glabratis canaliculatis; inflorescentiis paucifloris laxe paniculatis 7-10 cm. longis glaberrimis folia paullo superantibus; pedunculis gracilibus 4-7 em. longis; pedicellis 4-6 mm. longis; floribus ca. 4 mm. longis; calycis lobis ovatis acutis ca. 2. 5 mm. longis | subsequalibus extus glaberrimis intus paullo strigosis medium i parte decidua; staminibus exterioribus introrsis 2 mm. longis, antheris oblongo- linearibus ca. 0.9 mm. longis 4-locularibus, loculis inferioribus latis longioribus superiores lateraliter oblique tangentibus, filamentis sparse villosis angustis; staminibus introrsis exteriores simulantibus sed basem versus biglanduliferis, glandulis ovatis sessilibus; staminodiis ca. 9 mm. longis, capitulis acute sagittatis quam stipes duplo longiori- bus, stipitibus villosis crassis capitula latitudine aequantibus; ovariis glabris globosis.—Mexico: Batopilas, Chihuahua, April 1892, C. V. Hartman 1029 (type, Gray Herb.).—Although in gross habit suggest- ing a species of Phoebe, in technical characters this seems clearly a Persea. It is well marked in the latter genus appearing to be without immediate relatives. The outstanding features of the plant are its glabrous ovary and flowers, and peculiar calyx-lobes. These latter have a transverse medial line of abscissign. P podadenia Blake, var. glabriramea, var. nov.; caulibus et pedunculis glaberrimis castaneis—Mexico: Orizaba, Boftteri 81 (typr, Gray Herb.), 1166. ps Aides of P. podadenia comes from Durango and is very densely st Phoebe longipes, sp. nov., uhetina: foliis lanceolatis 12-22 cm. longis 4—5.5 cm. latis longe faleate-actaninpatis subtriplinerviis supra laevibus subtus paullo pallidioribus delicate reticulatis: bast rotun- datis vel centibus fuscis; inflorescentiis laxissime paniculatis ca. 1 dm. longis quam foliae valde brevioribus paucifloris; pedunculis gracilibus ca. 6.5 cm. longis ascendentibus; pedicellis gracilibus 15-25 mm. longis; bracteolis deciduis; floribus ca. 4 mm. longis; calycis segmentis ovatis acutis 2.5 mm. longis subaequalibus extus glaberrimis intus paullo strigosis; staminibus exterioribus 1.8—2 mm. longis introrsis, antheris oblongis ca. 1 mm. longis, loculis 4 valde 2-seriatis; staminibus in terioribus exterioribus similibus sed extrorsis et cum filamentis saepe villosis basin versus abrupte dilatatis et cum glandulis disciformibus 70 JOHNSTON munitis; staminodiis ca. 1.6 mm. longis, capitulis ca. 4 mm. longis deltoideo-ovatis stipite compresso strigoso duplo longioribus; ovariis glabratis globoso-ovoideis; stylo glabro quam ovarium 14 longiori.— Mexico: without precise locality, Pringle 8829 (rypr, Gray Herb.).— Apparently most closely allied to P. salicifolia Nees, but differing in having longer petioles and pedicels, as well as larger leaves and flowers. The type was probably collected in south-central Mexico. Misanteca costaricensis, sp. nov., paniculata arborea 7-10 m. alta; ramis glabris; foliis elliptico- vel obovato-oblongis glabris concoloribus 10-14 em. longis 3-5 em. latis, subtus sparse delicateque reticulatis; supra inconspicue reticulatis, basi cuneatis, apice breviter acuminatis, petiolis ca. 15 mm. longis quadrangularibus; inflorescentiis paniculatis 5-10 em. longis dense brunnescenterque villoso-velutinis; floribus ad apices pedunculorum laxe pyramidaliterque congestis; pedicellis 1-2 mm. longis sparse villosis; calyce ca. 1.5 mm. longo infundibuliformi vel subobconico brunneo extus glabrato; lobis 6 late ovatis obtusis intus sparse pubescentibus, exterioribus ca. 0.8 mm. longis, interioribus ca. 0.6 mm. longis; staminibus 3 connatis ca. 1 mm. longis conspicuis; antheris glabratis crasse bilocularibus; filamentis latis dense villosis; glandulis 3 distinctis semiorbicularibus vel ellipticis basi ad andro- phorum adfixis; staminodiis absentibus: ovario glabro ovato quam stylus 14 breviori—Costa Rica: Santiago Hills near San Ramon, 1100 m. alt., June 1, 1901, A. M. Brenes 14403 (vypr, Gray Herb.).— Related to the West Indian M. triandra (Sw.) Mez, but differing in its_villose-velutinous ‘inflorescence, distinct entire glands on the androphore, and larger more elongate less gradually acuminate leaves. It is readily separated from the remaining species of the genus, all close allies of M. capitata C. & S., by its velutinous loose paniculate inflorescence and smaller less prominently reticulate eaves. Misanteca Peckii, sp. nov., arbor parva cum floribus racemosis; petiolis 9-12 mm. longis, inflorescentiis racemosis 2-3 em. longis; paucifloris brunnescenti-velutinis;. floribus atro-coeruleis valde diametro 1.5 mm. altis subcarnosis; pedicellis 0.2-1 mm. longis lobis 6 valde 2-seriatis, exterioribus latissime tri- angularibus ca. 1.5 mm. latis 0.6 mm. altis, interioribus deltoideis obscuris ca. 0.3 mm. altis; staminibus 3 in’ androphorum latum pyra- TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 71 midale connatis; antheris 2-locellatis late obovatis glabratis, locellis supra medium impositis; filamentis nullis (?); staminoidiis nullis; glandulis confluentibus ad basem androphori in annulum undulatum confluentibus; ovario globoso glabrato; bacca ellipsoidea ca. 2.2 em. longa 11 mm. diametro coccinea; cupulo cyathiformi conspicue duplo- marginato, basi incrassato.—British HonpurRas: locality not given, 1905-7, M. E. Peck 826 (ryer, Gray Herb.).—A very distinct species and perhaps representing an undescribed genus since it differs from the other members of Misanteca in its few-flowered racemose in- florescence, depressed broad rather pulpy flowers, strongly bicolored leaves, and stamens which appear to lack filaments and are connate to form a very broad pyramidal rather than columnar androphore. The type was probably collected in middle eastern British Honduras. Calliandra socorrensis, sp. nov., fruticosa glabra; caulibus rigidis laxe multiramosis; foliis glaberrimis numerosis; pinnis bijugis 15-30 mm. longis; foliolis 5-7-jugis congestis 4-12 mm. longis 24.5 mm. latis firmis oblongis apice obtusis basi valde obliquis; stipulis per- sistentibus oblongo-lanceolatis ca. 2-3 mm. longis acutis; pedicellis 1-2 mm. longis; pedunculis axillaribus 2-3 em. longis ascendentibus; floribus ignotis; leguminibus glabris ca. 5-spermis, valvis membran- aceo-coriaceis reticulatis incrassato-marginatis 5-6 cm. longis 8-9 mm. latis—Mexico: Socorro Island, 1897, A. W. Anthony (TYPE, Univ. Calif. Herb. no. 83534).—Most nearly allied to C. formosa (Kunth) Benth., a species ranging widely over northern and western Mexico, from which it differs in its much branched compact wood habit, and much smaller crowded leaflets. cidia acuminata (Blake), comb. Ichthyomethia acuminata Blake, Jour. Wash. Acad. ix. 249 (191 Piscidia communis (Blake), comb. oh Ichthyomethia communis ye Jour. Wash. Acad. ix. 247 (19 ‘ dia grandifolia (Donn. Sm.), enti nov. Derris grandifolia i ri Bot. Gaz. lvi. 55 (1913). Fagonia cretica, var. canariensis, var. nov., prostrata ramosa F. creticae, var. typicam simulans sed differt floribus et apn perspicue minoribus; fructibus 5-6 mm. crassis: petalis 6-7 mm. longis—Canary IsLAnps: roadside near Guia, Gran PO Dec. 1893, Cook 77; arid open lower parts of Teneriffe, Jan.1845, Bowrgeau 46; littoral belt near Santa Cruz, Teneriffe, April 1888, J. Ball (typz, Gray Herb.).—This plant of the Canary Islands has passed as F. cretica L., but that is a plant of the Mediterranean basin which has fruit 6-7.5 mm. thick and petals 7-10 mm. long. The variety 72 JOHNSTON cana rensis, because of its smaller fruit and flowers, is very suggestive of the common American forms of the genus, and it was perhaps this fact that gave Engler, Veg. der Erde ix. Band 3, Heft 1, 731 (1915), his reasons for treating the plants of America as F. eretica var. chilensis (H. & A.) Engler, var. californica (Benth.) Engler, and var. asper (Gay) Engler. However, Standley, Proc. Biol. Soc. Wash. xxiv. 244 (1911), has pointed out that the American plants are distinguishable from true F. cretica by the possession of a fruit which is noticeably smaller in size, and which has a shorter, scarcely thickened beak. As the American plants are widely separated geo- graphically from their closest ally, F. cretica, of the Old World, and since the collections from the two regions are distinguishable by their fruit, it seems best to consider them specifically distinct, and to take up F. chilensis H. & A. as the comprehensive name for the American relatives of F. cretica. Some writers, such as Anderson, Jour. Linn. Soc. v. suppl. 11 (1860), and the editors of the Index Kewensis, have’ referred prac- tically all the described forms of Fagonia to F. eretica. This treat- ment, however, is extreme, for there appears to be a goodly number of species in northern Africa and southwestern Asia which are well defined entities with characteristic habit and natural ranges. Most of the species have several synonyms, and in a number of cases the current name is not always the oldest. For example, F. indica Burm. f., founded on a good description and plate, has been disre- garded for such comparatively recent synonyms as F. persica DC., F. mysorensis Roth, F. microphylla Boiss., F. myriacantha Boiss., and F. parviflora Boiss. — Protium panamense (Rose), comb. nov Icica panamensis Rose, N. Am. Fl. xxv. 260 (1911). Euphorbia (§Tithymalus) Hinkleyorum, sp. nov.. perennis e radice erecta crassa oriens 1-2 dm. alta glaberrima; caulibus compluribus erectis herbaceis medium versus laxe dichotomeque ramosis; foliis caulinis inferioribus alternatis sparsis, caulinis superioribus et rameis oppositis valde conspicuis, omnibus glabris concoloribus late adfixis 8-24 mm. longis et latis palminerviis distinctis lacerato-dentatis orbicularibus vel obovatis vel subflabelliformibus basi truncatis vel cuneatis edentatis symmetricis; stipulis nullis; involucris solitariis axillaribus glabris cyathiformibus ca. 2 mm. longis ca. 1 mm. longe pedunculatis, fauce cum lobis subulatis vel anguste deltoideis hirsutis ca. 0.5 mm. longis munita; glandulis 5 ca. 1.3 mm. longis et latis, subtus pallidis rariter hirsutis, supra laminato-rugosis brunneis, TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 73 margine retusis vel truncatis undulatis; stylis ca. 1.5 mm. longis, lobis bipartitis glabris; stigmatibus paullo incrassatis; capsula glabra ca. 2 mm. longa ca. 4 mm. pedunculata; seminibus oblongis obscure quadrangularibus ca. 2 mm. longis ca. 1.3 mm. crassis ecarunculatis griseis delicate impresse atromaculatis——PERv: sandy pampa on the southern slopes of Chachani Mountain near Arequipa, alt. 3300 m., March 1920, Mr. & Mrs. F. E. Hinkley 7 (typ, Gray Herb.).—This species evidently belongs to Boissier’s subsection Ipecacuanhae, and probably has its nearest relative in the variable but very different E. portulacoides Spreng., from which it differs in its larger dark- colored coarsely toothed apparently connate (though really distinct) broader leaves, and much more loosely branched habit. In foliage E. Hinkleyorum suggests the Galapagean, E. amplexicaulis Hook. f., in the shape and attachment of its upper leaves. It is, however, not at all closely related to that species. According to its discoverers the plant is known as “esquera”’ and is used in the treatment of fevers. vastrum Hi yorum, sp. nov., annuum simplex erectum 1.5- 4.5 dm. altum glabratum vel apicem versus sparse villosum et setosum; caulibus stramineis gracilibus; foliis palmate 3-foliolatis 34.5 cm. longis; foliolis acutis oblanceolatis lacerato-dentatis vel lobatis gla- berrimis subtus pallidis, dentibus cum seta terminatis; petiolis 1-2.5 mm. longis gracilibus canaliculatis supra minute viscidulo-villosis; stipulis conspicuis oblique acuminateque lanceolatis vel oblongo- lanceolatis 8-13 mm. longis 2-4 mm. latis longe sparseque ciliatis; pedunculis gracilibus 2-6 cm. longis 2-4-floris unilateraliter congesti- floris, ab axillis superioribus orientibus; calycibus ca. 4 mm. longis conspicue ciliatis sparse setosis, fructiferis ca. 4.5 mm. latis, lobis ovatis acutis ca. 2.5 mm. longis; pedicellis 0.5—1 mm. longis; bracteolis 3 lineari-filiformibus 3-4 mm. longis; corollis ochroleucis (in sicco rosaceis) ca. 3.5 mm. longis; tubo staminali 2-2. 5 mm. - longo; stylo 6 1.2mm. longo, lobis 10-13 glabris non dilatatis; stigmatibus capitellatis lobo styli vix crassioribus minute pubescentibus; carpellis oe valde rugosis 1-1.5 mm. longis——Perv: hillsides on the southern slope of Chachani Mountain near Arequipa, 2100 m. alt., March 1920, Mr. & Mrs. F. E. Hinkley 43 (rype, Gray Herb.).—A very distinct annual species, well marked by its glabrous herbage, lacerate-dentate leaves, and conspicuous stipules. In Baker’s synop- sis, Jour. Bot. xxix. 168 (1891), it falls with i. peruvianum (L.) Gray, a species from which it is clearly distinct. It is a pleasure to be able to name this and the previous species in honor of Mr. and Mrs. F. E. Hinkley, who in 1920, during their hours of recreation while connected 74 JOHNSTON with the Arequipa Station of the Harvard College Observatory, made on the southern and lower slopes of the volcano of Chachani (about 10 km. north of Arequipa) a small but highly interesting plant-col- lection which has materially enriched the Gray Herbarium with well prepared material of many new or rare species. Malv: arequipense, sp. nov., annuum herbaceum simplex ~ erectum 5-20 cm. altum stellato-setosum gracile; foliis paucis oblongo- ovatis 3-lobatis sinuato-dentatis 1-3 em. longis 8-15 mm. latis stel- lato-setosis basi obtusis; petiolis 3-13 mm. longis pubescentibus; stipulis lineari-lanceolatis ca. 3.5 mm. longis acutis; pedunculis gracil- ibus 1-3 em. longis unilateraliter 2—6-floris ex axillis superioribus orientibus; pan 4—5 mm. longo stellato-setoso, fructifero 4-5 mm. lato, lobis 2-3 mm. longis ovatis.acutis; bracteolis 3 lineari-filiformibus ca. 2 mm. longis; pedicellis setosis 0.5-1 mm. longis; corolla 4 mm. longa in sicco rosacea, lobis obovatis rotundatis; tubo stamineo 2 mm. longo; antheris 8-10; stylis ca. 1.5 mm. longis ca. 0.5 mm. longe connatis glabris linearibus; stigmatibus capitellatis hirtellis quam styli ramus paullo crassioribus; ovarii loculis 15-18; carpellis oblongo- ovatis 1-1.2 mm. altis ca. 1.5 mm. longis, sinu angusto longo, lateribus medium versus reticulato-rugosis marginem versus crasse rugosis, dorso sparse stellatis marginem versus saepe breviter cristatis, apice cum appendicula aristata hirsuta 2-3 mm. longa subpersistenti or- natis; spermatibus brunneis laevibus compressis pedicellato-ovatis vel cum sinu obliquo.—Peru: hillside on the south slope of Chachani Mountain near Arequipa, alt. 2100 m., March 1920, Mr. & Mrs. F. E. Hinkley 43a (tyPr, Gray Herb.).—In Baker’s synopsis of the Mal- vaceae, Jour. Bot. xxix. 168 (1891), this species falls with M. boliv- anum Baker, but that has much larger petals, much larger pinnatifid leaves, biaristate carpels, and taller growth. cong: , Sp. nov., annuum herbaceum 1-3 dm. altum subsimplex; caulibus flexuosis decidue stellato-tomentosis apicem versus dense lanuginosis; foliis opacis sparse stellatis rhom- boideo-ovatis acutis elobatis triplinerviis grosse sinuato-dentatis basin versus integris 2-5 em. longis 1.5-3 em. latis basi cuneatis, rare foliis palmato-trilobatis 2.5—4 mm. latis basi obtusis lobis obovatis grosse pauceque obtusidentatis; petiolis 5-20 mm. longis stellatis; stipulis ca. 5 mm. longis lanceolatis; floribus in glomerulos densos sessiles terminales 4-8-floros congestis; calyci ca. 5 mm. longi tomen- tosi lobis ovatis sae ca. 3.5 mm. longis; bracteolis 3 linearibus 3-4 mm. longis; corolla pallida ca: 4 mm. longa calycem vix superante, lobis obovatis rotundatis 2.5-3 mm. longis; tubo stamineo 1.5 mm. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 75 longo sparse longeque villoso; columna styli glabra ca. 1.5 mm. longa, lobis ca. 8 filiformibus 0.5—-0.7 mm. longis sparse villosulis; stigmatibus capitellatis; ovario depresse globoso dense stellato-tomentoso; fructu ignoto.—PeEru: rocky ravines on the south slope of Chachani Moun- tain near Arequipa, alt. 2745 m., March 1920, Mr. & Mrs. F. E. Hinkley 37 (vypr, Gray Herb) —Anoandntig related to M. tara- pacanum (Phil.) Baker, but distinguished by its larger green, rather than tomentose, foliage, pale corollas, and non-aristate fruit. alvas Shepardae, sp. nov., annuum humile herbaceum stellato-setosum 2-6 cm. altum multicaule; caulibus erectis vel decumbentibus teretibus viscido-villosis setosis; stipulis scariosis ciliatis 3-4 mm. longis 0.8-1 mm. latis; foliis subsemiorbicularibus vel late ovatis 1-2.5 em. latis 1-2 em. longis palmate lateque trilobatis vel irregulariter dentatis setosis subtus pallidioribus prominenter nervatis basi truncatis vel obtusis; petiolis 1-3 em. longis; floribus axillaribus solitariis 2-5 mm. longe pedicellatis; calycibus ca. 5 mm longis setosis, lobis oblongo-lanceolatis ca. 3.5 mm. longis acutis, fruc- tiferis 5-6 mm. latis; bracteolis duabus 0.5-0.9 mm. infra calycem locatis; corollis purpureis ca. 5 mm. longis sepala ca. 1 mm. super- antibus; tubo stamineo 1.5 mm. longo; stylo ca. 1.5 mm. longo, lobis ca. | mm. longis-glabris vix dilatatis; stigmatibus capitellatis glabris; carpellis ca. 2.5 mm. longis 1.5 mm. latis glabris ovatis valde rugosis. —PeERv: in meadows, Puno, Dec. 1919, Mrs. R. S. Shepard 123 (tTyPE, Gray Herb.).—Apparently a relative of M. pygmaewm (Remy) Gray, but differing from the description of that species in its acute calyx- lobes, pedicellate flowers, more numerous stamens, and glabrous fruit. Perhaps also related to M. mollendéense Ulbrich, but differing from the diagnosis of that species in having larger stipules, shorter pedicels, smaller flowers, smaller glabrate calyces, and leaves which are merely dentate or have three broad. shallow palmate lobes. In naming this species for the collector it is a pleasure to give recognition to the botanical work of Mrs. Shepard who, in company with her husband, has for some years diligently prepared for the Gray Herba- rium excellent specimens of the floras visited during her missionary work i in the region, west of Lake Titicaca and in Tacns-Aniea. ’ sp. nov , annuum h 8-20 em. altum basin versus longe ascendenterque ramosum; foliis 2— 3.5 mm. longis 1.3-3 cm. latis rhomboideo-ovatis irregulariter inciso- serratis vel crenato-serratis, subtus re rugecnenee stellatis pre: minenter nervatis supra plerum umque setosis ni truncatis vel late subcordatis, apice | bane sips lanceolatis 23 bad gee ee is a | ar one ey 76 . JOHNSTON mm. longis; petiolis 1—2.5 cm. longis; inflorescentiis axillaribus, pedun- culis gracilibus 8-16 mm. longis 1-3-floris ibus ca. 4 mm. longis infra medium partitis lobia uminerviis stellato-setosis 2-3 mm. longis anguste acuteque ovatis; bracteolis lineari-filiformibus 2-3 mm. longis; petalis albis vel dilute rosaceis quam sepala 1/5-1/3 longioribus 3-3.5 mm. longis; tubo stamineo 2.5 mm. longo; stylo ca. 2 mm. longo, lobis 6-7 ca. 0.5 mm. longis glabris vix dilatatis; stig- matibus subcapitellatis glabris; ovario glabro; carpellis maturitate reniformibus glabris 2 mm. longis 1.2 mm. latis ca. 0.9 mm. crassis dorso convexis rugosis lateraliter laevibus; seminibus glabris reni- formibus 1.2 mm. longis 9 mm. latis minute faveolatis brunneis.— ARGENTINA: E] Candado, Dept. of Andalgala, Prov. of Catamarca, Sept. 2, 1916, Pedro Jérgensen 1388 (typK, Gray Herb.).—A very distinct addition to the rather small assemblage of annual Malvas- trums. In Baker’s synopsis of the Malvaceae, Jour. Bot. xxix. 168 (1891), the species falls with M. peruvianum (L.) Gray, from which it is unquestionably distinct differing in habit, cut of leaf, calyx, and pubescence. The conspicuous, slender, very few-flowered peduncles of M. catamarcense will usually serve to distinguish it. Jorgensenii, sp. nov., annuum; radice erecta simplici, radicellis fibrillatis divaricatis; caulo erecto gracili subsimplici 4-6 dm. alto ad 3 mm. crasso basin versus suffruticuloso apicem versus folioso decidue stellato-tomentoso; foliis ovatis paucis acutis sparse stellatis 3-lobatis 4-7 em. longis 3.5-5 cm. latis basi truncatis vel sub- cordatis subtus pallidioribus cum nervis prominentibus margine ser- rato-crenatis; petiolis 1.5-2.5 em. longis; stipulis lanceolatis 1-2 mm. longis; inflorescentia terminali 10-30-flora congeste corymbosa; calyce dense stellato-tomentoso ca. 5-6 mm. longo, lobis deltoideis acutis 3-4 mm. longis, bracteolis linearibus 2-3 mm. longis; petalis laete roseis (in sicco coeruleis) ca. 1 em. longis ca. 5 mm. latis oblongo- obovatis rotundatis quam calyx 2-3-plo longioribus; tubo stamineo ca. 6-7 mm. longo; stylo 2-2.5 mm. longo, lobis ca. 2 mm. longis stigma versus subdilatatis glabris; stigmatibus oblongo-capitellatis; ovario dense stellato-tomentoso cylindrato-globoso; fructibus ignotis.— ARGENTINA: near Suncho, Prov. of Tucuman, Sept. 5, 1915, Pedro Jorgensen 1389 (rypE, Gray Herb.).—Distributed as M. peruvianum (L.) Gray, and somewhat resembling that species in its subsimple habit. The new species has, however, smaller stipules, a tomentose rather than stellate-setose pubescence, and conspicuous corymbose gensenit suggests M. capitata (Cav.) Griseb. in ie flora] structures, but that is a branched and densely tomentose shrubby perennial. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 77 Monnina ramosa, sp. nov., fruticulosa 15-20 cm. alta laxe ramosa paullo canescens adpresse compresseque albo-villosa ut videtur fur- furacea (?) annua; foliis lanceolato-linearibus 15-25 mm. longis 1.5- 2.5 mm. latis obtusis integerrimis concoloribus basi attenuatis; pore olis 0.5 mm. longis; stipulis nullis; inflorescentia racemosa 5-10 ¢ longa ca. 1 cm. longe pedunculata ramulos laterales seteieiiga pedicellis ca. 0.7 mm. longis; sepalis ciliato-fimbriatis duobus in- ferioribus ovatis acutis basim versus connatis ca. 1.7 mm. longis, superiori lanceo-ovata 2 mm. longa; alis albis oblique obovatis 44.5 mm. longis 2—2.2 mm. latis basi deltoideis; carina (in sicco) aurea emarginata orbiculari plicata a latere visa oblonga 4-5 mm. longa ca. 2.5 mm. alta obtusa; petalo superiori albo obtuso ligulato;androecio paullo ciliolato; staminibus 8; ovario glabro; stigmate inferiori tuber- culato horizontali; samara oblique obovata membranacea in lateribus medium versus dense albo-strigosa, ala undulata evidenti—Perv: on sandy pampa at 3300 m. alt. on south slope of Chachani Mountain near Arequipa, March 1920, Mr. & Mrs. F. E. Hinkley 13 (vypx, Gray Herb.).—Related to M. Weberbaueri Chodat, Engler Jahrb. xlii. 102 (1908), a species of central Peru, from which it differs in its more elongate narrower leaves, glabrate stigma, and differently shaped floral keel. Oenothera verrucosa, sp. nov., annua herbacea pilosa erecta simplex foliosa 5-25 em. alta; foliis 2.5-6 cm. longis 5-9 mm. latis inconspicue undulatis vel denticulatis concoloribus marginem versus pubescenti- bus, infimis spathulato-oblanceolatis petiolatis, superioribus lance- olatis sessilibus basi rotundatis vel cordatis; floribus axillaribus; alabastro erecto ellipsoideo villoso ca. 1 cm. longo; hypanthio in- fundibuliformi usque ad ca. 3 mm. longo paullo longiori quam lato; sepalis in alabastro apice libris, anthesi separatim reflexis lineari- lanceolatis; petalis luteis obcordatis ca. 5 mm. longis ca. 3 mm. latis; staminibus glabris ut videtur 2-seriatis; i ag anguste oblongis dorsaliter medium versus adfixis; stylo glabro ca. 3 mm. longo; stigmate 4-lobato lobis ligulatis 1-1.3 mm. longis; ovario viscido- villoso; capsulis 15 mm. longis 2-3 mm. crassis sparse villosis oblance- olato-fusiformibus obtusi-quadrangularibus, valvis verrucosis medi- aliter longitudinaliterque costatis; seminibus erectis uniseriatis ellip- soideis pallidis 1-1.4 mm. longis vix rugosis.—PERv: ravines and rocky slopes at 3600 m. alt. on southern slope of Chachani Mountain near Arequipa, March 1920, Mr. & Mrs. F. E. Hinkley 17 (r¥Pe, Gray Herb.) .—Certainly a near relative of Oe. coquimbensis Gay, from which it differs in its narrower more elongate leaves, and short 78 JOHNSTON hypanthium. From descriptions it seems to be also closely related to the Chilian Oe. bracteata Phil., but it differs from that in having a lower stature, more elongate leaves, smaller flowers, and a sparse villous pubescence. The new species is locally known a “chochillo,” and is used for urinal troubles. Jarrilla, nom. nov. Mocinna La Llave, Registro Trimestre i. 351 (1832); not Mocinna Lag. Gen. et Sp. Pl. Nov. 31 (1816); nor Mozinna Ort. Nov. Pl. Descr. Decad. 104 (1798).—The generic name given by La Llave is not only later than the exactly homonymous one proposed by Lagasca, which is now referred to Calea, but is later also than the practically homonymous one published by Ortega. La Llave and Ortega both dedicated their names to Don José Mocifio, latinizing the latter's name slightly differently, Ortega’s genus has been accepted as valid by some recent authors and has appeared in print, at least once, Benth. Pl. Hartw. 9 (1839), spelled “ Mocinna.” Ia therefore discarding the generic name given by La Llave, since it is essentially homonymous, being the same in origin and intent, and since its acceptance would lead to’confusion. The generic name sub- stituted is formed from one of the vernacular names given by La Llave. The genus Jarrilla evidently belongs to the Caricaceae and jts recog- nition adds a fourth well-marked genus to the family. It appears to be most closely related to Carica, but differs in having low herba- ceous stems, tuberous-thickened roots, and cornutely appendaged fruit. Ramfrez, Anal. Inst. Med. Nac. Mexico i. 209-211 (1894) and Naturaleza ser. 2, ii. 449-450 (1894), has excellently shown the numerous characters of the genus by means of a tabular comparison. Jarrilla ylla (La Llave), comb. nov. Mocinna heterophylla La Llave, Registro Trimestre i. 351 (1832): Naturaleza ser. 1, vii. append. 70 (1887). (?) Carica nana Benth. Pl. Hartw. 288 (1848). M. heterophylla, var. Sesseana Ramirez, Anal. Inst. Med. Nac. Mexico i. 207, t. 1-4 (1894); Naturaleza ser. 2, ii. 447, t. 25-26 (1894). C. caudata Brandg. Zoe iv. 401 (1894).—The type of this very variable species is said to have come from Guanajuato, Mexico. Its descrip- tion agrees well with the plates of Ramfrez’s variety Sesseana which presumably was based upon cultivated material from Guanajuato Brandegee’s species Lower California. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 79 Jacaratia costaricensis, sp. nov., dioica arborea pyramidalis ; ramis brunneis aculeatis; aculeis acutis durin ca. 2 mm. altis conicis vel lateraliter compressis; foliis herbaceis longe petiolatis digitatis ambitu rotundatis 10-15 cm. diametro; petiolis 10-16 cm. longis; foliolis 5-7 oblanceolatis integerrimis acuminatis, supra viridibus sub- tus densissime glaucis; foliolo centrali 6-11 cm. longo 2-3 em. lato cum petiolulo 2-3 mm. longo; inflorescentiis masculis axillaribus petiolo brevioribus 6-12 cm. longis laxe racemosi-paniculatis multi- floris, femineis axillaribus erectis longe pedunculatis unifloris; floribus masculis ca. 17 mm. longis extus glaucissimis, in alabastro tenuiter clavatis, lobis ca. 7 mm. longis linearibus subobtusis, tubis tenuiter cylindratis ca. 1.2 mm. crassis intus sparse villosis, staminibus extus strigosis valde inaequalibus, filamentis conspicue connatis, antheris similibus sine connectivo ligulato, processis erectis spiniformibus; floribus masculis ca. 2.7 em. longis glaucis, tubo ca. 2 mm. longo ca. 6 mm. crasso.—Costa Rica: forest of Las Vueltas, Tucurrique, 650- 700 m. alt., Dec. 1898, Adolfo Tonduz 12881 (rypr, Gray Herb.).— Jacaratia costaricensis is evidently a close relative of the Guianian, J. spinosa (Aubl.) DC., but differs from that species in its longer (5-9 mm. long) fruiting peduncles, and very strikingly bicolored foliage. The new species is quite unique in the genus on account of its excessively glaucous flowers and lower leaf-surfaces. The vernacular name given by Senior Tonduz is “ papayillo de venado.” Pileus mexicanus (A. DC.), comb. nov. Jacaratia mexicana A. DC. Prodr. xv. pt. 1, 420 (1864); Calg. Fl. Mex. t. 1163-1164 (1874). J. conica Kerber, Jahrb. Bot. Gart. Berlin ii. 282, t. 9 (1883). Carica heptaphylla Sesse & Moc. Pl. N. Hisp. 172 (1887); not Vellozo (1827). P. heptaphyllus Ramirez Anal. Inst. Med. Nac. Mexico v. 28, t. 1-4 (1901); Naturaleza ser. 2, iii. 707, t. 41-45 (1903).—In its alate or strongly costate fruit and in its quite anomalous habit this species differs strikingly from all other members of the Caricaceae, and appears to merit well the generic rank accorded it by Dr. Ramirez. There are, however, two names older than the particular one taken up by Ramfrez; the above combination being the proper one. The species has been almost universally referred to Jacaratia, but in habit it is anomalous in that genus and in addition differs in having unarmed branches and non-connate stamens. ‘The plant differs from Carica in its habit, divided leaves, costate or angled fruit, and oppositi- sepalous corolla-lobes. It differs from Jarrilla in its habit, 5-celled costate or alate non-cornute fruit, divided leaves, oppositisepalous corolla-lobes, and linear (not lingulate) glabrous stigmas. Ramfrez 80 JOHNSTON reports the plant from the state of Morelos and Guerrero. Kerber collected it in Colima and mentions a collection from Campeche. I have seen a specimen (Millspaugh 96 in part, mixed with Jatropha Gaumert Greenm.) from between Merida and Izamal, Yucatan. Caiophora (§ Bicallosae) Jorgensenii, sp. nov., volubilis; foliis oblongis vel late lanceolatis acutis pinnatisectis vel pinnatilobatis 2.5-5.5 em. longis 1-3 em. latis 3-6 mm. longe petiolatis, lobis acutis oppositis vel subalternis serratis; floribus axillaribus solitariis 5- meris; pedicellis florigeris tortuosis 10-15 mm. longis, fructiferis ca. 5 mm. longis apicem versus cernuis; hypanthiis hirsutis graciliter cylindratis ad basin sensim attenuatis ca. 8 mm. longis ca. 2 mm. crassis, lobis linearibus ca. 4 mm. longis ca. 0.5-0.7 mm. latis; petalis patentibus 8-10 mm. longis late navicularibus extus setosis; squamis pubescentibus 3-4 mm. longis ca. 2-2.5 mm. latis a latere visis semi- lunatis a dorso ovato-oblongis sub collo callos 2 valde incrassatos decurrentes transversos gerentibus apice valde bicalloso-incrassatis, nervis 3 ad 2 altitudinis squamae filum emittentibus, filis 3 filiform- ibus basin versus paullo dilatatis squamas paullo vel vix super- antibus; staminodiis 2 subulatis gibbere papillosis tortuosis 4-5 mm. longis quam squama sensim longioribus; filamentis numerosis; cap- sulis cylindro-claviformibus inferne in pedicellum attenuatis 2-2.5 em. longis 4-5 mm. crassis nutantibus setosis conspicue spiraliterque contortis; seminibus 0.6-1 mm. longis ca. 0.6 mm. diametro pallide brunneis faveolato-reticulatis irregulariter alatis—ARGENTINA: El Candado, Dept. of Andalgald, Prov. Catamarca, Jan. 16, 1916, Pedro Jorgensen 1163 (typr, Gray Herb.) —This species is evidently a close ally of C. stenocarpa Urb. & Gilg, and C. Arechavaletae Urb., and has a range geographically intermediate between these species. It differs from C. stenocarpa in its smaller solitary axillary flowers, elongate pedicels, and evidently contorted capsules. From C. Are- chavaletae it differs in its shorter petioles, smaller less deeply cut narrower leaves, solitary axillary flowers, and longer more slender capsules. In C. Jorgensenii the scales in the flower appear to be nearly intermediate between the conditions illustrated by Urban & Gilg, Monog. Loasace. t. 8, fig. 4 and 7 (1900), for its near relatives. e staminodia in the proposed species are subulate, tapering very gradually from the broadened base. About 1 mm. above the base, and slightly below the first bend, the staminodia are provided with a pair of spreading linear coarsely and densely villous appendages about 1 mm. long. Above the attachment of the appendages the pale staminodia is covered to the very tip with gibbous tuberculations. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 81 The staminodia in C. Jorgensenii is hence quite different from those of C. wpakin = and C’. Arechavaletae as illustrated by Urban & Gilg, l. c. fig. 6 and 8. Gilibertia amplifolia, sp. nov., arborea glaberrima; ramis cortice griseo spongioso tectis; foliis paucis ovatis vel oblongo-ovatis coriaceis undulatis vel sinuatis obtusis 15-18 cm. longis 9-11 cm. latis conspicue nervatis bicoloribus basi rotundatis vel abrupte cuneatis; petiolis 6-7 cm. longis; inflorescentia terminali pyramidaliter paniculata, rachibus crassis ca. 3.5 cm. longis, bracteis conspicuis latis persistenti- bus saepe ad 2-3 mm. longis; ramulis primariis patentibus 2.5-3.5 em. longis inarticulatis bracteatis; pedunculis 1-3 em. longis saepe medium vel basin versus bracteolatis; capitulis 10-20-floris; pedicellis 1.5-2.5 mm. longis; hypanthiis anguste obconicis; corollis 5-meris; alabastris subglobosis; petalis saepe ca. 2.5 mm. longis anguste tri- angularibus crassiusculis acutis quam stamina paullo brevioribus; ovario glabro; baccis ca. 5 mm. longis 4 mm. latis valde 5-costatis; stylis medium versus connatis apice valde recurvatis—CoLoMBIA: Santa Marta, alt. 1200 m., Feb. 1898-1901, H. H. Smith 1778 (TYPE, Gray Herb.).—Related to G. arborea (L.) March., but differing in having short pedicels, paler broader leaves, and a more compound inflorescence with a thicker main axis. Gilibertia Smithiana, sp. nov., arbor glaberrima; foliis oblongis vel elliptico-oblanceolatis 2 dm. eons 4-7 cm. latis margine 0 crenatis et involutis, basi cuneatis vel rotundatis, apice valde acu- minatis, subtus pallidioribus valde nervatis; petiolis saepe 3-8 cm. longis; racemis terminalibus, rachibus 2-6 cm. longis gracilibus erectis, pedunculis gracilibus elongatis non articulatis ima basi et medium versus bracteas breves ovatas concavas acutas gerentibus; calycis late obconici lobis acutis; corolla sphaerica paullo depressa, petalis del- toideo-ellipticis acutis uninerviis crassiusculis purpureis ca. 1.5 mm. longis stamina paullo superantibus; stylis sub anthesi in conum rugosum diametro disci circiter aequantem connatis maturitate valde recurvatis; ovariis glabris; baccis globosis laevibus 5—6 mm. longis.— GuaTEMALA: Cubilquitz, Dept. Alta Verapaz, alt. 350 m., Aug. 1907, von T' tirckheim 11407 (typeE, Gray Herb.); Aug. 1901, von Tiirckheim > : Pedro 5917; (?) Cerro del Boqueron, Chkacas: June 1914, Purpus 7363.— Related to the Antillean G. arborea (L.) March. under which name it bis key, Bot. Gaz. lv. 436 (1913), to the Central American species of 82 JOHNSTON the genus. ‘Gilibertia Smithiana, however, differs from its West Indian relative in having smaller flowers in a more slender less pro- jected inflorescence, and in possessing less firm usually more elongate leaves. The only specimen of G. arborea in the Gray Herbarium from Central America is Fendler 131 from Panama. Gilibertia eurycarpa, sp. nov., arborea; ramis cortice griseo longi- tudinaliter sulcato tectis; foliis glaberrimis saepe obtuse breviterque acuminatis integerrimis, subtus pallidioribus costa prominenti et nervis ascendentibus arcuatis ad basin inferioribus cum costa de- currentibus, foliis ramulorum juvenum 3-5-palmato-lobatis cum petiolis 2 dm. longis, eis ramulorum maturorum ovatis acutis vel obtusis 10-15 em. longis 5-11 em. latis basi cuneatis; umbellis 25—40- floris in racemum 2-5 em. longum axillarem digestis; pedunculo 4-5 cm. longo gracili medium versus bracteis brevibus ovatis concavis acutis instructo, apice in receptaculum hemisphaericum mediocre dilatato; floribus 5-meris, pedicellis filiformibus ca. 7 mm. longis bracteolis scariosis ferrugineis brevissimis basi stipatis; tubo calycis obeonico, limbo anguste triangulari 1.52 mm. longo; corolla hemi- sphaerica paullo depressa; petalis ovatis vel triangulari-ovatis acutis ca. 2 mm. longis; staminis filamento flexuoso petalis longiore; stylis valde arcuato-reflexis basin versus connatis; bacca oblate globosa ca. 5 mm. longa cum exocarpio corrugato; pedicellis 8-10 mm. longis.— Mexico: Tamasopo Canyon, San Luis Potosi, Pringle 3723 (TYPE, Gray Herb.); Gémez Farias, Tamaulipas, ca. 350 m. alt., Palmer 275, 312.—The species is related to G. stenocarpa Donn. Sm. of Guatemala and Salvador, but differs in its globose fruit, shorter style, and more uniform leaves. It is also a close relative of G. insularis Rose of the Tres Marias Islands, but differs from the description of that species in having larger looser infl with longer pedicels and peduncles, and in having larger fruit and stron gly recurved styles. _ Gilibertia alaris (C. & S.), comb. nov. Hedera alaris C. & §. Lin- naea ix. 605 (1835). oe : 7 _ Gilibertia darienensis (Seem.), comb. nov. Dendropanax dari- enense Seem. Jour. Bot. ii. 300 (1864). _ Gilia chachanensis, sp. nov., glabra glandulosa 15-20:cm. alta stricte vel ascendenter ramosa suffruticulosa (?) annua; ramis nu- merosis, inferne pallidis nudis, superne brunnescentibus vel viridibus gracilibus rigidusculis foliosis; foliis 1-2 em. longis integris vel pal- matifidis vel varie pinnatifidis, lobis paucis linearibus 0.5-1 mm. latis 5-18 mm. longis compressis vix dilatatis firmiusculis cuspidatis sparse glandulosis; floribus foliis oppositis solitariis; pedicellis ascendentibus TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 83 1-3 (saepius 2-3) em. longis: glandulosis; calyee campanulato viscido- glanduloso 4-5 mm. alto ca. 2 mm. crasso, lobis triangulari-lanceolatis ca. 1.5 mm. longis; corolla glaberrima ca. 9 mm. longa hypocrateri- formi, tubo calycem paullo superanti albo, lobis ovalibus ca. 2.5 mm. longis purpurascentibus; staminibus infra medium tubo corollae adfixis; filamentis tubum corollae superantibus filiformibus 3.5 mm. longis; antheris orbicularibus 0.3-0.4 mm. diametro basi adfixis; stylo 2 mm. longo; stigmatibus lanceolatis ca. 1 mm. longis; loculis ovarii multiovulatis; capsula cylindrato-ovoidea 4-5 mm. longa sepalis paullo breviori polysperma; seminibus brunneis oblongis sub aqua spirilliferis—Prru: on a sandy pampa at 3300 m. alt. on the south slope of Chachani Mountain near Arequipa, March 1920, Mr. & Mrs. F. E. Hinkley 25 (rypr, Gray Herb.).—The above species belongs to the § Eugilia and is clearly a member of that small group of South American species centering around G. laciniata R. & P. In Brand’s monograph, Pflanzenr. iv. Fam. 250, 95 (1907), it keys out with G.laciniata. It differs from that, however, in its very slender glabrous habit, long filaments, solitary long-pedicellate flowers, and linear or simply cleft very slender elongate glabrous leaves. The stems of G. chachanensis are very slender, rather rigid, and towards the ground slightly fruticulose. The plant is absolutely glabrous but is sprinkled over with short-stipitate glands, and the upper parts are somewhat viscid. It bears the local name of “romerito del cerro.’ Patima formicaria, sp. nov., fruticosa ca. 2.4 dm. alta; caulibus 1-5 erectis simplicibus fitaloas tetraguais obtuse angulatis 8-12 mm. crassis in lateribus longitudinaliter sulcatis; foliis herbaceis obovato- ellipticis vel late oblongeque obovatis 30-45 cm. longis 12-17 cm. latis integerrimis apice cum acumine 15-25 mm. longo falcate acu- minatis basi cuneatis vel attenuatis subtus pallidis sparse puberulentis, nerviis 30-40 arcuate laxeque ascendentibus; petiolis 4-7 cm. longis lateraliter compressis angulatis; stipulis 4—5 mm. longis lineari-attenu- atis vel lanceolatis crassis; inflorescentia glabra axillari; axillis cum 1-3 uni- vel quadri-floris 1 mm. longe pedunculatis umbellis ornatis; pedi- cellis 4-8 mm. longis; hypanthio glabro hemisphaerico ca. 2 mm. alto; ovario multiovulato multiloculari (?); calyce glabro integro 1.5-2.56mm. alto cupuliformi; corolla 2.5-2.8 mm. longa firma citrea vel (in saltem sicco) rubiginosa extus glabra, tubo cylindrico ca. 1.5 mm. longo ca. 4 mm. crasso intus glabro supra in fauces infundibuliformes 8-10 mm. longas usque ad 8-10 mm. crassas intus aureo-villosas ampliato, lobis deltoideis ascendentibus graciliter acuminatis in alabastro paullo lateraliter imbricatis ca. 3.5 mm. latis; staminibus glabris inclusis in 84 JOHNSTON faucibus ad corollam affixis; filamentis ligulatis paullo attenuatis 2-2.2 mm. longis; antheris lanceo-linearibus ca. 4 mm. longis basi emargin- atis apice acuminatis; stylo corolla breviori tereti glabro ca. 1.5 mm. longo, stigmatibus ligulatis ca. 2.7 mm. longis; bacca 6-7 mm. crassa 5-8 mm. alta; seminibus subglobosis brunneis ca. 0.5 mm. diametro foveolatis—BritisH Guiana: Tumatumari, Potaro River, lat. 5° 20’ N., June 1920, Hitchcock 17348 (tyPE, Gray Herb.); clayey hills at cisatk of Merumé Creek, 5° 53’ N., 59° 52’ W., Dec. 1922, H. 0. Lang 330 and 339.—Evidently related to P. guianensis Aubl., PI. Guian. 196, t. 77 (1775) and Hoffm. in Mart. FI. Brasil. vi. pt. 6, 309 (1889), but having taller tetragonal stems, larger long-acuminate leaves, larger flowers, and smaller apparently many-celled fruit. Ac- cording to Dr. W. M. Wheeler, to whom the Gray Herbarium is in- debted for excellent material of the species, the plant is myrmeco- philous, its fistulous stems being the habitation of at least five species of ants GNAPHALIUM MICROCEPHALUM Nutt. Trans. Am. Philos. Soc. ser. 2, vii. 404 (1841). G. bicolor Bioletti, Erythea i. 16 (1893).—A photo- graph of the type of Nuttall’s Gnaphalium microcephalum has been recently received from the British Museum. A study of this photo- graph shows conclusively that the species has been universally mis- interpreted and that it clearly represents the species well known to Californian botanists as G. bicolor Bioletti. The species of Southern California which has been mistaken for the Nuttallian species is hence without a name and may be called:— Gnaphalium albidum, sp. nov., perenne 5-9 dm. altum; caulibus numerosis dense tomentosissimis laxe ramosis, foliis oblanceolatis vel spathulatis late adfixis saepe subdecurrentibus 3-5 cm. longis 5-10 mm. latis acutis tomentosis supra paullo viridioribus; capitulis 5-6 mm. altis ca. 4 mm. diametro in glomerulos lanugineos terminales congestis; tegulis ca. 18 albidis 3-seriatis, exterioribus ovatis acumi- natis, interioribus oblongis acutis; floribus perfectis 3.5 mm. longis 5-6, imperfectis 35-40; pappi setis ca. 20 distinctis ca. 4 mm. longis barbatis; achaeniis glabris oblongis ca. 0.6 mm. longis.— ALIFORNIA: in the chaparral, Granite, San Diego Co., 550 m. alt., July 11, 1916, Mary F. Spencer 69 (rYPE, Gray Herb.); grassy Vile. Laguna Mts., San Diego Co., Spencer 1034; border of stream, San Bernardino Mowtatie. Peis Bros. 579; Pasadena, Grant 518; Mission Canyon, Santa Barbara Co., Eastwood 124.—This species is confined to the coastal drainage of Sauthern California where it is most common in gravelly places on the alluvial fans along the base of TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 85 the mountains, and in open places in the chaparral belt in the foothills or at low altitudes in the mountains. It is probably most related to G. Wrightii Gray, a species of northern Mexico and New Mexico, which has smaller heads, more acute tegules, more slender stems, and a much less dense tomentum on its stems and foliage. GNAPHALIUM BENEOLENS Davidson, Bull. So. Calif. Acad. Sci. xvii. 17 (1918).—Lower Catirornta: San Vicente, Orcutt 1241; Cate FORNIA: Saratoga, San Diego Co., Spencer 166; Palomar, S pencer 1043; Crescenta, Burlew 3275 (IsoTYPE); Converse Basin, Fresno Co., Dudley 3397 ; Yosemite, 1877, Hooker & Gray; Del Monte, Elmer 4088; Black Mt., Santa Clara Co., Baker 1542; Mt. Tamalpais, Eastwood 1601; near Forest Ranch, Butte Co., Heller 12659. Nevapa: Bowers, Heller 10659; near Carson City, Anderson 19.—This species seems properly to include almost all that material from middle and northern California, and most of that from Southern California, which in the past has gone as G. microcephalum. It differs from G. albidum in its narrow linear elongate leaves and larger usually stramineous heads. In Southern California where G. albidum and G. beneolens both occur, the latter appears to flower later, to be more common, and to reach its optimum development at lower altitudes growing in the broad warm valleys and low foothills. The two species have very different habits, at least in Southern California. Gnaphalium albidum branches more, particularly towards the base, and has somewhat decumbent stems, and consequently has a decidedly loose habit. On the other hand G. beneolens is stiffly erect, with little branching towards the base, and forms close trim clumps. The difference in foliage between the species, while occasionally obscure in very mature specimens, is remarkably accentuated on the vigorous young flowerless shoots. GNAPHALIUM THERMALE E. Nels. Bot. Gaz. xxx. 121 (1901).— Wyoming: on the formations, Norris Geyser Basin, Yellowstone Park, Nelson 6139 (IsotyPE). Ivano: Lake View, Kootenai Co., Sandberg, MacDougal & Heller 872; Trinity, Elmore Co., Machride 529. British CoLuMBIA: vicinity of Nanaimo, Vancouver Isl., Macoun 430. Wasuineton: New Port, Kreager 454; Peshastin, Sandberg & Leiberg 830; Friday Harbor, San Juan Islands, Zeller 1213. OrEGoNn: Kamela, Peck 4672; Detroit, Nelson 1912. CALiFoRNIA: near Deetz Station, Siskiyou Co., Heller 11702; Truckee, Heller 7092; Bear Valley, San Bernardino Mts., Abrams 2888.—Although grossly inappropriate, the above name seems the only one correctly applicable to the plant of northwestern United States now current as G. microcephalum. The plant in question is a well marked one and ranges from western 86 JOHNSTON Wyoming to Oregon and southern British Columbia and thence south- ward along the Sierra Nevadas to Southern California. It appears to be most related to G. Wrightii of Mexico and New Mexico, but dif- fers in its very different distribution, smaller heads, and sharply acute usually stramineous tegules. Gnaphalium thermale is readily distinguished from G. beneolens by its very conspicuously smaller heads and northern range. In California where the two latter species both occur, G. thermale grows on the mountains in the pine belt at altitudes much higher than G. beneolens. Gnaphalium texanum, sp. nov., perenne 3-4 dm. altum; caulibus compluribus laxe ramosis dense tomentosis foliis numerosis oblance- olatis vel oblanceo-linearibus 3-7 mm. latis 1.5—4 em. longis acutis late affixis, subtus tomentosis, supra sparse decidueque tomentosis; capitulis 4-5 mm. altis 2-2.5 mm. diametro in glomerulos lanu- ginosos densissimos congestis; tegulis 18-20 albidis 3-seriatis interior- ibus oblongis acutis, exterioribus ovatis acuminatis tomentosis; floribus perfectis 2.5 mm. longis 8-10, imperfectis ca. 40; pappi setis ca. 18 distinctis ca. 3 mm. longis subnudis; achaeniis glabris oblongis ca. 0.5 mm. longis.—TeExas: mouth of Tarlinga Creek, Brewster County, Sept. 1883, V. Havard 26 (rype, Gray Herb.); rocky soil, lower slopes, Davis Mts., Aug. 20, 1915, without collector.—Although passing as G. Wrightii Gray, this species seems unquestionably dis- tinct, differing in its stiff habit and branching, narrow leaves, and particularly in having its small heads crowded into tight capitate clusters terminating short leafy branches. The new species appears to replace G. Wrightii in Texas. Gnaphalium viridulum, sp. nov., bienne 2-4 dm. altum; caulibus compluribus erectis simplicibus vel rariter laxe ramosis tomentosis; foliis oblanceolatis superioribus gradatim reductis late affixis 2-6 cm. longis 4-10 mm. latis, apice acutis cuspidatis, subtus dense tomen- tosis, supra paullo viridioribus; inflorescentia dense corymbosa; capitulis 4-5 mm. altis 2-3 mm. diametro in glomerulos densos con- gestis; tegulis ca. 18 albidis basem versus viridulis 3-seriatis exterior- ibus ovatis acutis sparsissime laxeque tomentosis, interioribus oblongis acuminatis; floribus perfectis 5-6 ca. 3.5 mm. longis, imperfectis 30-35; pappi setis ca. 15 distinctis solitariter deciduis ca. 4 mm. _ longis antrorse hispidulis; achaeniis glabris oblongis 0.6 mm. longis.— New Mexico: Bear Mts. near Silver City, 2400 m. alt., Sept. 19, 1903, O. B. Metcalfe 742 (typ, Gray Herb.). Arizona: Gooseneck, Paradise. iricahua Mts., 1950 m. alt., Blumer 2207; Tucson, 1907 ? Loyd.—This species has been confused with G. Wrightii although dif. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 87 fering in its lower stature, smaller cylindrical greenish heads, and less sharply acute partly green tegules. The plant has a characteristic habit. It appears to range more to the westward than does G. Wrighti and perhaps replaces it in Arizona. GNAPHALIUM UsTULATUM Nutt. Trans. Am. Philos. Soe. ser. 2, vil. 404 (1841).—It has been the current practice to apply the name - purpureum L. to an indigenous plant which ranges along the Pacific Coast from British Columbia to Southern California and seemingly reappears in Arizona and northern Mexico. Comparison of this plant with the true G. purpureum of eastern United States shows the former to be uniformly coarser, to have a much looser tomentum, and a broader dense stout somewhat leafy spike of com- monly brown heads. The plants from the two coasts are readily recognized and may be quickly separated merely on their difference of gross habit and aspect. A study of the photograph of Nuttall’s type of G. ustulatum, recently received from the British Museum, shows that both of the specimens cited by Nuttall clearly belong to the Pacific Coast plant and hence that his name is properly applicable to the “G. purpureum” of Californian authors. Although Nuttall cites one specimen as collected “On the plains of the Platte, towards the Rocky Mountains’’ the species is not known there, and it seems almost certain that the locality is the result of mislabeling. In Southern California G. ustulatum is not known south of Santa Barbara ‘and. Santa Rosa Island. The specimens from San Diego and Clare- mont cited by Hall, Univ. Calif. Pub. Bot. iii. 111 (1907), represent the amphigean weed, G. spathulatum Lam., a plant readily recognized by its distinctly annual weedy habit, floccose greenish foliage, and yellowish or greenish heads embedded in tomentum. "3. A. NEGLECTED Paper BY JEAN Lovis BERLANDIER. It was recently discovered that the library of the Gray Herbarium contained copies of a sixteen page brochure, apparently written by J. L. Berlandier, which seems to have escaped the notice of recent authors. The title-page of the paper reads: Memorias | de Comision | de | limites | a | las ordenes del | General Manual de Mier y Teran. The first numbered page bears at its head: Memorias de la Comision | de limites. | Historia Natural | Botanica. | por | El General Teran y L. Berlandier. The preface is dated thus: Mata- moros de las Tamaulipas Junio.20 de 1832. The commission referred to is evidently that which, according to Bancroft, Hist. Mex. v. 154 (1885), was appointed “in 1827, with General Manual Mier y Teran 88 JOHNSTON as its chief, to ascertain the boundary between the Mexican and American republics under the treaty of 1819.” According to Alcocer, Naturaleza ser. 2, iii. 556-557 (1901), the commission operated as late as 1830 or 1831. Since the date to the preface of the botanical report is over a year after the disbanding of the commission, it seems probable that the time mentioned is within a year, if not a month, of the exact date of publication, since the paper is a small one and especially since certain crudities in printing suggest that it was printed (in all proba- bility very promptly) on some frontier press such as would have been found at that period in Matamores. The brochure is hence taken as having been published in 1832. In substance the paper consists of dual Latin and Spanish descrip- tions of eleven newly proposed species and four new genera. The descriptions are carefully prepared and are evidently the work of Berlandier. The identification of the proposed genera and species, which appear to have been completely neglected and which are not listed in the Kew Index, has been greatly facilitated by the specimens from the Berlandier herbarium now preserved in the Gray Herbarium. These in a number of cases are labeled in Berlandier’s handwriting with the names published in the brochure. Further help has been derived from the volume of Berlandier’s unpublished plates, now preserved in the library of the Gray Herbarium. In the present paper these plates are cited: Berl. Icon. Ined. It seems probable that the brochures and the Berlandier manuscripts all came to the Gray Herbarium through the gift of Dr. Short under the conditions men- tioned by Gray, Am. Jour. Sci. ser. 2, xxxv. 16 (1863). Rivina vernalig Teran & Berl. Mem. Comision Limites 1 (1832); Berl. Icon. Ined. no. 1, pt. 1, t. 3, fig. 3— Crescit in locis sterilibus siccis humidisve, prope Matuarn de las Tamaulipas.” = R. HUMILIS L. Sp. Pl. 121 (1753). _ Cactus bicolor Teran & Berl. Mem. Comision Limites 1 (1832); Berl. Icon. Ined. no. 1, pt. 2, t. 4; no. 6, t. 4.—“ Crescit in locis siccis sterilibusque prope Matamoros de las Tamaulipas.” to- cactus bicolor (Teran & Berl.), comb. nov. Petccadia setispinus Engelm. Bost. Jour. Nat. Hist. v. 246 (1845). —Britton & Rose, Cactaceae iii. 105 (1922), say that E. setispinus is “ very different from Cactus bicolor Berlandier.’”’ Berlandier’s description, however, agrees very closely with their description of Engelmann’s species, while pencilings on Berlandier’s ae which certainly represent the species as s illustrated by Britton & Rose, show vy that ertcneage also con- — P a y cation of the latter. TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 89 Cacrus CHLORANTHUS Teran & Berl. Mem. Comision Limites 3 (1832); Berl. Icon. Ined. no. 1, pt. 2, t. 3—“ Crescit in locis sterilibus siccisve prope Matamoros de Tamaulipas.” This species is a Mam- millaria apparently belonging to the genus Escobaria of Britton & Rose, Cactaceae iv. 53 (1923). CACTUS TAMAULIPENSIS Teran & Berl. Mem. Comision Limites 3 (1832).—“ Crescit in locis sterilibus inter ri pam fluvii dicti Rio Bravo del Norte et locum dictum arroyo colorado prope Matamoros de las Tamaulipas.” Unrecognized. Terania frutescens Berl. in Teran & Berl. Mem. Comision Limites 4 (1832); Berl. Icon. Ined. no. 6, t. 2.—“ Habitat in regionibus septen- trionalibus et orientalibus Reipublicae Mexicanae; prope San An- tonio de Bejar in Texas, circa Monterrey in Nuevo Leon, prope Matamoros et Victoria in Tamaulipas.” Berlandier’s genus, 7erania, is evidently a late synonym of Leucophyllum, but his species is some years older than L. texanum Benth. in DC. Prodr. x. 344 (1846). The correct name forthe plant is, therefore, Leucophyllum frutescens (Berl.), comb. nov. Gaza anacua Teran & Berl. Mem. Comision Limites 5 (1832); Berl. Icon. Ined. no. 4, t. 9—*“ Habitat in locis siccis humidive; in Texas prope la Bahia del Espiritu Santo; in Tamaulipas prope Matamoros ubi vulgo adpellatur Anacua.” The genus Gaza, de- scribed at the above citation, is a synonym of Ehretia. The species, . anacua, however, is an older name for the plant current as E. el- liptica DC. Prodr. ix. 503 (1845), which therefore should be called ‘Ehretia anacua (Teran & Berl.), comb. nov. hrysodendron tinctoria Teran & Berl. Mem. Comision Limites 7 (1832).—“ Habitat in locis umbrosis montium dictorum Sierra de Tamaulipas del norte, prope San Carlos et San Nicolas, et forsan in omnibus montibus del Nuevo Leon prope Monterrey. Vulgo Palo amarillo.” The generic name Chrysodendron was proposed at the above reference. It is a synonym of Mahonia. As the specific name is much older than the synonymous Berberis chochoco Schlecht Bot. Zeitg. xii. 652 (1854), the plant should properly be called Mahonia tinctoria (Teran & Berl.), comb. nov. Three varieties of C. tinctoria were published, namely oblongifolia, latifolia, and longifolia. Au- thentic material in the Gray Herbarium, however, shows them to be trivial leaf-forms. JATROPHA CATHARTICA Teran & Berl. Mem. Comision Limites 9 (1832).—“ Crescit in locis sterilibus fere totius Tamaulipae, praecipue 90 JOHNSTON ad margines fluvii dicti Rio de las Nuecas et prope Matamoros urbem.” = Jatropha Berlandieri Torr. Bot. Mex. Bound. 198 (1858). Mimosa pseudo-Echinus Teran & Berl. Mem. Comision Limites 11 (1832); Berl. Icon. Ined. no. 1, pt. 2, t. 7.—* Habitat in sylvis pro- vinciarum internarum Orientalium prope San Luis Potosi ad ripam fluminis dicti Rio Grande usque ad Texas. Habitus Echini mollis.” = Prosopis JULIFLORA (Sw.) DC. Prodr. ii. 447 (1825). It is evident that the specific name was meant to be “ pseudo-Schinus” and it is so spelled on herbarium specimens. In the brochure and in the un- published plates, however, it is spelled as given above. Standley, Contr. U. S. Nat. Herb. xxiii. 353 (1922), has commented on the spelling of the specific name.. Calia erythrosperma Teran & Berl. Mem. Comision Limites 13 (1832).—* Habitat in Texas, prope S. Antonio de Bejar urbem, in Tamaulipas inter Palmillas et Jaumave locis ruderatis.” Calia was described as a new genus. It is, however, much later than the homonymous and universally recognized genus of the Compositae. The species is clearly a synonym of SOPHORA SECUNDIFLORA (Ortega) Lag. in DC. Cat. Hort. Monspl. 148 (1813). Lantana rubra Berl. in Teran & Berl. Mem. Comision Limites 15 (1832); Berl. Icon. Ined. no. 1, pt. 1, t. 5, fig. 1—*“ Habitat in Ta- maulipas circa Matamoros urbem, in Tejas ad ripam maris, in loco dicto Bahia de Matagorda et prope San Antonio de Bejar urbem.” This is a synonym of either L. Camara L. or L. HORRIDA HBK. 4. ON THE VALIDITY oF MOLINa’s SCIENTIFIC NaMEs. In the most recent addition to his list of critical papers on the flora of the Argentine, Professor Lucien Hauman, Physis vii. 67-76 (1923), has reviewed and discussed the validity of the plant-names proposed ty Molina in thefirst (1782) edition of the “Saggio sulla storia naturale del Chili.” It is surprising that Professor Hauman, admitting the priority of many of Molina’s names, has refused to accept the neces- sary name-changes, considering the names published in the Saggio ~ to be “nomina nuda” because they are more frequently identifiable | by items of folk-lore or by philological or pharmaceutical details mentioned by Molina, than by thé formal latin diagnoses. It must be admitted that Molina’s botanical descriptions are very brief and are frequently inaccurate. The descriptions, however, are always accompanied by a discussion in the text of the Saggio relative to the habit and to the native names and uses of the plant treated. Such discussion, supplementing the poorly prepared technical matter, TAXONOMIC RECORDS CONCERNING AMERICAN SPERMATOPHYTES 91 commonly serves to identify the species positively. In the brevity of his descriptions Molina was not at all exceptional in his age, and his names are about as readily identified from their diagnoses as are the species described by Linneus in the “Systema Naturae.”’ If we are to accept the principle upon which Professor Hauman rejects Molina’s names, treating all species as nomina nuda which are uni- dentifiable from their original description alone, and refusing to admit any a posteriori evidence as to the identity of such species, then botanical nomenclature is destined to a highly undesirable and caty- clismic reorganization, since most of the early species, and even a goodly proportion of the recent ones, are more readily, if not ex- clusively, identifiable from some locality, use, or local-namementioned, or from a study of the original material, than they are from the whole of the formal portion of the description. Consistency, therefore, demands that we accept Molina’s names when they are satisfactorily identified, as most of them have been through the efforts of Philippi, Anal. Univ. Chile xxii. 699-741 (1863). It is also to be noted that most of Molina’s names were considered sufficiently well identified to be cited in the synonymy of Reiche’s Flora de Chile. When the identity of such names has been determined with accuracy sufficient to warrant citing them, unquestioned, in synonymy, it would seem that we are logically obliged, if our acceptance of priority be more than a pretense, to take them up if they fulfill in other respects the necessary nomenclatorial requirements. Although Molina’s names were disregarded in the past, their neglect was not from their inade- quate definition, but because the earlier workers on the flora of southern South America followed, more or less closely, the principle finally formulated as the “ Kew Rule” which calls for the acceptance, not of the oldest specific name, but of the first specific name under the accepted generic name. While it is regrettable that well known specific names should be displaced by the neglected ones of Molina, there seems no other course for those who find their guide in the modern International Rules of Nomenclature formulated at Vienna, which call for the acceptance of the oldest specific name irrespective of the genus under which it was described. With the exception of the few changes made at this time, most nomenclatoria! adjustments oc- casioned by the priority of Molina’s names, have been made by pre- vious writers. tzroya cupressoides (Molina), comb. nov. Pinus cupressoides Molina, Sagg. Chile 168 (1782). Libocedrus cupressoides Kuntze, Rev. Gen. iii. 375 (1893). F. — Hook. f. ex Lindl. Jour. Hort. Soe. vi. 264 (1851). 92 JOHNSTON ' Frankenia salina (Molina), comb. nov. Ocymum salinum Molina, Sagg. Chile 239 (1782). F. Berteroana Gay, FI). Chile i. 247 (1845). Gomortega keule (Molina), comb. nov. Lucuma keule Molina, Sagg. Chile 187 (1782). G. nitida R. & P. Syst. Fl. Peruv. i. 108 (1798). Larrea balsamica (Molina),comb. nov. Mimosa balsamica Molina, Sagg. Chile 165 (1782). L. nitida Cav. Anal. Hist. Nat. Madrid ii. 120 (1800). : Myrceugenia luma (Molina), comb. nov. Myrtus wma Molina, Sagg. Chile 173 (1782). Myrceugenia apiculata (DC.) Niedz. in E. & P. Nat. Pflanzenf. iii. Abt. 7, 74 (1893). Nierembergia minima (Molina), comb. nov. Nicotiana minima Molina, Sagg. Chile 153 (1782). Nierembergia repens R. & P. FI. Peruv. ii. 13, t. 123 (1799). Statice guaicuru (Molina), comb.nov. Plegorhiza guaicuru Molina, Sagg. Chile 164 (1782). S. chilensis Phil. Anal. Univ. Chile xx. 58 (1861). PROSOPIS CHILENSIS (Molina) Stuntz, U. S. Bur. Pl. Indust., Invent. xxxi. 85 (1914). Ceratonia chilensis Molina, Sagg. Chile 172 (1782). P. juliflora (Sw.) DC. Prodr. ii. 447 (1825).—As the above combination was made in an obscure place it is repeated here to call attention to the fact that it is the proper name for the widely distributed, variable, and much. named species known as “ mesquite”’ in Mexico and southwestern United States, and generally current as Prosopis juliflora, The combination is not given in the Kew Index, nor is ARISTOTELIA CHILENSIS (Molina) Stuntz, |. c. [= Cornus chilensis Molina], nor VILLARESIA CHILENSIS (Molina) Stuntz, l. c. xxxii. 39 (1914) [= Citrus chilensis Molina]. [Reprinted from Ruopora, Vol, 26, No. 304, April, 1924.] CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY New Series.—No. LXXI NOTES ON NORTH AMERICAN SCUTELLARIAS By c WILLIAM PENLAND Date of issue in RHopora, 2 June, 1924. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY. | New Serises.—No. LXXI. NOTES ON NORTH AMERICAN SCUTELLARIAS. C. WILLIAM PENLAND. (Plates 140, 141.) Dr. Asa Gray (Proc. Am. Acad. viii. 370), after remarking as to the placing of the genus Scutellaria in a subtribe, goes on to state that “The winged nutlets of Perilomia, however, are curiously imitated in one or two species of Scutellaria, only obscurely so in S. parvula, as has been noted by Dr. Torrey, . . . but strikingly in S. ner- vosa, pier and in a Japanese species not otherwise very similar The presence of this membranaceous wing, together a a footed difference in shape of the corolla, is used as a basis of separation of the above-mentioned genus Perilomia from Scutellaria by Humboldt, Bonpland & Kunth (Nov. Gen. & Sp. Amer. ii. 326). Prof. Fernald (Ruopora, xxiii. 85), after examining sheets of the Old World 8S. galericulata and of the North American form, which for many years had passed as S. galericulata, came to the conclusion that the Old World plant is really not found as such on this side of the Atlantic, but that it has a closely related representative here. This was demonstrated chiefly by nutlet-characters, the exact significance of which will be clarified in the sequel. These facts have combined to indicate that perhaps a critical inspection of the group in question, from the standpoint of the fruit, might serve to throw light on, if not to clear up, some well known taxonomic difficulties existing here. From an examination of over 62 Rhodora [APRIL two thousand sheets of herbarium material, it was found that the nutlet-characters indicated in a striking way the relationships of species and sections within the genus. At this point it will be convenient to establish the generic and tribal relations of Scutellaria. In 1832-36, Bentham (Labiatarum Genera et Species) put Scutellaria and Perilomia—a South American genus—into a single tribe, but later, in 1848 (De Candolle’s Prodro- mus), he added Brunella and Cleonia to these two and threw them into a sub-tribe Scutellarieae of the tribe Stachydeae. Briquet, in his treatment of the genus in Engler & Prantl’s Die Natiirlichen Pflanzen- familien, assigns Scutellaria and Salazaria Torrey to the tribe Scutel- larioideae, placing Perilomia in a separate tribe, Stachyoideae. It is noteworthy that in his grouping of the Perilomia entities he includes certain forms which other authors have put with Scutellaria, e. g., S. Mociniana Benth., and which apparently do not differ from that genus except by the alleged upright position of the seed. This special portion of the genus even has the scutellum on the calyx,—a character lacking in other sections. Undoubtedly the position of the seed is important in the classification, but its infallibility is questionable when it separates such apparently closely related forms. It is to be regretted that more material of these forms is not at hand. It is thought that the presence of the scutellum might serve as a more reliable generic character. Seutellaria then, is characterized by its bilabiate calyx, with lips entire and closed in fruit. From the upper lip is a projection known as the scutellum. Its co-genus Salazaria is set off on account of its calyx which becomes swollen in fruit. As somewhat detailed accounts of the genus, we have the earlier work of Arthur Hamilton, “Monographie du Genre Scutellaire,”’ Bulletin Seringe, 1832, and Bentham’s treatise in the work above mentioned. Both these earlier writers have seen fit to make a number of sections, separated according to the nature of the inflorescence. Bentham confessing on his part the inadequacy of these characters. Hamilton makes but three sections and includes in them fifty-two species, fourteen of which are assigned to North America. Bentham makes five sections into which sixty-three species are placed, fifteen of which are North American. Inasmuch as in the present investi- gation the fruit has been specially studied, and since the Old World — —— available were too iasiren tresithy fruiting for accurate L yy observation and too scarce for authorit judgment, it was thought 1924] Penland,—Notes on North American Scutellarias 63 best to confine attention to North American material, exclusive of Mexico. Interestingly enough, the sections earlier adopted in classification are, with a few exceptions, confirmed by the reproductive characters. Briquet, following the work of Gray, was evidently the first to expand on the more conservative, and therefore more reliable reproductive characters, to indicate taxonomically the natural relationships. The presence or absence of a membranaceous wing does not establish DIAGRAM. 1. Showing relation of species of Scutellaria. (Forms with winged nutlets are in the circle. generic differences, as indicated by Bentham, for at least three well defined American species of the genus have this wing. It is also well to remark here that the two sheets referred to Perilomia (P. ocymoides) which are found in the Gray Herbarium do not have the winged achenia attributed to the genus. Apparently, therefore, we have double evidence as justification for the discard of this character as a generic distinction. However, the membranaceous wing is important within the genus Scutellaria and might afford a basis for the first division into component parts in an artificial key. Se Roa It is believed that even this nutlet-character is artificial, since it apparently has been secondarily acquired, appearing as it does in several different groups. It seems beyond a doubt, that Scutellaria 64 Rhodora [APRIL lateriflora and S. epilobiifolia are nearest the ancestral branch of the phylogenetic tree (diagram 1). Both are very wide-spread in distri- bution. According to the diagram, as I have conceived the situation, they represent the first forks of the tree, and both have given rise to groups showing the character in question. Further, in a hasty review of nutlet-characters of the other genera of the Labiatae, it appears that the character referred to is absent. This is another justification for believing it is of more recent development. The first division was taken up by Gray, later adopted by Briquet and is now in use by authors of the larger manuals in this country. _ Briquet introduced names for these divisions or sections. The section Scutellariopsis should be extended to include at least two more North American species. Hitherto Scutellaria nervosa has been classed as our sole representative of this group. Yet it is difficult to understand why S. parvula, the achene of which has a conspicuous band, amazing in its constancy, should be left out. Here too we should place the S. angustifolia group, one species of which has nutlets with wings rivaling even those of S. nervosa. The present status of the work and the limited area covered do not justify the grouping of the species of the genus into sections. It is hoped that later a survey of the world representatives may be made, at which time such sectional characters may be properly adjudged. It is hoped also, that a study of the immediately related genera may be taken up, following the more critical reproductive characters. The arrangement of species given below follows a more or less natural system. As arranged, numbers 1, 2 and 3; 4 and 5; 6, 7, 8, 9, 10 and 11; 12, 13, and 14; and 15, 16, 17, 18, 19, 20 and 21, form fairly well defined groups. It will be seen that they do not agree with treatments given by other students of the group. The fact that militates against this kind of grouping, of course, is that it indicates an impossible linear evolution. True relationships are better shown in the tree (diagram 1). There remains to say a word or two upon the acceptance of nutlet- characters as a basis for the separation of sections. This can best be brought out by means of an example: Perhaps one of the best- marked groups of species, the interrelations of which are shown by nutlet-characters, is the series consisting of S. lateriflora, S. galeri- culata and S. epilobiifolia. All three have canary-yellow achenes which are not duplicated elsewhere in the genus. Of these three S. lericulata (I am using this Ol 1 World species for purposes of illustra- 1924] Penland,—Notes on North American Scutellarias 65 tion) comes between the others, both from a superficial standpoint and from the characters of the nutlets. S. lateriflora has small papillae or wart-like protuberances on the surface of the achene, while S. epilobiifolia has the surface of the achene merely scabridulous. he European 8. galericulata has nutlets closely resembling those of S. lateriflora, while the vegetative characters are very similar to those of S. epilobiifolia. The corolla of the latter exceeds by about 1 cm. that of S. galericulata, which attains a length of no more than 1.5 em. ~ From this comparative examination, based on reproductive characters, it would seem anything but following the natural system to break off S. lateriflora and put it into a separate section as both Bentham and Briquet have done. I desire to record my thanks to Prof. M. L. Fernald, who has supervised this work and given invaluable advice; to Dr. B. L. Robin- son, Dr. J. M. Greenman, and Mr. Bayard Long, for the loan of herbarium material of their respective institutions; to Miss Mary A. Day and other members of the Gray Herbarium staff for help rendered along various lines during the execution of the work. The key, based upon nutlet-characters, is artificial, since, as already explained, species which seem to have come through different lines of descent are thrown together. In the citation of specimens all which are not in the Gray Herbarium are indicated by initials in parentheses: N, Herbarium of New England Botanical Club; M, Missouri Botanical Garden; P, Academy of Natural Sciences of Phila- delphia. ANALYTIC KEY BASED ON NUTLET-CHARACTERS.! A. Nutlets banded or winged, or conspicuously compressed rso-ventrally B. B. Nutlets conspicuously winged C. : C. Nutlets not over 1 mm. in diameter; wing more or less D.’ Nutlets with muriculate papillae; wing relatively wide 7. S.nervosa. . Nutlets with bluntish papillae; wing relatively narrow 6.58. parvula. C. Nutlets mostly over 1 mm. in diameter; wing somewhat basal, giving the nutlet a flattened appearance F. B. Nutlets with relatively narrow, unexpanded and ~ o oe Re ee ea ee ew 'In the absence of tlet: pecies h to be omitted ee key. Though in the succeeding descriptions they are placed in what appea their natural sequence. 66 Rhodora [APRIL I, Nutlets dull black, with somewhat tuberculate, flattish WO a a 10. S. Brittonii. J. Nutlets merely rugose or finely granulate-papillate K. . Nutlets canary yellow L. L. Nutlets merely scabridulous, 7. e., with inconspicuous ads pac DS MR aS ra 2S the 1. S. epilobiifolia. L. Nutlets with minute, wart-like papillae........ 3. S. lateriflora. <. Nutlets black or brown, finely granulate-papillate....12. S. resinosa i M. - Nutlets yellowish or orange..................... 4. S. Bolanderi. M. Nutlets brown or black N. - Nutlets with conical, more or less sharp-pointed pillae O. O. Nutlets with very slender elongate papillae... .21. S. sazatilis. tlets strikingly rosulate in appear nee....16. S. integrifolia. P. Nutlets not conspicuously rosulate Q. Q. Nutlets with thin, laminate papillae...... 13. S. Drummondii. Q. Nutlets with thick papillae R. R. Nutlets with irregularly disposed muriculate DONG ce eg 5. S. tuberosa. R. Nutlets with papillae regularly disposed S. S. Nutlets about 2mm. in diameter......._.. 19. S. serrata. S. Nutlets 1-1.5 mm. in diameter........._ 15. S. pilosa. J. Nutlets with short, truncate papillae +> prillote Dingle oo ae oe ek 14. S. cardiophylla, tlets brow : Nutlets with processes flat on top, frequently with achenes yellow, merely scabridulous or slightly pebbled, about 1.5 in di 1832). 8. galericulata of American authors, not L.—Newfoundland to British Columbia and south- ward to Arizona. The following specimens are representative. NeEwrounpianp: M. L. Fernald & 1924] Penland,—Notes on North American Scutellarias 67 O. E. Lansing, no. 2829. Minnesota: M. A. Barber, no. 7. Sours Dakota: P. A. Rydberg, no. 948. NEBRASKA: P. A. Rydberg, no. 1490. Asstnipora: J. Macoun, no. 5848. Montana: F. L. Scribner, no. 216. Wyomine: A. Nelson & E. Nelson, no. 6558. Urtan: S. Watson, no. 835. Arizona: J. T. Rothrock, no. 245. CALIFORNIA: C. F. Sonne, no. 287 (P). Orrcon: W.C. Cusick, no. 1939a. BritisH Cotumsra: 8. Brown, no. 748. Forma rosEa (Rand & Redfield) Fernald, RHoporA xxiii. 86 (1921). a adam, forma rosea Rand & Redfield, Fl. Mt. Desert, 137 1894). Maine and adjacent New Brunswick. The following specimens are characteristic. Marine: M. L. Fernald & B. Long, nos. 14455 (P), 14452, 14454 and 260; M. L. Fernald, no. 2088; W. W. Eggleston & : 6, 00; J. A. Cushman, no. 2134 (N); J. R. Churchill, Fort Kent, July 19, 1908 (N); Kate Furbish, East Livermore, July & August 1896 (N), South Poland, 1893 (N), Great Diamond Island, 1888 (N), Great Chebeague Island, June 11, 1902. New Brunswick: M. L. Fernald & B. Long, no. 14453. 3. S. LaTERIFLORA L. Fig. 7. Erect, mostly branched, 2-6 dm. high, smooth: leaves thin, oblong-lanceolate, acuminate, serrate, and New Mexico. The following are representative. NEWFOUND- LAND: M. L. Fernald & K. M. Wiegand, no. 6102. Quesec: S. F. Blake, no. 5619; M. L. Fernald & H. B. — no. 12152. Prince Nova Scotta: H. St. John, no. 1443. MassacHuserts: : Robinson, no. 588. New Jersey: Gershoy,no 589. PENNSYLVANIA: A. A. Heller & E. G. Heller, no. 653. West Viratnta: J. M. Green- man, no. 480. Ontario: J. Macoun, no. 21859. Onto: J. M. Green- - man, no. 1399. ILitots: Chase (P), no. 514177. Mussissrpptr: S. M. Tracy, no. 8752. Missouri: E. J. Palmer, no. 2651. Ne- BRASKA: F, Clements, no. 2781. New Mexico: C. Wright, no. 1538. 68 Rhodora [APRIL OrEGon: J. C. Nelson, no. 1832. Wasuincton: W. N. Suksdorf, no. 1001. British Cotumsia: J. Macoun, no. 54677. Forma RHODANTHA Fernald, RHopora xxiii. 86 (1921). A form with pinkish flowers, found in Quebec. Forma ALBIFLORA (Farwell) Fernald, Ruopora xxiii. 86 (1921). S. lateriflora var. albiflora Farwell, Mich. Acad. Sci. Ann. Rep. xix. 249 (1917). A form with white flowers. sometimes tinged with blue, 15-18 mm. long.—S. californica Gray, Syn. Fl. ii. 381 (1878). Including S. viarwm Heller, Muhlenbergia, i, 32 (1904)—In Cairornta from Amador County northward to Tehama County. The following represent the variety. J. D. Hooker & A. Gray, Calaveras County, 1877; A. A. Heller, nos. 5786, 7020, 12379; E. Braunton, no. 1047 (M); T. Bridges, no. 304; J. Torrey, no. 3 C. F. Sonne, no. 286 (M); A. Eastwood, no. 1530; M. E. Jones, no. 13485 (M); J. P. Tracy, no. 2297. This variety includes a number of forms intermediate between S. Bolanderi and S. antirrhinoides. In fact it might well be regarded as a hybrid of these two. In the main it has more of the characters of the former than of the latter and hence is referred here rather than to S. antirrhinoides as at first done by Gray. Fruiting material is too poorly represented for good comparison. 5. S. TUBEROSA Benth. Fic. 12. Mostly simple and erect, fre- quently with many leafy stems from a common base, or even decum- bent, villous to nearly glabrous, 2.5-15 em. high, or, when trailing, up to 3.5 dm. long: leaves villous or nearly glabrous, petioled, broadly ovate, with few coarse crenations; base rounded, truncate or sometimes cuneate: flowers short-pedicelled, solitary in the upper axils: corolla dark-blue or purple, rarely curved, 1-2 em. long; the lips subequal: calyx villous: nutlets with projections somewhat muriculate, dark- brown when fully mature—Lab. Gen. et Sp. 441 (1832-1836).— Foothills or valleys throughout California and southern Oregon. The following specimens are representative. Cairornta: W. H. 1 One sheet is from Indian Valley, Plumas County, J. G. Lemmon, autumn, 1886: the rest from Amador Co. southward. 1924] Penland,—Notes on North American Scutellarias 69 Brewer, no. 357; C. C. Parry, no. 338; W. W. Jones, no. 285; J. G. Lemmon, no. 338: A. A. Heller, no. 7288: C. R. Orcutt, no. 1343 (M). Orecon: H. S. Prescott, Grants est: April 5, 1912; 7. Howell, no. 1252 (MD; E. W. Hammond: no. {). S. PARVULA Michx. Fic. 2. aieket simple or branching, 1-3.5 showing close relation with S. nervosa.—F1. ii. coe 803). S. campestris Britton, Mem. Torr. Club, v. 283 (1894). S. parvula, var. mollis Gray, Syn. Fl. ii. 380 (1878), in part.—On richer soils of the Great Lake region and the Mississippi valley, Quebec and Ontario south- ward to Georgia and Texas. The following are characteristic. QueE- BEC: W. F. d Mi acrae, “in river St. Lawrence opposite Montreal” (M); A. 8S. Pease, no. 12940; Mrs. Shepard. Ontario: J. Macoun, Pelee Point, Lake Erie, August 6, 1901. Micuiean: ex. Herb. E. F. Smith, Tonia, 1877. Vermont: L. R. Jones & W. W. cage Burlington, July 12, 1894. New York: A. Wood, Dexter. Oxnto: Marblehead, May 25, 1895. Inunors: H. M, Smith, no. "5924; A. . Pease, no. 12555; J. M. Greenman, no. 2614 (M). TENNESSEE: E. J. Palmer, no. reine ); S. M. Bain, no. 38. Virernta: J. Q. A. Fritchey, Richmond, calor 28,1888. Grorata: T..J.Wray, Augusta os S. Boykin (P). ourt: O. E. Lansing, no. 3033; J. Davis, no. 4 ( : : ‘ >. ae nee (M); J. Reverchon, no. 3245 (M). Loutstana: E. J. Palmer, no. 7217 (M). emcee oF E. J. Palmer, no. pe Re H.W. Houghton, no. 358214. ARKA B. F. Bush, n Var. AMBIGUA (Nutt) Fernald. Similar e- pn but slightly more rigid in aspect; stem and under surface of leaves usually pur- plish, glabrous or with very slight appressed puberulence; the upper surface of the leaves sometimes with spreading aa the leaves nar- rower and rarely toothed—Ruopora iii. 201 (190 a Nutt. Gen. ii. 37 (1818). S. parvula, authors day drier habitats, oe to Minnesota and southward to Texas and westward ristic (P). w Jersey: P. Dowell, no. 6027. pve: W.M. ccabe, eae July, 1878 (P). Inptana: A. H. Young, Lafayette, June, 1879 (P). Tennessee: A. Ruth, Knoxville, June, 1897 (P). Inunors: S. B. Mead, Augusta, 1845 (P); H. N. Patterson, Oquawka, August, 1873; H. A. Gleason, Rantoul, July 5, 1907. Missourt: B. F. Bush no. 336; 0. E. — no. 2977. Loutstana: J. Hale, Alexandria 70 Rhodora [APRIL (P). Texas: E. Hall, no. 453 (M). Oxtanoma: G. W. Stevens, no. 2235. Iowa: R. Coombs & C. R. Ball, no. 569 (M). Kansas: J. B. Norton, no. 411 (M). Nesrasxa: H. J. W eber, Lincoln, June, Minnesota: FE. P. Sheldon, Princeton, July, 1892. - 9. NERVOSA Pursh. Fie. 1. Very slender, mostly simple, but frequently branched, 1.5 dm. high: internodes separating widely the opposite pairs or leaves: leaves ovate or oblanceolate, toothed, yellowish or buff, with muriculate papillae.—Fl. ii. 412 (1814). S. gracilis Nutt. Gen. ii. 37 (1818).—Pennsylvania to Virginia, west- ward and southward to Missouri and Alabama. The following speci- mens are representative. PENNSYLVANIA: E. B. Bartram, Home- wood near Pillsbury, June 15, 1907 (P); S. Brown, Quakertown, June 3, 1894 (P); A. A. Heller, Lancaster, May 28, 1889. New Jersey: ra% LAND: E. B. Bartram, Conowingo, May 30, 1907 (P). Virernta: C. 8S. Williamson, Great Falls of Potomac, May 28, 1909 (P). West Virginia: W. M. Pollock, May 30, 1896. Kentucky: 8. F. Price, orma ternata, n. f., foliis ternatis—One sheet collected by C. W. Short at Lexington, Kentucky, 1835 (P). h. Fie. 3 somewhat fibrous base, purplish below: leaves linear to oblong- ovate, obtuse, practically sessile, commonly conduplicate, entire, 1-3 em. long, except the lower; these when present, petioled, few-toothed, cordate, small, commonly purplish below: flowers in the upper axils, on pedicels 5-8 mm. long: corolla curved at the slender base, commonly 2-2.5 cm. long but reaching 3 cm., not broadening until 5-10 mm. above the calyx: lips subequal or the lower slightly longer, various as to color, deep blue to purple: nutlets large, with slender papillae and faintly banded at the base.—Fl. ii. 412 (1814). Including S. ustineae Eastw., Bull. Torr. Bot. Club, xxx. 493 (1903) and S. linearifolia Eastw., 1. e. 493 (1903).—Stream-bars or on rocky hill- sides from British Columbia and Idaho to southern California. The following are representative. IpanHo: W. Trelease, no. 4883 1924] Penland,—Notes on North American Scutellarias 71 (M); A. A. & E. G. Heller, no. 3150 (P); J. H. Sandberg, no. 8689 (M); J. B. Leiberg, no. 1548; J. F. Macbride, no. 104. Bririsn Cotum- bias Ji M. Macoun, no. 67887. WasHINGTon: F. O. Kreager, no. 10; C.V. Piper, nos. 1570, 1571; A. D. E. Elmer, no. 900; R. M. Horner, no. R178B411. Orecon: EL. W. Hammond, no. 330 (M); W. C. Cusick, nos. 75, 2145; M. Spalding, Ae 28; E. P. Sheldon, no. 8007. CaLt- FORNIA: Culbertson, no. 4446; C. C. Parry & J. G. Lemmon, no. 337; A. Gray, Chico (field), February to May, 1885; C. A. Purpus, no. 5605 (M); S. B. Parish, no. 3122 (M); L. Abrams, no. 2778; H. M. Hall & H. P. Chandler, no. 949; 4 A. Heller, no. 7889; A. East- wood, no. 1015. Var. canescens Gra ay. In manyrespects similar to the above, mony simple, virgulate, canescent; leaves linear-oblanceolate, 2-4.5 ¢ _ long, nearly tomentose, firm, not conduplicate, ascending le. corolla slender and gracefully curved.—Bot. Calif. i. 603 (1876). S. siphocampyloides Vatke, Bot. Zeit. xxx. 717 (1872).—In CaLirorNiIa rom Sierra County southward on foot-hills and mountains. following are representative. H. N. Bolander, nos. 4946 (M), 3947; W. R. seed no. 4131; A. D. E. Elmer, no. 4434 (M); W. H. Brewer, no. 1285 . S. ANTIRRHINOIDES Benth. Fic. 4. Mostly — somewhat spindling, simple or branched, from a chiefly fibrous base, 1-3.5 din. high, minutely pubescent: leaves ovate-oblong, meat mostly short-petioled, never over 2.5 cm. long; all but the very lowest entire: flowers in the upper axils, on pedicels 5-10 mm. long, com- monly 1-1.5 cm., never over 2 cm. long; the corolla-tube flaring from immediately above the calyx; the lower lip usually longer, deep blue to violet in color: nutlets with slender emg i wg winged, suggesting those of S. parvula a eter) vill. 1493 (1822). S. sanhedrensis Heller, Muhlenberg, ae (1904). S. nevadensis Eastwood, Bull. Torr. Bot. 49> (1903). —Rocky banks and slopes, Idaho and Utah to California and Oregon. The following are characteristic. Ipano: J. F. Mac- bride, nos. 937 (M), 482 (M); A. Nelson & J. F. Macbride, no. we) Uran: L. H. Pammel & R. E. Blackwood, no. 3778. Nevapa: A. ag no. 11120; M. E. Jones, no. 4036; A. Nelson & J. F. Machride, 1927. CALIFORNIA: A. ce He ller &é P. B. Kenned, y, nos. 8798, 8843 (M); J. P. Tracy, no. L. E. Smith, no. 316: A. A. Heller, nos. 5894, 12111. OrEcon: *D ial no. 398; Le. Robes: no. 2668; - Howell, no. 1253 10. Britroni Porter. Erect, simple or branched, minutely —. or puberulent, usually 1-2 dm. high: underground stems frequently with ee tubers: is obovate, acuminate at each end or obtu tip, somewhat viscid, sometimes purplish beneath, 1.5-3 cm. ee hardly ae) margins subrevolute; nerves prominent dorsally: corolla slender a , gradually dilated to ampliate throat and lips, blue, 2-3 cm. long: nutlets dull black, mostly angled with tuberculate processes.— . Torr. Bot. Club, xxi. 72 Rhodora [APRIL H. N. Patterson, no. 296; C. C no. B 413 (P); E. B. Payson, Eldora, July 7, 1919. Wyomine: A. Nelson, nos. 94 and 7009. 11. S. Nana Gray. Fic. 5. Small, usually 4—6 em. high, cinereous- puberulent, several branches from the base: rootstocks yellow, with moniliform tubers or simply tuberiferous: leaves entire, ovate to obovate, with long attenuate base, or sometimes spatulate, 1-2 em. long; nerves usually conspicuous on the lower surface of the leaf: corolla white, the lips equal: nutlets dull yellow with conical protuber- ances, sometimes angled or compressed.—Proc. Am. Acad. xi. 100 (1876). S. Footeana A. I. Mulford, Bot. Gaz. xix. 118 (1894).— Dry sandy hillsides in northwestern Nevada, northern California, southern Oregon, and southwestern Nevada. The following are representative. CALIFoRNIA: L. E. Smith, no. 390; A. A. Heller, no. 8086; H. E. Brown, no. 613; E. Palmer, no. 2602 (P). Nevapa: J. G. Lemmon, no. 538; M. E. Jones, Wadsworth, June 16, 1897 (M); P. B. Kennedy, no. 1028 (M). Orrcon: M. E. Peck, no. 6747; J. B. Leiberg, no. 472; W. C. Cusick, no. 1980. Ipano: A. I. Mulford, Black Cafion, June 18, 1802. 12. S. REstNosa Torr. Fic. 8. Stems erect, few to many, from a woody base, 1-3 dm. high, puberulent: leaves ovate, entire, or the lower subcrenulate, attenuate at the base, slightly petioled or sessile, minutely pubescent, resiniferous, 1-2 cm. long; nerves prominent on the upper surface of the leaves: corolla bluish to violet, 12-20 mm. long; upper lip forming an arch with the tube; lower lip occasionally lightly blotched, equal to or exceeding the upper: nutlets about 1 prairies. The following sheets are representative. Kansas: J. M. tholomew, June 6, 1889 (M). OKLAHOMA: Rk. L. Clifton, no. 3025; G. W. Stevens, no. 1291: P. J. White, no. 28 (M); W. H. Emig, no. : A. 67 (M). Arizona: L. N. Goodding, nos. 848, 224 (M); J. G. Lemmon, no. 2860; D. Griffiths, no. 4809; C. G. Pringle, no. 15956. 5 m __ folia, var. brevifolia Gray in E. Hall, Pl. Tex. _ only. S. brevifolia Gray, Syn. FI. ii. 380 (1878).—Northeastern Texas, on dry calcareous hills or rocky bluffs. The following specimens 1924] Penland,—Notes on North American Scutellarias 73 represent the range and habit of the variety. FE. Hall, no. 458 (P); J. Reverchon, nos. 440, 2059, 771 (M) and 2126 (M); G. W. Letterman, Texarkana, October 15-24, 1894, and Dallas, August, 1882. This form was first given recognition by Dr. Gray as a variety of S. integrifolia, but was later raised to specific rank. It is, however, merely a variety of S. resinosa. An inspection of the range of S. wntegrifolia shows that it has no representative in Texas except var. hispida. Finally, the nutlet of S. brevifolia is an exact duplicate of that of S. resinosa. Although with leaves somewhat longer than those of the ordinary S. resinosa (about 1 cm.) a quirk in the nomen- clatorial rules gives to this plant the somewhat paradoxical name of S. resinosa, var. brevifolia. Originally connected with S. integrifolia, the application of the name was apparent. : 13. S. DrummMonpi Benth. Fic. 11. Stems many, erect, branching at the base, villous-pubescent, mostly 1-2.5 dm oF leaves slightly petioled or subsessile, oblong-ovate or oval, entire, 1-2 em. long, or commonly less, undulate-crenate: calyx lists. teontiontty tinged with purple: corolla short, 8-12 mm. long, bluish purple or violet; the lower lip exceeding the upper, flaring, notched, violet-spotted: nutlets yellowish to dark brown, or blackish with age, with conical or frequently laminate processes.—Lab. Gen. et Sp. 441 (1832-1836). —S. Helleri Small, Fl. So. U. S. 1024 (1903) ys | on rich plains or in open woods throughout Texas, in adjacent Oklahoma and New Mexico, and running into Mexico e following ee are Sate sentative. New Mexico: C. Wright, nos. . and 1540. LAHO . W. Stevens, nos. 695 and 1078. Texas: S. M. Tracy, no. 7994; c. x hurber, San Antonio, April, 1853; F. gy eae nos. 1095 5 and se. Hall, no. 456; B. F. Bush, no. 627; E. J. Palmer, nos. (MD and 13482 (M); G. Jermy, no. 101 (M); W.M. Canta i nd: 201 CM); J. Reverchon, no. H. A. Pilsbry, New Braunfels, April raed 1903 (P); a A. Heller, no. 1503 (P); H. C. Hansen, no. 542 Somewhat variable in the size of respective parts, this species is yet constant enough not to warrant any breaking up into subdivisions. The shape of the corolla is especially constant. The nutlets serve to 4s a connection with the S. nervosa—S. parvula group. _ S. CARDIOPHYLLA Engelm. & Gray. Stems slender, — set io 3-6 dm. tall, puberulent, often purplish-tint the leaves petioled, deltoid or subcordate, mostly tisk: crenate, 1.5-4 cm. long: 1 It slender, 12-17 mm. long, blue: nutlets —_ tuberculate with truncate cesses, never oe granular as in S. restnosa.—PI. Lindh. 3 19 (1845) — Arkansas and Texas, on andes hillsides or sandy woods. The following sheets 74 Rhodora [APRIL — are representative. ARKANSAS: F. L. Harvey, no. 114; E. J. Palmer, no. 10503 (M); G. W. Letterman, Hot Springs, August 4, 1879 (M), and July or August, 188-(M). Texas: J. Reverchon, no. 3910; R. A. Dixon, no. 339; F. Lindheimer, no. 144; E. Hall, no. 454; FE. J. Palmer, no. 7887 (M); H. Eggert, Palestine, June 10, 1899. A distinct species with achenes that show closer relation to S. pilosa than to any other Scutellaria. The leaves also suggest a relation. It is interesting to note that toward the west S. pilosa runs out in Arkansas, though no intermediate forms between these two are found. 15. S. prrosa Michx. Stem erect, mostly slender and simple, the inflorescence sometimes branched, finely pubescent or hirsute, 2-7 dm. high; internodes frequently 10 cm. long: leaves petioled, crenate; the lower ovate or oblong-ovate, obtuse, abrupt or even truncate at the base, 2-8 em. long; the upper subsessile and somewhat oblong- spathulate: corolla slender, bluish, 1-1.5 em. long; lips subequal; the upper arched: calyx villous: nutlets with protuberances as in S. serrata (fig. 13) but smaller (1-1.5 mm. in diam.).—Fl. Bor.- Am. ii. 11 (1803). ?S. caroliniana Walt. Fl. Car. 163 (1788). Includ- ing var. 8 Benth. (S. ovalifolia Pers. Syn. ii. 136 (1807) ) and S. altamaha Small, Fl. So. U. S., 1022 (1903).—River-banks or wood- lands, New York to Michigan and southward to Georgia and Texas. June 28, 1878 (M). Prennsyivanta: T. C. Porter, Easton, July 10, 1868 (P); B. Long, Lancaster, June 22, 1909 (P). New Jersey: 18 ry, : - 8. Bt ; Virernia: A. H. Curtiss, Bedford County, June 15, 1871; H. D. House, no. 1046. West Vireinia: J. M. Greenman, no. 478 (M). Norra Caroutna: J. R. Churchill, Hot Springs, June 1, 1899; A. A. Heller, ickory, June 23, 1893 (P). Sourn Carona: W. Stone, no. 546 (P); Rev. J. Backman, Charleston. Grorcia: A. H. Curtiss, no. 6826 ; R. M. Harper, no. 1358. ALABAMA: C. Mohr, Mobile, May, 1884 (M); F. S. Earle & C. F. Baker, Auburn, June 5, 1897 (M).. Misstssrppr: J. Skehan, no. 77; 8. M. Tracy, no. 4451. Lovtst- ANA: J. Hale, Alexandria (P); C. R. Ball, no. 656 (M). Texas: C. Wright. Oxtanoma: G. W. Stevens, no. 2732. ARKANsas: N. M. Glatfelter, Eureka Springs, July 17, 1898 (M). Mussourt: E. J. Palmer, no. 5802 (M); B. F. Bush, no. 667 (M). TENNESSEE: H. Eggert, Sherwood, June 6, 1897 (M); 7. H. Kearney, no. 871 (M). Kentucky: C. W. Short, Lexington (P). Inptana: C. C. Deam, no. 20417. Micutcan: (M) no. 109374. Inuinois: E. J. Palmer, no. 15399. Var. nirsuTA (Short) Gray. Like the preceding but taller and with leaves having coarser crenations, and longer pubescence.—Syn. FI. il. 379 (1878). S. hirsuta Short, Cat. Pl. Ken. 8 (1835).—Not a very marked variety; found only in northern Kentucky. We have examined the following sheets, all collected by C. W. Short: Louis. or 1924] Penland,—Notes on North American Scutellarias 7 ville, September, 1835 (P), 2 sheets; 1842, 2 sheets; 1842, no. 10939 M); 1846, no. 109395 (M); 1848, nos. 109394 (M), 109396 (MD), and 109397 (M). 16. S. INTEGRIFOLIA L. Fic. 10. Erect, ae or branched at & sippi. The following are representative. PENNSYLVANIA: E. — no. 1078; F. W. Patel no. 69 (P); H. W. Pretz, no. 7559 (P). New JERSEY: B. Long & 8. Brown, no. 17 (P); W. Stone, Med- ford, July 4, 1910 (P). DELAWARE: E. Tatnall, Wilmington, 1886, Porter, June 6, 1874. Mary.ann: W.R. Mazon, no. 5918. Virernta: . A. Heller, no. 930; B. L. Robinaon, Buckroe, May 21,1912. Nort CaRrorina: T. G. Harbison, Waynesville, July 1, 1897; Biltmore Herbarium, no. Ae SoutH Carouina: J. Perkins, Summerville, April 29-May 1 , 1918; J. Davis, no. 8381 (M). Geroretra: H. Eggert, Belair, Mey 22, 1899 (M). Frorma: 8S. M. Tracy, no. 9162; AA: : : Eaton, : “seater May, 1891 (M). age ask C. W. Short, Flats of Red River (P). TEenNeEssEE: A. Ruth, 120; ei a Tullahoma, rns 9, 1897 (M). hegre ne i "Ske han Var. MAJOR Cha More rigid in habit, a heavily pubes- cent, 2-8 dm. hiss wiih eign — of dentate, ieee basal leaves, these larger than the 1. So. U. S. 323 (1860). Inel. S. arenicola Small, Bull. Tore Bot. Club, xxv. 143 (1898).—Low sandy locations, Georgia and Florida to Mississippi. The leet) rk mens are characteristic. Frortma: A. H. Curtiss, nos. 2060 a G.V. Nash, no. 1316; A. A. Eaton, no. 1153; - J: Weber, no. 7 an) Misstssrppr: S. M. “Tracy, n o. 4453 (M); J. Skehan, Ocean Spri May 8, 1895 (M) and no. 62. ; @lonells Mrs. So ay (?), Sevuinel ( Var. nisprpA Benth. A slender form, reddish in color when dri stem and leaves distinctly pilose: the leaves thin: the crenate ones, when present, not exceeding the upper cauline leaves.—Lab. 435 (1832-1836). —Florida to Arkansas and Texas on moist sandy ground. The following are characteristic. FLorma: 5 no. 6645 (M); 8S. M. Tracy, no. bp (M), (M) no. 109080. GrorcIa: H. Eggert, DeKalb Co., July 24, 1897; (M) no. 788487. Lovistana: J. Hale, Alexandria (P); E ie vu Palmer, no. 7604 (M); C. R. Ball, no. 517; J. F. Joor, Madisonville, eet 4, 1888 (M); 7. spcasaorne. _ ve TEXAS: ff Reverchon, no. 2127 (M); E. Hall, no. ARK : E. J. Palmer, no. 10522 (M), H. Eggert, Jefferson a Saas 8, 1808 (M), 76 Rhodora [APRIL Var. MULTIGLANDULOSA Kearney. A short-stemmed leafy form, never exceeding 2.5 dm.; ; puberulent or pilose on the leaf-nerves and margins and on the stem: leaves oblong-ovate or spatulate, slightly, if at all petioled, entire, revolute-margined; occasionally 1-3 pairs of small, dentate petioled basal leaves, but these always shorter than the upper: flowers in the upper axils, the inflorescence appearing scarcely racemose: corolla as in the species.—Bull. Torr. Bot. Club, xxi. 482 (1894).—Pine barrens or dry open ground, Georgia and Florida to Louisiana. The following are characteristic. Fiormpa: A. S. Hitch- cock, Suwanee Co., June-July, 1898 MD), and no. 477 (M);. €.48. Williamson, Palatla, April, 1897 (P); Dr. Leavenworth, Fort King as GrorcIa: R. M. Harper, no. 822. ALABAMA: Gates & Jewett, Mobile. Lovuistana: FE. J. Palmer, no. 7959 (M). Var. glabriuscula (Fernald), n. comb. Much like the variety hispida: slender, 2-5 dm. high, simple or branched, very slightly pubescent or glabrous: inflorescence racemose: leaves on distinct petioles, practically glabrous; the lower sometimes dentate: corolla subglabrous.—S. glabriuscula Fernald, Bot. Gaz. xxxiii. 156 (1902). “S. hyssopifolia L.” on many herbarium sheets, nomen dubium.— ostly on sandy pine lands, Georgia and Florida to Mississippi. The following serve to represent the variety. Grorata: R. M. Harper, no. 885. Fiorma: A. H. Curtiss, nos. 6097, 3, and 13425 (M); G. V. Nash, no. 2277, ALABAMA: H. Eggert, Cullman, June 18, 1897 (M); F.S. Earle & C.F. Baker, ig a ae 1897 (M). Murssts- stppi: J. Skehan, no. 22603; 8. M. Trac 0 Var. floridana (Cha pm.), n. comb. _; ver Be slender Snparioaved Chae fae aan aaa pe of the une very broad.—S. floridana 1, So. Ul, 4 (1860).—Probably confined to the pine panes swamps near sfiaece vets in pice coastal Florida. The following are representative. (M) nos. 788588 oe Cee 108997 (D. U. Dean), 108999, 109000 (Herb. Chapm.); A. W. Cha man, no. 13207 (M), and Apalachicola (ex. Herb. J. Carey). 17. S. Busun Britton. Fie. 15. Erect, mostly caespitose in habit; stems simple, whitish-pubescent or puberulent, 1.5-3.5 dm. high: leaves entire, oblanceolate, obtuse, sessile: nutlets with more wart- like, less rosulate papillae: otherwise as in S. ae ae mul- tiglandulosa to which it is nearest related—Man. 01).— Rocky barrens or hillsides in Carter and Shannon area See The following are representative: B. F. Bush, nos. 49 (M), 189, 378, 7817 (M), 461 (M), 48 (M), 4737 (M); E. J. Palmer, no. 19496. 18. S. canescens Nutt. Fic. 14. Erect, tall, much branched at the top, canescent throughout except the upper surface of the leaves, 3-12 dm. high: leaves ovate-lanceolate, acute to cordate at the base, babies usually glabrous above, 5-12 cm. long, on petioles 1.5—-4 cm. - cate pene brownish, close to those of S. serrata but smaller and with blunter papillae.—Gen. ii. 38 (1818). S. incana Muhl. Cat. 1924] Penland,—Notes on North American Scutellarias 77 56 (1813), nomen subnudum.—Dry woods, river-banks, etc., Pennsyl- vania to Wisconsin and south to Arkansas and Georgia. The fol- lowing represent the plant and its range. PENNSYLVANIA: 0. E. Jennings, Glenshaw, August 3, 1918: 7. C. Porter, Huntingdon Inpiana: C. C. Deam, no. 5183. Iino: H. A. Gleason, no. 2617. Wisconsin: 7. J. Hale, Lake Pepin, 1861. Kansas: A. S. Hitchcock, no. 797. Muissourt: B. F. Bush, no. 6113; E. E. Sherff, no. 635. Arkansas: F. L. Harvey, no. 109. TENNESSEE: S. M. Bain, no. 323. Kenrucxy: C. W. Short, 1840 (M). West Vireinia: W. M. Pollock, Upshur Co., July 8, 1895 (M). Norru Carouina: 7. G. Harbison, Waynesville, May 30, 1897. Grorata: Rk. M. Harper, no. 1368. Var. puNcTATA Chapm. Like above but with foliage glabrate and densely punctate.—FI. So. U. S. 323 (1860).—North Carolina, Georgia and Florida. The following are characteristic. NortH CAROLINA: J. D. Smith, August 7, 1882, C. S. Williamson, Balsam, July, 1897. Grorata: J. K. Small, Rabun Co., August 11, 1893; C. S. Williamson, Atlanta, August, 1896. Fiortpa: G. V. Nash, Bellair, September 3, 1895, (M) nos. 108778 and 108875. 19. S. serraTA Andr. Fic. 13. Stem erect, 2.5-6 dm. high, nearly always simple, with from 3-5 pairs of leaves: leaves thin, mostly glabrous, ovate, acuminate at both ends or rounded at the base, serrate or crenate, 2.5-10 cm. long, on slender petioles 2.5 em. or less ENNSYL . W. Sm I - Co., June 23,— (P); J. Pennell, no. 2713 (P); E. B. Bartram, Darby Creek, July 21, 1907 (P); U. C. Smith, no. 1177 (P). Maryann: J.D. Smith, Patapsco Valley, Howard Co., May 25, 1881; J. J. Carter, Conowingo, June 1, 1906. Disrricr or CoLumsia: E. S. Steele, Washington, May 19, 1896; T. Morong, May 21, 1877 (M). West Vireinta: F. W. Hunnewell, July 4-6, 1914. Virernta: S. B. Buckley, June, 1838 (M); A. H. Curtiss, Bedford Co., June 6, 1872 D. House, no. 858 (M). Norru Carona: W. W. Ashe, 6445; Biltmore Herbarium, Biltmore, no. 1250b. TENNESSEE: |. Ruth, no. 116. Iurots: ex. Herb. G. Thurber. Missouri: Pilot Knob, June 17, 188-. es Var. montana (Chapm.),n. comb. Similar to above, but stems and leaves glandular-pubescent; -semaage oe goed art _— imply serrate: coro uish, strongly ampl — gars? BE i gth.—S. montana Chapm., Bot. Gaz. iii. 11 (1878). Inc. S. Mellichampti Small, Fl. 1022 (1903).— 78 Rhodora [APRIL In the mountains of northern Georgia, western North Carolina, South Carolina, southern Tennessee, and northern Alabama. Th following are representative. Sour CAROLINA: Mellichamp, no. 14 (M), and Bluffton, 1872 (M). Grorcta: A. W. Chapman, Rome (M); (MD) nos. 109124, 109125, ead and 109485. setaee.s CAROLINA: T. G. Harbison, Highlands, July 20 , 1904. TrEnneEssre: J. R. Church- all, Chattanooga, May 21, 1911; iF. James, Spring City, June 11, 883 BAMA: Hf. Eggert, Springville, July 7, 1898; S. Watson, Queensboro, 1857. 0. S. veERsIcoLor Nutt. Fic.9: Erect, are eS reeves in it inflorescence, which is commonly branched, leaves broad, cordate, rugose, crenate, 3-12 cm. ms lugeetinlad, corolla 1.7—2.5 em. long, slender up to the throat, ampliate at the lips, blue to purple at the limb, but whitish on the tube: nutlets buff to orange in color, the processes tuberculate-conical.—Gen. ii. 38 (1818). ? 8. cordifolia Muhl., Cat. 56 (1813), nomen subnudum. } Acad. Sei. eastward to North Carolina. The ra tc are rep aierigs Wisconstn: 7. J. Hale, Maiden Rock, 1861 (M); Psoka n (P). Iowa: A. S. Hitchcock, lows City, 1888; C. R. Ball, no. 1588 (M). Inuinots: F. C. Gates, no. 10828 (M); 0. E. Lansing, no. 62; H. C. Skeels, no. 388; E. E. Sherff, no. 320; H. A. Gleason, no. 1842. Missount: B. F. Bush, nos. 5851, 725 (M). ARKANSAS: no. 5912 (M). Lovutstana: Dr. Carpenter, Jackson: June; E. J. Palmer, no. 7601. Mussissrpp1: S. M. Tracy, no. 4896. Kenrucxy: S. F. site, Peaks of Otter, August 6, 1871. Ma ARYLAND: W. E. A. ives arper’s Ferry. Nort CAROLINA: R. Thaxter, Cullowhee, June 15— July 15, 1887; C. 8. Williamson, Weldon, August, 1892 (P). Grades gradually into Var. BRACTEATA Benth. This form as found in Texas is ver 'y distinct, swith floral bracts showy and much exceeding the combined length of pedicel and calyx. The leaves of the stem gradually merge into the bracts.—Lab. Gen. et Sp. 433 (1832-1836). S. cordifolia var. pulosissima Mack. & Bush, Trans. Acad. Sci. St. Louis, xii. 84 (1902), in part. Lindheimer’s no. 492, J. Reverchon’s nos. 769 and 253, and C. Wright’s no. 476 are typical. ‘It is probable also that certain plants from Arkansas and Missouri should be referred here, but one hardly knows where to stop. Var. MINOR Chapm. Very similar to typical S. versicolor, having 1 leprae size but with much smaller Ase rugose leaves.— FI. So. 323 (1860). S. rugosa Wood, Cl. Bk. 246 (1848).! In his iho description Chapman oRicaks this form to the “dry woods near Washington, Wilkes County, Ga.,” and — he sleds states “upper districts.” The plants which seem most logi- 1 In later editions of his Class Book, Wood reduces his S. rugosa to S. sazatilis. 1924] Penland,—Notes on North American Scutellarias 79 cally to be referred to this category occur rather scatteringly on the southern edge of the range of S. versicolor, and some of the specimens are not distantly related to S. saxatilis. However, Chapman had a different plant in mind. The following are taken as representatives. Missouri: W. ee | no. 721 (M); B. F. Bush, no. 791 (M); K. “gien ies Riddell. ‘Erect; or somewhat assurgent and weak, glabrous or somewhat hairy, 1-5 dm. long: leaves thin, with few spreading hairs on the upper surface, obovate or cordate, obtuse, Ppl centage the upper crenate-serrate; bracts entire: corolla about m. long, slender; the upper lip not arched, nearly as in S. versicolor: aperet resembling those of S. versicolor, but the protuberances some- what longer and more acute.—Supp. Ohio Pl. 14 (1836).— Rocky woodlands from Pennsylvania to ‘Kaus a Tennessee. Prnnsytvanta: J. A. Schafer, Jacobs Creek, Westmoreland County, July 20, 1900 (P); C. S. Williamson, Ohio Pyle, aoe = — N N EW JERSEY: C. F. Parker, no. 6619 . Dera atnall, Wilmington, ee 8, 1858. District oF Cour Bias WR: Mazon, no. Onto: H. N. Mertz, Steubenville, ee 5, 1880 (P). West oa Mr. & Mrs. E. 8. Steele, no. 31 (M). KeEn- Tucky: C. W. Short, no. 2 (P). TENNessEE: A. Ruth, no. 545 (M); Paes A 054 Var. arguta (Buckley), n. comb. This is doubtfully given rank as a variety of the above. It is small, es with ovate, sharply dentate leaves scatteringly piss and is to the mountains of North Carolina and Tennessee.—S. a eats Buckley, Am. Journ. Sci. xlv. 170, 177 (1843).—The Rais fepreacnt the plant. Nort Carona: Biltmore Herbarium (G), no. 7171; TENNESSEE: A. gee Colorado Sores Colorado. EXPLANATION OF PLATES 140 AND 141. Nutlets of Scutellaria, about X 25. Fig. 1, S. nervosa; fig. 2, S. parvula; g. 3, S. angustifolia; fig. 4, S. antirrhinoides; fig. 5, S. nana; fig. 6, S. epi- lobiifolia; fig. 7, S. lateriflora; fig. 8, S. a fig. 9 9, S. versicolor; fig. 10, } lia; fig. 11, S. Driimmondii fig. 12, S. tuberosa; fig. 13, S. serrata; fig. 14, S. er fig. 15, 8. Bus Rhodora Plate 140 yak ry rt ’ 4.4.4. €3:*. AAO ese vote at er ely ste vg ait a C. W. P. del. NUTLETS OF SCUTELLARIA. . 4 - * - _ | f ‘ + Aes : ie J i Ay! ; ney salsa ; , : ; 2 a Plate 141 Rhodora e * > ae eh A eee .75e * W. P. det. C. NUTLETS OF SCUTELLARIA. Reprinted from Ruopora, Vol. 26, May and June, 1924 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF . Polystichum mohrioides and some other subantarctic or HARVARD UNIVERSITY . New Serisgs, No. LX XII I Andean Plants in the northern Hemisphere... ...... 89 II. The dwarf Antennarias of northeastern America ...... 95 III. The eastern American Representatives of Arnica alpina . 103 IV. Some Senecios of eastern Quebec and Newfoundland... 113 V. New or restudied Plants of eastern America......... 122 ( By M. L. FERNALD Dates or Issue Pages 89 to 107, plates 142 and 143.........------+++++--: 18 June, 1924 #118 to 107, piste 144 icc re rik ree eee see e5 July, 1924 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY.—New Sertes.—No. LXXII. M. L. FERNALD. I. POLYSTICHUM MOHRIOIDES AND SOME OTHER SUBANTARCTIC OR ANDEAN PLANTS IN THE NORTHERN HEMISPHERE. PoLysTICcHUM MoHRIOIDES (Bory) Presl, var. sonia Ter (DEC. Eaton), n. comb. Aspidium aculeatum, var. scopulinum D. C. Eaton, Ferns of N. A. ii. 125, t. Ixii. fig. 8 (1880). P. Spake (D. C. Eaton) Maxon, Fern Bull. viii. 29 (1900). P. aculeatum scopulinum (D. C. Eaton) Gilbert, List N. A. Pterid. 20 (1901). P. Lonchitis, ie — (D. C. Eaton) Jones, Bull. Univ. Mont., Biol. Ser. xv (1910). After several times collecting and studying during a period of nearly twenty years this anomalous plant, which has been treated as a variety of Polystichum aculeatum or of P. Lonchitis, as P. mohri- oides and as a distinct species standing midway between P. aculeatum and P. mohrioides, I find myself firmly convinced that its relationship is primarily with the latter plant and that it is at best a northern variety of that wide-ranging austral species. P. mohrioides was dis- covered on the Falkland Islands (les Malouines) by D’Urville & Lesson during the voyage around the world of the French corvette, La Coquille. It was described by Bory de St.-Vincent as Aspidium Mohrioides and beautifully illustrated! Gradually our knowledge of the plant has been extended and P. mohrioides in one form or an- other has been found to have a wide and very distinctive range, one which, with some modifications, is duplicated in several other groups of plants. i Bory in Duperrey, Voyage autour du Monde sur La Coquille, Bot. pt. 1: 267, t. 35, fig. 1 (1829). 90 Rhodora [May West of the Falkland Islands it is found on Tierra del Fuego, whence it follows slightly northward on the high Andes as P. mohrioides, var. plicatum (Poeppig) Christensen! or P. andinum Phil., which differs from the Falkland plant only in its dwarf size (growing in exposed alpine habitats) and in the paler and thinner scales of the stipe. Skottsberg got a form of P. mohrioides on South Georgia, 800 miles (1290 km.) southeast of the Falklands; Moseley, during the voyage of the Challenger, collected an extreme form on Marion Island, 1200 miles (1930 km.) southeast of the Cape of Good Hope and more than 3000 miles (4800 km.) northeast of South Georgia; and the next year De I’Isle discovered the species on Amsterdam Island, more than 2000 miles (3200 km.) east of Marion Island. Northward, in the Andes, it occurs as P. mohrioides, var. elegans (Remy) Christensen? or P. elegans Remy, an extreme with longer and more divided pinnae. In North America two members of this alliance are recognized: P. Lemmoni Underwood and P. scopulinum (D. C. Eaton) Maxon, both confined to arid regions of the Sierra Nevada-Cascade axis, with the exception of four isolated stations for the latter, one each in the Teton Mts. of Idaho, the Mission Mts. of Montana, the Wasatch Mts. of Utah and the Shickshock Mts. of Gaspé Co., Quebec. So close are these North American plants to those of the southern Andes and the Falklands that the late D. C. Eaton, after trying to find specific differences, described and illustrated* as typical P. mohrioides (or Aspidium mohrioides) the Californian P. Lemmoni and at the same time, in describing his Aspidium aculeatum, var. scopulinum, he surmised that it belonged with P. mohrioides, saying: “T have some doubt about the plant here named var. scopulinum, as it differs more from all the rest than any of them do from each other. It has a little the habit of A. mohrioides, but, though the specimens I have seen are old, they still keep in a degree the aculeate points of the present species.”* And again he wrote that his var. scopulinum was “almost as much like A. mohrioides as it is like A. aculeatum, but as it has the lobes of the pinnae somewhat aculeate it is better to leave it with the latter species.”® As already stated, Eaton had tried to find nse characters for P. Lemmoni but was unable to do 1 Christensen, Arkiv fér Bot. x. No. 2: 17 (1910). 2 Christensen, 1. c. pagina 3 Eaton, Ferns of N On kas ag cena 4-9 (1880). 4 Eaton, |. ¢. 127, £ § Eaton, |. c. 254. 1924] Fernald,—Polystichum mohrioides 91 so, saying clearly of the Lemmon plant: “‘At first I believed it to be a distinct species.”! Similarly the great Swiss specialist upon the ferns, Christ,? in monographing Polystichum § Mohrioides had no hesitation in treating P. Lemmoni as identical with P. mohrioides (which, as understood by him, was chiefly var. elegans). In his discussion of the ferns of temperate South America collected by Skottsberg, Christensen* points out the important characters - which separate P. mohrioides from P. aculeatum and its allies; namely, the fleshy texture, scaleless surfaces of the fronds, thick and flat ribs and the large and immersed stomata (so immersed that under a good lens the lower surfaces of the fronds appear pitted or punctate). The fleshy texture, flat ribs and punctate lower surfaces are all obvious enough in P. scopulinum, and the fronds are either with or without some scales on the lower surface, but this latter character is inconstant, some sheets of perfectly good P. mohrioides, var. typicum Christensen from the Falkland Islands (coll. Cunningham, January 21, 1868) before me showing numerous slender scales among the sori. The Cunningham specimens are quite like the original plate of Aspidium mohrioides and in outline, size, texture and punctation they are so close to several North American specimens of P. scopulinum that only the keenest inspection reveals slight differences. Thus Parish’s material from Snow Canyon, San Bernardino Co., California, is a very close match in all these characters for the Cunningham plant and for the original plate. All the Falkland material I have seen, however, has the basal scales of the stipe darker and firmer than in P. scopuli- num, although the Fuegian specimens (var. plicatum) have them as pale and thin. P. scopulinum is regularly defined as differing from P. mohrioides and P. Lemmoni in the sharper and more acicular tips of the upper lobes or teeth of the pinnae. In general this character holds, but in the Cunningham material from the type-region of P. mohrioides the teeth of the lower pinnae are quite as sharp as in some of the North American plants, while in the Parish material above cited only the lowest pinnae show the sharp teeth, the upper having them quite as blunt as in the most ideal P. mohrioides or in P. Lemmonz. Further- more the plant of Marion Island has some of the pinnae quite as spinulose-toothed as in the most extreme P. scopulinum. 1 Eaton, 1. 128. 2 Christ, ke die australen Polystichum-Arten. Arkiv for Bot. iv. No, 12: 1-3 92 Rhodora : [May Reference has been made to the fact that neither D. C. Eaton nor Dr. Christ could distinguish P. Lemmoni from P. mohrioides, var. elegans. Well developed fronds of the two are almost identical and - Lemmoni has the scaleless surfaces of theoretical P. mohriordes, aad the pits in its lower faces are unusually conspicuous. In only one character, apparently, can the two be distinguished with satis- faction: in var. elegans the scales at the base of the stipe are castaneous and subcoriaceous as in typical P. mohrioides; in P. Lemmoni paler - and thinner as in vars. plicatum and scopulinum. In their extremes P. Lemmoni and P. mohrioides, var. scopulinum are well differentiated, but certain small plants of the former too closely simulate plants of the latter with unusually pinnatifid pinnae; and in view of their both possessing the essential characters of P. mohrioides and the failure of the spinulose teeth of var. scopulinum to retain the constancy one might wish them to, it seems the part of sound classification to treat P. Lemmoni as P. MOHRIOIDES, var. Lemmoni (Underw.), n. comb. Aspidium mohrioides D. C. Eaton, Ferns of N. A. ii. 251, t. xxx. figs. 4-9 (1880), as to Lemmon plant figured. P. Lemmoni Underw. Our Nat. Ferns, sat 6: 116 (1900). Besides agreeing in all their fundamental specific characters, typical P. mohrioides and its vars. scopulinum and Lemmoni are amazingly similar in their selection of habitat. Bory’s statement, based upon the observations of the original collectors of the Falkland plant, was that “elle croit dans les fentes des rochers”’ and the latest Spetenent by Skottsberg, i is similar: “Rocky places, often deep down in crevices, ” while in the “stone-runs” “Two ferns are found in solitary tufts be- tween the blocks, Blechnum magellanicum and Polystichum mohrioides, both finely developed.”! How strikingly like Eaton’s account of the type-station of var. Lemmoni: “Mr. Lemmon- writes that this fern grows in loose and moist granitic soil, the root-stocks hidden under rocks;’” or like Bradley’s account of the type-region of var. scopu- linum in the Teton Canon of Idaho: “we climbed a sharp slope of stumbling rubbish, and then found ourselves on a narrow crest, overlooking an immense cafion, the Great Téton Canon . The descent from this crest is very steep; and, in dodging falling masses of rock, started by those behind him, Mr. Bechler unfortunately 1Skottsberg, A Botanical Survey of the Falkland ne. Kungl. Svenska Veten- skapsakad. Handl. 1. No. 3: 9, 114 (1913). ~ * Eaton, 1. c. 252 (1880 1924] Fernald,—Polystichum mohrioides 93 got a severe sprain.’’! At the isolated stations at the head of Snow Brook and on the ragged walls of Devil’s Gulch on Mt. Albert, Quebec, where I have several times collected the plant, var. scopuli- num is in dry rock-crevices (serpentine) or under broken rock whence its tough roots are most difficult of extraction. In the latter situations, where the tallest and least plicate fronds naturally develop, the fronds are often badly broken by the shifting rock-debris. The range of the aggregate-species, Polystichum mohrioides, is, as already stated, similar to the ranges of several other plants, although differing, naturally, in many details. Thus M yriophyllum elatinoides Gaudichaud occurs on New Zealand, Chatham Island, Tasmania, the Falkland Islands, in the Ardean region from Cape Horn to Ecua- dor, locally in Mexico, and it is known in the western United States in Arizona and Oregon.? Empetrum rubrum Vahl, characterized by white-woolly branchlets, leaves not reflexed in age and red drupes, occurs on the Falklands, along the Andes from Tierra del Fuego into Chile, on Masafuera (the western island of the Juan Fernandez group), and 2500 miles (4025 km.) east of Patagonia on Gough Island and on the islands of the Tristan da Cunha group. Outside the Subantarctic and southern Andean regions the only Empetrums are the Arctic circumpolar E. nigrum L. with branchlets at most minutely puberu- lent, the leaves reflexed in age and the berries black or purplish; and two species centering on the Gulf of St. Lawrence, E. Eamesii Fernald & Wiegand and E. atropurpureum Fernald & Wiegand, both of which have the white-woolly branchlets, non-reflexed leaves and red berries as in the Subantarctie E. rubrum but differ from it in more trailing habit and in seed-characters.? The quaint little genus Lilaeopsis of the Umbelliferae has three strongly marked species or groups of species. One, L. lineata (Michx.) Greene,‘ with the linear-clavate broadly round-tipped 3-6-jointed leaves scattered and solitary along the creeping filiform stem and 1 Bradley in Hayden, U. 8. Geol. Surv. of Terr. 6 Ann. Rep. 219 (1873). Mei clue Sicdcnat uk Pu A Wc. HODORA, XV. 213-217 (1913). ‘It is probable that, when the original plant of Hydrocotyle chinensis L. Sp. Pl. i. 234 (1753) is critically examined, it will prove to be Lilaeopsis line he characteristi 94 Rhodora : [May without obvious stipular margins, and with the fruits constricted at base or pyriform, is confined to saline mud of the Atlantic coast of the United States and of southwestern Nova Scotia. The second species L. carolinensis Coult. & Rose, a plant with long-petioled leaves having spatulate or oblong blades up to 2.5 cm. long and 1.5 cm. broad and comparatively large globose fruits, was originally described from the southeastern United States but it is apparently found also in Paraguay (for example, Hassler, no. 12,271) and elsewhere in temperate eastern So ; The third and most widely spread group of species is typified by L. attenuata (Hook. & Arn.) Fernald,! a plant characteristic of southern South America and the Andes and represented northward by L. occidentalis Coult. & Rose and L. Schaffneriana (Schlecht.) Coult. & Rose and southward by the plants of the Falkland Islands, New Zealand, Tasmania and Australia which have erroneously passed as the Atlantic North American L. lineata or Crantzia lineata (Michx.) Nutt. The published illustrations? of fruits indicate considerable differences and it is possible that the austral series contains other species than L. attenuata and L. Schaffneriana but without better material than is now at hand it would be unwise to attempt further subdivision. The essential point in regard to the Australian, New Zealand, Tasmanian, Falkland, Argentine and Andean plants is, that they as well as the Mexican and Pacific North American plants all differ in fundamental characters from the Atlantic North American L. lineata; for in them all the more elongate and slender or often attenuate leaves are tufted along the comparatively stout creeping stem, not solitary and scattered as in L. lineata; when well developed they show 6-13 joints instead of only 3-6 (rarely 7) and they often have scarious stipular margins which frequently persist as old shreds. Whether they finally prove to be a single species, L. attenuata, or several, the plants of Subantarctic regions and of temperate and Andean South America constitute, with the Mexican and Pacific North American plants, a distinct section of Lilaeopsis. It is thus evident that, although differing in details of distribution, Polystichum mohrioides, Myriophyllum elatinoides, the red-berried —— and the species of Lilaeopsis centering about L. attenuata ta (Hook. & Arn.),n.comb. Crantzia attenuata Hook. & Arvo in —- Bot. Misc. iii. 346 a 833). ? Hook. Fl. Antarct. ii. 287, t. C. (1847); Weddell, Chloris Andina, ii. t. 68 ye Coult. & Rose, Bot. Gaz. xxiv. 48, 49, figs. and 4 (1897); Jepson, Madrofio, fig. 25 (1923). ee 1924] Ferna!d,— Dwarf Antennarias 95 are similar in having interrupted Subantarctic and Andean ranges and in occurring in the northern hemisphere only in western America or in the region of the Gulf of St. Lawrence or in both areas. Il. THE DWARF ANTENNARIAS CF NORTHEASTERN AMERICA. (Plate 142.) The larger species of Antennaria of temperate eastern America are reasonably understood, but there is another series of highly localized species, chiefly of the calcareous areas from northern Labrador to Newfoundland, eastern Quebec and the James Bay region, which have not been so clearly defined. These are the plants which have passed chiefly as A. alpina (L.) Gaertn. and which, with it, form a rather natural group of species. In the cordilleran region of North America these plants have attracted much attention and a large number have been proposed as species. In eastern America they occur wholly beyond the habitations of resident botanists, unless we include in eastern America botanically quite similar Greenland. The Antennarias of Greenland have been carefully treated and beautifully illustrated by Porsild;! but in the region immediately to the west and southwest of Greenland these plants are collected only by the chance botanical visitor and our knowledge of them, like our knowledge of the whole vast region from Gaspé and Newfoundland northward, is in very rudimentary condition. However, the necessity of properly identifying two quite distinct species discovered in 1923 on the Shickshock Mountains makes it desirable to draw into convenient form our knowledge to date of these plants. The following synopsis of the species occurring south of Hudson Straits is therefore presented, not because it is final but because it may draw attention to a group about which much more information is needed. Further exploration of Newfoundland, Anti- costi Island, the Gaspé Peninsula and the Labrador Peninsula will surely bring to light many additional species; and, to judge from our experience to date, they may be most hopefully looked for on barrens and mountains of limestone, basic schists and traps. 1 Porsild, On the Genus Antennaria in Greenland (Arbejder fra den Danske Arktiske Station paa Disko, Nr. 9), Meddel. om Groenl. li. 267-281 (1915). 96 Rhodora [May In order properly to orient the small-leaved boreal series here spe- cially considered, the leading features of the other sections of the genus represented in eastern America are indicated in the key. a. Basal — erect, oblanceolate to elliptic-acuminate, 2-16 cm. long, similar to the cauline ones: involucres of the pistil- late ee brown to blackish: plants sparingly to not at all st A. eucosma Fernald & beg a and A. pulcherrima se ) Greene. a. Basal leaves spreading, forming depressed rosettes, strong contrasting in outline og the cauline leaves: Seti humifuse or heeds stoloniferous b. b. — — — Pome .O-5 mm. wide, blunt or barely short-m c. All Pond ae daces bracts of the pistillate heads? deep- rab or blackish d. d. Bracts hoe age narrow; the inner linear or lanceolate and acute: heads 1-6 e é. Rossteeleives terminated by a short but distinct glabrous mucro: corollas 3.5-5 mm. lon Jf. Rosette-leaves Bo ac and brig ight -green or grayish-pubescent above: cauline leaves dis- ppendag denuded receptacle 20-30, 0.3-0.4 mm. broa: much broader n the intermediate ridges. . . fe a ¢ alpina. f. Rosette-leaves annaend: cauline leaves crowded; he upper with a lanceolate pubescent scarious appendage: corollas age mm. a — of Poise Cas oe Via a pies pepe. . Sornborgert. é. Rostte leave canescent, biunt, t the terminal mucro obsolete or in old weater wort leaves barely visible: co colle 3—4 mm Z g. g. Cauline eaves 9-15; Hans median cae upper igri a slender subulate one the pi once wae us appendage. .3. A. cana. g- Cauline leaves 5-8; all but the ‘verareinge with a. ng-lance eolate flat scarious tip 1.5-3 ona sas a ae alee nen geet ee Pie hea fet 4. A. vexillifera. d, Bracts in op se 5 outwardly shorter series, oblong, co Meg ik MOE er eae Sen 5. A. pygmaea. c. At deen: the inner bracts of the pistillate heads with white or whitish tips: basal leaves whitened above with minu te pubescence h. h. Thel 7 “2%. 7 7 appendages: rosette-leaves blunt or short- reticent FB sees ee -7 mm. high, with ¢. Tends o. of 4-6 series oe distinctly unequal pale- peowa WG ee a ee os ek 6. A. straminea. __ By basal leaves are meant not only those of the basal rosette ee ne ing On account of weathering the bruised and uncharacteristic. uri apecies of chit series of which stamiinate plants are known is 4. nitida. The descriptions of Su abi germ rab die aes gm pist plan 1924] Fernald,— Dwarf Antennarias 97 4: inyaiows of 2-3 (-4) series of subequal or obsctrely cated whitish, creamy or roseate bracts j. : Trdyeaaed of the rosette-leaves close and lustrous as if varnished: cauline leaves 9-18: inflores- cence glomerulate: corollas 3- .5 mm. long k. gs ooet leaves with scarious tips 2-2. 5 : involucre not viscid, with thin bracts; the thin tips of the inner series lacer- Sick pieind SaaS & OSA Cava ee : albicans. k. All Sat the uppermos line leaf merely short- mucronate: citar viscid-hirsute, with thick and firm bracts; the thick tips entire Or merely orenuiate 65s. ies oh cue bys pan 8. A. nitida. j. Indument of the rosette-leaves a loose tomentum: eo leaves 5-10: orescence a corymb: corollas 3. 3-8 mii ee l. paneer gate 3-7 ¢ high, not glandular: the upper cauline awe with oblong-lanceo- late scarious appen nda; ges 2 -3 mm. long: heads th. .9. A. Peaset t. Flowering stems 0.5-1.5 dm. high, glandular- appendages: rosette-leaves mucronate: involucre api ™. Rowman comparatively mg 02-2 le wide, : dri ae Pena . Middle and upper cauline leaves terminated Mid a flat or merely involute sc ete appendage. ge: gagenet hers Greene, A a enguunces F A. spathulata Fernald, A Gree n. Middle. ‘and upper cauline leaves subulate-tipped or mucronate, without a scarious appendage (except sometimes on thie bracteal leaves of t the mantis torre Fhe Zune a ernald, m. Rosette-leaves comparatively e, 0.7-5. 5e¢ with 3-7 somew t prominen om bees eath . A. "plntapiniolia, o ) Bishardones A. occidentalis ea A. fallax Gree reene, Fernald, A. Parlinii Fernald, A. solitaria ete 1. A. auprna (L.) Gaertn. Fic Humifuse, with trailing sublig- le 4-13, beco: distant by elo tion te stem; the 1 es — 2 ie ps: pistillate heads 1-5, ses- about 3-seriate, subequal; the outer lanceolate to oblong, acuminate inner attenuate: corollas 4-5 mm. long: longer pappus- bristles 5-6 mm. long: achenes glabrous, 1.3-1.5 mm. long: pits of 98 Rhodora [May the denuded receptacle 20-30, 0.3-0.4 mm. broad, much broader than the intermediate ridges. Three varieties with us: Rosette-leaves green and glabrous above.................. Var. typica. osette-leaves canescent-tomentose above............... Var. canescens. Flowering stems elongating to 2-2.3 dm., with #7 13 leaves: he sette-leaves green and glabrous above: hea ~2, the low (when present) on a pedicel 1-1.5 cm. lon ve Sse as ached ungavensis. Var. typica. Gnaphahum alpinum L.-Sp. Pl. ii. 856 (1753). A. alpina (L.) Gaertn. Fruct. ii. 410 (1791).—Arctic America, south to Kangalaksiorvik Bay, Labrador (Owen Bryant) and mountains of ae h Columbia; also northern Eurasia. Fl. July-September. . CANESCENS Lange, FI. Dan. xvi. (fase. xlvii.) 9, t. 2786, fig. 1 (1869); Fernald, Ruopora, xviii. 337. (191 6). A. angustata Greene, Pittonia, iii. 284 (1898). Var. cana Fernald & Wiegand, Ruopora, xiii. 24 (1911), in part—The commoner extreme in eastern America, ae south to Port Manvers, Labrador (Delabarre). . UNGAVENSIS Fernald, Ruopora, xviii. 238 (916); ws tee only ei the type-region, Stillwater River, Ungava Dist . A. SoRNBORGERI Fernald. Fic. 2. Humifuse, with orc kien eous branches up to 1 dm. long: stolons very short and crowded: ret a oblanceolate, 6-12 mm. long, 1.5—-2 mm. wide, narrowed - summit to the short-mucronate tip, canescent-pannose: flowering em stiffly erect, 0.4-1.1 dm. high: cauline leaves rather crowded, O13, linear, 5-15 mm. long, 0.5-1.5 mm. wide; the upper with villous ceolate scarious tips: Sepia heads 1-3, sessile, campanulate: involucre 6-7 cm. high, lanate at base: bracts about 3-seriate, sub- equal; the outer lanceolate, Toe the inner linear-attenuate, yellowish-brown, erose-serrulate: corollas 3.5-4 mm. g: longer ps. agar sige 4-5.5 mm. long: achenes glabrous, 1.2-1.4 mm. long: pits of the denuded receptacle 60-100, 0.1 mm. broad, about as wide as the blunt-edged intermediate ridges —-Ruopora, Sas 237 (1916).— ve Lasrapor (J. D. Sornborger). Fl. Augu 3. A. caNA (Fernald & Wiegand) Fernald. Frc. 3. Humifuse, the ae leafy stolons very short (up to 2 cm. long): rosette-leaves narrowly cuneate-obovate or broadly oblanceolate, obtuse, not mucro- nate, broad-based, 3-11 mm. long, 2-5 mm. broad, white above with dense minute tomentum: flowering stems 2.5-12 cm. high, slender: cauline leaves 9-15, rather crowded, linear; the lower 8-15 mm. long, subequal, v fas aes outer ea brown; the i inner lanceolate, tawny, slightly prow esa corollas 3.5-4 mm. long: longest pappus 1924] Fernald,—Dwarf Antennarias 99 4-5.5 mm. long: achenes glabrous, 1.2-1.6 mm. long: pits of the mature denuded receptacle 60-100, 0.1 mm. broad, much broader than the acute Dry limestone barrens, western NEWFOUNDLAND. Fl. Ju y, early c; cum apice scarioso Lass masa 1.5-3 mm. Sea's nibiaes capitulis femineis 1-5 corymbosis hemisphaerico-campanulatis basi rotundatis; involucro 6-7 mm. alto basi lanato; bracteis 2-3-seriatis subaequalibus tenuissimis, exterioribus anguste oblongis obtusis vel subacutis basi castaneis, interioribus lanceolatis fulvis plerumque acuminatis; corolla 3-4 mm. gree vi lobis 26 BY SPREE stylo purpurascenti; planta mascula Matane County, QuEBEC dry open gravel on che? sablitead- auadle cdenie: about 1070 = between Mt. Mattaouisse and pie _ Collins, July 8, 1923, M. L. Fernald, Ludlow Griscom, K. K. Mackenzie, A. S. Peas e & L. B. Smith, no. 26,056 (TYPE in "Gray Herb.). Closely simulating A. cana but at once distinguished by its cauline leaves. In A. cana these are 9-15, the median and upper ones tapering to a slender subulate tip, only the very uppermost with a scarious tip. In A. vexillifera the cauline leaves are fewer (5-8) and all but the very lowermost bear somes broad pennant-like scarious tips (whence the specific nam A. vevillifera has the ae leaf-tips of A. alpina, but that species has much narrower and distinctly mucronate basal leaves, larger heads with narrower and more fuscous involucral bracts and longer corollas (4-5 mm. long). 5. A. pyGMAEA Fernald. Fic. 5. Dwarf, 3-4.5 em. high, mono- cephalous, humifuse; the assurgent stolons very short, not ly elongated: basal leaves oblanceolate, mucronate, 8-14 mm. long, 2.5- 3.5 mm. wide, glabrous or glabrate above; the cauline about 9, crowded linear-oblanceolate, 6-14 mm. long, lanate beneath, glabrous or glabrate above, with a lanceolate or narrowly deltoid glabrous flat scarious tip 1.5-2 mm. long: pistillate involucre hemispherical, 7 mm. high, 12-13 mm. broad (in — dried an gee lanate at base: bracts in about 6 series, definitely imbricated, oblong, obtuse; the outer fuscous, with short stramineous tip; the inner with a long obtuse stramineous ti corollas 4 mm. long: staminate plant unknown.—RuHopora, xvi. 139 100 Rhodora [May (1914)—Known only from northeastern Laprapor! (Weitz et al.). Fl. August. 6. A. STRAMINEA Fernald. Fic. 8. Plant humifuse, the leafy stolons very short or slightly elongated (up to 7 cm. long): Jeaves of the rosette spatulate, subacute, barely mucronate, 5-12 mm. long, 2 mm. broad, white above with dense fine tomentum: flowering stem 3-14 em. high, slender, remotely leafy: cauline leaves 8-10, linear, 6-14 mm. long, 1-2 mm. wide; the median attenuate to a dash subulate tip; the upper with a linear scarious tip: pistillate ape 1-7, usually in a close corymb, hemispheric-campanulate, rounded at base: involucre 5.5-7 mm. high, 4.5-8 mm. broad (in the dried ok aS with 4-6 series of very agpeetas imbricated bracts: the outer bracts ovate or oblong, brown, slightly lanate at base, with a thin chartaceous stramineous obtuse or “subacute do; the median oblong, with a deltoid obtuse or tip; the inner with a lanceolate erose stramineous tip: corollas 3.7-4.2 mm. long: achenes glabrous, 1.4 mm. long: longer pappus-bristles 4. 55 mm. long: style. yellowish, becoming rown: staminate plant unknown.—Ruopora, xvi. 130 (1914).— Rocky or turfy caleareous barrens and headlands bordering Notre — and Ingornachoix Bays, NEwFouNDLAND. Fl. July, early. Augu 7 re ALBICANS Fernald. Fic.6. Plant humifuse, the leafy stolons very short (up to 2 cm. long): basal leaves spatulate, subacute or ob- tuse, scarcely mucronate, 3-8 mm. long, 2-3 mm. wide, white above with dense minute somewhat shining tomentum: flowering stem 4.5-9 em. high, slender, somewhat remotely leafy: cauline leaves 9-15, _— , 6-12 mm. long, 1-2 mm. wide; the median attenuate, subulate / tip; the upper with a glabrous linear scarious tip 2-2.5 mm. long: Distillate heads (1—)2-5 in ee turbinate-campanulate: involucre 1 As noted when A. pygmaea was publis this is the Fags which Gray in = Synoptical Flora treated as A. carpathica wile’ R. . ador (a mo cephalous form!).”” But Gray, of course, was writing oe ‘we the intensive pat oup of ies ‘leaves (the section including A. eucosma Fernald & Wiegand, A. pulcherrima (Hook.) Greene, A. lanata (Hook.) Greene, etc.), Peiapend Gray’s ‘‘monocephalous form”’ is a humifuse p y leaves. Dr. Theodor Holm, lamenting of the name A. rpahien from American literature, cites Gray's erstood b ‘Having examined a number of specimens ens of A. Janata Greene I I find it Pega to distinguish them afrom A. carpathica.”’ see & f A pathica : in A. lanata inthe former 2 4 he median Sind upper caulin leaves end in hoe. oFtnear-subulte te scarious tip, only 1 or 2 of the uppermost leaves ever showing dilated a appendages; co oheedt aac dane of the cauline leaves have broad and appendages. This obvious character, supplementing the broader leaves, ‘the smaliet heads, other will different A. pupmaca back into the European 1924]. Fernald,—Dwarf Antennarias 101 4.5-6 mm. high, 4.5-6 mm. wide (in dried specimens): bracts in 2-3 series, subequal, thin; the outer 3-4 mm. long, oblong or lanceolate, obtuse or subacute, straw-color or brown, green rod a little aaa at base; inner oblong, obtuse, licerate-crose, milk-w corollas ° ie mm. long: achenes glabrous, 0.8—1 mm. long: aber pecbericatee 44.3 mm. long: staminate plant unknown.—Ruopora, xvi. (1914).—Dry pnetne cee, northern domes of Table Mt., Port & Port Bay, Newrounp Fl. July. 8.. A. NITIDA ate ‘Fre. 9. Resembling A. albicans. Pist1- inwalucrs 6-7 mm. high, viscid-hirsute: Beate about. 4-seriate, very unequal; the outer 34.5 mm. oak oblong, obtuse, whitish, green and densely lanate at base; the inner white, gradually narrower, obtuse, entire or barely erose: corollas 3-3.5 mm. long: achenes 0.8-1.2 mm. long: longest pappus-bristles 5 mm. ong. Sialic: PLANT with glomerules 1-2 em. in diameter: involuere with creamy or Pate og ree ie obovate entire to merely ae bracts: corollas lon appus 44.5 mm. long, upwardly barbellate; A “lightly dilated (rately broad) tips crenate. a Fitton: | iii. 283 (1898 A, arida catia E. Nelson, Proc. U. S. Nat. Mus. xxiii. 710 (1901). a Rydberg, FI. Colo. 369 (1906).—Dry oe soil, Charlton "ek. James Bay, QuEBEc to ATHABASCA and Uran. Fl. late June, J A. Peasei, spec. nov. (Fie. 11), humifusa, stolonibus foliosis con- fertis perbrevibus (ad 2 cm. longis) ; oli basilaribus late oblanceolatis vel anguste cuneato-obovatis 7-12 mm. _ longis 2+4.5 mm. latis mucro- natis supra albidis, tomento denso minuto; caule florifero 3-7 cm. alto gracili; foliis caulinis 5-7 lineari-lanceolatis 0.8-1.5 cm. longis, 2-5 dense corymbosis hemisphaerico-campanulatis basi —— invo ucro 6-7 mm. alto basi lanato; bracteis 2—3-seriatis chartaceo-petaloideis, exterioribus oblongis subacutis pallide i basi subcastaneis, interioribus similibus acuminatis is; corollis 3.8-5 mm. longis; purpurascenti; achaeniis glabris; planta — ignota.—Ma c: talus of mica-schist, Ful 13, 1923, A. S. Pease & L. B. Smith, no. 26057 (ryPeE in Gray Herb.). Related to A. albicans of western Newfoundland and A. subriscosa of Rimouski County, Quebec. From the former distinguished by the distinctly mucronate and loosely tomentose basal leaves, the few cauline leaves, the larger heads, the acuminate inner involucral bracts and the longer corolla; A. subsiereie ia:= conteer png ith, ais 102 Rhodora [May branches often 4 or 5 dm. long. It has 7-9 cauline leaves, the upper- most with subulate or involute tips; its heads are more numerous (3-9) and its glandular-viscid involucres have more definitely imbri- cated bracts. Various cordilleran species, A. wmbrinella Rydberg, A. pulrinata Greene, etc., are related to A. Peaset but I am unable to find any species to which it can be satisfactorily referred. 10. A. susviscosa Fernald. Fie. 10. Plant densely humifuse, — trailing branches subligneous, often 4-5 dm. long; stolons very ort and crowded: leaves of the rosettes spatulate, obtusish, scarcely cacsonls or with a very short mucro, 0.5-1.5 cm. long, 2-5 mm. broad, densely white-tomentose: flowering stems 0.5-1.5 dm. high, white-tomentose throughout, glandular-hirsute above: cauline leaves 7-10, tomentose; the lower linear-oblanceolate, mucronate, 1.5-2. cm. Jong; the upper linear-attenuate, with a subulate or involute sub- scarious pubescent tip: pistillate heads 3-9, densely or loosely corym- bose: involucre soared nahh apt 5-6.5 mm. high: bracts about 3-seriate; the outer 3-4 mm. long, oblong, subherbaceous, greenish or stramineous, often rose-tinged, glandular-viscid, with a thin obtuse cream-colored or rose-pink tip; Sasa narrower, acutish: corollas 3.8-4.3 mm. long: achenes papillose, 1-1.2 mm. long: longer pappus- bristles 4.5-5 mm. long: staminate hat unkno own.—RHOoDoRA, Xvi. 131 (1914).—Cold ogres limestone sea-cliffs east of Bic, press Co., QUEBEC . July. 1. A. ISOLEPIS aa F IG. 7. Humifuse, the leafy stolons up to 7 em. long: rosette-leaves oblanceolate or obovate, subacute or obtuse, short mucronate, 0.8—2.5 cm. long, 2-7 mm. broad, densely white-tomen- tose above: flowering stems 0.7-2.5 dm. high, flexuous: cauline leaves na 8, linear-oblanceolate, tomentose, 1.5-3 em. long, 2-4 mm. broad; the lower with subulate tips; the 5-8 upper with broad flat scarious tips: heads 5-9, corymbose: pistillate involucres 6-8 mm. high: bracts of about 3 lengths, fuscous except for the oblong erose gee or whitish widely spreading petal-like tips: corollas 3-3.7 mm. long: achenes sparingly papillose, 11.2 mm. long: longest Pride Seige 4.5-5.5 mm. long: staminate plant unknown.—Ottawa Nat. 41 (191 b= saosin of Hudson Strait, Uncava District, and sich near the outer coast at least to Port Manvers, Lasrapor. Fl. July, August. sc paparrene ae? OF elena 142. A, a. 3 3a, Serer ss ie, Hp “4, one aD ie vendigees EE ere id cauline leaf x 4. fie. 3 5, A. pygmaea X 1; 5a, tip of median cauline leaf x 4. Fic. 6, A. albicans X 1; 6a, tip of median cauline leaf < 4; 6b, ae tee of i at } cauline : Fic. 8, A. straminea X 1; 8a, tip of median eaf < 4. Fic nitida X 1; 9a, tip of medi leaf X 4; 9b, inner bract of involucre x 4. Fie. 10, A. subviscosa X 1; 10a, tip of median cauline leaf x 4; 10c, central flower and , achene X 4. Fig. 11, A. Peasei X 1; lla, tip of median cauline leaf X 4; 11c, central flower and achene X 4. 1924] Fernald,—Representatives of Arnica alpina 103 II. THE EASTERN AMERICAN REPRESENTATIVES OF ARNICA ALPINA (Plate 143.) In 1905 I published! a synopsis of the species of Arnica known in northeastern America and at that time predicted that “doubtless many other species will be found, especially on the mountains and cliffs of eastern Quebec.” Although the prediction has not yet been wholly fulfilled, two very distinct new species have been discovered, one on the Long Range of western Newfoundland, the other in the Shickshock Mountains. In determining the exact relationships of one of these plants, discovered in 1923 on Mt. Logan and the neigh- boring Mt. Mattaouisse in Matane County, Quebec, it has seemed appropriate to present a new treatment of the eastern species with few and small cauline leaves and with white and merely barbellate pappus (A. alpina and its allies) and to publish Miss Brackett’s drawings of these species. In order to show the contrast between these plants and the well known A. mollis Hook. of the Rocky Moun- tains and of the Gaspé Peninsula and of northern New England and northern New York, the latter species is also included in the key, although it seems unnecessary to discuss it in detail. a. — 14 pairs (sometimes scattered small bracts ve), at omy the Rae entire or only sparingly toothed: y rom se leaves entire or with only remo ie Gragele ts of involucre and summit of peduncle saeor pened ee e. Leaves nearly glabrous or only minutely oe baa hispid: involucral hemos 15-20, late, 2-3 mm. only sparingly pubescent above the middle: ligules "with blunt lobes only : Oe ee ee eer er eee 1, A. alpina, e. Leaves loosely villous: involucral bracts 8-10, narrowly rhombic, 3-3.5 mm. broad loosely Sie to the tip: ligules with h acutish lobes 3 A. wehiciin involuere and pst glabrous or merely minutely pilose on nerves f and margn, entire; the cauline lanceolate to y pilose or glab- wanes parmny Babes Se eb long. con eeprom Een oe fF plantaginea, ane ular-puberulent visci ene b saeelity = anes He the cauline linear or linear- 1 RuopvorRA, vii. 146-150 (1905). 104 Rhodora [May lanceolate: stems viscid-villous at base: _— cral bracts linear- to lance-attenuate, 1.2-1.5 ¢ pa, uniformly viscid-villous: ligules with srotak 4, ee ee ee eer ie ar A. Sornborgeri. e 0 Dal leaves a dentate g . Disk-corollas 6-8 m “che are glabrous or only sparingly setose ry the densely pilose portion - tube and throat 3.5-6 ng: mature us 5-8 mm. long, bright-white........... 5. A. chionopappa. g. Disk-corollas 9-10 mm. long, with copiously setose ; the ion cies Sartion o of the tube and sicat -9 mm. long: mature pappus 9-10 mm. WO TI ois oak es es eo Oe s Fo 0 6. A. gaspensis. b. Achenes glabrous or nearly so below the middle........ 7. A. Griscomi. a. Stem-leaves 3-5 pairs, all des dentate: pappus yellowish- rown or + alive tinged: its bristles subplumose............ 8. A. mollis. 1. A, atprna Olin & Ladau. Fic.1. Stem 0.5-3 dm. high, sparingly short-villous below, densely so above, copiously villous-lanate at tip: leaves nearly glabrous or minutely glandular hispid; the basal linear- lanceolate, 3-9 cm. long, 0.4—1.6 cm. broad, entire or with few remote small teeth; cauline 1-3 pairs, entire, sessile, the lowest usually much longer than the reduced upper ones: heads solitary (rarely 2-4), 4-6 em. broad: involucre densely villous-lanate at base; its 15-20 narrowly lanceolate long-acuminate mere! purplish bracts only sparingly pubescent t above the middle, 1.2-1.5 cm. long, 2-3 mm. broad: ligules 10-15, with 3 blunt lobes 1-2 mm. long: achenes hirsute: pappus cream-white; its bristles barbellate. at 11 (1799), acc. to Ind. Kew. A. montana, B alpina L. Sp. Pl. ii. 884 (1753). A. angustifolia Vahl, Fl. Dan. ix. t. 1524 (1818).—Arctic regions, south to rocky slopes of the Torngat Mts., Labrador, Mosquito Bay, Ungava Distr., northern er and ’Alaska:; arctic and alpine Eurasia. FI. August. 2. A. PULCHELLA Fernald. Fic. 2. Stem 1-1.8 dm. high, villous, especially above, with white hairs 1-2 mm. long mixed with minute stipitate glands: leaves loosely villous on both surfaces; the basal i asblate’ or oblanceolate, 3-15 cm. long, 0.5-1.5 cm. broad, entire or remotely callous-dentate; cauline 1-2 pairs, entire: head solitary, 4.5 cm. broad: involucre densely villous-lanate at base; its 8-10 nar- white, epee —RuHopora, xvii. 18 (1915). Dre expo ned lime- stone ledges an dt shingle, ‘Table Mt., Port A Port Bay, Nerfoundland. Fl. July. 3. A. PLANTAGINEA ne Fic. £ ‘Stem pe 5 dm. high, glabrous be or only very sparingly villous-hispid below, glandular-puberulent at summit: Habeous Ut minutely on the nerves and margins, entire; the basal « to oblanceolate, 0.6—1 ig long, 1.2-2 cm. broad; pairs, pobre the lower 7-13 cm. long, the upper much reduced: bee 5 4-5 cm. broad: involucre glandular-villous at 1924] .Fernald,—Representatives of Arnica alpina 105 base; its 10-12 oblanceolate bracts 1 em. long, sparingly pilose or glabrate above the base: ligules 10-15, with 3 blunt lobes 1-2 mm. long: achenes hispid: pappus white, barbellate.—Fl. 527 pte Fernald Ruopora, vii. 147 (1905). A. angustifolia Torr. & Gray, FI. ii. 449 (1843), in part, not Vahl. A. alpina yeinch Syn. FI. i. pt. 2: $2 ‘(1886), in part, not Olin & ma, var. Lessingiit Fernald & Sornborger, brooks, ete., Ramah to — Labrador, west to Kooksoak River, nee ene District. Fl. Augu A. SORNBORGERI Fernald Fic. 3. Stem 1.5-2 dm. high, viscid- villous throughout, densely so at summit: leaves glandular-puberulent and viscid-rillous, irregularly dentate; the basal ee eae to narrowly oblong, 6-9 em. long, 0.8-1.6 cm. broad; cauli pairs, linear to lanceolate: head solitary, 4.5-6- cm. babii: “inecliitie uniformly viscid-villous throughout; its 10-12 linear- to lance-attenuate bracts 1.2-1.5 em. long: ligules 10-15, with acutish lobes 3-5 mm. ong: achenes hirsute: pappus white, barbellate——Rwopora, vii. 147 (1905).—Rocky banks of streams, Torngat Mts., Labrador. Fl. August, September 5. A. cutonopappa Fernald. Fic.5. Stem 0.7-4dm. hi gh, loosely or sparingly villous: leaves sparingly villous or glabrate; the basal lanceolate narrowly ovate, regularly dentate, 0.3-1.8 dm. long, 0.5-3.3 cm. broad; cauline 24 pairs, the lower slender-petioled and resembling the basal, the upper sessile and much smaller: heads 1-4, 3-4 cm. broad: involucre villous throughout; the 1 0-15 linear- to lance-atienuate bracts 7-13 mm. long: ligules 10-15, with short bluntish lobes: disk- corollas 6-8 mm. long, with glabrous or only sparingly setulose lobes; the densely pilose portion of the tube and throat 3.5-6 mm. long: achenes densely shore eotalese, 4-6 mm. long; the ascending white hairs about equaling one-half the diameter of the achene: pappus bright-white, in ruit 5-8 mm. long.—Ruovora, vii. 148 (1 905).—Calcareous ledges, shingle, cliffs and river-gorges, Table Mt., Port 4 Port Bay, New- foundland; Cap-Rosier, Gaspé Co., Quebec to the Tobique River, New Brunswick. Fl. late June-ear y August. _ 6. A. GASPENSIS ficone Fic. 6. Stem 3-4 dm. high, villous the achene: ppus cr eine in fruit 9-10 mm. gta —RHODORA, vii. 148 (1905) Galan ledges and cliffs, Cap Tourelle, Gaspé “2 Bp peach Fl. July. oe A. Griscomi, n. sp. (F1¢. 7), —— 106 Rhodora {May horizontali vel oblique adscendente; caule solitario (rarissime 2-3) piloso supra sparse villoso-hirsuto 0.8-2.5 dm. alto; foliis rosulatis lanceolato-ellipticis 3-10 cm. longis 0.5-2 cm. latis subacuminatis basi angustis plus minusve petiolatis supra papillosis vel sparsissime villosis glabratisque ciliatis margine calloso-serratis, dentibus 3-7- jugis; foliis caulinis 1—2-jugis, inferioribus oblanceolatis vel oblongis acuminatis ciliatis dorso glabris, paginis interioribus apice villosis; ligulis 9-12 luteis, lamina 1-1.5 em. longa 4-5 mm. lata 7-9-nervata apice 3-dentata, dentibus longioribus 0.5-1 mm. longis; corollis disci 6-7 mm. longis, tubo villoso 2.5-3.5 mm. longo; achaeniis 3-3.5 mm. longis glabris vel ad apicem sparse breviterque hirsutis; pappo 6-7 mm. longo albo, setis barbellulatis—Matane County, QUEBEC: moist rock-walls and shelves of hornblende-schist at head (altitude about 1070 m.) of Big Chimney, Mt. Mattaouisse, J uly 8, 1923, M.L. Fernald, Griscom, K. K. Mackenzie, A. S. Pease & L. B. Smith, no. 26,082; same station, August 20, 1923, Fernald & Smith, no. 26,085; dry schistose crests and talus of Razorback Ridge (altitude 850-1000 m.), Mt Logan, July 13, 1923, Pease & Smith, no. 26,083; cold chim- neys in the schist at about 900-1000 m. altitude, south of Fernald Pass, Mt. Mattaouisse, August 20, 1923, Fernald & Smith, no. 26,084 (TYPE in Gray Herb.). Named for Ludlow Griscom, professional ornithologist, keen amateur botanist, tireless explorer and choice companion, who first discovered the plant which subsequently proved to be somewhat generally distributed on the cold walls of Mts. Mattaouisse and Lo, gan. Arnica Griscomi belongs to the group of species nearly related to A. alpina. From all of them it is distinguished by its achenes being glabrous at base or often nearly throughout; A.alpina, A. plantaginea, A. chionopappa, A. gaspensis and A. arnoglossa Greene, the five Plate 142 Rhodora 7% aus KA if oe E. Brackett del. - DwarF ANTENNARIAS. jae : j ‘ Aa 4 r > * + e t i i fi . * ey " " x > Vers a rea oa m3 | es : as * : eo % A S 3 ‘a s : ; ithe 3 hm , * * : 3 eet ; t ‘ Sei ; : i 2 iS Ne 5 3 M ‘ { a A q 2 a. wy sod ei s ae Plate 143 Rhodora A. E. Brackeit del. - REPRESENTATIVES OF ARNICA ALPINA. * f * 1Y: a5 * Ne ee aes ei A 3 a) : | , ote 7 . : | 3 Pe ee | ‘ | i | | ; 5 Saif . ‘ | 1924] Reports on Flora of Boston District, —XLVI 107° known only at the type-station where it has thrice been collected) has the involucral bracts glandular as well as pilose, its disk-corollas slender and longer than in others of the group and with the villous tube more than twice as long as in A. Griscomi, its achenes nearly twice as long and its pappus creamy- or sordid-white instead of snow- white. A. arnoglossa, described from the Black Hills, is densely glandular-puberulent above and the very young achenes (the only ones seen) are definitely longer than the mature ones of A. Griscomt. EXPLANATION OF PLATE 143. Fia. 1, Arnica alpina X 1; disk-flowered X 4; 1b, tip of ligule x 4 Fia. 2, A. pulchella X 1; o>, Ds of ligule X 4. Fie. 3 ” head of A. Sornborgeri x 1. Fia. 4, A. plan ntagine aX. Fie. 5, A. brs do xX 1; 5a, disk- pl x4. Fie. ba, —— of A. gaspensis sis X 4. Fig. 7, A. Griscomi X 4; 7a, disk-flower x 4 (To be continued) Aa = hia hs tae ee ' sande of CONTRIBUTIONS FROM THE GRAY HERBARIUM OF RVARD UNIVERSITY.—New Series.—No. LXXII. (Continued from page 108.) IV. SOME SENECIOS OF EASTERN QUEBEC AND NEWFOUNDLAND. (Plate 144.) SENECIO RESEDIFOLIUS.—In 1818 Ledebour described from the Altai of Siberia a monocephalous Senecio under the name Cineraria lyrata.. The plant was subsequently collected by Bunge, Fischer, Tiling, Turczaninow, Charles Wright and others in the Altai and Baikal regions or at other points in Siberia eastward to Amur and the coast of Bering Straits. It is a well marked but highly variable plant, a single variation of which was illustrated in Reichenbach’s Iconagraphia2 In 1831 Lessing gave a beautifully detailed description of one of Chamisso’s plants from Bay St. Lawrence or Gulf of St. Lawrence on the Asiatic side of Bering Straits as Senecio resedifolius,* and as a syncnym he cited Cineraria lyrata Less., Lessing’s specific name being unavailable on account of the older Senecio lyratus L. f. Gradually the knowledge of Senecio resedifolius was extended until, in 1867, Herder‘ recognized it as a wide-ranging species, extend- 1 Ledeb. Mém. Acad. Pétersb. v. 576 (1818). Cecpiecbeaia Ic. Bot. Crit. ii. 1, t. 101 (1824). , Linnaea Lessing’s scons fate clearly that the original in Chamisso’s hand, in the Gray Herbarium, is equally explicit: “ from “Alaska: St. Lawrence Island, Cha as co: 4 Herder, Reisen in den Siiden von Ostsibirien, iii. Hot 2: 116 (1867). 114 Rhoedora (JUNE ing from Bering Straits across arctic Siberia and south to the Altai, and along the coast of Alaska to “Brittisch Nord-Amerika.” In 1916 its range, as understood by the monographer of the genus, Green- man,! was essentially the same, he citing specimens from as far south as the Aleutian and Shumagin Islands and “Mucklung River”’ (sup- posed to be in British Columbia). At the same time Senecio Fernaldi Greenman,” from a spur of the Long Range in Newfoundland, was described as a new species. The type of the latter was a single rather exceptional individual,* but it showed characters in the foliage and in its lack of ligules which seemed to separate it from the ordinarily radiate S. resedifolius. Subsequently, however, the Newfoundland plant (figs. 1% and 14) has been twice collected in quantity at the type- locality, in July, 1914 (Fernald & St. John, no. 10,873) and in July, 1921 (Mackenzie & Griscom, no. 10,483). A few individuals of these later collections are good matches for the type; but of the 47 plants of these two collections preserved in the Gray Herbarium all but these half-dozen individuals show such departures in the toothing and shape of the basal leaves, development and cutting of the cauline leaves, and development of ligules that, by those who have not seen them growing in one area of limestone shingle, and who might depend too reliantly upon the key-characters used by Greenman, “heads iscoid,”” an Lower leaves sharply dentate,” they titghe readily be mistaken for several species. The basal leaves may be sharply dentate, rounded-dentate or shallowly undulate, oblanceo- late, rhombic, elliptic, orbicular or reniform, and at base from nar- rowly cuneate to cordate. e middle and upper cauline leaves likewise vary from linear- or lance-attenuate to deltoid and from entire to pinnatifid; and the heads are either discoid or radiate. S: Fernaldit, growing as it does upon a dry exposed limestone tableland without shelter from wind and brilliant light, is a dwarf plant with strong tendency to purple coloring in leaves and involucre, and in all these features, as well as in its disk-flowers, the individuals with crenate-dentate basal leaves seem quite inseparable from Chamisso’s original material from the west side of Bering Strait of S. resedifolius, 1Greenman, Ann. Mo. Bot. Gard. iii. 99 (1916). 2 Greenman, is c. 90 (1916). 3 “but after y t back the b cn order o reach she stent before dak Pp g 1 to my very +2 Sd tnihnatl +h feet. TI p a ae Ruovora, xiii. 131 (1911). tinted 1924] Fernald,—Some Senecios of Eastern Quebec 115 while others are inseparable in foliage, involucre and disk-corollas from material collected by Charles Wright on Arakamtchetchene Island in Bering Straits or by Blaisdell at Cape Nome; while taller and looser-growing individuals are almost identical with Weinmann’s specimens from the Kuskokwim Valley in Alaska. The only difference I can make out is that the material from about Bering Sea usually has well developed rays; the Newfoundland plant is usually discoid, though occasionally with short ligules; but the ligules of the Chamisso material are nearly as short as in the radiate plants from Newfound- land, while S. resedifolius, var. columbiensis Gray is a lax and luxu- riant shade-form separated because of its rayless heads. I am, therefore, unable to find any stable characters by which S. Fernaldii can be maintained as specifically distinct from S. resedifolius. During the summer of 1923, the party! exploring the schistose mountains centering on Mt. Logan in Matane County, Quebec, found a very handsome Senecio with large commonly solitary and radiate heads. This plant (figs. 14, 1? and 1a), which abounds in turfy chimneys and upon cliffs and talus at altitudes of 850-1150 m. on Mts. Fortin, Logan and Pembroke, and which in its well developed green leaves (basal leaves up to 3 cm. broad, cauline leaves in luxuriant plants up to 8 cm. long and 3.5 cm. broad), tall stature (often 2-3 dm.) and showy elongate ligules, did not suggest in the field the little purplish plant of the dry barrens of western Newfoundland already discussed; but at one station, the dry upper talus and cliffs at the head of Hanging Valley on Mt. Pembroke, plants clearly conspecific with the large-leaved specimens of the moister chimneys had the low stature, reduced foliage and occasionally the discoid heads of the Newfoundland plant. Altogether the plant of the Shickshock Mts. shows almost every conceivable variation, even for so protean a group as Senecio, and in this series collected at three closely adjacent stations all the recorded variations of S. resedifolius can be matched. The variation of S. resedifolius which was described by Ledebour from the Altai as Cineraria lyrata is so much taller and has so much better-developed basal leaves than Chamisso’s plant that it has been treated by DeCandolle, Herder and others as a distinct variety, var. lyratus (Ledeb.) DC., supposed to be confined to the Altai and Baikal region; but some of the individuals from the Shickshock Mts. are 1J. F. Collins, C. W. Dodge, M. L. Fernald, Ludlow Griscom, K. K. Mackenzie, _ $. Pease and L. B. Smith. 116 Rhodora [JUNE scarcely distinguishable from authentic Altai and Baikal material, except in their slightly shorter and broader rays. Others again closely match in their angulate-toothed leaves Tiling’s original specimens from Ajan (on the west side of Ochotsk Sea) of var. multicaulis Herder; while others have the cauline leaves even more developed than in the type of var. columbiensis. In the Newfoundland plant the unexpanded disk-corollas and the outsides of the lobes are deep orange, in the Shickshock plant usually orange-yellow; but in the latter region a single individual was found with both disk- and ray-flowers decidedly reddish. And, as already noted, in the plants of Newfoundland the heads are commonly discoid but sometimes radiate, while in the Shickshock area they are com- monly radiate but occasionally discoid. Exactly the same variations of the heads are found in Siberia, Ledebour explicitly saying in his description: “Capitula majora v. minora, saepissime radiata nec nisi rarissime discoidea. Ligulae plerumque aureae variant flavae et licet rarissime rubicundae.’! In brief, Senecio resedifolius, long supposed to be restricted to the region from Alaska across Siberia, is now known in two extremely isolated regions, the Shickshock Mountains of Gaspé and the Long Range of Newfoundland, where in its numerous modifications it not only parallels but seems to exceed the Alaskan and Siberian plant. Miss Brackett’s drawings (fig. 1) show only a fcw of the variations in eastern America. SENECIO PAUCIFLORUS AND. S. INDECORUS.—Two of the most characteristic species of Senecio on the Gaspé Peninsula are strikingly different plants which, in his M. onograph of the North and Central American Species of the Genus Senecio, are treated by Greenman? as a single species, S. pauciflorus Pursh. One of the plants, true S. pauciflorus (fig. 2), is in eastern Quebec confined to the alpine and subalpine chimneys and meadows of the Shickshock Mts. and to the calcareous terraces bordering the Straits of Belle Isle. North of the Straits of Belle Isle it is characteristic of the mountainous region from Cape Mugford to the northern limits of Labrador, while in western America it extends from near the Arctic Circle in Mackenzie, Yukon and Alaska to the mountains of Wyoming and the Yosemite region of California. In the field S. pauciflorus is at once distinguished by its thick and fleshy foliage, the few cauline leaves with blunt teeth 1 Ledebour, Fl. Ross. ii. 632 (1844-46). 3 *Greenm. Ann. Mo. Bot. Gard. iii, 91 (1916). 1924] Fernald,—Some Senecios of Eastern Quebec 117 and lobes; by its comparatively low stature (1 to rarely 5.5 dm.), few (1-6, more rarely -11) broadly. campanulate heads with usually purple involucres, and disk-corollas with orange-red or deep-red lobes. The other species (fig. 3) is the plant figured and described in Britton & Brown, Illustrated Flora, iii. 479, fig. 4042, as S. discoideus.! This plant should, however, be called S. indecorus Greene, since Hooker’s S. aureus, 8 discoideus, upon which S. discoideus rests, proves to be S..pauciflorus.. Unlike S. pauciflorus, S. indecorus is not an arctic-alpine plant, but it occurs in the Canadian forest at low levels, from Gaspé County, Quebec to the Lake Superior region and at comparatively low altitudes from southern British Columbia to Idaho and northeastern California. In Quebec, where I have been familiar with the two species for twenty years, S. indecorus occurs in thickets and woodland-swamps or on talus and cool ledges of limestone; and in the field it is at once distinguished from the alpine and subalpine S, pauciflorus by its thin or membranaceous basa. leaves sharply lacerate-pinnatifid and rather numerous vauline leaves, tall stature (up to 9 dm.), more numerous (6-40) slenderly cylindric- urceolate heads with green or merely purple-tipped involucres and yellow flowers. In S. indecorus the filiforsi tube of the corolla is 3.6-6 mm. long, in S. pauci, only 3-4 mm.; and in the former the slender, strongly costate achenes (fig. 3d) are 2-3 mm. long, the plumper, less promin- ently ribbed and darker achenes of S. pauciflorus (fig. 2d) being 3-3.5 mm. long, When the heads are quite mature the bracts of the involucre of S. pauciflorus (fig. 2c) spread or become only slightly reflexed, exposing a smooth and nearly flat denuded receptacle (fig. 2b), but in S. indecorus the mature and narrower bracts (fig. 3c) become strictly reflexed and the convex receptacle (fig. 3b) is con- spicuously alveolate, the walls of the pits very thin and jagged or almost fimbriate Although the esas of S. indecorus are normally discoid, they, like those of S. pauciflorus, are sometimes radiate and plants of both forms sometimes occur in the same colonies. These radiate. individuals - discoideus Hook. ex Torr. & Gray, FI. ii. 442 (1843) is usually cited as S. dis- caiea (Hook.) Britton. A s (1898), but the eombination was actually published and eae cemratane* “radiate,” is gales unsatisfactory in Senecio. In end ptrlcapnges ie dn” he ordinarily r ate S. aureus. Li, S. Robbin. Oakes, S. pauperculus Michx., S. obovatus Muhl. ots resedifolius Less., in to those above discussed, are occasionally discoid. vote 118 Rhodora [JUNE which occur in Quebec side-by-side with the commoner discoid form of S. indecorus are quite inseparable from the type from Idaho and other authentic specimens from British Columbia of S. Burkei Green- man, Ott. Nat. xxv. 114 (1911) and Ann. Mo. Bot. Gard. ii. 626, t. 20, fig. 1 (1915); and it is significant that many of the British Colum- bian localities for discoid S. indecorus are identical with those cited by Greenman for S. Burkei. In the development of its foliage S. indecorus varies as much as S. pauciflorus, 8. aureus, S. pauperculus and S. resedifolius, the plants showing quick responses to degrees of moisture, exposure and soil- fertility. In different habitats of the same region may be collected specimens with greatly reduced cauline leaves and others with them remarkably dilated. Plants of the latter extreme from Rimouski County, Quebec, are quite inseparable from S. idahoensis Rydberg, Bull. Torr. Bot. Cl. xxvii. 183, t. 6, fig. 5 (1900), which Greenman recognizes as a species distinct from S. pauciflorus. S. idahoensis has all the technical characters of S. indecorus,—membranaceous basal leaves, laccrate-pinnatifid cauline ones, tall stature, numerous heads, slender green involucre, characteristic corolla and achene, and the alveolate receptacle——and the type-number is, except for lack of ligules, a good match for the type of S. Burkei. None of the characters of disk-corolla, achene and receptacle are mentioned by those who maintain the distinctness of S. idahoensis; but the plant is considered remarkable in having the “stem branched from near the base” and “large broad stem-leaves” (Greenman, Ann. Mo. Bot. Gard. iii. 96). The foliage of the type-number, Sandberg, MacDougal & Heller, no. 803, is, as already implied, closely matched by that of luxuriant individuals from eastern Quebec and by the type of S. Burkei; but, as shown by the sheet of the type-number of S. idahoensis in the Gray Herbarium (one of the sheets cited by Greenman), the basal branching of this Idaho specimen is due to the fact that the primary stem had been broken or bitten off, causing the development of few-headed basal branches. The type of Senecio discoideus is Hooker’s S. aureus, § discoideus from Fort Franklin in Mackenzie. An excellent photograph of Hooker’s type is preserved at the Field Museum of Natural History and I am under obligation to Messrs D. C. Davies and J. Francis Macbride for an opportunity to examine it. It is certainly not the plant (S. indecorus) which has been passing as S. discoideus in eastern 1924] Fernald,—Some Senecios of Eastern Quebec 119 America. In the lobing of its leaves the iype cf S. discoideus is like S. pauciflorus, but it has broadly deltoid instead of oblong cauline leaves and much more numerous heads than in the typical S. pauci- florus of Labrador and Quebec. Material from British Columbia closely approaches the Fort Franklin plant, however, and S. discoideus is best treated as an extreme development of S. pauciflorus. Greene’s description of S. indecorus at once suggests the plant which has been passing as S. discoideus and a beautiful photograph of the type which has been generously furnished me by Professor George W. Albertson of the University of Notre Dame settles beyond question this identity. Briefly summarized the main characters and the bibliography of S. pauciflorus and S. indecorus are stated below, and a few character- istic specimens are — S. PAUCIFLORUS Pur: eyes ii. 529 (1814); Greenman, iscoideus Ann. Mo. Bot. Gard. iii. ee in part. S. aureus, Hook. FI. Bor. hs, i. 333 (1834). S. Sicilian Hook. ex Torr. & Britton in Britton & Brown, IIl. Fl. iii. 480 (1898).—Stems rather stout, 1-5.5 dm. high: leaves thick and fleshy; the basal long-petioled, elliptic to reniform, cuneate to cordate at base, coarsely dentate; the median and upper sessile, 2-5 below the inflorescence, lanceolate, —_— or oblanceolate, with mostly obtuse lobes and coarse teeth: heads 1-6, rarely -11, broadly campanulate, usually discoid ; —— cres usually purple; the blunt to acute bracts 1-2 mm. broad: corollas with filiform tube 3-4 mm. long; the lobes orange-red to deep-ted achenes dark red-brown, plump, 3-3.5 mm. long: mature involucre rotate or oosely reflexed; denuded receptacle smooth and nearly flat.—Calcareous meadows, alpine slopes and wet rocks, northern Labrador to Matane County, Quebec; Mackenzie, Yukon and Alaska to Wyoming and California. The following are in the Gray Herbarium. Lasrapor: hills back of Okkak, Moravian Brothers, August, 1911, F. C. Hinckley; Hebron, Moravian Brothers; Rama, July, 1894, J. D. Sornborger, August, 20-24, 1897, Sornborger, no. 67; Mugford, August 14, 1900, Ears barrens, Forteau, July 30, 1910, Fernald & Wie- QueEBEc: shore of Seal Lake, ngava, August 3 1396, Spread- borough, no. 14,387 in part; little swales near summit of limy ridge, Pointe seas Brest, Saguenay Co., July 29, 1915, St. John, no. 90,779; alpine and subalpine meadows and caleareous wa itude 1000- 1200 m., Tabletop Mts., Gaspé Co., many collections in August of 120 ? Rhodora [JUNE 1906 and 1923, Fernald & Collins, nos. 261, 263, Fernald, coe & Smith, nos. 26,109, 26,110; wet hornblende-schist at base of Big imney, north slope of Mt. Mattaouisse (called o ane a at schistose talus and wet shelves at base (altitude 400-600 m.) of Big Chimney, Mt. Mattaouisse, July 10, 1923, Dodge, Griscom & Pease, no. 26,106; moist rocks and turfy chimneys (alt. ate 800-1050 m.), Pease Basin, between Mts. Logan and Pembroke, Matane Co, July 138, 1923, Pease & Smith, no. 26,107; out bushy ta'us and mouths of chimneys at about 900-1125 m. alt., Hanging Valley, Mt. Pembroke, July 16, 1923, Griscom & Sais no. 26,108, August 24, 1923, Fernald & Smih, no. 26,111. Mackenzie: Fort Franklin, Richardson (tracing from photograph of S. aureus 8 discoideus Hook.). ALBERTA: Flatts, junction of North Fork and West Branch of Saskatchewan, June 21, 1908, S. Brown, no. 1008; Malique Lake, July 8 and 18, 1908, Brown, nos. 1173, 1247: Blind Valley aa Lakes, alt. 7000-8000 ft., July 6, 1906, Brown, n no. 414, British COLUMBIA: below Pipestone summit, ; July 6, 1904, J. Macoun, no. 65,018; _ Spillamacheen Valley, alt 6500 ft., August 3, 1904, C. H. Shaw 440; Cornwall Hills, July 28,—J. "McEvoy, no. 5100; Alberni, Fa couver Island, June 27, 1907, Rosendahl, no. 1971. Wyomre: small willows, Middle Ten Sleep Creek, Big Horn Co., August 1, 1901, yin ae no. 2 Little Goose Cafion, Sheridan oe July 1, 1901, A. Nelson 2383. Ipano: Grand Saline (Salt River?), "Burke. Casavonsta: 3 pres ‘rae 1893, J. B. Lembert (S. Lemberti Greene); Soda Springs, Tuolumne Ce July, 1894, Lembert (distri- buted by a as C 71). S. orus Greene, Fl. Francise. 470 (1897). 8S. discoideus Hook. ex. Torr. & Gr. Fl. N. A. ii. 442 (1843) as to plant described; Britton in Britton & Brown, IIl. FL. ii. 479, fig. 4042; not S. aureus 8 discoideus Hook. Fl. Bor.-Am. i. 333 (1834). S. idahoensis Rydb. Bull. Torr. Bot. Cl. xxvii. 183, t. 6, fig. 5 (1900); Greenman, Monogr Selieiia, © Teil, 24 (1901), in Engler, Bot, Jahrb. xxxii. 20 (1902), Aen: Mo. Bot. cheats iii. 96 (1916). S. Burkei Greenman, Ott. Nat. xxv. 114 (1911), Ann. Mo. Bot. Gard. ii. 626, t. 20, fig. i. (1915), ibid, iil. 136 (1916). —Slender, 2-9 dm. high: leaves membranaceous; the basal ee ee, oblong, elipic or ovate, cuneate to edits at base, sharply or often incisely toothed or sometimes pinnatifid; the middle and appr usually ae. 4-11 below the inflorescence, oblong, lanceolate or oblanceolate, incisely and variously cut into sharp teeth or lobes; the uppermost leaves greatly reduced or obso- lete; heads 6-40, slenderly cylindric-urceolate (when fresh), dis- coid or radiate: involucres usually green or merely with purple tips: the acute bracts 0.5-1 mm. broad: filiform tube of corolla 3.6-6 mm. long; lobes yellow: achenes drab or gray-brown, slender, strongly _costate, 2-3 mm. long: mature involucre strictly reflexed: denuded receptacle conve x, conspicuously alveolate; ee ey 1924] Fernald,—Some Senecios of Eastern Quebec 121 thin and jagged or almost fimbriate—Rich thickets, swales and cal- careous rock, Gaspé Co., Quebec to the Lake Superior Region; British Columbia to Idaho and northern California. The following are in the Gray Herbarium. Quesec: sandy alluvium of Riviare Ste. Anne July 17, 1922, Fernald & Pease, no. 25,340; openings in Arbor Vitae swamps, Carleton, Bonaventure Co., July 24 and 27, 1904, Collins, Fernald & Pease; edge of Arbor Vitae swamp, July 18-21, 1905, Wilhams, Collins & Fernald; dry spruce woods at base of limestone- conglomerate cliffs, Bic, Rimouski Co., July 16, 1904, Collins & Fernald; meadows and swamps in slaty region south and east of Bic, July 17, 1907, Fernald & Collins, no. 1202 (some heads radiate); dry limestone-conglomerate ledges, south of Bic, July 28, 1907, Fernald & Collins, no. 1203; pastures, Riviére du Loup, Temiscouata Co., July, 1913, Victorin, no. 40; Roberval, Lake St. John, July 15, 1892, G. G. Kennedy; sur les caleaires paléozoiques de l’Ouiatchouan inféri- eur, Val-Jalbert, Lake St. John, July 22, 1921, Victorin, no. 15,436; Spirit Lake (Abitibi), Pontiac Co., July 9, 1918, Victorin, no. 8517. Micuican: Champion, July, 1889, E. J. Hill; Keweenaw Peninsula, 1863, Robbins: Keweenaw Co., July, 1890, Farwell, no. 776. Cotumsia: Kicking Horse Valley, alt. 4000 ft., July 15, 1906, S. Brown, no. 487a; river-margins, Silver City, August 7, 1885, J. Macoun; Griffin Lake, July 6, 1889, J. Macoun; Carbonate, altitude 2700 ft., July 7, 1904, Heacock, no. 185; Wapta Lake, August 4, 1904, J. Macoun, nos. 65,015, 65,020; Skagit Valley, altitude 2500-3000 ft., various dates in July and August, 1905, J. M. Macoun, nos. 69,358- 65,360; Chilliwack Valley, altitude 3000-4500 ft., various dates, 1901, J. M. Macoun, nos. 26,682—26,685; Sophie Mt., altitude 5500 ft., July 17 and 20, 1902, J. M. Macoun, nos. 64,990, 64,991; Creston, Kootenay, June 22, 1914, J. K. Henry, no. 4; gravel-bed of mountain- stream, Armstrong, July 4,—, E. Wilson, no. 247. IpAno: meadows at Granite Station, Kootenai Co., July 30, 1922, Sandberg, Mac- Dougal & Heller, no. 803 (type-number of S. idahoensis Rydb.); Forks of St. Mary’s River, altitude 1100 m., July 3, 1895, Leiberg, no. 1158; Grand Saline (? Salt River, tributary to the Snake), August 16, ——, J. Burke (type of S. Burkei). Wasnincton: Big Meadows, Stevens or Spokane Co., August 6, 1902, Kreager, no. 428; damp forest, ewman Lake, July 9, 1916, Suksdorf, no. 8797; Mt. Constitution, Orcas I., San Juan Islands, August, es! Henderson, satire t 1, 1917, Zeller, no. 1252. ALIFORNIA: wet meadows, i fonk, Lassen Co., July 9, 1894, Baker & Nutting (photograph of tyPE of S. indecorus). Most of th t pecimens : : the western as S. pauciflorus, 8S. Burkei and S. idahoensis. + + etic aa discoideus; 122 Rhodora [JUNE EXPLANATION OF PLATE 144. Fig. 11, Senecio resedifolius x 1, from Mt. Pembroke, Matane Co., Quebec, Fernald & _ no. 26, = 1?, from Mt. Fo ortin, Matane Co., Quebe ec, Fer- nald & Sm 6, 101; 13 and 14, from Table Mt., Port & Port Bay, New- foundland iS Fernaliit), Fernald & St. John, no. 10,873; 1a, young plant X 1 rom Mt. Fortin, Quebec, Fernald, Griscom & Mackenzie, no. 2! 9; 1b, de- rier receptacle x 4. Fig. 2, 8. pauciflorus < —) 2b, denuded receptacle 4; 2c, ee involucre x wt 2d, achene X< ig. >! gee half of small V. NEW OR RECENTLY pense PLANTS OF EASTERN AMERICA. Carex Hostrana DC., var. laurentiana (Fernald & ts aa Fernald & W iegand, n. comb. C. fulvescens Mackenzie, Bull. Tor Bot. Cl. xxxvii. 239 (1910). C. yet aire cia var. se cetinns Fernald & Wiegand, Ruopora, xiii. 130 (1911). C. Hostiana DC. Cat. Pl. Hort. Bot. Monsp. 88 (1813) was well described and the plate of C. fulvua Host, not Goodenough, was cited: Although Kiikenthal retains for the species the name C. Hornschuchi- ana Hoppe (1824) he definitely cites the perfectly described C. Hostiana (as C. Hosteana) as a synon We have again studied the plant of Rewfoandisaid and Anticosti and can find no characters not shown by the European C. Hostiana, merely a tendency to slightly larger size of the parts; and several European specimens so closely approach var. laurentiana that it might be urged with good reason that the variety is scarcely worthy of separation. Sattx Bepsrana Sargent, var. luxurians (Fernald), n. comb. S. rostrata, var. luxurians Fernald, Ruopora, ix. 223 (1907). S. Beb- Cony. forma luxurians (Fernald) Schneider. Journ. Arn. Arb. iii. 75 Schneider shows! that Salix rostrata Thuill. (1797) is not a perfect synonym of S. repens L. but that it is identified with two generally recognized varieties of that species and by those who would treat eitler of these variations as species the name S. rcstrata Thuill. would have to be used. Under these circumstances the name S. rostrata Richardson (1823) for the characteristic American species must give way to S. Bebbiana Sargent (1896). Schneider considers var. luxurians as “nothing but a forma luxurians,” which “should be looked for in other localities of the type.” In the vast accumulation of American 1 Schneider, Journ. Arn. Arb. ii. 68 (1920). Plate 144 Rhodora A. E. Brackett del. Fic. 1,SENECIO RESEDIFOLIUS. Fic.2, S.pauctFLorus. F1G.3,S.INDECORUS. wied : és : + " | * A | | peachy SONS Pi 1924] Fernald,—New Plants of Eastern America 123 Salix which Schneider studied, however, he saw var. luxurians only from the “banks of the St. Lawrence from Rimouski Co. to Gaspé Co.,’”’ Quebec. There it is the characteristic extreme of S.Bebbiana and on account of its very long capsules (9-12 mm.) and long pedicels (5-8.5 mm.), its plane rather than rugose leaves and its great stature it might almost be considered a species. In S. Bebbiana (typical) arid S. Bebbiana, var. perrostrata (Rydb.) Schneider, the mature capsules are usually 5-8 mm. long though in very exceptional speci- mens they reach 1 cm. and the mature pedicels range from 2-6 mm. in — BIANA var. capreifolia (Fernald), n. comb. S. rostrata, var. cuuteitolin Fernald, Ruopora, xvi. 177 (1914).. SaLrx hebecarpa aieoryy n. comb. S. fuscescens, var. hebecarpa Fernald, Ruopora ix. 224 (1907). When S. hebecarpa was originally published it was taken for an extreme of the Alaskan S. fuscescens Anderss. with pubescent capsules and the shrub growing with it but having the capsules glabrous: (though the pedicels pubescent) was identified with S. fuscescens. Collections made on Mt. Albert in July, 1923 by Messrs. Griscom, Mackenzie, Smith and myself at the type station for S. fuscescens, var. hebecarpa show that the pubescence of the fruit is by no means constant but in all other characters the Mt. Albert material forms a consistent series; and a second sheet of true Alaskan S. fuscescens which has recently come to hand shows that it and the Mt. Albert shrub have quite different leaf-venation, and that the Alaskan shrub has darker scales and smaller capsules. S. hebecarpa stands between S. fuscescens and S. pedicellaris Pursh. From the latter it is separated by its more obovate leaves, the more fubescent and darker scales of the ament, the much shorter, thicker and pubescent pedicels (about equaling to twice as long as the scales), the long nectary and the defi- nite style. The following specimens, all from alpine bogs on Mt. Albert, Gaspé County, Quebec, are representative, The form with glabrous or glabrescent capsule is marked “gl.” August 8-15, 1905, Fernald & Collins (gl); July 21, 1906, Fern ald & Collins, no. 206 (gl), no. 207 (TYPE in Gray Herb.); July 24, 1923, Fernald, Griscom, Mackenzie & Smith, no. 25,686 (gl). OpLOPANAX HORRIDUM (Sm.) Miquel. The Devil’s Club is known -in our eastern manuals as Fatsia horrida (Sm.) B. & H. or as Echino- panazx horridum (Sm.) Decaisne & Planchon. It seems to be a mono- 124 Rhodora [JUNE typic genus (though the Japanese plant is sometimes separated) clearly distinguished from Fatsia, and by the International Rules of Botanical Nomenclature its name should be Oplopanazx horridum (Sm.) Miquel. Echinopanax Decaisne & Planchon, Rev. Hort. Sér. 4, ili. 105 (1854) was started as a genus merely by the statement, “Un autre type trés remarquable (Echinopanax Nob.) est fourni par le Panax horridum.”’ Absolutely no differential character was given by which any one else could tell on what points Echinopanazx rested as a genus; that highly important detail being left for others to struggle with, they, if successful, giving the credit for the solution to the authors who had not taken sufficient pains themselves to point out the differential characters. The International Rules wisely exclude these names, the authors of which, to use the slang of the day, “pass the buck,”’ stating (Art. 38) that “The mere indication of species as belonging to a new genus or of genera as belonging to a higher group, does not allow us to accept the genus or group in question as charac- -terized and effectively published.””, The American Rules, however, accept such irresponsible publication and by that code Echinopanax is a properly published genus. So far as a prolonged search has re- vealed, the first satisfactorily published generic name for the Devil’s Club is Oplopanazx Miquel, Ann. Mus. Bot. Lugd.-Batav. i. 16 (1863), which was based upon Torrey & Gray’s Panaz § Oplopanax, a section set off by them with a real characterization. Nelson & Macbride, Bot. Gaz. Ixi. 45 (1916), have revived the name “epg and have made the combination R. horridum (Sm.) Macbr.; but a moment’s examination of the original publi- sce would has shown that Ricinophyllum was merely indicated in synonymy as an herbarium-name found on an old sheet of Pana horridum Sm. The name Ricinophyllum, therefore, has no nomen- clatorial status. HACKELIA DEFLEXA (Willd.) Opiz, var. americana (Gray) ina & Johnston, n. comb. Echinospermum deflexum, var. americanu Gray, Proc. Am. Acad. xvii. 224 (1882). Lappula deflexa, var. ameri- cana (Gray) Greene, Pittonia, ii. 183 (1891). L. americana Rydb. Bull. Torr. Bot. Cl. xxiv. 294 (1897). When the genus Hackelia was recently taken up! the American plant was merged with the European since the character relied upon _ by Gray in separating it, namely, the presence of a few prickles on 1 Johnston, Contrib. oa n. 8. He, Ixviii. 43-48 (1923). 1924] Fernald,—New Plants of Eastern America 125 the back of the fruit, was found to be very inconstant. Recent examin- ation of the American material, however, shows that it differs from the European in having the leaves elliptic-lanceolate or narrowly ovate and acuminate, the leaves of the European type being spatulate- oblanceolate. In the latter plant the back of the fruit is deltoid-ovate; in the American plant more definitely deltoid. Typical L. deflexra extends eastward into China while var. americana is present in Amur (Mazximowicz). AGOSERIS gaspensis, spec. nov., perennis; foliis lineari-lanceolatis flaccidis integris vel remote pinnato-incisis glabris basi ciliatis; seapo deinde folia superante apice villoso; involucri bracteis lanceolatis attenuatis 1.5-2 cm. longis, exterioribus viridibus apice recurvatis basi valde villoso-ciliatis, ciliis 1-2 mm. longis, bracteis interioribus scariosis valde purpureo-maculatis; floribus aeneis; pappo albo deinde 1.2-1.5 em. longo, setis minute serrulatis: achenio lineari-fusiformi 7-9 mm. longo glabro, rostro 5-7 mm. longo firmo vix filiformi glabro. —Tabletop Mountains, Gaspé County, QuEBEc: dry stony subalpine meadows, at 1100-1200 m. altitude, east of Lac Perrée, August 14, 1923, M. L. Fernald, C. W. Dodge, & L. B. Smith, no. 26,119; alpine and subalpine meadows and brooksides, northern and northeastern slopes (alt. 950-1300 m.) of Mt. McNab, August 7, 1923, Fernald, Dodge & Smith, no. 26,120 (rype in Gray Herb.); subalpine meadows on southern slope (alt. about 1200 m.) of Mt. Au Clair, August 10, 1923, Fernald & Smith, no. 26,121. Strongly simulating Agoseris gracilens (Gray) Kuntze of the region from southern Alberta and British Columbia to Wyoming, Utah, Nevada and northern California. That species, however, as shown by Gray’s criginal material of Troximon gracilens, has the body o the achene only 5-7.5 mm. long and hispid at summit (in A. gaspensis 7-9 mm. long and glabrous) ; the delicate filiform beak 6-8 mm. long (the stouter and firm beak of A. gaspensis 5-7 mm. long); and the mature pappus only 8-10 mm. long (in A. gaspensis 1.2-1.5 cm. long). Furthermore, in A. gracilens the disk at the tip of the beak is about 0.5 mm. in diameter, in A. gaspensis nearly twice as broad. A. gaspensis is of peculiar interest because it is the first member of this predominantly cordilleran genus to be found in the Shick- shock Mountains, a region in which the cordilleran relationship is very pronounced and already well known through such plants as Polystichum mohrioides, var. scopulinum (D. C. Eaton) Fernald (discussed on p. 89), Salix brachycarpa Nutt., Dryas Drummondii Richardson, Lonicera involucrata (Richardson) Banks, and hundreds of others. 126 -Rhodora [JUNE The genus which was long called in America and which is maintained by Dalla-Torre & Harms as Trorimon Nutt. should be known as Agoseris Raf. The question was cleared by Greene in 1891,! but more recently Hall,” who has been followed by other western botanists has urged the retention of Troximon in the following words. “The reinstatement of the genus Agoseris Raf. (1817) has been proposed for those species of Troximon in which the achenes are beaked. But the two groups are best received into one genus, being connected by the thick-beaked 7. glaucum. Troximon was first used as a generic name by Gaertner (1791) but since his genus is not sustained, we may properly write Troximon Nutt. (1813) as the name of the present group.” In Hall’s argument several fundamental facts seem to have been overlooked. In 1791 Gaertner* published Troximon as a genus based upon three Linnean species of T'ragopogon, Trag. Dandelion, T. vir- ginicum and T.lanatum. The three species were, according to Index Kewensis, formally named Troximon Dandelion, T. virginicum and T. lanatum by F. W. Schmidt in 1795 and they were certainly de- scribed under these binomials by Persoon in 1807. The first two are species of Krigia Schreber (1791), belonging to the subgenus Cynthia (D. Don) Gray, which is often maintained as a distinct genus Cynthia D. Don (1829). The third species, Tox. lanatum, is generally referred to Scorzonera L. (1753). By whichever principle we reasor, whether we typify Troximon Gaertn. by its first species, by the greater number of species or by the exclusion from it of the species (7. lanatum) which belongs to the earlier-published Scorzonera, we arrive at the same conclusion, that Troximon Gaertn. (1791) = Cynrnia D. Don (1829) and should be used by those who maintain Cynthia as generically distinct from Krigia Schreb. (1791). Whether Krigia (1791) has priority over Troximon (1791) I have as yet been unable to determine. Hall, apparently considering Troximon Gaertner as a “genus - . . not sustained” would retain “Troximon Nutt. (1813).” But what was Troxrimon Nutt. (1813)? The reference ordinarily given is to Nutt. in Fraser’s Cat. nos. 83 and 84. Fraser’s Catalogue 1Greene, Pittonia, ii. 176 (1891). ? Hall, Univ. Cal. Pub. Bot. iii. No. 1: 276 (1907). panera Fruct. ii. 360 (1791). D. Don, Edinb. New Phil. Journ. vi. 309 (1829), 1924] Fernald,—New Plants of Eastern America 127 was merely a list of names and the complete treatment under nos. 83 and 84 was as follows: “83* Troximon ciliatum. t M. 84* —————slaucum. ¢ M.” This certainly does not constitute satisfactory publication of a genus. The next reference commonly given is to Pursh, Fl. Am. Sept. ii. 505 (1814). But Pursh gave absolutely no generic diagnosis and ascribed the genus without question to Gaertner and to Persoon. That Pursh had no thought of setting up a new genus is perfectiy evident from the fact that he has two species, 7’. glaucum which is an Agoseris, and 7’. virginicum which was one of the original species of Troximon Gaertn. Nor did Nuttall in his Genera, ii. 127 (1818) indicate any intent to set off a new genus 7'roximon as distinct from Gaertner’s genus; for he ascribed Troximon directly to Gaertner and he so far accepted Gaertner’s definition as to include the oriental I’. lanatum which belongs to the genus Scorzonera. In 1817 Rafinesque properly published Agoseris' and again, in 1819,’ he gave a good characterization of the genus. There is no question, then, that Acosrrts Raf. (1817) is the correct name for the genus which still passes erroneously as Troximon “Nutt.,” and that Troximon Gaertn. (1791) is the correct name for Cynthia D. Don (1829). ' Raf. Fl. Ludov. 58 (1817). ? Raf. Journ. de Phys. Ixxxix. 100 (1819). snd os ik * * = ia Mie i: atl tigre P bad OEE . ie # : ee ee ay i ea at “ y ah ee CONTRIBUTIONS FROM THE GRAY HERBARIUM sen HARVARD UNIVERSITY. New Series.—LX XIII. _ “rp of sa ISSUED SEP 24 1924 I. Records preliminary to a general aprons of the Eupato- reeae,—lV. By B. L. Roprnson . Il. New Plants clas by Mrs. Richard C. Curtis 3 in Porc guese West Africa. By I. M. Jounston III. On some South American Proteaceae. By I. M. JouNSTON 41 IV. Studies in the Boraginaceae,—III. By I. M. Jonnston. 42 1. The Old World Genera of the Boraginoideae . 2. Notes on miscellaneous American Boraginaceae . . 73 THE GRAY HERBARIUM OF » HARVARD UNIVERSITY CAMBRIDGE, ps — U.S.A. I. RECORDS PRELIMINARY TO A GENERAL TREATMENT OF THE EUPATORIEAE,—IV. By B. L. Rorprnson. 4 y SINCE publishing about a year ago the preceding installment of similar records, the writer has had opportunity to examine several extensive series of specimens throwing further light upon the classi- fication of the Eupatorium tribe. The chief sources of this material have been as follows: 1) A third loan of more than a hundred speci- mens from the Museum of Natural History in Paris, including a few additional plants of the very interesting Brazilian collection of Auguste de Saint-Hilaire and an admirable suite of Triana plants from Colombia, the latter being unusual from the fact that they bear the original detailed geographic and ecological data of collection. 2) The later gatherings of Dr. H. A. Gleason and Mr. J. S. De La Cruz in British Guiana. 3) The Compositae-Eupatorieae secured on the second expedition of the Field Museum of Natural History for the botanical exploration of Peru. These plants collected by Mr. J. Francis Macbride had the same high excellence as those of the preceding expedition and included many novelties of unusual interest. 4) The Compositae-Eupatorieae of Prof. Hitchcock’s expedition to Ecuador, admirably supplementing the earlier collecting of Dr. J. N. Rose and including material from some regions, such as the valley of the Pastaza River to the east of Mt. Chimborazo, scarcely if at all explored previously and yielding noteworthy novelties. 5) Mis- cellaneous Eupatorieae referred to the writer by the United States National Herbarium from Panama, Costa Rica (collections of Dr. W. R. Maxon and assistants), and Honduras (collection of Mr. M. A. Carleton); by the New York Botanical Garden from the West Indies; by the Missouri Botanical Garden from Panama and Central America (collections of Dr. J. M. and Mr. M. T. Greenman). 6) Previously unstudied portions of large loans from the Royal Gardens of Kew and from the University of Geneva. 7) A con- siderable series of excellent photographs obtained by purchase from the Botanical Maseum at Berlin and showing the precise identity of many species hitherto imperfectly known in America Such new data as have been obtained from bee b sources relative to the further classification of the Ewpatoricae may be put on record as follows: (§ Ophryochaeta) apricus, spec. nov., suffrutescens atinatn gracilis supra patenter ramosus; caulibus etl: ca. 3 4 ROBINSON mm. crassis dense pubescentibus pallide brunneis; internodiis 6-12 cm. longis; ramis curvatis; foliis oppositis ovatis vel ovato-oblongis acuminatis basi rotundatis vel subacutatis basin versus integris aliter serratis (dentibus usque ad 1 mm. altis et 4 mm. basi latis) supra parce puberulis subtus paullo pallidioribus breviter praecipue in nervis brunneo-pubescentibus supra basin pinnatim 5-7-nervatis inter nervos delicate reticulatis 3.5-4.5 em. longis 1.8-2.8 cm. latis membranaceis; petiolo ca. 8 mm. longo dense brunneo-pubescenti; paniculis terminalibus et lateralibus patenter ramosis et laxiusculis foliaceo-bracteatis 1-1.5 dm. longis 6-12 cm. diametro saepe ad apicem caulis in inflorescentiam magnam laxam compositam conjunctis; capitulis ca. 7-floris breviter pedicillatis ca. 4.5 mm. altis et 3.5 mm. diametro; involucri squamis ca. 7 longitudine subaequalibus oblongis apice rotundatis ciliolatis dorso glabris vel subglabris ca. 3 mm. longis et 1 mm. latis, intimis 1-2 angustioribus oblanceolatis; corollis albis ca. 3 mm. longis; tubo proprio 1.5 mm. longo glandulari-puberulo; faucibus paullo ampliatis ca. 1 mm. longis; dentibus limbi 5 deltoideis patentibus 0.5 mm. longis; achaeniis immaturis glabris ca. 1.3 mm. longis; pappi s@tis ca. 25 vix’ scabratis corolla paullo brevioribus. —PERU: sprawling in sunny montafia, about 1830 m. alt., July 17- 26, 1923, J. F. Macbride, no. 5204 (rypr, Field Mus.; tsorype, Gr.). This species most nearly approaches the Ecuadorian 0. Sodiroi Hieron. but has much more sharply and deeply toothed leaves, narrower and oblong rather than obovate involucral scales, and much smaller stigmas. O. (§ Ophryochaeta) Machridei, spec. nov., fruticosus gracilis ca. 1.5 m. altus oppositirameus; caulibus teretibus juventate crispe uberulis mox glabratis ca. 3 mm. crassis medullosis; medulla alba; internodiis 4-9 cm. longis; foliis oppositis breviter petiolatis oblongis longe acuminatis basi acutis supra mediam partem utroque obscure 3—4-serrato-dentatis (dentibus 0.5 mm. altis et 4 mm. inter se distan- tibus) vel subintegris plerisque deflexis 5-6 cm. longis 1.5—2 em. latis utrinque praecipue in costa nervisque paullo hirtellis supra basin 3-nervatis in aetate bullescentibus et subcoriaceis subtus vix pallidi- oribus; petiolo 3-5 mm. longo; paniculis et axillaribus et terminalibus foliaceo-bracteatis compositis sordide villosulis; capitulis breviter pedicellatis 5~7-floris in corymbis convexis subdensis gestis ca. 5.5 mm. altis; bracteolis lanceolatis vel spatulato-oblanceolatis vel linearibus 1-3 mm. longis; pedicellis 1-1.5 mm. longis; involucri squamis ca. 6 oblongis ciliolatis dorsaliter subglabris ca. 3.5 mm longis et 1 mm. latis apice rotundatis, extimis 4 basin versus cum nervis 2 calloso-incrassatis munitis; corollis tubulatis ca. 2.5 mm. TROPICAL AMERICAN EUPATORIEAE 5 longis sine faucibus distinctis; dentibus limbi patentibus ca. 0.7 mm. longis; styli ramis leviter clavatis; achaeniis ca. 1.5 mm. longis glabris; pappi setis ca. 20 corollam subaequantibus albidis vix scabratis. —Peru: slender shrub of sunny thicket-edges, Yanano, alt. about 1830 m., May 13-16, 1923; J. F. Macbride, no. 3717 (rypx, Field Mus.; tsoTyPe, Gr.). Falling into the group with the Bolivian 0. Cumingii (Sch.-Bip.) Benth. and 0. Kuntzei Hieron., this species is readily distinguished from both by its much larger aiid more open inflorescence and its subentire leaves of firmer texture. Eupatorium (§ Hebeclinium) beneolens, spec. nov., suffruticosum vel fruticosum 1-3 m. altum regulariter patenterque ramosum; caule tereti 7 mm. vel ultra crassitudine medulloso; medulla alba; indumento pallide brunneo copioso et patenter lanato-velutino vel breviori et subappresso-sericeo; internodiis 5-9 cm. vel ultra longitudine; ramis curvato-adscendentibus; foliis oblongis vel rhombeo-ovatis gradatim acuminatis basi acuta et apice exceptis serratis (dentibus 0.3-2 mm. altis 3-5 mm. basi latis apice callosis) firmiter membranaceis supra viridibus inconspicue villosis subtus pallidioribus subsericeo-pubes- centibus pinnativeniis 12-17 cm. longis 5-7.5 cm. latis; costa media densissime sericeo-velutina in facie inferiori exserta, venis lateralibus numerosis utroque ca. 12-19 prorsus curvatis; petiolo 1-3 cm. longo; panicula terminali pyramidali 13-19 cm. alta 17-27 cm. diametro oppositiramea; bracteolis linearibus; capitulis sessilibus ca. 6 mm. altis et 3.5-4.5 mm. diametro ca. 17-floris; involucri subcampanulati squamis ca. 20 ut dicitur roseis sed post exsiccationem brunnescentibus margine pallidis conspicue ciliatis, subuninerviis apice subacutis vel rotundatis 3—4-seriatim imbricatis, extimis ovatis multo brevioribus, intermediis oblongo-linearibus, intimis linearibus plerisque acutis in sicco patenter recurvantibus; disco convexo hispido-villoso; flosculis ut dicitur fragrantibus; corollis albis glabris ca. 4 mm. longis; tubo proprio 2.5 mm. longo; faucibus cylindratis 1.5 mm. longis vix amplia- tis; dentibus limbi deltoideis patentibus 0.5 mm. longis; antheris subliberis apice conspicue appendiculatis; achaeniis ca. 1.3 mm. longis glabris stramineis vel tardius nigrescentibus cum costis pallidis; pappi setis ca. 30 albis corollam subaequantibus.—PERv: a liana-like shrub with whitish flowers, Mufia, trail to Tambo de Vaca, June 27, 1923, alt. about 2440 m., J. F. Macbride, no. 4912 (rypr, Field Mus.; ISOTYPE, Gr.); a itragpling shrub 3 m. high with fragrant decom white flowers, at same place and altitude, June 5-7, 1923, J. F. Macbride, no. 4270 (Field Mus., Gr.). oe This species from its hairy receptacle is clearly of § Hebeclinium and differs much both in habit and technical features from FE. ne- 6 ROBINSON morosum Klatt and E. macrophyllum L., the only other species of this section thus far known from Peru. In E. beneolens two marked pubescence-forms are to be noted in the admirable series of specimens secured by Mr. Macbride, namely: Forma typicum, caule densissime patenterque velutino-tomentoso sublanato.—Distrib. and exsicc. as above. Forma ctenotum, forma nova, caule subsericeo-pubescenti, indu- mento denso quam apud formam typicam breviori et quasi pectinato subappresso.—PeErv: sunny brush-lands, a half shrub 1-3 m. high, uniformly branching; inflorescences terminal; involucral scales pink; fls. white, very fragrant, Mujfia, trail to Tambo de Vaca, about 2440 m. alt., June 27, 1923, J. F. Macbride, no. 4917 (ryPE, Field Mus.; ISOTYPE, Gr.); also slender-stemmed open-branched shrub about 3 m. high at edge of a bamboo thicket, alt. about 1982 m., Huacachi, near Muia, 20 May to 1 June, 1923, J. F. M acbride, no. 4085 (Field Mus., Hh eee E. (§ Conoclinium) bullatissimum, spec., nov., herbaceum perenne fortasse frutescens (basi ignota); caule tereti flexuoso incurvo-pu- beru mm. diametro albo-medulloso; foliis oppositis petiolatis ovatis acuminatis basi rotundatis plus minusve crenatis valde supra basin pinnatim 5-nervatis grosse reticulato-venosis utrinque in nervis et venis principibus puberulis supra atroviridibus valde rugosis subtus laete viridibus cum venis valde exsertis reticulatis (areolis plerisque 1-3 mm. diametro) textura membraceis 8-13 cm. longis 4.7-6 cm. latis; petiolo puberulo tereti ca. 12 mm. longo; appendicibus stipuli- formibus nullis; panicula terminali basi foliaceo-bracteata divaricatim oppositiramea ovoidea ca. 2 dm. ‘longa et crassa fulvido-tomentella; capitulis ca. 22-floris sessilibus ad apices ramorum subglomeratis ca. 5 mm. longis; involucri campanulati squamis regulariter gradatis ca. 35 ovatis apice rotundatis dorso viridibus glabris 2-3-striatis margine ciliolatis; receptaculo subhemisphaerico glabro; corollis graciliter tubulatis ca. 3 mm. longis ut videtur atropurpureis saltim post exsiccationem fuscescentibus sine faucibus distinctis; limbi dentibus brevibus plus minusve puberulis; achaeniis nigris crassius- culis ca. 1.3 mm. longis lucidulis in angulis paullo scabratis; pappi setis ca. 30 albis barbellatis corollas subaequantibus.—Ecuapor: Province Tungurahua: edge of woods, Valley of Pastaza River between Bafios and Cashurco, 8 hours east of Bafios, alt. 1300-1800 m., Sept. 25, 1923, Prof. A. S. Hitchcock, no. 21,814 (tyPE, Gr.). This species on account of its much elevated indeed nearly hemispherical re- ceptacle falls technically in the § Conoclinium, and is of course very different from either of the two Ecuadorian species heretofore known TROPICAL AMERICAN EUPATORIEAE 7 to belong to this section, namely, E. lamiifolium HBK. and E. Turc- zaninowit Robinson both of which have leaves which are densely gray- or white-tomentose beneath, longer petioles and many other differences. E. bullatissimuwm in many respects resembles the still obscure FE. smilacinum HBK. of Colombia, but that is said to have angulate branches, much smaller leaves hispid beneath and provided with a petiole about a quarter as long as the blade E. (§ Cylindrocephala) caldense, spec. nov., arboreum oppositira- meum; ramis primo distincte sexangularibus deinde subteretibus post exsiccationem multicostulatis purpureo-olivaceis villosulis et minute granuliferis; pilis patentibus delicatulis articulatis; internodiis 6 cm. vel ultra longis; foliis oblongis acuminatis integris basi subacutis breviter petiolatis supra basin 3-nervatis supra primo inconspicue . pubentibus tardius glabratis et subbullatis subtus vix pallidioribus persistenter pubescentibus crassiuscule membranaceis 5-9 cm. longis 17-32 mm. latis; petiolo 4-7 mm. longo fulvido-tomentello; corymbis terminalibus valde convexis ca. 1 dm. diametro; capitulis ca. 12-floris sessilibus ad apices ramulorum plerumque trinis vel quinis ca. 12 mm. longis 5-6 mm. diametro; involucri cylindrati squamis ca. 21 regular- iter ca. 5-seriatim gradatis arcte adpressis apice obtusissimis vel rotundatis conspicuiter ciliatis dorso subherbaceis 3-nervatis, extimis suborbicularibus ca. 2.2 mm. longis et 1.7 mm. latis, intermediis oblongis ca. 7 mm. longis et 2 mm. latis, intimis oblanceolato-linearibus subacutis ca. 9 mm. longis et 0.8 mm. latis; corollis roseo-albis glabris tubulatis ca. 5 mm. longis gradatim sursum ampliatis sine faucibus distinctis; dentibus limbi 0.7-0.9 mm. longis; antheris apice bene ap- piculatis; achaeniis valde immaturis gracilibus deorsum attenuatis angulis minute scabratis ca. 4 mm. longis; pappi setis ca. 30 albidis 4—5 mm. longis——Cotomsta: tree on bushy flat, alt. 1400-1600 m., on Rio Quindio above Armenia, Dept. Caldas, in the Cordillera Cen- tral, July 25, 1922, Pennell, Killip & Hazen, no. 8736 (tyPE, Gr.). This species approaches in many of its characters E. punctulatum DC., which, however, has a flatter-topped more fastigiately branched inflorescence, somewhat fewer-flowered heads, and distinctly toothed cauline leaves, longer petioles, and nearly smooth stem. While these differences are not individually strong they are in combination suf- ficient to make it very unlikely that the present plant of the Central Cordillera can represent the species of Eastern Brazil. E. (§ Eximbricata) Carletonii, spec. nov., herbaceum verisimiliter perenne (basi ignota) decumbens 1.5-2 dm. altum obscure adpresseque puberulum aspectu glabrum adscendenter ramosum; caulibus sub- hexagonis purpureis vix 2 mm. crassis; foliis oppositis petiolatis pal- 8 ROBINSON matim 3-fidis 3.5-6 cm. longis (petiolo anguste alato 1-2 cm. longo incluso); lobis bifidis vel trifidis vel pinnatim dissectis; rhachibus et segmentis linearibus 1-1.5 mm. latis; lobis ultimis apice obtusis; inflorescentiis terminalibus fastigiatim ramosis irregulariter cymosis planiusculis 3-6 cm. diametro; pedicellis 2-7 mm. longis capillaribus; capitulis ca. 4-5 mm. altis et 4 mm. diametro ca. 20-floris; involucri campanulati squamis anguste lanceolatis acutis viridibus sub- aequalibus laxe imbricatis plerisque 2-costulatis ca. 3 mm. longis et 0.5 mm. latis minute puberis, extimis 2-4 multo minoribus; corollis laete roseis ca. 2.5 mm. longis glabris; tubo proprio ca. 0.9 mm. longo; faucibus cylindratis vix ampliatis ca. 1.3 mm. longis; dentibus limbi ca. 0.3 mm. longis; achaeniis nigrescentibus 1.2 mm. longis deorsum decrescentibus subglabris; pappi setis 13-15 delicatule - capillaribus albis 1.5-2 mm. longis.—Honpvuras: on rocks, Cuyamel, March 30, 1923, M. A. Carleton, no. 466 (tyPE, Gr.). This highly interesting species was among plants recently sent from the United States National Herbarium for examination at the Gray Herbarium. It is well marked by its deeply cleft linear-lobed leaves which give it a habit very unlike that of any other Central American species of the genus. E. (§ Subimbricata-Cylindrocephala) T y Glaziou, Bull. Soc. Bot. Fr. Ivi. Mém. 3, 384 (1909), where listed as a probable exceptis) 4—5-floris; involucri cylindrati ca. 7 mm. longi squamis ca. 12 brunnescenti-stramineis 3—4-seriatim laxe imbricatis caducis dorso glabris 2-3-nervatis obtusis vel apice rotundatis ciliolatis, extimis ovatis ca. 2 mm. longis et 0.8 mm. latis, intermediis oblongis ca. 4-5 TROPICAL AMERICAN EUPATORIEAE 9 mm. longis et 1.3 mm. latis, intimis lineari-oblongis ca. 6 mm. longis et 1.2 mm. latis; corollis roseis vel albis graciliter tubulosis ca. 5 mm. longis glabris; achaeniis nigris gracilibus ca. 3 mm. longis deorsum decrescentibus; pappi setis ca. 32 flavescenti-albis barbellatis basin versus paullulo incrassatis.—BraziL: in forest of Tijuca near Cor- catimba, Rio de Janeiro, Glaziou, no. 11,002 (isotype K., Copen- hagen, fragm. Gr.); summit of Tijuca, E. W. D. & M. M. Holway, no. 1426 (Gr.); Brazil without locality, Glaziou, no. 536 (Gr., from herb. of Klatt). This species somewhat recalls the following, from all of which it can be readily distinguished: E. roseum Gardner which has shorter much blunter and clearly pinnate-veined not pinnately nerved leaves. E. ttatiayense Hieron. which has much longer internodes, strongly _ angled stems, greenish-stramineous involucre and 3(not 5-7)-nerved leaves. Symphyopappus polystachyus (DC.) Baker which has at least part of its heads much more clearly pedicelled, its involucral scales broader and more scr asens its achenes shorter and thicker, and its leaves of firm E. (§ Susbsiveintnatae cacao spec. Nov., fruticulosum ca. 3 dm. altiim griseo-puberulum vel -t habitu | erectum a basi decumbenti adscendenter oppositirameum; caule tereti fusco-griseo ad 3 mm. diametro; internodiis 5-12 mm. longis; ramis gracilibus a basi et praecipue in media parte foliosis superne subnudis; foliis oppositis dense fasciculatis linearibus integerrimis sessilibus subacutis margine valde revolutis uninerviis utrinque canes- centi-tomentellis 7-15 mm. longis 0.8-1.2 mm. latis; corymbis ter- minalibus plerisque 3—5-capitulatis ca. 3 cm. diametro; capitulis ca. panulati squamis ca. 24 pluriseriatim imbricatis scariosis brunnes- centibus 5-7-striato-costulatis ciliatis apice rotundatis dorso parce griseo-puberulis vel glabriusculis, extimis brevissiniis orbiculari- ovatis, intermediis late oblongis, 4 mm. longis 2 mm. latis, intimis oblongo-linearibus; receptaculo plano glabro; corollis ut videtur albis vel albidis tubulatis 6.5 mm. longis glabris gracilibus sine faucibus distinctis, dentibus limbi lanceolato-oblongis ca. 0.6 mm. longis; achaeniis nigrescentibus minute scabratis ca. 3.2 mm. longis basi eT pappi setis ca. 30 corollam subaequantibus flavescenti-albis. u: dry banks above the Rio Marafion, Chavanillo, alt. about 2440 m., Aug. 17-26, 1922, Macbride & Featherstone, no. 1975 (TYPE, Field Mus:: IsoTYPE, Gr.). A species of marked habit (suggesting a labiate rather than a composite), obviously related to E. lavandulae- folium DC. but differing in its larger more numerously flowered heads, . 10 ROBINSON broader multistriate involucral scales of firmer texture and brownish color, smaller leaves, ete. E. (§ Conoclinium) dichroum, spec. nov., ut videtur herbaceum et perenne 5 dm. vel ultra altitudine (basi ignota); caule erecto opposi- tirameo paullo flexuoso post exsiccationem costulato pallide brunneo araneoso-subtomentello; ramis gracilibus adscendentibus foliosis; foliis oppositis rhombeo-ovatis utroque acutis basin cuneatam versus integerrimis aliter inciso- vel crenato-dentatis conspicuiter bicoloribus supra viridibus paullo arrachnoideo-puberulis subtus dense griseo- lanulatis membranaceis 2.8-4.2 cm. longis 1.2—2.4 em. latis multo supra basin subtrinervatis; petiolo gracili griseo-puberulo 8-12 mm. longo; inflorescentiis terminalibus densis primo subglobosis 1-2 cm. diametro maturitate irregulariter subcorymbosis ca. 4 em. diametro; pedicellis 2-3 mm. longis; capitulis ca. 20-floris 8 mm. altis 4 mm. diametro; disco subconico; involucri squamis ca. 20 subtriseriatis argute acuminatis leviter striato-costulatis dorso pubescentibus, ex- timis lanceolatis, interioribus anguste oblongo-lanceolatis; corollis gracilibus; tubo proprio 2 mm. longo; faucibus paullo ampliatis ca. 1.2 mm. longis, dentibus limbi angustis recurvatis extus minute hispidulis; achaeniis 2.2 mm. longis 0.6 mm. crassis atrobrunneis glabris basi calcaratis; pappi setis ca. 32 flavescenti-albidis tenuibus acutis barbellatis—CoLomBia: Andes of Bogotd at El Boqueron, alt. 2700 m., October 20, 1852, J. Triana, no. 1239 (ryPr, Par., phot. and » small fragm. Gr.). This species in pressed specimen singularly recalls in habit and its gray and green coloration Verbesina encelioides though its leaves are much smaller. It is somewhat intermediate between L. ballotaefoliuin HBK. and E. lanulatum Robinson. From the former it may readily be distinguished by its cuneate-based and more incisely toothed leaves of rhombic-ovate outline. From the latter it differs in having much larger, thinner, more incisely toothed and somewhat triplinerved leaves. In E. lanulatum the throat of the corolla was found.to exceed the proper tube while in E. dichrowm the reverse is the case. In this group and probably from the same region is the slightly known E. parvulum HBK. which, however, has leaves truncate at base. E. preLopictyon Robinson, Proc. Am. Acad. liv. 325 (1918). This species was founded upon a specimen collected in Colombia by Lobb and now in the Kew Herbarium. This was unhappily by more precise locality. It is Siucwedcine.; of interest to recognize in Dr. Pennell’s recent ouieetion a specimen which can be confidently identified with this species and therefore establishes its geographic position. Dr. Pennell’s pees bears the following data: climbing herb TROPICAL AMERICAN EUPATORIEAE 11 on rocky stream-bank in forest, Cordillera Occidental below Cerro Tatama, Rio San Rafael, Department of Caldas, alt. 2500-2800 m., Sept. 7-11, 1922. This new material, agreeing with that of Lobb in all important respects, differs in having leaves papillo-scabrid and dull above rather than subglabrous and shining. It is probable that these differences are merely individual and perhaps due to the stage of development, the plant of Pennell being still in bud, that of Lobb being fully mature. Re-examination of the species suggests that it is rather closely related to though distinct from the Ecuadorian E. lamiifolium HBK. and E. Turczaninowii Robinson and would better be referred to § Conoclinium rather than § Cam puloclinium. E. (§ Subimbricata) eximium, spec. nov., fruticosum carnosum, in inflorescentia obscure strigilloso-puberaiai, aliter glaberrimum; caulibus pluribus 1-2 m. longis teretibus molliter lignosis medullosis ad inflorescentiam terminalem foliosis; internodiis 1-2 em. longis; foliis oppositis petiolatis oblongis acuminatis basi abruptius acutatis paullo undulatis vel integerrimis margine anguste revolutis carnoso- coriaceis ca. 5-6 em. longis et 2-3 cm. latis utrinque glaberrimis subtus vix pallidioribus; costa magna conspicui, venis lateralibus paucis pinnatis supra obscuris subtus saepius omnino invisibilibus; petiolo crasso 1-1.2 em. longo; panicula laxe corymbosa valde convexa transverse ellipsoidea ca. 1.5 dm. alta et 2.5 dm. diametro; pedicellis 9-14 mm. longis gracilibus cum bracteolis parvis filiformibus 1-3 munitis; capitulis numerosis sed nunquam congestis ca. 9-floris 8.5 mm. altis 6 mm. diametro; involucri subcylindrati squamis ca. 10 inaequalibus subtriseriatis, exterioribus lanceolatis acutiusculis, interioribus oblongis ciliolatis apice rotundatis ca. 4 mm: longis et 1.3 mm. latis; corollis anguste infundibuliformibus laete lilaceo-roseis glabris 5 mm. longis; tubo proprio 1 mm. longo; faucibus gradatim ampliatis 3 mm. longis; dentibus limbi 5, lanceolato-deltoideis 1 mm. longis; achaeniis (immaturis) glabris 1.5-2.2 mm. longis; pappi setis ca. 40 albis ca. 4.5 mm. longis——Costa Rica: vicinity of La Palma, on the road to La Hondura, alt. 1500-1700 m., July 17, 18, 1923, W. R. Maxon & A. D. Harvey, no. 7941 (type, Gr., U. S.). “Growing in a bushy mass spreading 1-2 m. in all divhetican from a large stump.” This exceedingly handsome species is most closely related to the Colombian E. sciaphilum Robinson, which however differs in having subsessile much more obtuse leaves and somewhat larger heads, flowers, and achenes. FE. eximium differs from E. araliacfolium Less. in having smaller leaves, fewer-flowered heads, much less un- equal involucral scales and much shorter proper tube of the corolla. 12 ROBINSON Other fleshy species with which E. eximium might possibly be con- fused are E. parasiticum Klatt, which has fewer-flowered heads, more ovate leaves, and a scarcely imbricated involucre; E. arthrodes Robinson and E. carnosum Kuntze, both of which have broad and obtuse outer involucral scales as well as many minor differences in foliage, ete. E. eximium must be an extraordinarily beautiful object when in full bloom and would seem worthy of horticultural attention. E.. GRANDE Sch.-Bip. ex Bak. in Mart. FI. Bras. vi. pt. 2, 347 (1876). To the synonymy of this species may be added with fair confidence E. angulosum Glaziou, Bull. Soc. Bot. Fr. lvi. Mém. 3, 390 (1909). Glaziou’s species, merely listed as a doubtful novelty and never given a published diagnosis, was founded upon Glaziou’s no. 21,578, which, to judge from an isotype in the herbarium of the Royal Gardens at Kew, is merely very young material of E. grande, with the inflorescence as yet too immature to show its long pedicels and openly branched nature which at a later stage become characteristic of the species. In all other respects, so far as the stage of development permits com- parison, the correspondence is perfect. eek imbricata) hidrodes, spec. nov., suffruticosum oppositi- rameum ad 12 dm. altum (Wright) vel fruticosum 2-3 m. altum (Leon); ramis teretibus post exsiccationem leviter striato-costulatis; ramulis gracilibus a pilis tenuibus moniliformibus atropurpureis in- dutis; foliis oppositis petiolatis ovatis vel ovato-lanceolatis plerisque subintegris apice subcaudato-attenuatis basi acutis supra glabris vel glabriusculis subtus vix pallidioribus praecipue in nervis hirtellis membranaceis vel aetate subchartaceis paullo supra basin pinnatim 3(-5)-nervatis 6-12 cm. longis 2.4-6 em. latis; petiolo 1.3-2 em. longo purpureo-villosulo; panicula terminali corymbiformi basi foliaceo-bracteata patenter ramosa purpureo-pubescenti; pedicellis subfiliformibus ad 1 em. longis bracteolas 1-3 minimas gerentibus; capitulis ca. 20-floris 6 mm. altis; squamis involucri ca. 12 linearibus 1-2-nervatis acutiusculis vel apice vero subobtusis vel etiam rotun- datis vix imbricatis dorso puberulis et cum glandulis globularibus lucidulis instructis; corollis primo purpureis deinde viridescentibus; achaeniis 2.7 mm. longis maturitate atrobrunneis cum glandulis globularibus sessilibus ornatis deorsum decrescentibus apice cum basi styli persistenti paullo incrassata nigra (verisimiliter nectarifera) coronatis; pappi setis corollam subaequantibus saepe purpurascenti- bus.—E. celtidifolium, var. quintuplinerve Griseb. Cat. Pl. Cub. 146 (1866), not E. quintuplinerre Less. E. nervosum Griseb. 1. c., not Sw. E. nervosum, var. quintuplinerve (Griseb.) ‘Gomez, Ann. Hist. Nat. TROPICAL AMERICAN EUPATORIEAE 13 Madrid. xix. 270 (1890).—Cusa: Proy. Oriente: Los Organos, Sept. 20, Wright, no. 2805 (typx, Gr.); Loma del Gato and vicinity, alt. 900-1000 m., July 11 to Aug. 14, 1921, Bros. Leon, Clement & Roca, nos. 10,184 (N. Y., fragm. Gr.) and 10,374 (N. Y., fragm. Gr.); banks of river, Santa Cruz, Aug. 27, Wright without number (Gr.); Cuba without precise pret Wright, no. 2804. (Gr.); Sierra Maestra, Bro. Leon, no. 11,048 (N. Y., fragm. Gr.). For many ot ee Cuban plant has been a puzzle. It is nearly related to several species long known in the more southern West Indian islands, but can be satisfactorily identified with none of them. Grisebach in his Catalogue referred material of it both to the Jamaican E. nervosum Sw. and to a newly proposed variety of E. celtidifolium Lam. of the Lesser Antilles. The real E. nervosum of Jamaica has its leaves rounded or even retuse at the base, more coarsely and sharply dentate, shorter-petioled and somewhat more firmly nerved. E. celtidifolium is an arborescent or even arboreous species with much smoother and lighter-colored twigs and with leaves tending to firmer texture, rounded:at base and more falcately attenuate at the tip. There is still a third species to which this Cuban material has been at times referred, namely FE. nudiflorum A. Rich. Fl. Cub. ii. 41 (1853), but this also, though certainly a closely related species, fails to give a satisfactory resting place for the specimens in question since these have cuneate-based leaves while Richard distinctly states regarding his E. nudiflorum that its leaves were obtuse or even rounded at base. As to the varietal name quintuplinerve employed by Grisebach it seems unwise to use it in the specific category since it has already been published in connection with a very different plant of Brazil. The specimens here referred to E. hidrodes, while agreeing with each other in all essentials, show considerable variation in the size ~ and texture of the leaves and in the size of the heads and the number of florets. In the presence of close resemblance in other respects it is believed that these differences are individual and developed in response to environment and stage of development. A single speci- men (P. Wilson’s no. 9303 from rocks in the River San Miguel, Prov. Pinar del Rio, at the other end of Cuba and thus separated by a distance of some 400 miles) has coarsely toothed leaves, in this respect showing some approach to the Jamaican E. nervosum, but the leaves are cuneate at the base. Without more copious material, this form with toothed leaves—probably a mere leaf-variant of E. hidrodes—cannot be satisfactorily placed. The name hidrodes alludes to the perspiration-like drops which are abundant on the involucral scales, corollas, and achenes. 14 ROBINSON E. (§Subimbricata) Hitchcockii, spec. nov., herbaceum perenne exaltatum ca. 3 m. altum; caule tereti 1.3 cm. vel ultra diametro glabro fistuloso; foliis oppositis magnis longe petiolatis ovatis acu- minatis grosse dentatis basi acutis vel acuminatis in parte proximali penninerviis utrinque glabris subtus pallidioribus textura subcarneo- membranaceis, caulinis 2-2.5 dm. longis 1-1.5 dm. latis, dentibus usque ad 4 mm. altis et 15 mm. basi latis cuspidatis, nervis lateralibus utroque ca. 4 prorsus curvatis; petiolo 4-6 cm. longo; appendicibus stipuliformibus nullis; corymbis terminalibus planiusculis vel modice convexis saepe alternirameis vel irregulariter di- vel trichotomis obscure incurvo-puberulis ca. 2 dm. diametro nudis; bracteolis mini- mis; pedicellis 3-12 mm. longis filiformibus; capitulis ca. 1 em. longis et 6 mm. diametro ca. 20-floris; involucri 3—4-seriatim imbricati campanulato-cylindrati squamis ca. 20 gradatis primo paullo puberulis deinde glabratis, exterioribus lanceolato-ovatis ca. 2.5 mm. longis e 1.5 mm. latis obtusis, intermediis similibus sed 3-4 mm. longis et 2 mm. latis, interioribus oblanceolati-linearibus ca. 7 mm. longis et 1 mm. latis; corollis roseis 6 mm. longis gracilibus sursum paullo ampliatis sed sine faucibus distinctis; dentibus limbi deltoideis ca. 0.3 mm. longis puberulis; achaeniis gracilibus atrobrunneis glaber- rimis deorsum paullo decrescentibus 2.8 mm. longis; pappi setis ca. 60 brunneo-fulvis corollas aequantibus.—Ecuapor: Province Tun- gurahua: Valley of Pastaza River between Bafios and Cashurco, 8 hours east of Bafios, alt. 1300-1800 m., Sept. 25, 1923, Prof. A. S. Hitchcock, no. 21,876 (type, Gr.). A notable species not closely related to any hitherto described from Ecuador. Its leaves are as large as those of E. morifolium Mill. and the inflorescence recalls that of a Verbesina. It is a pleasure to name this well marked new species for its discoverer who, in addition to his special work on the grasses, has, through his copious collecting in difficult tropical regions, contributed much to the knowledge of other groups. E. (§Subimbricata) lithopl , spec. nov., f decumbens 3-5 dm. altum aspectu oculo nudo glabrum sed sub lente brevissime puberulum; caule tereti ad 5 mm. diametro firmiter lignescens cur- ad inflorescentiam simplicibus 2-4 dm. longis teretibus es beeps brunneis superne puberulo-granulatis; internodiis plerisque 1-8 c longis; foliis oppositis petiolatis lanceolatis attenuatis basi ieee integra — | serratis 3-4 cm. jongis: 1-1. 8 cm. latis mem- ai Py ge hg i Gilat obscure in netvis § Venioque puberulis a basi trinervatis; petiolo gracili TROPICAL AMERICAN EUPATORIEAE 15 saepius curvato 6-10 mm. longo; capitulis ca. 30-floris ca. 9 mm. altis et 6 mm. diametro in glomerulos paucos 2-3 cm. diametro con- gestis; pedicellis saepius 14 mm. longis; involucri squamis pluri- seriatim gradatis lanceolatis vel anguste lanceolato-oblongis purpureo- brunneis acutis vix costatis minute resinoso-granulatis et aliquando vernicosis; receptaculo planiusculo glabro; corollis tubulatis roseis vel albis glabris ca. 5.5 mm. longis sine faucibus distinctis; dentibus limbi deltoideis ca. 0.5 mm. longis; antheris apice bene appendiculatis; achaeniis gracilibus praecipue in angulis sursum hispidulis ca. 3 mm. longis apice ab annulo pallido pappifero coronatis; pappi setis ca. 34 flavescenti-albis minute scabratis apice non incrassatis.—PERv: stone fences and rocky cafion-sides, alt. about 2135 m., Aug. 21, 1922, Macbride & Featherstone, no. 2235 (typE, Field Mus.; ISOTYPE, Gr.). This species, in habit recalling a glomerate-headed Stevia, is without close relatives in the Peruvian flora. It bears, however, some slight resemblance to the eastern Brazilian E. Vauthierianum DC., but that has much greater stature, much larger leaves, more diffuse inflores- cence, more distinctly costulate involucral scales, and puberulent E. (§Subimbricata) mesoreopolum, spec. nov., fruticosum ali- quando scandens; ramis patentibus teretibus post exsiccationem multicostulatis fulvido-tomentellis; foliis oppositis petiolatis ovatis 1. = Lend “i a . eee Tens we iuerneed n ag Se gel ic vel Pp g vel obtusis membranaceis supra brevissime puberulis scabridulis subtus fulvido-tomentellis prominulenter reticulato-venosis 4-11 cm. longis 2-6 em. latis, venis lateralibus principibus ca. 3-jugis rervi- formibus pinnatim a costa 2-11 mm. supra basin divergentibus; petiolo gracili fulvo-tomentello 10-16 mm. longo; paniculis terminali- bus multicapitulatis oppositirameis pyramidalibus ca. 1 em. altis 10-18 cm. crassis fulvido-tomentellis; pedicellis filiformibus tomentellis saepe curvatis 4-6 mm. longis; capitulis 18-20-floris ca. 8-9 mm. altis 6 mm. diametro; involucri campanulati 4 mm. alti squamis 22-25 ovatis vel oblongis acutiusculis 3—4-seriatim imbricatis regulariter gradatis dorso 2-3-striato-costulatis tomentellis; recep- taculo planiusculo calvo; corollis glabris subtubulatis apicem versus paullo ampliatis ca. 4 mm. longis limbum versus violaceis vel pallide roseis; faucibus vix ullis distinctis; limbi dentibus ovato-deltoideis ca. 0.5 mm. longis; achaeniis stramineis gracilibus glanduloso-atomiferis praecipue ad costas deorsum decrescentibus 2.5-3.5 mm. longis; pappi setis 47-68 inaequalibus laete albis laeviusculis achaenium subaequantibus.—CotomBia in the Cordillera Central: shrub in forest at “Pinares” above Salento, Dept. of Caldas, alt. 2600-2900 16 ROBINSON m., Aug. 2-10, 1922, Dr. F. W. Pennell, no. 9209 (rypx, Gr.), flowers said to be violet; climbing shrub in moist valley at “Calaguala,” Coconuco, Dept. of El Cauca, alt. 2500-2800 m., June 14, 18, 1922, Dr. F. W. Pennell, no. 7179 (Gr.), flowers said to be pale pink. This species differs from E. Stuebelii Hieron. in its more numerous florets, much looser and more pyramidal inflorescence and thinner less rugose leaves which are pointed at base. From E. Mayorii Robinson it differs in its larger and longer-pedicelled fewer-flowered heads, relatively shorter involucre, and longer stramineous gland- sprinkled achenes. E. (§ Eximbricata) parvulum Glaziou, Bull. Soc. Bot. Fr. lvi. Mém. 3,390 (1909). This species, named as a doubtful novelty by Glaziou, l. e., but without diagnosis, seems never to have been described. To give the name legal status the following characterization is here put on record, the characters having been drawn from isotypes (Glaziou’s no. 15,150 seen in loans from Kew and from the Botanical Museum of Copenhagen). Herbaceum perenne suberectum 2.5-3 dm. altum; caulibus gracilibus teretibus dense foliosis breviter purpurascenti- pubescentibus ad inflorescentiam simplicibus; foliis plerisque alternis oblanceolatis apice rotundatis basi attenuatis leviter crenatis utrinque -viridibus pubescentibus punctatis vix petiolatis 1-1.7 em. longis 4-6 mm. latis; costa media inconspicua; nervis venisque lateralibus obscuris; corymbis alternirameis terminalibus planiusculis ca. 6 cm. diametro; pedicellis filiformibus purpurascenti-pubescentibus 3-8 mm. longis; capitulis ca. 30-floris ca. 7 mm. altis et 7 mm. diametro; involucri turbinati-campanulati squamis ca. 25 sub-biseriatis vix imbricatis lineari-lanceolatis acutis plerisque subaequalibus, exteri- oribus subherbaceis purpurascenti-pubescentibus obscure 3-nervatis, interioribus substramineis 1-nervatis subglabris; corollis roseis ca. 3.5 mm. longis; tubo proprio brevi ca. 1 mm. longo in fauces sub- cylindratas ca. 2 mm. longas gradatim paullo ampliato; dentibus limbi ca. 0.5 mm. longis; achaeniis (immaturis) gracilibus pallidis deorsum attenuatis in angulis hispidulis ca. 3 mm. longis; pappi setis ca. 32 sordide flavescenti-albidis ca. 2.5 mm. longis.—BraziL: between Ouro Preto and Queluz, Minas Geraés, Feb.—Mar., Glaziou, no. 15,150 (isotype, K., Copenhagen, phot. and small fragm. Gr.). Of char- acteristic habit and seemingly very distinct from any Brazilian species hitherto characterized. E, PENINSULARE Brandegee, Erythea, vii. 4 (1899). This species of Lower California is subject to conspicuous variation in its pu- bescence, as follows: Var. a. obscure pubescens; foliis utrinque viridibus supra glabrius- TROPICAL AMERICAN EUPATORIEAE 17 culis subtus praecipue in nervis venisque obscure pubescentibus; petiqlo puberulo—Lower Ca.tirornia, San José del Cabo, Sierra de la Laguna, San Francisquito, San Bartolomé, Cafion Hondo, Mesa Verde. Var. 8. epipolimum, var. nov., griseo-tomentellum; foliis utrinque griseis minute papilloso-puberulis; petiolo dense patenterque griseo- tomentello.— Lower CaLirornia: Aguaje de Santana, 56 km. north of San Ignacio, alt. 1037 m., Oct. 4, 1905, EF. W. Nelson & E. A. Goldman, no. 7186 (U. S., fragm. Gr.). In habit, leaf-form, inflorescence, involucre, ete., agreeing with the smoother typical variety and clearly conspecific with it, but cer- tainly meriting varietal distinction. Material collected at San Juleo Cafion, Brandeyee, Apr. 20, 1889 (Mc., fragm. Gr.) is intermediate. E. (§ Subimbricata) procerum, spec. nov., fruticosum elatum ca. 4.5 m. altum; caulibus teretibus dense crispeque fulvido-tomentellis tardius glabratis post exsiccationem leviter costulatis medullosis 3-8 mm. diametro; internodiis 2-10 cm. longis; foliis oppositis lanceo- lato- vel ovato-oblongis vel rhombeo-ovatis petiolatis apice acutis vel acuminatis basi subacutis vel obtusis vel etiam rarius rotundatis margine irregulariter grosseque serrato-dentatis vel undulatis vel aliquando etiam integris supra scabris bullatis subtus vix pallidioribus molliter tomentosis textura crassiusculo-membranaceis 8-15 cm. longis 3-6.7 em. latis penniveniis vel paullo supra basin pinnatim subnervatis; petiolo fulvido-tomentoso 8-13 mm. longo plerumque erecto vel suberecto; corymbis terminalibus et in axillis superioribus orientibus pedunculatis convexis subdensis multicapitulatis 1-3 dm. diametro; pedicellis 4-9 mm. longis tomentellis; capitulis ca. 12-floris 10-12 mm. altis 6-7 mm. diametro; involucri squamis ca. 15 laxe subtriseriatim imbricatis lanceolato-oblongis valde inaequalibus obtusis vel subacutis, extimis ca. 3 mm. longis et 1.2 mm. latis subherbaceis griseo-tomentellis, intimis 6-7 mm. longis ca. 1 mm. latis pl jue 3-costulatis saltim apicem versus puberulo-tomentellis purpurascentibus; corollis rubro-lilaceis glabris 5-6.5 mm. longis glabris; tubo sine faucibus distinctis ca. 5 mm. longo vix sursum ampliato; dentibus limbi patentibus deltoideis 0.7-0.9 mm. longis; achaeniis gracilibus deorsum attenuatis fuscescentibus granulato- scabratis ca. 5 mm. longis; pappi setis ca. 35 flavico-albidis ca. 5.5 mm. longis.—Prrv: on steep shrubby slopes, with flowers red-lilac and fragrant, alt. about 2745 m., at Mito, July 8-22, 1922, Macbride d- Featherstone, no. 1551 (tvPE, Field Mus.; tsoryre, Gr.); also 3 sheets from herb. of Pavon from Peru but without recorded locality (herb. Univ. Geneva). 18 ROBINSON This species must be rather closely related to the little known E. helianthifolium HBK., which came from Ayavaca toward 1000 km. northward from the Mito material here described. FE. helian- thifolium is known to the writer from a photograph which he made of the type material at Paris some years ago. Microscopic examination of this photograph and close study of the brief and rather unsatis- factory original description seem to leave no doubt as to the probable distinctness of the species here characterized. In E. helianthifolium the leaves are much more regularly and definitely serrate, their lateral veins leave the midnerve at a somewhat greater angle and are less curved forward in their course, the upper surface is scabrid, but not at all, or at least by no means to the same extent bullate, the petioles tend to spread widely instead of rising parallel to the stem, and ac- cording to the original description the leaves are reticulate-veiny and are beneath punctate with very numerous orange-glands—a feature not observable in the material of E. procerum. Finally the inner involucral scales of E. helianthifolium are said to be glabriusculous while in E. procerum the inner as well as the outer scales are at least | toward the tip clearly pubescent. These differences, though indi- vidually of no great strength, and, especially when taken in conjunction with the remoteness of the stations, seem to show the plants pretty certainly distinct. E. sotmacinoweEs HBK. Nov. Gen. et Spec. iv. 126 (1820). This widely distributed and variable species has long been known from Mexico and Central America, Venezuela, Colombia, Ecuador,.and Bolivia, but until now does not appear to have been recorded from Peru. It has, however, been recently collected at the following points in that country; as a slender shrub 1.5 m. high with few erect branches and greenish-yellow flowers, at a hacienda 9 miles above Huertas, Cabello, alt. about 2440 m., June 25, 1922, Macbride & Featherstone, no. 1321 (Field Mus., Gr.); as a slender open shrub-liana of dense montafia and with greenish flowers, at La Merced, alt. about 610 m., Aug. 10-24, 1923, J. F. Macbride, no. 5258 (Field Mus., Gr.), a form with large subentire leaves; also on a disintegrating granite river- bluff, at the same place, date, and elevation, Macbride, no. 5333 (Field Mus., Gr.), a form distinctly lignescent but with a root suggesting that of an annual. E. (§ Subimbricata) sphagnophilum, spec. nov., fruticosum erec- tum; caulibus paucis strictis simplicibus vel adscendenter ramosis teretibus medullosis fusco- vel brunnescenti-tomentellis ca. 9-12 dm. altis ad 8 mm. vel ultra crassitudine; internodiis 2-6 em. longis; foliis oblongis utroque acutis argute serratis (dentibus 1-1.5 mm. altis 2-3 TROPICAL AMERICAN EUPATORIEAE 19 mim. basi latis) 6-10 cm. longis 2.5-4.7 cm. latis chartaceo-coriaceis penniveniis et utrinque prominulenter reticulato-venosis supra granuliferis subtus paullo pallidioribus crispe fulvido-pubescentibus; venis lateralibus principibus utroque 8-11 prorsus curvatis; petiolo 7-12 mm. longo tomentello; corymbis terminalibus planiusculis vel modice convexis 4-9 cm. diametro densis; pedicellis ca. 2 mm. longis; capitulis 5—6-floris 8-10 mm. longis 5-7 mm. diametro; involucri campanulati squamis ca. 11 valde inaequalibus gradatis obtusis vel apice rotundatis vix costulatis dorso pubescentibus, extimis lanceolatis 2.5 mm. longis 0.7 mm. latis, intermediis oblongis ca. 5 mm. longis et 2 mm. latis, intimis oblongis ca. 5.5 mm. longis; corollis purpureis albis glabris; tubo sursum gradatim ampliato ca. 4.3 mm. longo sine faucibus distinctis; dentibus limbi ovato-deltoideis ca. 1.2 mm. longis; achaeniis immaturis glabris 2.5 mm. longis; papp! setis ca. 35 inaequalibus sursum barbellatis attenuatis flavidulo- albis plerisque 3.54 mm. longis.—-PERv: a few-stalked strict bush on sphagnum slopes, alt. about 1830 m., Villacabamba, Hacienda on Rio Chinchao, July 17-26, 1923, J. F. Macbride, no. 4968 (ryPr, Field Mus.; isotype, Gr.). Probably best placed near E. Macbridei Robin- son, from which, however, it is readily distinguished by its much lower stature, shorter leaves with different indumentum, and purple flowers. | E. THEAEFOUUM Benth. PI. Hartw. 199 (1845). The type of this species from the slopes of Mt. Sotaro in the Province of Popayan had leaves 1-114 inches in length, 4-7 lines wide, oblong in form and, as described by Bentham, acute at both ends. This form may be known as var.typicum. From it all material subsequently referred to the species differs in varying degree. The following seems worthy of varietal distinction Var. subellipticum, var. nov., foliis subellipticis vel elliptico- oblongis apice rotundatis vel obtusis vel rarius subacutis basi plerum- que rotundatis 18-25 mm. longis 13-16 mm. latis; aliter formae typicae speciei simillimum.—CoLomBIA: in mountainous districts, Province of Bogota, alt. 2700 m., Feb. 1854, Triana, no. 1221 (TYPE, Par., phot. and small fragm. Gr.); Prov. Pamplona la Baja, alt. 2440 m., L. Schlim, no. 1056 (hb. Univ. Geneva). VENEZUELA: Pdéramo del Tamboi, Merida, Jahn, no. 747 (U. S., phot. and fragm. Gr.); also Linden, no. 464 (Gr., Univ. Geneva). This material secured by Linden bears his Caracas label but error is suspected since the species not otherwise known from the coast of Venema is abundant in parts of Colombia traversed by this collecto E. uBgR Robinson, Proc. Am. Acad. lv. 37 (1919). After publish- 20 ROBINSON ing this Peruvian plant, which had been carefully differentiated from other Andean species of the genus, the writer noticed its likeness to E. rufescens Lund of Atlantic. Brazil. This resemblance, though striking, did not appear to reach specific identity. It has recently been possible, through the kindness of Capt. John Donnell Smith, of Baltimore, to borrow from his herbarium a sheet of excellent Brazilian material of EF. rufescens. This, carefully compared at the Gray Herbarium with a fragment of the type of FE. wher, dis- closes the following differences. E. rufescens: leaves tomentose beneath; inflorescence distinctly pyramidal; involucral scales 3.54 mm. long, soft in texture, dorsally tomentose except for the hyaline margin, also sometimes glandular- atomiferous; achenes glabrous or merely with minutely scabrid angles (sometimes with a few scattered sessile glands). . uber: leaves finely and somewhat sparsely piliosnent beneath; Satidresbunies flattish or only moderately convex; involucral scales ~ 4.5-5 mm. long, of firmish texture, dorsally granular and very spar-. - ingly puberulent; achenes setulose. These distinctions, while none of them strong, are in combination sufficient to make probable the distinctness of these species, especially as they are separated by such distance and belong to floras in general so different as those of the Andean portions of Peru and the coastal parts of Brazil. E. (§ Subimbricata) variolatum, spec. nov., ut videtur herbaceum perenne 3-4 dm. altum erectum pluricaule a caudice paullo lignescenti; caulibus cheers teretibus rectis virgatis 1-2 iam. diametro griseo- puberulo p2 | has DisePusatclos crassiusculis integerrimis apice obtusiusculis 3-7 - mm. longis 0.4-1.1 mm. latis supra subglabris subtus sparse griseo- puberulis utrinque qu ariolatis cum glandulis (post exsiccationem) depressis; foliis primariis te omnibus proliferis et in axillis folia plura paullo minora gerentibus; corymbo terminali alternirameo convexo multicapitulato griseo-tomentello; pedicellis filiformibus ca. 2 mm. longis; capitulis ca. 5-floris ca. 6 mm. altis et 4 mm. diametro; involucri squamis ca. 11 valde inaequalibus 2—3-seriatis erubescentibus griseo-tomentellis, exterioribus 4-5 lanceolatis subacutis 1-2 mm. longis, interioribus elliptico-oblongis conspicue ciliatis 3-4.5 mm. longis 1.7 mm. latis; corollis ut videtur rubritinctis ca. 3.5 mm. longis paullo granulatis, tubo proprio ca. 1 mm. longo, faucibus sub- cylindratis paullo ampliatis ca. 2 mm. longis, dentibus limbi ca. 0.6 mm. longis; achaeniis glabris crassis vix 1.5 mm. longis sublucidis brunneis; pappi setis ca. 40 albidis barbellatis. —Braziu: Prov. St. TROPICAL AMERICAN EUPATORIEAE 21 Paul and Rio Grande do Sul, Auguste de Saint-Hilaire, no. 871 Catalogue C? No. 1553 (type, Par., phot. and small fragm. Gr.). E. variolatum forms a new member of the little series Disynaphia (Hook. & Arn.) Bak. In this group it differs from E. ericoides DC. and £. angustissimum Spreng. by its obtuse puberulent and pitted leaves; from E. calyculatum Hook. & Arn. (Stevia calycina DC. and Disynaphia montevidensis DC.) in its smaller heads, elliptical and obtuse inner involucral scales, very small leaves, ete., from E. ligulae- folium Hook. & Arn. (E. gnidioides DC.) and E. spatulatum Hook. & Arn. (E. halimifolium DC.) in its much smaller and less pubescent but (under lens) distinctly variolate leaves, from E. multifolium Robinson in its very much smaller, relatively narrower leaves, more open infl d smaller heads. In its characteristic habit E. variolatum recalls some species of Petalostemon. Two sheets of this species were found in a recent loan from the Paris Museum of Natural History sent to the writer for study and identification. While both were from the collection of Auguste de Saint-Hilaire and bore identical numbers, one had the label form for “ Province de Saint-Paul” and the other for “ Province de Rio Grande do Sul.’”’ It seems probable that the material, which is identical in form, development and condition from drying, was collected at one time and place, and that one of the labels was correct and the other used by error. The species is so characteristic that its recognition on rediscovery should not offer any difficulties. Mikania (§ Spiciformes) aquaria, spec. nov., robusta scandens; caulibus teretibusad 12 mm. diametro usque ad inflorescentiam glabris fistulosis et cum aqua impletis post exsiccationem multicostulatis; ra- mis oppositis patentibus; internodiis 2 dm. vel ultra longitudine; foliis oppositis ovatis acuminatis integris basi vel profunde cordatis vel subtruncatis et cum acumine in petiolum paullo decurrentibus sub- coriaceis supra atroviridibus impresse reticulatis lucidulis glaberrimis subtus pallidioribus opacis olivaceis minutissime pulverulenti- puberulis prope basin palmatim vel pinnatim 5-9-nervatis 20-23 cm. longis 15-17 em. latis; petiolo 7-9 em. longo; paniculis magnis lateral- ibus vel terminalibus 1,5-4 dm. altis 1-3 dm. diametro foliaceo-brac- teatis crispe puberulis; bracteis ovatis vel ovato-oblongis utroque acutatis 3-8 em. longis 1.54 em. latis petiolatis; spicis laxis 3-7 cm. longis; rhachi gracili plerumque leviter curvata; capitulis sessilibus late patentibus vel divaricatis; bracteolis ovatis acuminatis ciliatis ca. 1.6 mm. longis; involucri squamis anguste oblongis ciliolatis apice rotundatis et mucronulatis ca. 3.8 mm. longis et 1.2 mm. latis 3-5- nervatis; flosculis fragrantibus; corollis glabris albis; tubo proprio 22 ROBINSON ca. 2 mm. longo apice sensim constricto; faucibus anguste turbinatis ca. 1.5 mm. longis; dentibus limbi deltoideis 0.5 mm. longis; achaeniis nigrescentibus 2.6 mm. longis scabridulis; pappi setis ca. 36 sordide albidis vix apicem versus incrassatis ca. 3.8 mm. longis.—PERU: a liana with succulent hollow and water-filled stems, growing on the montafia along open stream, alt. about 610 m., Aug. 10-24, 1923, J. F. Macbride, no. 5242 (typ, Field Mus.; tsotypr, Gr.). A well marked species, probably related to the Bolivian M. cinnamomifolia Lingelsh. and M. dioscoreoides (Rusby) Robinson, but having larger heads and different foliage. M. (§ Thyrsigerae) cryptodonta, spec. nov., volubilis subherbacea glabriuscula; caule subtereti post exsiccationem purpurascenti- runneo flexuoso pluricostulato primo obscure puberulo; internodiis 1-2 dm. vel ultra longitudine; foliis oppositis graciliter petiolatis - ovatis acuminatis basi subtruncatis vel late subcordatis vel distincte cordatis margine revolutis remote obscureque cuspidato-dentatis (dentibus vix 0.2 mm. altis 5-10 mm. inter se distantibus) mem- branaceis 8-14 cm. longis et 5-10 cm. latis a basi subpalmatim 5- nervatis utrinque obscurissime in nervis sordide puberulis aliter glabris subtus paullo pallidioribus; petiolo 3-4.5 cm. longo puberulo, eis ejusdem jugi ab annulo prominulenti conjunctis; paniculis pyra- midatis laxiusculis axillaribus pedunculatis 1.5—2 dm. longis ca. 1 dm. diametro; capitulis 8-9 mm. longis ad apices ramulorum irregulariter subcymosis aliis subsessilibus aliis graciliter pedicellatis; pedicellis 0.5-3 mm. longis; bracteolis oblanceolato- vel spathulato-linearibus; involucri squamis lineari-oblongis obtusis apice puberulis; corollis albidis glabris ca. 5 mm. longis; tubo proprio gracili 2.3-2.7 mm. longo; faucibus campanulatis 1.2-1.4 mm. altis; dentibus limbi del- toideis 1-1.3 mm. longis; achaeniis gracilibus fuscis minutissime sca- bratis deorsum attenuatis 3.2-4 mm. longis; pappi setis ca. 36 albidis apicem versus non incrassatis 4.5-5 mm. longis.—-PERv: herbaceous or nearly so, clambering over stumps or in sunny woods, about 610 m. alt., Pozuzo, June 20-22, 1923, J. F. Macbride, no. 4683 (TYPE, Field Mus.: IsOTYPE, Gr.), and no. 4552 (Field | Mus., Gr.). Most nearly related to M. decora Poepp., but diff more deltoid-ovate rather than oblong leaves, longer petioles and cy- n of the heads in the inflorescence. M. pecora Poepp. in Poepp. & Endl. Nov. Gen. ac Spec. iii. 53 (1845). Seas with the technical characters of M. decora but differing clearly i in several minor features may be varietally appended to this species and distinguished from the typical variety as follows: Var. —_ foliis subtus-distincte in nervis puberulis; nervis TROPICAL AMERICAN EUPATORIEAE 23 lateralibus ab costa saepuis ca. 5-8 mm. supra basin divergentibus; petiolo ca. 1.5 cm. longo. Var. heteroneura, var. nov., foliis (ad maturitatem) subtus glabris; nervis lateralibus 3-jugis distincte pinnatis, jugo tertio 1.5-2.5 supra basin oriente; petiolo 2-2.5 cm. longo.—-PERu: liana with green- ish flowers and. pinkish pappus, Villacabamba, Hacienda on Rio Chinchao, alt. about 610 m., July 17-26, 1923, /. F. Macbride, nos. 5167 (Type, Field Mus.; tsorype, Gr.) and 5191 (Field Mus., Gr.). The distance of the upper pair of lateral nerves from the leaf-base is so great that the leaf becomes truly pinnate-veined. The leaves also are somewhat less sharply acuminate than in the typical variety and the cross-veining between the nerves is less regular. M. (§ Corymbosae) Gleasonii, spec. nov., volubilis elata verisim- iliter fruticosa saltim suffruticosa inflorescentiis exceptis aspectu glaberrima; caule leviter hexagono post exsiccationem pluricostulato glabro 3-4 mm. vel ultra crasso; internodiis usque ad 14 cm. longis; foliis oppositis petiolatis oblongis acuminatis integris basi acutis vel subobtusis paullo supra basin. pinnatim 5-nervatis supra nitidis lucidis impresse reticulatis subtus concoloribus minutissime ob- scureque puberulis 9-15 cm. longis 4-7 em. latis coriaceis margine revolutis; petiolo 2-2.5 em. longo plerumque curvato supra canali- culato, eis ejusdem jugi ab annulo prominulento obscuro transverso conjunctis; bracteis foliaceis petiolatis; corymbis saepe binis vel trinis ex eodem axillo graciliter pedunculatis valde inaequalibus (verisimiliter partialibus in corymbo arcte sessili) 4—6 cm. diametro laxis valde convexis fusco-puberulis conjunctim inflorescentiam terminalem interruptam paniculatam ad 2.5 dm. vel ultra longitudine formantibus; pedicellis plerisque 4-8 mm. longis; bracteolis lanceolatis vel anguste oblongis 3-4 mm. longis plerumque ad vel prope summam partem pedicelli orientibus; capitulis ca. 9 mm. longis; involucri squamis anguste oblongis 7-8 mm. longis 1.3-1.6 mm. latis apicem versus minutissime puberulis aliter glabris brunneis; corollis glabris (immaturis) ca. 5 mm. longis tubulatis paullo sursum ampliatis sine - faucibus distinctis in parte superiori vinaceo-purpureis; dentibus limbi ca. 0.8 mm. longis; achaeniis valde immaturis glabris ca. 2.8 mm. longis; pappi setis ca. 50 albidis ca. 5 mm. longis vix sursum incrassatis -—BritisH GutIaNa: dense upland forest, Rockstone, July 15-Aug. 1, 1921, Dr. H. A. Gleason, no. 681 (type, in herb of N. Y. Bot. Gard., sk. and fragm. Gr.). A well marked species with thick shining leaves recalling those of many Lauraceae. In in- florescence this species is somewhat vaguely intermediate between § Corymbosae and § Thyrsigerae, its apparently sessile corymbs being 24 ROBINSON loose and very convex and disposed along the branches so as to form a lax and interrupted panicle. It is a pleasure to commemorate in the naming of this species the very notable work accomplished by its collector in the exploration of British Guiana. M. 1ncastna Robinson, Contrib. Gray Herb., n. ser. Ixiv. 11, 82 (1922). This species has hitherto been known only from the original material collected by Alexander Mathews eighty or ninety years ago. It was scarcely in anthesis and showed only the upper, probably rameal leaves. Specimens which appear to be specifically identical have recently been collected by Mr. J. F. Macbride (no. 5778, Field Mus., Gr.) on sunny brush at La Merced, Hacienda Schunke, Peru, alt. about 1220 m. These are in advanced anthesis and exhibit not merely the rameal but the considerably larger cauline leaves which are rounded at the sides of the base but have the same narrowly deltoid-ovate form. They are sometimes as much as 18-22 em. long and 9-10 cm. wide with petiole 4 cm. long; the floral characters agree closely with those of the type though from greater maturity the dimensions are somewhat larger than those described in the original characterization. A noticeable trait is found in the tendency of the mature involucral scales to bend rather sharply backward and out- _ward at the point where the deltoid tip begins. This spreading tip appears to be of softer texture than the rest of the scale, which in full maturity becomes firmish in character. M. NIGROPUNCTULATA Hieron., var. denticulifolia, var. nov., foliis margine distincte Yaipnlatosdestintnie: dentibus ca. 0.5 mm. altis plerisque 2-9 mm. inter se distantibus; aliter var typicae a nike: —Coromsia: Dept. of El Cauca: climbing shrub in bushy forest, alt. 2400-2700 m., Cordillera Occidental, June 28, 1922, Pennell & Killip, no. 7430 (Gr. Fe M. (§ Globosae) hesperia, spec. nov., fruticosa volubilis; caule sub- tereti post exsiccationem costato patenter piloso vel subsetoso et dense cum glandulis nigrescentibus globosis sessilibus vel brevissime stipitatis ornato ca. 3 mm. vel ultra diametro; internodiis plerisque 4-5 em. longis; foliis oppositis breviter petiolatis oblongo-lanceolatis - attenuato-acuminatis basin versus gradatim angustatis sed ad basin veram abrupte rotundatis penniveniis margine cuspidato-denticulatis (dentibus ca. 0.4 mm. altis et 5-7 mm. inter se distantibus) mem- branaceis utrinque viridibus supra in costa obscure puberulis subtus in costa et nervis principibus appresso-pilosis in superficie minute nigropunctatis utrinque prominenter reticulato-venosis 12-14 cm, longis 2.5-4 em. latis; petiolo crasso setoso-piloso fuscescenti 4—5 mm. longo; panicula terminali 1-1.5 dm. longo vix 4 cm. diametro TROPICAL AMERICAN EUPATORIEAE 25 (immatura) oppositiramea setoso-pilosa; ramis paniculae adscen- dentibus brevibus a bracteis foliaceis multo superatis; capitulis ca. 8 mm. longis subsessilibus in glomerulos subglobosos ca. 2 em.. di- ametro dispositis; bracteolis subherbaceis pilosis nigropuncticulatis ca. 2-3 mm. longis et 1.5 mm. latis; involucri squamis oblongis obtusis apicem versus puberulis et saepissime nigropunctatis ca. 44.5 longis et 1.2-1.3 mm. latis; corollis purpureis ca. 4 mm. longis; tubo proprio gracili ca. 1.5 mm. longo; faucibus subcylindrato-campanu- latis ca. 1.5 mm. longis; dentibus limbi ca. 1 mm. longis; achaeniis (valde immaturis) glabris 2 mm. longis non punctatis; pappi setis ca. 25 albis.—Cotomsi1a: Dept. of El Cauca; in bushy forest, “San José,” San Antonio, Cordillera Occidental, = 2400-2700 m., June 28, 1922, Pennell & Killip, no. 7430 (tyPE, G This plant, at first suspected to be a variety of M. nigropunctulata Hieron., is found to differ from that species (of which a photograph of the type has recently been obligingly supplied to the writer from the Botanical Museum at Berlin) in the following points: its leaves are longer and relatively much narrower, their margins instead of being entire are distinctly denticulate, they are more regularly feather- veined and the veinlets instead of being impressed are prominulent on the upper surface; the internodes are shorter, not exceeding 4-6 cm., and the petioles are scarcely half as long as in M. nigropuncticu- lata. The corollas are deep purple while Hieronymus states the flowers of M. nigropuncticulata to be “cinereo-albidi.” M. lanceolata Hieron. of Ecuador with somewhat similar habit is a smoother plant with entire subcoriaceous less gradually attenuate leaves and some- what more loosely spreading inflorescence, the branches equalling or exceeding the subtending bracts. M. (§ Thyrsigerae) Hitchcockii, spec. nov., fruticosa reclinans; ramis elongatis teretibus dense fulvescenteque villoso-tomentosis 4 mm. vel ultra crassis; internodiis usque ad 14cm. longis; foliis oppositis dato-dentatis (desitzbius baie brevissimis 4-13 mm. inter se distan- tibus) basi late rotundatis paullo supra basin pinnatim vel subpal- matim 5(-7)-nervatis membranaceis ‘supra, juventate appresse villosis maturitate (nervis V reticulatis subtus fulvescente villoso-tomentosis li- 13 em. longis 5.5-6.5 cm. latis; nervis per venas transversas subparallelas conjunctis; petiolo 2-2.5 em. longo dense fulvescenteque tomentoso; paniculis axillaribus oppositis vel terminalibus longe pedunculatis ovoideo-pyramidatis oppositirameis 1-2.5 dm. altis 12-20 em. crassis; bracteis linearibus dense sericeo-velutinis 3-10 mm. longis; capitulis ca. 4 mm. altis 26 ROBINSON graciliter pedicellatis in ramulis ultimis subracemose dispositis; pedicellis plerisque 1.5—2 mm. longis; bracteolis lanceolatis ca. 1 mm. longis ad basin pedicelli affixis; involucri squamis anguste oblongis maturitate brunneis apice rotundatis vix nervatis apicem versus dorsaliter puberulis et margine ciliatis aliter glabris; corollis albis ca. 2.5 mm. longis; tubo proprio ca. 1 mm. longo gracili; faucibus modice ampliatis subcylindratis ca. 1-1.2 mm. longis; dentibus limbi ca. 0.4 mm. longis; achaeniis (immaturis) atrobrunneis glabris ca. 2 mm. longis; pappi setis ca. 30 corollas subaequantibus albidis-—Ecuapor: Prov. Tungurahua: Valley of the Pastaza River between Bafios and Cashurco, 8 hours east of Bafios, alt. 1300-1800 m., Sept. 25, 1923, Prof. A. S. Hitchcock, no. 21,818 (tyP8, Gr.). This species has rather closely the habit of the widely distributed Cordilleran M. Ruiziana Poepp., but differs in having decidedly smaller -heads and florets, and in having the bractlets borne at the base of the pedicel and not as in M. Ruiziana at the summit and closely subtending the head. M. sigerae) inornata, spec. nov., volubilis verisimiliter fruticosa glabra; caule hexagono ad 5 mm. diametro; internodiis 5-23 em. vel ultra longitudine; ramis Recuiien tortis; foliis oppositis oblongo-ovatis acuminatis basi acutis vel subrotundatis margine integerrimis vel obscure remoteque undulato-dentatis utrinque glaberrimis viridibusque concoloribus textura subcarnosis prope basin pinnatim 5-nerviis 8—11 cm. longis 3-6 em. latis; petiolo 2-4 cm. longo, eis ejusdem jugi basi ab annulo suberoso prominenti transverso con- junctis; appendicibus stipuliformibus nullis; panicula laxa interrupta ramosa elongata; inflorescentiis partialibus breviter pedunculatis patentibus 5—20-capituliferis corymbiformibus 2-4 cm. diametro sub- densis; bracteis ovatis 1-3 cm. longis; bracteolis oblongo-ovatis her- baceis ciliatis acutis involucrum subaequantibus ad apicem pedicelli orientibus; pedicellis gracilibus plerisque ca. 2 mm. longis; involucri anguste ovoidei squamis oblongis vix acutatis obscure plurinerviis ca. 6 mm. longis 1.6—2 mm. latis erosis glabris; corollis viridescentibus ca. 5 mm. longis glabris; tubo proprio ca. 2.5 mm. longo basin versus paullo constricto; faucibus brevissimis campanulatis; dentibus limbi oblongis acutiusculis ca. 1.8 mm. longis et 0.8 mm. latis; achaeniis griseis maturitate crassiusculis glabris ca. 3 mm. longis; pappi setis numerosis ca. 100 carneo-albidis corollam maturitate superantibus.— Peru: “on sunny shrubs,” alt. about 1067 m., July 19-25, 1923, J. F. Maebride, no. 5061 (rypx, Field Mus.; tsorypre, Gr.). In its slightly fleshy or succulent character recalling M. carnosa Muschler and M. montana described below, but differing from both in its hexagonal TROPICAL AMERICAN EUPATORIEAE 27 branches, much larger leaves, broader involucral scales, stouter achenes, and in the disposition of its heads in an elongated loose panicle with much less corymbose tendency, a feature which imposes the necessity of placing the species in § Thyrsigerae. Flowers said to be greenish. M. (§ Thyrsigerae) lasiopoda, spec. nov., fruticosa scandens; cauli- bus flexuosis tortis lignescentibus medullosis teretibus 4-5 mm. diametro praecipue nodos versus subsetoso-pubescentibus aliter maturitate subglabratis sed scabratis post exsiccationem leviter costulato-angulatis; internodiis ad 2 dm. vel ultra longitudine; ramis oppositis patentibus; foliis lanceolato-oblongis attenuatis integris basi angustatis distincte cordatis (vel apud folia ramea rotundatis vel obtusis) pinnatim nervatis vel subregulariter penniveniis supra glabris lucidulis minute impresseque reticulatis post exsiccationem fusces- centibus subtus laxe sordideque pubescentibus olivaceis minute punctatis firmiter membranaceis 8-14 cm. longis 3-5.5 cm. latis; petiolo brevissimo robusto -dense setoso-piloso 1-4 mm. longo; eis ejusdem jugi basi annulo obscuro plus minusve incrassato setoso transverse conjunctis; paniculis ramos terminantibus ovoideo-pyra- midatis 1.5-3 dm. longis 8-15 cm. diametro sordide pubescentibus; pedicellis gracilibus plerisque 1-3 mm. longis; bracteis bracteolisque ovato-lanceolatis acutis laxe pilosis parvis; involucri squamis anguste oblongis subacutis 3.5 mm. longis ca. 1 mm. latis maturitate subglabris tenuibus brunnescentibus; corollis glabris, tubo proprio gracili ca. 1.5 mm. longo; faucibus campanulatis ca. 1.4 mm. altis; dentibus limbi deltoideis ca. 0.7 mm. longis; achaeniis gracilibus glabris deorsum decrescentibus acute angulatis fusco-griseis ca. 1.7 mm. longis; pappi setis ca. 37 tenuibus albis sublaevibus corollam subaequantibus.— Peru: a “liana of montafia river-slope,” with flowers odorless and whitish, alt. about 1220 m., Aug. 10-24, 1923, La Merced, J. F. Macebride, no. 5567 (ryPE, Field Mus.; isoryPe, Gr.); also “on sunny brush,” flowers dirty-greenish, alt. about 1220 m., Aug. 27 to Sept. 1, 1923, La Merced, Hacienda Schunke, J. F. Macbride, no. 57: (Field Mus., Gr.). This species keys out to the neighborhood of M. lancifolia Robinson but differs in having broader thinner leaves which are sublucid above and (at least the cauline) distinctly cordate at the narrowed base; furthermore its involucral scales are more obtuse, its panicle less leafy, and its achenes shorter than in M. lancifolia. 'There can be no doubt that it is a distinct species. The name, of course, alludes to the shaggy petioles. M. . Macbridei, spec. nov., fruticosa verisimiliter scandens vel in fruticetis inclinans robusta; caule 6-8 mm. crasso 28 ROBINSON tereti a basibus pilorum delapsorum quasi granulato-scabrato; ramis oppositis patentibus griseo- vel albido-lanatis; internodiis 8-14 cm. longis; foliis oppositis, caulinis ovatis acutis integris vel obsc undulato-dentatis (dentibus 1-1.5 em. inter se distantibus) Sa rotundatis textura coriaceis supra maturitate glaberrimis viridibus rugulosis subtus niveo-tomentosis pinnatinerviis 1.6-2 dm. longis 8-12 cm. latis; nervorum jugis principibus 2-3 supra basin 140 mm. orientibus prorsus curvatis supra sulcatis subtus prominentibus; petiolo 2.8-3.5 cm. longo dense albo-tomentoso, eis ejusdem jugi basi ab annulo crasso conjunctis; foliis rameis ovato-oblongis 7-10 cm. longis 3.5-6 cm. latis; paniculis ramos terminantibus ovoideis magnis oppositirameis 2.5-3 dm. longis 1.5 dm. crassis divaricatim ramosis albo-tomentellis; capitulis ad apices ramulorum plerisque binis vel ternis subsessilibus ca. 8 mm. longis; involucri squamis oblongis ca. 4 mm. longis 1-1.4 mm. latis apice rotundatis dorso subsericeo-puberulis maturitate rigidiusculis; corollis glabris pallide viridibus ca. 4 mm. longis; tubo proprie gracili ca. 1.5 mm. longo; faucibus campanulatis ca. 1.5 mm. longis; dentibus limbi ca. 1 mm. longis deltoideis; achaeniis gracilibus fusco-griseis glabris deorsum decrescentibus 3.2-3.7 mm. longis; pappi setis ca. 30 sordide albidis corollam subaequantibus.—Prru: “on sunny brush, flowers light green,” alt. about 1220 m., Aug. 27-Sept. 1, 1923, J. F. Macbride, no. 5728 (TypPx, Field Mus.; tsotyps, Gr. Available material of this very well marked species is rendered the more striking from the fact that all but the y V beneath by a very curious white porous veil-like coating which on ex- amination by my cryptogamic colleagues, Thaxter and Dodge, proves to be undoubtedly a fungus-mycelium. U ‘nfortunately no fruiting portions have been found and identification is impossible. This close membranelike network completely covers and obscures the normal pubescence in the mature leaves without however greatly changing the shade, for both are a very light gray, in fact nearly white. The Mikania seems to have developed normally and to have matured well ripened achenes notwithstanding the presence of the fungus, which appears to be confined to the lower leaf-surface. M. Machridei may perhaps be best placed next or at least near to M. lanuginosa DC. but differs from that species in its subsessile heads which are cymose rather — subracemose i in disposition, and in its leaves whicl arger, firm ase and nearly glabrous above. In naming this striking species it is a pleasure to commemorate Mr. Macbride’s notable progress in the exploration of the Peruvian flora. TROPICAL AMERICAN EUPATORIEAE 29 M. (§ Corymbosae) montana, spec. nov., scandens paullo lignescens infra inflorescentiam glaberrima; caule tereti articulato ca, 4 mm. diametro flavido-brunneo laevi; internodiis 2-4 cm. longis; nodis ab annulo prominulo suberoso: tenui circumcinctis; foliis oppositis petiolatis ovatis glaberrimis carnosis acutis basi rotundatis et breviter ad insertionem petioli acutatis a basi 5-7-nervatis subtus paullo pallidioribus impresse reticulato-venosis margine utrogue distanter 1-4-cuspidato-dentatis 3-5 cm. longis 1.5-3 em. latis; petiolo ca. 1.5 cm. longo gracili glabro, eis ejusdem jugi ad basin annulo conjunctis; corymbis terminalibus et lateralibus trichotomis ca. 8 cm. diametro conjunctim paniculam laxam foliaceo-bracteatam ena iee granu- loso-puberulis; bract ti m. longis; bracteolis oblongis acutis, a summa parte pedicelli at et in- volucrum aequantibus; pedicellis gracilibus 3-4 mm. longis; capitulis plerisque in cymulos 3 ispositis 1 cm. longis; involucri squamis herbaceo-substramineis oblongis obtusis vel acuti- usculis apice ciliatis dorso glabris vel apicem versus puberulis ecostatis basi paullo calloso-incrassatis; corollis gracillimis glabris; tubo proprio 3 mm. longo; faucibus brevissimis ca. 0.5-0.7 mm. longis; dentibus limbi linearibus ca. 3.5 mm. longis et 0.6 mm. latis; achaeniis glabris gracilibus 4.5 mm. longis; pappi setis ca. 80 Firisteitiilibaas plerisque 4.5-5 mm. longis alkidis vix scabratis——PrErv: a liana on shrub- montafia, Huacachi near Muna, alt. about 1982 m., May 20—-June 1, 1923, J. F. Macbride, no. 4169 (typr, Field Mus., phot. and fragm. Gr.). This slightly succulent species somewhat recalls M. carnosa Muschler, but differs in having its bractlets equal to the involucre instead of one-half its length, involucral scales obtuse instead of acute, corolla-lobes much longer and throat much shorter, and leaves (trian- gular rather than rhombic-ovate) remotely cuspidate-dentate. In many of its characters M. montana resembles the Brazilian M. inro- lucrata Hook. & Arn. but that has firmer unreticulated leaves, broader and obtuse or rounded yy and denser inflorescence. Corymb , spec. nov., volubilis ut videtur her- rete verisimiliter perennis pe ignota); caulibus saltim ramis basi gradatim vel abruptius angustatis vel subrotundatis et ad in- sertionem petioli breviter acuminatis usque ad 8 em. longis et 4 cm. latis paullo supra basin 3-nervatis utrinque glabris membranaceis; petiolis ca. 2 cm. longis supra canaliculatis, eis ejusdem j jugi ab linea paullo prominenti conjunctis; corymbis numerosis oppositis pedun- culatis 6-12 cm. diametro planiusculis vel convexis vel valde rotunda- 30 ROBINSON tis apicem caulis versus inflorescentiam magnam compositam foliaceo- bracteatam formantibus; ramulis ultimis et pedicellis sordide pu- berulis; capitulis ca. 1 em. longis 5 mm. diametro sessilibus vel breviter pedicellatis; bracteolis ovatis acuminatis 3-nerviis ca. 6 mm. longis 3.1nm. latis; involucri squamis anguste lanceolato-oblongis subacutis ca. 8.5 mm. longis et 2 mm. latis dorso tenuiter ‘striato-costulatis apicem versus paullo patentibus; corollis glabris albis; mr p o 2.5 mm. longo gracili; faucibus campanulatis ca. 1-1.2 m siites: Aeabibin limbi ovatis ca. 1.7 mm. longis; achaeniis glabris griseis vel olivaceis vel brunnescentibus ca. 3 mm. longis deorsum decres- centibus; pappi setis sordide albidis ca. 80 capillaribus apicem versus non incrassatis sublaevibus—Ecuapor: Province Tungurahua; Valley of Pastaza River between Bajios and Cashurco, 8 hours east of Bafios, alt. 1300-1800 m., ner 25, 1923, Prof. A. S. Hitchcock, no. 21,875 (TYPE, Gr.). M. puncrata Klatt. Bull. oe Bot. Belg. xxxi. 195 (1893). This species, hitherto known from Southern Mexico to Venezuela and Bolivia, seems never to have been recorded from Peru. It has now been found by Mr. Macbride at two Peruvian stations, namely La Merced, on a sandy valley floor, alt. about 610 m., Aug. 10-24, 1923, no. 5461 (Field Mus., Gr.) and “on sunny shrubs,” alt. 1040 m., at Pampayacu, Hacienda at mouth of the Chincao River, July 19-25, 1923, no. 5062 (Field Mus., Gr.); also in less characteristic form as a liana on trees of open wood at about 1220 m., Rio Huallago Cafion below Rio Santo Domingo, June 3, 1923, no. 4219 (Field Mus., Gr.). The last mentioned specimen, while agreeing with the species in important features, has leaves rounded rather than cordate at base _ and obtusely angled rather than lobed laterally. It appears, however, to be only a leaf-variant. Indeed the leaves are chiefly those in or very near the inflorescence and similar though smaller floral leaves may be sometimes found on the more northern material of the species. Thyrsigerae) rivularis, spec. nov., fruticosa vel vix lignescens sequent glabra; caulibus teretibus fuseo-brumneis post exsiccationem multicostulatis; internodiis 9-17 cm. longis; foliis oppositis oblongis longe acuminatis integerrimis margine revolutis basi acutis I glabris subtus paullo —— minute papillosis supra basin 3 nervatis (cum venis 2 obscuris intra-marginalibus) 9-11 em. uel 3-4 em. latis membranaceis; petiolo 1—-2.2 cm. longo; paniculis axil- laribus et terminalibus oppositirameis foliaceo-bracteatis; capitulis ca. 9 mm. longis in glomerulis globularibus ca. 3 cm. diametro dis- positis, capite terminali arcte sessili, lateralibus distincte pedicellatis pedicellis 1-3 mm. longis; bracteolis ovato-lanceolatis acuminatis NEW PLANTS OF PORTUGUESE WEST AFRICA 31 2.5 mm. longis ca. 0.9 mm. latis; involucri squamis ovato-oblongis apice rotundatis ciliolatis ca. 4.8 mm. longis 1.2 mm. latis dorso glabris leviter costulatis; corollis ca. 6 mm. longis glabris; tubo proprio 2.5 mm. longo; faucibus cylindratis vix ampliatis 3 mm. longis; dentibus limbi acutis deltoideis patentibus 0.5 mm. longis; achaeniis maturis 2 mm. longis crassiusculis argute angulatis griseo-olivaceis glaber- rimis lucidis; pappi setis rubellis ca. 45 apice paullo incrassatis obtusis penicillatis—Prru: liana-like, on stream-brush; flowers white, turning greenish, about 610 m. alt., Aug. 10-24, 1923, J. F. Macbride, no. 5414 (typr, Field Mus.; tsoryps, Gr.). This species in foliage has some resemblance to M. Stuebelii Hieron. but that has a different inflorescence (the lateral as well as the terminal heads being subsessile), longer and pubescent achenes, less blunt pappus-setae and other differences. II. NEW PLANTS OF PORTUGUESE WEST AFRICA COL- LECTED BY MRS. RICHARD C. CURTIS. By I. M. JoHNsTon. THE species here described are a part of an excellently prepared and very interesting collection of African plants made by Mrs. Richard C. Curtis, of Boston, during the summer of 1923, when ac- companying her husband as well as his father, Mr. Charles P. Curtis, and brother, Mr. Charles P. Curtis, Jr., on a hunting trip to the interior of Kenya Colony and Angola. The plants are not only carefully chosen and well preserved, but are accompanied by full ecological, habital and color notes. Hence they form a most welcome addition to the iaicaes representation of the tropical African flora available in Ameri The most interesting of the plants collected by Mrs. Curtis were those secured in central Angola, for here the expedition pene-trated a region which iat to have been previously unexplored bot- anically. A considerable part of the Angola collection was made in the range of the Giant Sable Antelope, between the Cuanza and Loando Rivers, about 100 miles south of their confluence. The party started from Dondé or Bella Vista, in Bailundo, near the end of the railroad east of Angola Bay, from there going by auto north- westward 120 miles to Capango. The march towards the Sabel Country started September 11th; the Cuanza River was crossed on the 14th, and a hunting camp formed on the 15th. The return march began September 22nd, and Capango was reached on the 26th, oo JOHNSTON the Cuanza having been crossed on the 24th. The Angola collection includes 311 numbers and was given by Mrs. Curtis to the Gray Herbarium. The Kenya collection is not as yet fully identified. So far as studied, it seems likely to contain many plants not previously represented in American herbaria, but relatively few, if any, that have not been scientifically described. Of the Kenya collection the ligneous plants have been deposited in the herbarium of the Arnold Arboretum and the herbaceous material, amounting to some 442 numbers, in the Gray Herbarium. Mrs. Curtis also made for the Gray Herbarium small collections during brief stops at Zanzibar, Mozambique, Lorenzo Marques, Walfisch Bay and Liideritz Bay, containing characteristic plants of these localities. Acrospira Curtisiae, sp. nov., glabra herbacea 20-24 cm. ‘alta e cormo obliquo depresso 15-25 mm. crasso rugoso squamoso apicem versus radiculoso producta; foliis 3—5 linearibus nervosis 4—5 mm. latis; pedunculo foliis destituto gracili suleato infra flores 13-15 em. longo; racemo 8-10-floro simplici stricto 10-15 mm. crasso bracteis ovatis acuminatis scariosis instructo; floribus alkis vel ochroleucis stricte et 2-6 mm. longe pedicellatis; perianthii lobis lanceolatis acutis ca. 10-11 mm. longis, exterioribus ca. 3 mm. latis in carina 3-nervatis; staminibus 6 subaequalibus, filamentis albis ca. 5 mm. longis linearibus, antheris luteis linearibus introrsis 44.5 mm. longis basifixis —-* —* basi profunde cordatis; capsula ovoidea lobata glabra 4-5 . crassa trivalva loculicide dehiscenti transverse delicateque - al deeeiac a. seminibus multis atris dense minute acuminate muriculatis irregulariter lobulatis ca. 2 mm. latis.— ANGOLA: east of the Cuanza River, Sept. 23, 1923, Anita G. Curtis 333 (TYPE, Gray Herb.).—Acrospira Curtisiae differs from A. aspho- deloides Welw., the other Angolan species, in its much smaller size, very much narrower leaves, and longer filaments; from A. —— DeWild., of the Belgian Congo, it differs in lack of puk leaves, iad much less developed bracts; while from A. lilioides ¢ rae of French Equatorial Africa, it may be readily distinguished by its narrower leaves and narrower lanceolate perianth-lobes. The plant has a depressed oblique corm with dark wrinkled and scarred surface bearing over the upper portion numerous fleshy, often fusiform, root- lets. Russelliae, sp. nov., suffruticosa; caulibus pluribus rhizomate crassissimo lignoso erumpentibus 3-4 dm. altis subsim- plicibus glaberrimis basi squamis imbricatis instructis; foliis paucis imparipinnatis 12-18 em. longis 4-5 cm. latis glaberrimis 9-12-jugis 1.5-2 em. longe petiolatis; foliolis coriaceis oppositis subsessilibus NEW PLANTS OF PORTUGUESE WEST AFRICA 33 glaberrimis oblongo-lanceolatis 6-11 mm. separatis 3-4.5 mm. longis -10 mm. latis superioribus paullo gradatim reductis apice obtuse rotundatis basi oblique truncatis vel subcordatis subtus paullo pal- lidioribus costa paullo excentrica; stipulis deciduis inconspicuis reniformi-deltoideis 3-4 mm. latis venosis; inflorescentia pauce laxeque ramosa 3-6 cm. longa ca. 3 cm. crassa fulvo-tomentulosa _folio multo breviori; bracteis ovatis vel oblongo-ovatis 2-3 mm. longis deciduis, pedicellis 2-4 mm. longis tomentulosis; bracteolis obovatis ca. 6 mm. longis 44.5 mm. latis coriaceis extus tomentulosis apice rotundatis; sepalis 5 inaequalibus imbricatis oblongis vel ovatis 2-3 mm. longis 1-1.5 mm. latis pallidis conspicue dense fusco- ciliatis; petalis 3-5 inaequalibus filiformi-subulatis 0.5-1.5 mm. longis pallidis; staminibus 10-15 ca. 10-12 mm. longis glabris; ovario dense villoso ca. 2 mm. longe stipitato; legumine ignoto.—ANGoLA: locality uncertain, but probably from the Sable Country, Anita G. Curtis 428 (typE, Gray Herb.).—This remarkable species is an undershrub with annual shoots from a woody rootstock and in that regard resembles the common habit of Cryptosepalum. In floral structures, however, the new species is clearly typical of Brachy- stegia, and I see no good reason for not considering it a member of that genus. It is noted that Burtt Davy & Hutchinson, Kew Bull. 1923, 130, in their recent revision of Brachystegia recorded no species with such a habit. Arborescent plants and undershrubs are ac- commodated, however, in the closely related Cryptosepalum. In the revision referred to, B. Russelliae keys out with B. longifolia Benth. of Nyasaland, a species which differs in its loosely branched ar- borescent habit, larger leaves, narrower bracteoles and sepals, and more stipitate, less villous ovary. Fie. 4. At the request of Messrs. C. P. and R. C. Curtis this plant has been named in memory of the late Margaret P. (Mrs. Robert S.) Russell of Boston at whose suggestion the botanical collecting of the expedition was undertaken. Cryptosepalum Curtisiorum, sp. nov., humilis; caulibus multis erectis simplicibus gracilibus e rhizomate crassissimo lignoso erum- pentibus dense puberulentis fuscis 1-2 dm. altis basi a squamis latis | scariosis instructis; foliis imparipinnatis breviter petiolatis 14—20- jugis 4-5(-9) em. longis 15-25 mm. latis oblongis vel rare ovatis; rhachi puberulenta; foliolis subsessilibus oblongis vel lineari-oblongis 10-18 mm. longis 2-3(-—4) mm. latis glaberrimis margine integerrimis et sparse ciliatis apice rotundatis vel obtusis basi obliquis; stipulis lanceo-linearibus scariosis 9-14 mm. longis libris; racemis terminalibus 10-25-floris 4-6 cm. longis 20-25 mm. crassis puberulentis, bracteis 34 JOHNSTON obovato-oblanceolatis scariosis, iedtteitbe strictis 8-15 mm. ‘longis; bracteolis petaloideis albis obovatis 5-8 mm. longis apice rotundatis extus puberulentis; petalo 1 oblongo vel obovato 5-8 mm. longo albo; staminibus 3 glabris; ovario compresso stipitato in lateribus glabro in marginibus plus minusve dense ciliato; legumine stipitato valde obliquo plano oblongo vel ovato 2-3 cm. longo glabro in margini superiori angustissime bialato.— ANGora: dry hard soil near Cuanza River, Sept. 13, 1923, Anita G. Curtis 186 (rypE, Gray Herb.); Giant Sable Country, east of Cuanza River, 900 m. alt., Anita G. Curtis 207, 212, 27 1a, 281, 290.—Closely related to C. ee Welw., but iéacialy distinguished by its more numerous flowers, better ae veloped inflorescence, and much longer pedicels. This species is named in honor of Messrs. Charles P. Curtis and Richard C. Curtis, whose careful planning and skillful direction of their difficult j journey into the interior of Angola rendered possible the discovery of the plants here described. It is learned that Cryptosepalum Curtisiorum is one of the plants chiefly grazed by the Giant Sable Antelope. Fic. 3. Crotalaria bicolor, sp. nov., perennis; caulibus numerosis sub- simplicibus strictis 15-25 cm. altis minute sulcato-striatis breviter villosis canescentibus e radice crassa orientibus; stipulis nullis; foliis trifoliolatis 2-5 mm. longe petiolatis, inferioribus gradatim reductis; foliolis lineari-oblanceolatis acutis, brevissime petiolulatis supra glabris subtus strigose pubescentibus, terminali 8-16 mm. longo 0.7-2.0 mm. lato; racemis terminalibus densiusculis 4-8 cm. longis 10-13 mm. crassis basem versus cum foliis normalibus bracteatis aliter minute setaceo-bracteatis; pedicellis ca. 3 mm. longis ascenden- tibus; calycibus 3-5 mm. longis, dentibus lanceolatis tubum exceden- . tibus; corolla calycem fere duplo superante; carina flava 5-6 mm. longa margine inferiori rectangulari-inflexa inde in rostrum rectum at- tenuata; vexillo sparse pubescenti rubro-brunnescenti-lineato 5-6 mm. longo; ala flava brunnescenti-lineata; leguminibus obliqui-depresse globosis sessilibus erecto-pedicellatis 3-4 mm. longis monospermis sparse adpressi-pubescentibus.—ANGOLA: just east of Cuanza River, . alt. 900 m., Sept. 18, 1923, Anita G. Curtis 211 (TypE, Gray Herb.), Giant Sable Country, alt. 900 m.; Sept. 21, 1923, Anita G. Curtis 252.—Ap hieba Welw., but having smaller Giese nts hath standusd and wings : streaked with purple or reddish brown, more numerously flowered inflorescence leafy-bracted only at very base, inconspicuously pubescent larger leaves, and lower decidedly perennial habit. The plant has a strong, thick, deep tap- root which distinguishes it from most of the other members of the NEW PLANTS OF PORTUGUESE WEST AFRICA 35 § Sphaerocarpae. The keel and standard of the corolla are subequal. The keel is yellow, and the standard and wings are yellow with many longitudinal brownish-purplish lines. Fie. 2. Indigofera latipinna, sp. nov., fruticosa erecta; caulibus pluribus ca. 4 dm. altis lignosis teretibus maturitate dilute brunnescentibus glabris basem versus 6 mm. crassis juventate adpresse villosis canes- centibus; petiolis ca. 1 ca. longis canescenti-villosis; foliis impari- pinnatis in rhachi breviter villosis canescentibus, primariis trifoliolatis ceteris quinque-foliolatis; foliolis integerrimis cuneato-obovatis 10-25 mm. longis 7-16 mm. latis ca. 1 mm. longe petiolulatis bicoloribus supra glaberrimis subtus dense canescenter tomentosis basi cuneatis apice truncatis vel late obcordatis vel retusis conspicue mucronatis, lateralibus oppositis, terminalibus paullo minoribus; stipulis lineari- lanceolatis 5-6 mm. longis deciduis; racemis 5-7 cm. longis axillaribus multifloris ascendibus quam folia paullo brevioribus vel ea paullo excedentibus; rhachi et pedicellis et calyce canescentibus adpresse villosis; floribus 0.5—-1.0 mm. longe pedicellatis; calycis 15 mm. longi obliqui tubo brevi, dentibus obtusis deltoideis vel late ovatis; corolla calycem fere quadruplo superante ca. 12 mm. longa flavescenti vel ochroleuca; vexillo 9 mm. longo elliptico plicato apice rotundato brevissime unguiculato extus adpresse sparseque fusco-pubescenti quam carina multo breviori; alis lineari-oblongis ca. 8 mm. longis; carina 11-12 mm. longa acuta (in rostrum non producta); legumine ignoto.—ANGOLA: near Cuanza River, a day’s march from Copango, Sept. 24, 1923, Anita G. Curtis 360 (typE, Gray Herb.)—A member of Baker’s group Tinctoriae, Jour. Bot. xxxii. 263 (1903), in which it is readily distinguished from most of the species by its woody stems and particularly by its few very large conspicuously bicolored leaflets. It is most nearly related to J. rhynchocarpa Welw. and I. nambalensis Harms, but differs from both in the number, size, outline and pubescence of the leaflets Aeschynomene Curtisiae, sp. nov., » perennis herbacea ~ 4 one, Po hat RES setulosis stricte ramosis; cite: “bnew acuminatis coriacels paralleliter nervatis basi truncatis; foliis paucis strictis 10-12 mm. longe petiolatis 4-9 em. longis ca. 1.7 cm. latis; rhachi compressa ulata glabrata; foliolis 8-20 oppositis vel rariter alternis coriaceis valde nervatis glabris 0.5-1.0 mm. longe petiolulatis oblongis vel elliptico-oblongis 15-20 mm. longis 4-6 mm. latis integerrimis apice —- conspicue mucronatis basi obliquis subcordatis; panicula termi ca. 15 em. crassa; pedicellis ascendentibus glabris 5-10 mm. cae calycis glaberrimi 5 mm. longi labio superiori profunde 36 JOHNSTON (ca. 1 mm.) emarginato 3-3.5 mm. lato ovato, labio inferiori integer- rimo ca. 3 mm. lato obtuso; vexillo oblongo ca. 1 mm. longe unguicu- latis, lamina 5 mm. longa 2 mm. lata flavescenti conspicue purpureo- nervata apice rotundata carinam subaequanti; alis flavis oblongo- lanceolatis acutis ca. 6-7 mm. longis 2 mm. latis 1 mm. longe ungui- culatis quam petala cetera longioribus; carina 2.5 mm. lata flavescenti purpureo-nervata; ovario glabro stipitato 2-ovulato; legumine glabro valde compresso venoso saepe 2-loculato 10-14 mm. longe stipitato; loculis 7-9 mm. latis 10-14 mm. longis ca. 0.7 mm. crassis marginatis sed non alatis—ANnco xa: Giant Sable Country, 900 m. alt., Sept. 23, 1923, Anita G. Curtis 321 (rypr, Gray Herb.).—Apparently most nearly related to A. Baumii Harms but glandless, much less pubescent, and with a lower herbaceous habit, entire lower calyx-lip, smaller flowers, and wingless pods. It is probably also related to A. siifolia Welw., but that has more pubescence and has many more leaflets. The new species is a handsome one because of its firm clean yellowish- green seemingly glabrous herbage, and loose panicles of small yellowish or ochroleucous flowers and later large conspicuously stipitate broad glabrous fruit. Fic. Dolichos linearifolius, sp. nov. perennis erectus 30-45 cm. altus; caulibus paucis herbaceis in sicco angulatis et striatis simplicibus vel ascendenter ramosis pallide viridibus sparse adpresseque villosis e radice crassa lignosa conico-fusiformi orientibus; stipulis triangulari- oblongis sessilibus glabris striatis subpersistentibus 4-5 mm. longis 24 mm. latis; foliis pinnatim trifoliolatis 8-16 mm. longe petiolatis, foliolis linearibus longitudinaliter trinerviis ca. 1 mm. longe petio- lulatis basi cuneatis apice acutis juventate dense argenteo-villosis maturitate glabratis terminalibus usque ad 5 cm. longis et 6 mm. latis; inflorescentia ante folia evoluta vel cum foliis novellis vel nondum plane evolutis coaetanea; floribus 5-7 mm. longe pedicellatis in axillis bractearum vel foliorum fasciculatis; calyce campanulato sparse villoso 4-5 mm. longo irregulari, dentibus brevibus 2 superiori- bus in unum excisum connatis, lateralibus lanceolato-ovatis acutis, infimo ceteris paullo longiori lanceolato acuto tubo breviori; pro- phyllis 2 infra calycem insertis filiformibus 2-3 mm. longis; corolla glabrata 7-9 mm. longa (in sicco lilacea) calyce pluries longiori; vexillo suborbiculari emarginato ca. 1 mm. longe unguiculato 9-10 mm. lato; alis carinam paullo excedentibus oblique ca. 8 mm. longis; carina curvata obtusa ca. 3 mm. longe unguiculata, laminis 6 mm. longis; ovario lineari-lanceolato supra basin dilatato apicem versus _attenuato apice 3-4 cm. longo ca. 8 mm. lato compressissimo.— AnGoLa: Between Cuanza and Loando Rivers, alt. ca. 900 m., Sept. NEW PLANTS OF PORTUGUESE WEST AFRICA 37 19-22, 1923, Anita G. Curtis 230 (TypE, Gray Herb.), 216, 288.— This species belongs to Harms’s § T'rinervati and appears to be most nearly related to D. trinervis De Wild. of Katanga. It differs, how- ever, in being less pubescent and in having short calyx-lobes that do not surpass the length of the calyx-tube. Fic. 8. Polygala congestiflora, sp. nov., annua herbacea; caule erecto striato in sicco atrato puberulo 4 dm. alto, ramis numerosis 10-15 cm. longis ascendentibus subsimplicibus angulatis floriferis; foliis numerosis oblanceolatis firmis integerrimis uninerviis glabratis con- coloribus 1.5—2 em. longis 3-4 mm. latis in sicco atropurpurascentibus apice breviter acuminatis basi attenuatis, racemis congestis sub- denseque ciliato; sepalis anterioribus oblongis acutis ciliatis 3-nervatis glabris supra medium connatis 2.5 mm. longis; alis oblique ovatis ca. 8 mm. longis ca. 5 mm. latis flavis glabris basi abrupte lato-ungui- culatis apice acutis margine basin versus inconspicue ciliatis; carina flavescenti 6 mm. longa ampla abrupte 1 mm. longe unguiculata; cristis plurifidis ca. 2 mm. longis; petalis superioribus obovato- oblongis ca. 3.5 mm. longis apice retusis basin versus abrupte breviter unguiculatis; antheris 6 oblongis glabris quam filamenta 4—6-plo brevioribus; filamentis glabris duobus interioribus sterilibus; stylo curvato glabro ca. 5 mm. longo apicem versus vaginato abrupte recurvo cucullato; stigmate refracto papilloso; ovario glabro oblongo- elliptico vix alato-—ANGOLA: vicinity of Capango, Sept. 27, 1923, Anita G. Curtis 419 (typ, Gray Herb.).—A well marked species be- longing in the “Subsect. Deltoidea § Tetrasepala” of Chodat. Its closest relation appears to be P. Dekindtii Girke, from which it differs in its broader leaves, larger decidedly acute wings and oblong upper petals. Its style is almost exactly like that of P. liniflora Boj. and P. acicularis Oliv. as they are illustrated by Chodat, Monog. Polya t. 29, fig. 2 & 8 (1893). Fie. 5. Triumfetta glabrata, sp. nov., herbacea perennis; caulibus numer- osis erectis 20-35 cm. altis glabritis gracilibus in basi a squamis latis obtusis ornatis e radice crassa orientibus infra medium foliosis; foliis simplicibus 3-6 cm. longis 1.54 mm. latis anguste lanceolato-lineari- tis vel obtusis basin versus attenuatis subtus pallidioribus maturitate glaberrimis vel cum pilis stellatis paucis obtectis juventate canescentibus leviter stellato-tomentosis; stipulis filiformibus deciduis 2~4 mm. longis; inflorescentia inali paniculata ascendenter pauciramosa supra folia valde projecta ca. 38 JOHNSTON 15 cm. longa 5 em. crassa; floribus in cymas plurifloras distantes aggregatis; pedicellis 3-7 mm. longis sli laa sepalis ca. 7 mm. longis late oblanceolato-linearibus ca. 1.5 mm. latis extus stellato- tomentulosis intus glabris 5-7-nervis supra Bree reflexis, cornu 0.3—-0.5 mm. longo; petalis obovatis 5-6 mm. longis apice rotundatis basi late unguiculatis; staminibus 25-35; disco et ovario villosis; fructibus ignotis.—ANGoLA: east of Cuanza River, Sept. 20, 1923, Anita G. Curtis 231 (TypE, Gray Herb.)—Obviously related to T. Welwitschit Masters from which it is distinguishable by its lack of pubescence, inconspicuous stipules, and smaller leaves and flowers. Ochna angolensis, sp. nov., fruticula 1-5 dm. alta glaberrima e radice crassa lignosa oriens; caulibus pluribus 1-2 mm. crassis sub- simplicibus erectis; foliis subcoriaceis oblongo-oblanceolatis vel ob- lanceolatis integeirimis vel apicem versus parcissime obsoleteque sertulatis 5-8 cm. longis 12-24 mm. latis 2-4 mm. longe petiolatis apice cbtusis mucronatis basi cuneatis, nervis lateralibus majoribus 18-24 angulo acuto in costa insidentibus, venis numerosis in costa subrectangulariter impositis atque a nervis transversis reticulatis, nervis venisque supra prominulentibus; stipulis deciduis lineari- lanceolatis 3-5 mm. longis; floribus flavis in pseudoracemos 4—7- floros laxos 2-6 em. longos 2.5-5 em. crassos 2-5 mm. longe pedun- culatos axillares dispositis; racemis conjunctim inflorescentiam com- plexam laxiusculam multifloram 2.5-4.5 cm. crassam formantibus; pedicellis 5-8 mm. longis; pedunculis unifloris 3-5 mm. longis; bracteis inconspicuis deciduis subulatis 1-3 mm. longis; sepalis obovatis apice rotundatis petalis aurantiaco-flavis oblonge obovatis 9-14 mm. longis; staminibus 25-35 quam petala brevioribus; filamentis filiformibus 3-4 mm. longis quam antherae 2-3-plo longioribus; antheris lineari- oblongis longitudinaliter dehiscentibus; ovario 5-lobo glabro.—An- GoLA: Dondé, Bailundo, ca. 1500 m. alt., Sept. 8, 1923, Anita G. Curtis 158 (typ, Gray Herb.), 136; dry, hard soil at Cuanza River, Sept. 13, 1923, Anita G. Curtis 171; west bank of Cuanza River, Sept. 25, 1923, Anita G. Curtis 372—Related to O. Debeerstii De Wild., which also occurs in Angola, but differing in its broader, larger, obtuse, subentire leaves and larger, more floriferous racemes of distinctly larger flowers. Oil obtained from the leaves of the plant is said to be used in ointments. Sym: n articulatus, sp. nov., glaberrimus herbaceus verisimiliter annuus; caule erecto gracili laxe et ascendenter alterni- ramoso subtereti ca. 35 mm. alto infra medium sparse alternifolia; ramis ramulisque a bracteis lanceolatis caducis 2-3 mm. longis suf- fultis; foliis non visis ut videtur caducis saltim ab exemplis Dominae NEW PLANTS OF FORTUGUESE WEST AFRICA 39 Curtisiae deficientibus; inflorescentia racemosa terminali laxa 10-15 cm. longa 2-3 cm. crassa; pedicellis laxe ascendentibus 8-10 mm longis a bracteis minutis deciduis suffultis medium versus articulatis, parte superiori fusce et sparse stellato-setulosa decidua, parte inferiori persistenti glabrata; calyce campanulato extus glandulifero fructifero 6-7 mm. longo ca. 3.5 mm. crasso; lobo superiori late ovato breviori; lobis inferioribus anguste triangularibus acutis; corolla alba vel dilute rosea calyce 4—5-plo longiori glabra extus glandulifera, tubo cylindrico supra basem abaxillariter valde gibboso ca. 4 mm. longo 1.5-2 mm. crasso, faucibus late campanulatis ca. 4-5 mm. longis 3-4 mm. crassis, labio inferiori ca. 10 mm. longo ovato ca. 8 mm. lato obtuso, lobis labii superioris 4, exterioribus suborbicularibus 2 mm. longis, interi- oribus oblongo-ovatis ca. 1.1 mm. longis inconspicuis ab exterioribus imbricatis obscuratis; filamentis linearibus ca. 8 mm. longis basem versus ca. 1.5 mm. longe connatis; nuculis globosis ca. 1.8 mm. crassis rubescentibus subopacis sublaevibus.—ANGOLA: east of the Cuanza River, Sept. 23, 1923, Anita G. Curtis 309 (1ypE, Gray Herb.).—Differing from S. insolitus (Wright) Hieron. in its larger and differently proportioned, decidedly sigmoid, pale corollas which have a well developed proper tube and proportionately smaller upper lip. In S. articulatus the outer pair of lobes on the upper lip are much the largest, being roughly orbicular in outline and surpassing the oblong- ovate inner lobes, as well as strongly overlapping and somewhat hiding them. The peduncles of the reduced racemose panicles bear only a single flower apiece, and hence the pedicel appears about twice its real length and is apparently articulate near its middle. Upon maturing the pedicel and calyx fall away leaving the rhachis of the inflorescence beset with the persistent peduncles. Fic. 7. Ocimum cuanzae, sp. nov., perenne herbaceum 2.5-4 dm. altum; caulibus pluribus pauce stricteque ramosis tetragonis sulcatis albide septateque hispido-villosis; foliis firmis lanceolatis vel oblongo- lanceolatis 4-11 em. longis 1.5-3.5 cm. latis glaberrimis vel spar- basi abrupte attenuatis subtus pallidis a conspicue nervatis margine grossidentatis ciliatis; petiolis 10-13 mm. longis sparse villosis; racemis multifloris 3-8 cm. longis 10-15 mm. crassis 1-3.5 em. longe pedunculatis terminalibus vel axillaribus, verticel- lastris 4—5-floris; pedicellis 1-4 mm. longis dense villosis; calyce villoso purpurascenti nutanti, tubo ee commit 5-6 mm. longo, dente superiori ovato-orbiculari acuminato decurren i, dentibus lateralibus triangulari-ovatis sehen acuminatis, ack in- ferioribus subulatis elongatis; corolla 10-12 mm. longa extus pubes- 40 JOHNSTON centi, tubo calycem floriferum aequanti 5-6 mm. longo, labiis tubo brevioribus extus pallide glanduliferis pubescentibus, inferiori ob- longo-lanceolato ca. 5 mm. longo apice bilobato, lobulis ca. 1 mm. longis oblongo-ovatis, labio superiori subrectangulari 2.5 mm. longo et lato apice ca. 1 mm. profunde trilobatis, lobulis ovatis; staminibus 4 inclusis, filamentis exappendiculatis glabris 2 mm. longis linearibus. — ANGOLA: near the Cuanza River, Sept. 24, 1923, Anita G. Curtis. 342 (ryPE, Gray Herb.), 343.—Belonging to the § Gymnocimum and ap- parently most nearly related to 0. laxiflorum Baker, from which, to - judge from descriptions, it differs in its larger, more elongate, firm dentate leaves, and less open inflorescence. Fic. 6. Among the ferns collected by Mrs. Curtis and determined at the Gray Herbarium by Mr. C. A. Weatherby one appears to require a new combination of names to designate it in accordance with the International Rules of Botanical Nomenclature. At Mr. Weather- by’s request this nomenclatorial Serpe with explanatory synonymy may be here placed-on record as follow. THELYPTERIS PALUSTRIS Schott, var. squamigera (Schlecht.) Weatherby, n. comb. Aspidiwm Thelypteris 8. squamigerum Schlecht. Adumb. 23, pl. 11 (1825). A. squamuloswm Kaulf. ex Schlecht. 1. c. (in synonymy). Lastrea squamulosa Presl, Tent. 76 (1836). Ne- phrodium squamulosum Hook. f. Fl. New Zealand, ii. 39 (1855). A. Thelypteris 8. squamuligerum Mett. Abh. Senckenb. Naturf. Ges. ii. 112 (1858). N. Thelypteris B. squamulosum Hook. Sp. Fil. iv. 88 (1862). N. Thelypteris, var. 8. squamuligerum Sim, Ferns S. Africa, 180 (1892). Dryopteris Thelypteris, var. 8. sguamuligerum [sic] Sim, Ferns South Africa, ed. 2, 102 (1915). ’ peRes OF bis: sp. X 3; 2a. mat m ae tisiorum icc loa n. sp. X 4; da. petal < 1, 3b. submature pod x + 3c. m 2. Brachystegia —— Johnston, P a; 4a. flower after femacieel of bracteoles x 2; 4b. bracteole x if he Pick PuaTe2. Fig.5. Polygalacongestiflora Johnston n. sp. X 3; 4a. wing . A; Fig. 6. Ocimum fag meg Johnston, n. sp. X 4; 6a. a large leaf . aD iting calyx Fig. 7. Symphostemon articulatus Johnston, wae 7a. side view of Prists x 1. Fig. 8. Dolichos linearifolius Fabnetaa n. Sp., in very immature foliage x }. Conrris. Gray Hers. LXXIII. PLATE 1. A. Brackett del. 1. AESCHYNOMENE CuRTISIAE Johnston, spec. nov. 4. BracuysteciA RusseLuian Johnston, spec. nov. Contris. Gray Hers. LXXIII. } m | IZ \@ ‘in | YF oye 4, GZ Ih ey) A. Brackett del. 5. PoLYGALA CONGESTIFLORA Johnston, spec. nov. 6. OcimuM cUANZAE Johnston, spec. nov. 7. SyMPHOSTEMON ARTICULATUS Johnston, spec. nov. 8. DoLicuos LINEARIFOLIUS Johnston, spec. nov. III. ON SOME SOUTH AMERICAN PROTEACEAE. By I. M. Jounston. _ Euplassa popidiontalls, ep. nov., arborea; ramis Juv entatis ia oO alternis Henosle’ foliis al alternis 30— 35 em. longis 13-15 cm. latis patenti-erectis 4-jugis; petiolis cum rha- chibus 20-23 cm. longis minute ferrugineo-velutinis striatis; foliolis op- positis vel alternis breviter petiolulatis obovatis vel obovato-oblongis 6-15 em. longis 35-70 mm. latis obtusis integerrimis glabratis venosis bicoloribus supra nitidis subtus opacis, superioribus oppositis si bus; stipulis nullis; racemis 15-20 em. longis 3-3.5 cm. crassis foliae brevioribus dense strigosis in axillis superioribus siitartis: floribus albis geminatis patenter 4—5 mm. longe pedunculatis 1-3 mm longe pedicellatis; alabastris rectis cylindricis 10-12 mm. longis basem versus dilatatis apicem versus axillariter gibbosis strigoso-canescenti- bus; sepalis 4 inaequalibus lineari-spathulatis intus glabris apicem versus abrupte acute subcucullate ovato-dilatatis; antheris subsessili- bus ovato-oblongis acutis bilocularibus albis; ovario glabro hn -Spar- sissime villoso; stylo glabroapicem g , $e ] = r olandiuhe 4 earnncic depresso laterali rubescenti oblongo; glandulis 4 glabris; fructibus ignotis.—Ecuapor: gold mine ps Zaruma, between Portovelo and El Tambo, Prov. Oro, alt. 600-1000 m., Sept. 2, 1923, Hitchcock, 21,311 (vyPE, Gray Herb Appasaay most nearly related to E. inaequalis (Pohl) Engler, of eastern Brazil, from which it is readily distinguished by its well developed peduncles, straight style, longer and more distinctly lateral pubescent stigma, less pu- bescent longer sepals, and 4-jugate leaflets. It is the first species to be reported from west of the Andes and the only one outside of Brazil and Guiana. Euplassa bahiensis (Meisn.),comb.nov. Adenostephanus bahiensis Meisn. in Mart. FI. Brasil. v. pt. 1, 94 (1855). Euplassa incana ee. comb. nov. 06 JOHNSTON attachment of the nutlets, as well as its unappendaged corollas, are similar to those found in Pulmonaria and Nonea, particularly in the former genus. In the attachment of its nutlets Ixthodora is clearly characteristic of the Anchuseae. Besides Lithodora fruticosa (L.) Griseb., L. hispidula (Sibth. & Sm.) Griseb , and L. oleifolia (Lapeyr.) Griseb. there are the following species which have not been properly named. Lithodora consobrina (Pomel), comb. nov. Lithospermum con- sobrinum Pomel, Nouv. Mat. FI. Atl. 296 (1874) L. diffusa (Lag.), comb. nov. Lithospermum diffusum Lag. Varied. Ciene. iv. pt. 4, 39 (1805); Gen. Sp. Pl. 10 (1815); Rouy & Fouc. FI. Fr. x. 314 (1908). Lithospermum prostratum Lois. Fl. Gall. ed. 1, 105 (1806). Lithodora prostrata Griseb. Spicil. Fl. Rumel. ii. 531 (1844). L. rosmarinifolia (Ten.), comb. nov. Lithospermum rosmarini- folium Ten. Prodr., Suppl. ii. 66 (1811-13); Fl. Nap. i. t. 114 (1811- sen ii (Heldr. ), comb. nov. Lithospermum Zahnii Heldr. in oy Verh. Zool.-Bot. Ges. Wien xlix. 190 (1899); Consp. FI. Graecae. ii. 347 (1902). Titeipieaaen fruticosum of Sibth. & Sm. Fl. Graeca ii. 52, t. 161 (1813). 27. Elizaldia Willk. Strand- u hapa Iber. Halbins. 128 (1 85 2). The 3 species of this western Mediterranean genus have been usually referred to Nonea, but in having a slightly different habit, and stamens placed high in the corolla-throat and exserted, they seem generically distinct. The generic name, Phaneranthera, is a synonym of Elizaldia, but appears to have never been legitimately established. The founding of the genus is usually attributed to Meisner, Gen. ii. 189 - (1840), but he does not even mention the name. The following com- binations are n Elizaldia violacea "(Dest. ), comb. nov. Echioides violacea Desf. FI. gens i. 164 (1800). Nonnea violacea DC. Fl. Fr. iii. 626 (1805). G. Kunze, Flora xxix. 691 (1846). E. nonneoides ; we Strand- u. Steppengeb. Iber. Halbins. 129, cum tab. (1852). E. iv.), comb. nov. Nonnea phaneranthera Viv. Fl. Lib. Specim. 9, t. 1, fig. 3 (1824); Murb. Lunds Univ. Arsskr. xxxiv. Afd. 2, no. 7, 10, t. 7, fig. 10-12 (1898). : emon (Murb.), comb. nov. Nonnea_heterostemon — Lunds Univ. Arsskr. xxxiv. Afd. 2, no. 7, 7, t. 7, fig. 13-15 (1898). STUDIES IN THE BORAGINACEAE a7 28. Nonea Medik. Phil. Bot. i. 31 (1789). Oskampia Moench, Meth. 420 (1794). Echioides Desf. Fl. Atlant. i. 163 (1800). Onochilis Mart. Denkschr. Acad. Muenchen v. 177 (1817). Nephrocarya Candargy, Bull. Soc. Bot. France xliv. 150 (1897).—About 25 species, these chiefly in the Mediterranean Region. Although maintained here, I am of the opinion that Nonea should be treated as a section of Pulmonaria, for there is no character by which it can be decisively separated from that genus. Nonea and Pul- monaria have the same corolla and calyx. Though the nutlets of Nonea are usually rugose, smooth and shiny ones, similar to those in Pulmonaria, occur in N. obtusifolia (Willd.) DC. Nonea usually has a loosely racemose inflorescence, whereas Pulmonaria commonly has a loosely glomerate one, but when the inflorescence of the latter occasionally loosens it reveals an arrangement of flowers quite like that in Nonea. The most striking difference between the two genera seems to reside in the root. Pulmonaria is always perennial, the stems, in scaly-based tufts, terminating rather fleshy shallow rhizomes. Nonea is mostly annual, but in such species as and N. alpestris the subterranean development is apparently a long deep perennial tap-root. It seems probable that Nonea and Pul- monaria have common immediate ancestors, and that the former is a xerophytic adaptation to rather arid open situations, es the latter is a mesophytic adaptation t p y abitats 29. Pulmonaria [Tourn.] L. Pl. 135 (1753); Gen. Pl. 65 (1754). Bessera Schult. Observ. Bot. 27 (1809).—A genus with about 12 ill-defined Eurasian species. It centers in southern Europe. ERITRICHIEAE. N pects erect or suberect (or rarely horizontal), straight or slightly ' bent, smooth or verrucose or appendaged, margined or marginless; areola Gaal or basal, at side of cotyledons or near their tip; gynobase pyramidal or columnar or rarely even flat; style entire; stigma 1, entire or very rarely somewhat emarginate, capitate or disciform; corolla blue or white or less commonly yellow or orange.—As a group the Eritrichicae are intermediate in degree of development between the Lithospermeae and the Cynoglosseae. Such genera as Megastoma, Oreocarya, and Mertensia are among the most primitive members of the tribe, whereas such genera as Eritrichium, Lappula and Trigonotis are the most specialized, coming nearest the Cynoglosseae. The 58 JOHNSTON tribe reaches its great development in Asia and in western America. The endemic American genera properly referable to this tribe are Amsinckia, Cryptantha, Oreocarya, and Selkirkia. Key To GENERA OF ERITRICHIEAE. toothed or eve rim or flange, or some or all with an epidorsal annulate o aeewere crest, or pronouncedly pret and eyidleaithy fetiatodie — base flat; nutlets regularly ied: attached at one Style patina sr ee neers elon anthers fully exposed; leaves crowded above, subverticillate............ 30. Brachybothrys. Style and sithers included; leaves peanaeed, alternate... .31. Trig ipa bate or columnar; nutlets not obviously and Nutlets circumdorsally winged, or with a definite upturned y to rly tetrahedral. Nutlets about equalling the elongate gynobase, attached = nearly their whole length, (with rare exceptions) ex ceeded by style. Bracteate wendy annuals with narrow strigose canescent firm ed leaves; nutlet-margin toothed or lacerate, frequently glochidiate. See ee mo fe h broad glabrous paieiesrtitae veiny leaves; Ssinte with a broad merely t an os, 5 BRC ete Bee ee ow ee ss eo et 8 8 ee oe ee 32. Lappula. Myosotidium. Nutlets much surpassing the stout pyramidal gynobas attached obliquely supramedially by a deltoid or vate reola, surpassing the Nutlets — Be »_fircundorsl commonly lacerate or enta Fruiting talve eel reflexed; rank tufted plants with broad bertiquecn leavened ceeN Fruiting nee 2 iors ascending; iw wens caespi- ose plants with small firm leaves........... 5. Eritrichium. vaca eae or all of — iteers with an epidorsal upulate or annulate corolla; nutlets with a subapical cupulate crest. 4 ae +L, . Craniospermum. Stamens included within the appendaged corolla; some nutlets with a thick medial dorsal annulate, Nutlets not encircled by a wing, rim, or flange, never bearing epidorsal annulate or seueute en: somewhat lencecline oricoteular and never conspicuously angular or polspecte 1. nee ..38. Myosolis. or less eacned as Flowers yeti ahha Sakae? diminutive Hima- la On eee. 39. Microcaryum. Flowers loosely racemose; emt diminutive. » appendaged, blue; n several dawn as lo STUDIES IN THE BORAGINACEAE 59 Corolla unappendaged, yellowish; nutlets broadly and _ erally mina to gynobase, exceeding i it in —) nnual; i ie er ee Oe ns 41. Megastoma. Pericarpial walls Sas woke along cintial ventral line, oe a ial ventral keel extending from areola to of Nutlets attached firmly to gynobase for nearly whole CVO mt DATOS CONROEODG.. <. ss «sv ic.e us cae babes 6s 42. Rochelia. Ovules 4; n utiete not contorted... 03... 6 se cee 43. Oreogenia, gee atkachie to gynobase through a definite areola. 2 hsad ont ——— somes Peed pubescent; “ vinate yan perennials............. 44. Chionocharis. Flowers several ro aay, uloesicents various; nutlets abrous Fruiting calyx strongly accrescent, plicate, veiny, irregularly toothed and lobed; — a oobgteky atta sees to an elongate strong RYDODARC. «0 ee vos vous ns one ne ben awn ieens 45. Asperugo. Fruiting anys: moderately if at all parsipaenun not plicate, nor conspicuously vein rl or funnelform throat, tube surpassing calyx- lobes; : rhathetis usually smooth, attached basally Or suprabassily. oo. . os eek ene ae Ses ey 46. Mertensia. Corolla pa ies throat not differentiated, tube usually shorter bern ealyx-lobes; nutlets aout ghee ois ventrally supra- basally media: nh suneitat mrceciat: reflexed in stag tube exceeding nutlets; style very Si oe 47, Anoplocaryum. Calyx practically fe Soca sas not refiexed; olla- surpassed by calyx; style little if a all surpassing nutlets...... 48. Plagiobothrys. 30. Brachybothrys Maxim. in Oliver, Hook. Icones xiii. 43, t. 1254 (1878). With a single species, B. paridiformis Maxim., in southern Man- churia and adjacent Korea. Evidently a close ally of Trigonotis, and like it to be included i in the Eritrichieae. It differs ‘from its re- lative in its open corollas p yp ‘ and coarse habit. 31. Trigonotis ee Bull. Soc. Nat. Moscou xxiv. 1, 603 (1851). Endogonia Lindl. Veg. nti ed. 2, 656 (1847).—A genus with a dozen odd species in Asia and temperate Melanasia. Evidently a relative of Eritrichium. Though the nutlets appear to be basifixed, morphologically they are attached medially and laterally to a very 60 JOHNSTON depressed gynobase. The decidedly tetrahedral nutlets are very characteristic of this and the preceding genus. 32. Lappula Moench, Meth. 416 (1794). Lapula Gilib. Fl. Lithuan. i. 25 (1781), nom. nud. Echinospermum Sw. in Lehm. Asperif. i. 113 (1818). Rochelia R. & S. Syst. iv. pg. xi and 108 (1819). Heterocaryum A. DC. Prodr. x. 144 (1846). Cynoglossospermum [Siegesb.] Kuntze, Rev. Gen. ii. 436 (1891).—A natural, but difficult genus of 30 to 40 closely related species centering in central and southwestern Asia and having outlying species in Europe, Africa, Australia, and America. Lappula spinocarpa (Forsk.) Aschers. is aberrant in the arming of its nutlets. These agree with those of typical Lappula in size, outline and attachment, but differ in lacking a well defined toothed margin, being instead coarsely and irregularly muricate over the back. The habit of L. spinocarpa is so obviously that of the other species of Lappula that generic segregation of the species seems quite unacceptable. The first generic name applied to the genus was that published by _ Gilibert in his Flora Lithuanica. . Though he described the species, Lapula {sic!] echinata and indicated that it was synonymous with Myosotis Lappula L., Gilibert did not describe his genus, nor indicate any character whereby it might be recognized. Hence, according to the International Rules (Art. 38), it is a nomen nudum. The first characterization given the genus was by Moench. Although Gili- bert’s specific name was launched under a generic name not effectively published, it has, in recent years, been generally accepted for the common European representative of the genus. 33. Myosotidium Hook. Bot. Mag. Ixxxv. t. 5137 (1859). A monotype with the single species known only from Chatham Island, New Zealand. Myosotidiwm hortensia (Decne.) Baill. is a very decorative plant, and has its closest relatives in Selkirkia of Juan Fernandez and in the widely distributed genus Hackelia. Brand, Pflanzenr. iv. Fam. 252, 45 (1921), follows past authors in placing Myosotidium in the Cynoglosseae, but the erect nutlets and their attachment are clearly those of an Eritric 34. Hackelia Opiz in Bercht. Fl. Boehm. ii. pt. 2, 146 (1839). A genus of 25 to 30 species, ‘most of which occur in western _ North America. Four species are known from the cooler parts of STUDIES IN THE BORAGINACEAE 61 pate: The peoorat se of the Eritrichieae has only increased , Contr. Gray Herb. n. s. Ixviii. 43 (192 3), that Hackelia is acs inimiediately aie! to Lappula, but is related rather to Eritrichium, and that it merits the generic rank here accorded it. 35. Eritrichium Schrad. in Gaud. FI. Helv. ii. 5 28). Centering in Asia where there are perhaps 12 scarcely separable species. One species occurs on the high mountains of Europe, and several more are known from America. The. genus is a natural one, and contains canescent perennial herbs characteristic of bleak plateaus or cold mountain-slopes. In the past the genus was vastly amplified, and at one time contained a large proportion of the Eritrichieae. As here defined, however, it is reduced to homogeneity and is a relatively small genus. 36. Craniospermum Lehm. Asperif. ii. 336 (1818). Diploma Schrenk, Bull. Phys.-Math. Acad. St. Petersb. ii. 195 (1844).—A well marked genus with several closely related species in the Altai-Baikal region of Asia. 37. Microula Benth. in Benth. & Hook. Gen. PI. ii. 853 (1876). Tretocarya Maxim. Bull. Acad. St. Petersb. xxvii. 505 (1881); Mél. Biol. xi. 270 (1881). Schistocaryum Franch. Bull. Mens. Soc. Linn. Paris 930 (1891).—As Gray, Proc. Am. Acad. xx. 261 (1884), has suggested, this genus seems very closely related to Plagiobothrys. It differs, however, in normally having blue corollas, and at least several of the nutlets of each fruit bearing a characteristic epidorsal annulate crest. Except for the crest, the nutlets are quite indis- tinguishable from those of Plagiobothrys § Euplagiobothrys, and it may be that further study will give reasons for reducing Macroula. Schistocaryum is evidently a synonym. The peculiar dehiscence of the nutlets, which was the justification for proposing the genus, may have been caused by extreme pressure in drying fruiting material. It is to be noted, also, that the author later, Jour. de Bot. v. 105 (1891), admitted, because of habital similarity, other species to his: genus which lacked the peculiar dehiscence. The species of Schisto- caryum have the nutlets of Microula, and seem to be close relatives of M. sikkimensis. Indeed S: myosotidium, the type of its genus, seems so close to M. sikkimensis as to throw doubt on their specific distinct- ness. The following 9 species are referable to Microula, They occur 62 JOHNSTON at high altitudes in the Himalayas, in Tibet, and in the mountains of western China. 1. Microvta TIBETICA Benth. in Benth. & Hook. Gen. Pl. 11. 853 (1876); Maxim. Bull. Acad. St. Petersb. xxvi. 501 (1880); Mél. Biol. x. 682 (1880); Hemsley in Hook. Icon. xxvi. t. 2562 (1898); Hemsley, Jour. Linn. Soc. xxxv. 192 (1902). M. Benthami Clarke in Hook. Fl. Brit. India iv. 167 (1883); Oliver in Hook. Icon. xxiii. t. 2257 (1893). Tretocarya pratensis Maxim. Bull. Acad. St. Petersb. xxvil. 505 (1881); Mél. Biol. xi. 272 (1881).—Tibet and adjacent Himalayas. 2. M. rancutica Maxim. Bull. Acad. St. Petersb. xxvi. 500 (1880); Mél. Biol. x. 682 (1880).—Northwestern China. Doubtfully distinct from M. tibetica. . M. pustunata (Clarke) Duthie, Kew Bull. 1912, pg. 39 (1912). Seiicschis = Clarke in Hook. FI. Brit. India iv. 164 (1883). urvyeee . M. sIKKIMENSIS (Clarke) Hemsley i in Hook. Icon. xxvi. sub t. are (1898). Anchusa Clarke in Hook. FI. Brit. India iv. 168 (1883). Tretocarya sikkimensis Oliver in Hook. Icon. xxiii. t. 2255 (1893).—Himalayas and adjacent Tibet. 5. M. myosotidea (Franch.), comb. nov. Schistocaryum myo- sotideum Franch. Bull. Mens. Soc. Linn. Paris 930 (1891).— Yunnan. 6. M. ciliaris (Bur. & Franch.), comb. nov. Schistocaryum ciliare Bur. & Franch. Jour. de Bot. v. 105 (1891).—Szechwan. 7. M. ovalifolia (Bur. & Franch.), comb. nov. Schistocaryuwm ovalifolium Bur. & Franch. Jour. de Bot. v. 105 (1891).—Szechwan. 38. Myosotis L. Sp. Pl. 131 (1753); Gen. Pl. 63 - (1754). Scorpioides Gilib. Fl. Lithuan. i. 20 (1781). Ezarrhena R. Br. Prodr. 495 (1810). Strophiostoma Turcz. Bull. Soc. Nat. Moscou xiii. pt. 2, 258 (1840).—A cosmopolitan genus of about 50 species, having two important distributional centers, one in Europe and the other in New Zealand. Attempts have been made to segregate, as a genus Exarrhena, those 2g Zealand species having large flowers and conspicuous stamens. As Cheeseman, Man. N. Zealand Fl. 458 (1906), has indicated, however, there appear to be no characters whereby this ‘Segregation can be sharply and naturally made. The large fi y related to the E uropean ones, having quite similar nutlets, nutlet-attachment, style, corolla- attachment, uncinate tly event habit, wv and since there are undoubted transitional forms it seems best to _ treat Exarrhena as a section of Myosotis. Myosotis sparsiflora Mikan STUDIES IN THE BORAGINACEAE 63 and allies have been grouped to form the genus Strophiostoma, char- acterized by its conspicuous exserted strophiole. Though well developed in Strophiostoma, the character is somewhat developed in many species of true Myosotis. This tendency towards a strophiolate condition is very suggestive of the Anchuseae, and it is a notable fact that the only other member of the subfamily with contorted corolla-lobes (i.e., Trigonocaryum) has been placed in that tribe. Myosotis, however, does not have a habit at all suggestive of the Anchuseae, nor a tumid rim surrounding the nutlet-attachment, while furthermore its habit, somewhat margined nutlets, and peculiar solitary stigma make it much more at home in the Eritrichieac. Because of its practically basally attached nutlets, the genus in the past has been referred to the Lithospermeae, but its habit, com- pressed slightly margined nutlets, solitary disciform stigmas, and blue corollas seem to indicate stronger affinities in the Eritrichieae. 39. Microcaryum, gen. nov. Calyx 5-fidus, lobis angustis, fructifer immutatus. Corollae tubus cylindricus calyce pau!lo brevior, fauce fornicibus 5 emarginatis clausa, lobi 5 ovati imbricati obtusi breves patentes. Stamina 5 tubo. affixa inclusa, filamentis brevibus et antheris ovatis minimis obtusis. Ovarii lobi 4 in gynobasi elongata laterales, stylus inter lobos brevis, stigmate subcapitato, ovula lateraliter affixa. Nuculae 4 erectae immarginatae dorso convexae rugosae et tuberculatae medie longitudinaliter carinatae ventre obtusae medie sulcatae sulco longi- tudinali basi divaricatim furcato ab ima basi fere ad medium gynobasi columnari affixae. Semina recta, cotyledones planae indivisae— Herba annua pernana himalayana villosa. Folia alterna oblance- olata obtusa. Inflorescentia pseudo-umbellata, floribus in fasciculos plures spicate dispositis. Corolla alba minima hypocraterimorpha. Pedicellae elongatae erectae. (Name from utkodc, small, and kaguoy, nut, in reference to the minute size of the nutlets.) Although it has been associated with Eritrichium, this monotype is not at all closely related there, appearing rather to have its closest affinities in Cryptantha. Microcaryum differs from Eritrichium in habit, inflorescence, and in the shape and attachment of its nutlets. It agrees with Cryptantha in all floral and fruiting structures, and has quite similarly shaped and attached nutlets, but differs, however, in its habit, falsely umbellate inflorescence, villous pubescence, and widely separated range. MJicrocaryum occurs at very high altitudes in the Himalayas and adjacent Tibet, whereas Cryptantha grows on 64 JOHNSTON the deserts and warm mountain-slopes at relatively low altitudes, from Alaska to Mexico, and Peru to Chile. Microcaryum pygmaeum (Clarke), comb. nov. Eritrichium pyg- maeum Clarke in Hook. Fl. Brit. India iv. 165 (1883). E. Riae Winkl. in Fedde, Repert., Beiheft. xii. 473 (1922) 40. Amblynotus, gen. nov. Calyx 5-fidus, lobis angustis elongatis, fructifer paullo accrescens. Corollae hypocraterimorphae tubus brevis calyce brevior, fauce fornicibus 5 obtusis clausa; lobi 5 imbricati obtusi patentes. Stamina _ 5, tubo affixa inclusa, filamentis brevibus; antherae oblongae obtusae. Ovarii lobi 4 in gynobasi pyramidali erecti; stylus inter lobos brevis; stigmate subcapitato. disciformi; ovula erecta. Nuculae 4 erectae emarginatae, dorso convexae obscure rugulosae nitidae glaberrimae, ventre basin versus cum areola triangulari obliqua instructae supra areolam secus angulum interiorem sulcatulae. Semina recta; coty- ledones planae indivisae.—Herba perennis asiatica caespitosa serico- villosa strigosa. Folia alterna obtusa oblanceolata. Racemi sim- plices bracteati. Corolla coerulea—Fritrichium § Amblynotus A; DC. Prodr. x. 128 (1846). Amblynotus has been confused with Eritrichium which it somewhat suggests in 1 habit and pubescence, but from which it differs i in the fits nutlets by an areola which is prolonged above into a ‘usually closed groove, and are polished and rounded and quite lack any suggestion of the dorsal Marginal crests characteristic of Eritrichium. The proposed genus is probably related to Cryptantha, though not very closely so, differing in habit and pubescence, as well as in the color and texture of the corolla. Amblynotus seems to be a very distinct genus well worthy of recognition. It surely can not be left in Eritrichium if the latter is to be naturally defined. Amblynotus obovatus (A. DC.), comb. nov. Eritrichiwm obovatum A. DC. Prodr; x. 128 (1846); Ledeb. Fl. Ross. iii. 152 (1847-9); Herder, Act. Hort. Petrop. i. 540 (1872). Myosotis obovata Ledeb. FI. Altaica i. 190 (1829). 41. Megastoma Coss. & Dur. in Bonn. & Barr. _ igh Tunis 301 (1896). A single species, M. pusillum Coss. & Dur., occurring in Algeria and Tunis. Although referred to Eritrichium ve Bentham & Hooker, Gen. Pl. ii. 851. (1876), it is not at all closely related to that genus STUDIES IN THE BORAGINACEAE 65 though it does approach Crypiantha which Bentham & Hooker also reduced to Eritrichium. Megastoma differs from Cryptantha in its nude, unappendaged corolla-throat, very irregular calyx, and opposite branching; and of course has a widely separated range. A beautiful illustration of the plant has been given by Bonnet & Barratte, Il]. Phaner. Tunis t. 11, fig. 4-11 (1895). 42. Rochelia iene Flora vii. 243 (1824); Icon. Crit. i. 13, t. 123 (1824). Maccoya F. Muell. pee Austr. i. 127 (1859).—An anomalous genus of 10 to 15 species, centering in southwestern Asia and ranging from central Asia to the Mediterranean Basin. One species occurs in Australia. Although properly placed in the tribe Eritrichieae by Bentham & Hooker, Gen. P). ii. 836 (1876), Giirke, E. & P. Pflanzenf. iv. Abt. 3, 131 (1895), associated it with the American genus Har- pagonella to form the tribe Harpagonelleae. A study of the nutlets and their attachment in Harpagonella and Rochelia should convince anyone that Giirke’s tribe is clearly an artificial association of two only distantly related, anomalous, biovulate annuals. Rochelia fits clearly and naturally into the Fritrichieae, probably nearest Lappula, but Harpagonella, as pointed out on another page, is a Cynoglossea. The correct combination for the Australian species of Rochelia appears to never have been made. Rochelia plurisepalea (F. Muell.), comb. nov. Maccoya pluri- sepalea F. Muell. Frag. Austr. i. 127 (1859). R. Maccoya F. Muell. m Benth. Fl. Austr. iv. 408 (1869) 43. Oreogenia, gen. nov. Calyx 5-fidus, lobis angustis, fructifer immutatus. Corollae tubus cylindricus intus infra medium transverse plicatus, calyci subaequalis; fauce fornicibus 5 obtusis clausa; lobi 5 orbiculares imbricati obtusi breves patentes. Stamina 5 tubo affixa inclusa, filamentis brevibus; antherae ovatae minimae obtusae. Ovarii lobi 4, in gynobasi elongata erecti; stylus inter lobos brevis, stigmate truncato plano vel concavo; ovula lateraliter affixa. Nuculae 4 lanceolatae immar- tae erectae, in facie interiori longitudinaliter carinatae carina ab ima basi fere ad apicem gynobasi columnari firmiter affixae, dorso rugosae strigoso-hispidulosae convexae medie carinatae. Semina recta; cotyledones planae indivisae.—Herba annua pumila villosa himalayana. Folia alterna obtusa oblanceolata. Racemi bracteati. Corolla coerulea—(Name from ég0¢, mountain, and -eved, birth.) 66 , JOHNSTON An Himalayan monotype which has been mistakenly referred to Eritrichium, a genus which is obviously not a close relative and which differs much in its perennial habit and margined broadly attached nutlets. The nutlets of Oreogenia are attached along the length of the keeled inner face. Oreogenia Munroi (Clarke), comb. nov. Lritrichium Munroi Clarke in Hook. Fl. Brit. India iv. 165 (1883). E. densiflorum Duthie, Kew Bull. 1912, pg. 39 (1912). 44, Chionocharis, gen. nov, Calyx 5-partitus, lobis lineari-spathulatis, fructifer immutatus. Corollae tubus calyci subaequalis, fauce fornicibus 5 cbtusis clausa; lobi 5 imbricati obtusi patentes. Stamina 5 tubo affixa inclusa, filamentis brevibus; antherae ovatae obtusae, Ovarii lobi 4 in gyno- basi pyramidali erecti; stylus inter lobos brevis, stigmate disciformi mucronato; ovula erecta. Nuculae 4 vel abortu pauciores erectae immarginatae strigoso-pubescentes dorso convexae laeves ventraliter obtusae cum areola parva triangulari supra basin ornatae. Semina recta, cotyledones planae indivisae.—Herba perennis himalayana pulvinata. Folia alterna numerosissima confertissime imbricata obtusa lata. (Name from yu, snow, and ydotc, beauty.) A remarkably distinct monotype of the Himalayas. The species has passed as a Myosotis, and indeed it does suggest in gross habit such New Zealand species as M. uniflora Hook. and M. pulvinaris Hook., but it is readily distinguished by its spathulate calyx-lobes, imbricate corolla-lobes, and laterally (suprabasally) affixed lanceolate very pubescent nutlets Chionocharis Hookeri (Clarke), comb. nov. Myosotis Hookert Clarke in Hook. Fl. Brit. India iv. 174 (1883). 45. Asperugo L. Sp. Pl. 138 (1753); Gen. Pl. 67 (1754). A ee monotype now widely disseminated as a weed. The plicate accrescent calyces of this plant are highly characteristic, and its nutlets also distinctive. The latter are strongly flattened with the small circular areola placed aboye the middle and clearly to one side of the medial line. The gynobase is strongly compressed. 46. Mertensia Roth, Cat. Bot. i. 34 (1797). me -Pneumaria Hill, V: eg. Syst. vil. 40 (1764). Casselia Dumort. ; Serer: Bot. 21 (1823). Steenhammerea Reichb. Fl. Germ. Excur. STUDIES IN THE BORAGINACEAE 67 i. 337 (1831). Platynema Schrad. Ind. Sem. Hort. Goetting. (1835); Linnaea Litt. (1837) 89. “Winkleria Reichb. Nom. 236 (1841).” Hippoglossum Hartm. ex Lilja. Linnaea xvii. 111 (1843). Oreocharis Lindl. Veg. Syst. ed. 2, 656 (1847). Cerinthodes [Ludw.] Kuntze, Rev. Gen. ii. 436 (1891).—With 50 to 60 species in the temperate portions of Eurasia and North America, most of them in western United States. The genus exhibits a number of diverse trends, which in a few instances have been given generic recognition, e. g. Pneumaria for M. maritima and allies, and Oreocharis for the racemose Himalayan species related to M. echioides. The recognition of Pneumaria would set the generic values too low, and logically compel the recognition of monotypic genera for such species as M. virginica (L.)-DC., M. rivularia DC., M. bella Piper, etc. This seems quite inadvisable since Mertensia in the broad sense is a natural and readily recognized entity. On account of its usually basifixed nutlets Mertensia has been placed in the Lithospermeae. Its relations, how- ever, appear to be rather in the Anchuseae or the Eritrichieae. Mer- tensia has much in common with Pulmonaria, in fact the most im- portant difference seems to be the absence on its nutlets of a plug- shaped strophiole surrounded by a tumid rim. On account of this difference, and because the nutlets are frequently attached obliquely or suprabasally by a flat areola, as well as from its abundance in western North America,—a region without any indubitable native anchusoid borages, I have thought it best to place the genus in the Eritrichieae. The style of Mertensia is almost always single. 47. Anoplocaryum Ledeb. Fl. Ross. iii. 154 (1847). A monotype known only from Transbaikalia. The genus has re- ceived practically no recognition since its publication, being con- sidered at most only a section of Eritrichium or Echinospermum. It is certainly distinct, however, from Eritrichium and Echinospermum since it has a very different habit, glabrate herbage, firm reflexed calyces, and unmargined and very differently attached nutlets. Anoplocaryum compressum (Turcz.) Ledeb. appears to be a close ally of Mertensia, from which it differs in the corolla-shape and nutlet- attachment. 48. Plagiobothrys Fisch. & Mey. Ind. Sem. Hort. Petrop. ii. 46 (1835). Allocarya Greene, Pittonia i. 12 (1887); Johnston, Contr. Gray Herb. n. s. Ixviii. 64 (1923). Havilandia Stapf, Trans. Linn. Soe. 68 JOHNSTON ser. 2, Bot. iv. 209 (1894).—A very large American genus with a few representatives in the Old World. Of the section Allocarya there are two species within our limits, one in Australia and another in Kam- chatka, the former probably derived from South America and the latter evidently from North America. C. Macbridei. Calyx 3 mm. long, lobes erect ; nutlets granulate............ C. limensis. Amsinckia hispida (R. & P.), comb. nov. Lithospermum hispidum R. & P. Fl. Peruv. ii. 5 (1799).—Ruiz & Pavon’s specific name is the oldest in the genus, and hence should replace A. angustifolia Lehm., cf. Contr. Gray Herb. n. s. Ixx. 44 (1924). The type of A. Weaile was collected in the arid belt to the north of Lima, Peru. onella Palmeri, var. arizonica, n. var., cornibus calycis quam ea formae typicae gracilioribus longioribus ca. 4 mm. ngi- 76 JOHNSTON tudine; nuculis paullo grandioribus.—Arizona: plains, Lowell, May 3, 1884, W. F. Parish 162 (ryer, Gray Herb.); near Tucson, April 1881, Pringle 363; Tucson, Lemmon; Tucson, 1877, Greene 1110.—The plant of Arizona currently referred to H. Palmert Gray differs from that of California and Lower California in having more elongate cornute processes on the fruiting calyx and noticeably larger nutlets. The type of H. Palmeri, collected on Guadelupe Island by Palmer in 1875, is quite indistinguishable from the plant of coastal Southern California and adjacent Lower California. Harpagonella was placed in a special tribe by Giirke, E. & .P. Nat. Pflanzenf. iv. Abt. 3a, 130 (1895), and was so treated in a paper of my own, Contr. Gray Herb. n. s. Ixx. 5 (1924). A recent detailed study of Harpagonella, however, has convinced me that it has been treated with too much dignity, and that it clearly falls into the Cynoglosseae next to Pectocarya as first indicated by Gray, Proc. Am. Acad. xi. 88 (1876). Harpagonella has only two ovules. The nutlets which develop from these are somewhat different in shape, attachment, and pu- bescence. The axial nutlet (that next the distinct calyx-lobes) is pubescent on all faces, and is slightly shorter and more loosely affixed to the gynobase than is the abaxial nutlet. The abaxial nutlet is enclosed by the peculiarly modified abaxial calyx-lobes, and _ is pubescent only on the face proximate to the axial nutlet. In both nutlets the radicle-end of the seed is next the attachment-end of the nutlet, and hence, morphologically at least, the nutlets are apically attached. It is quite evident, therefore, that the nutlets must stand inverted with their backs juxtaposed. A close examination reveals a margin surrounding the pubescent inwardly facing side of the abaxial nutlet which further suggests that this is, indeed, morphologically the back of the nutlet. The style, gynobase, and corolla, the indur- ated recurved pedicels, and all the vegetative characters of Har- pagonella are those of Pectocarya, a genus which also has very elongate nutlets apically attached, and hence there seems to be every reason for returning to the treatment of Gray, l. c., and Bentham & Hooker, Gen. PI. ii. 846 (1876), placing the genus next to Pectocarya and con- sidering it no more than a highly specialized and anomalous member of the Cynoglosseae. Pcie se: LATERIFLORA (Lam.) DC. Prodr. x. 120 (1846). P gracilis, var. boliviana Johnston, Contr. Gray Herb. n. s_ Ixx. 37 (1924) -—The study of the material of Pectocarya, recently collected by Mr. J. F. Macbride in Peru, lias convinced me th that P. gracilis, boliviana is ym with P. lateriflora, since the habit- STUDIES IN THE BORAGINACEAE 77 difference, by which P gracilis and P. lateriflora were distinguished, has proved illusory. The three species constituting the § Eupecto- carya appear to be remarkably similar in habit, differing only in the shape and arming of the nutlets. Pectocarya lateriflora is character- ized by its obovate, rather than parallel-sided oblong-linear nutlets, and appears to represent its section in Peru and Bolivia. Weber- bauer, Engler & Drude, Veg. Erde xii. 136, fig. 5 (1911), has given a suggestive, but not accurately detailed illustration of P. lateriflora. CYNOGLOSSUM LIMENSE Willd. Sp. Pl. i. 762 (1798).—This species is based entirely upon a plate and description given by Feuillée, Jour. Obs. Phys. ii. 765, t. 49 (1714), who found the plant in the “vallée d’Ylo.” As clearly shown by Feuillée’s maps and discussion, the locality known to him as “Ylo” is the same as the port near the southern boundary of Peru now known under that name. The type of C. limense, hence, was not collected near Lima as Willdenow’s inappropriate specific name, and Brand’s statement, Pflanzenr. iv. Fam. 252, 142 (1921), would suggest. As the plant has not been reported from within 500 miles of Ylo it is possible that its reputed occurrence there is the result of some confusion of data. CynoGLossum TriaNaEuM Wedd. Chlor. And. ii. 90 (1859).—A fine plate of this was published by Oliver, Hook. Icon. xxv. t. 2458 (1896). Brand, Pflanzenr. iv. Fam. 252, 136 (1921), apparently intended to cite this illustration, but part of his citation has been omitted and the remainder, the plate-number, was added to the citation to Weddell’s Chloris Andina. CoRDIA GERASCANTHUS L. Syst. ed. 10, 936 (1759); not of Griseb. FI. Brit. W. Ind. 478 (1861), nor Chodat, Bull. Soc. Bot. Genéve ser. 2, xii. 209 (1920). Gerascanthus P. Browne, Hist. Jamaica 170, t. 29, fig. 3 (1756). C. gerascanthoides HBK. Nov. Gen. et Sp. iii. 69 (1818). —In 1910 Urban, Symb. Antil. iv. 516, indicated that, as then used, the binomial, Cordia gerascanthus L., was incorrectly applied to the widely distributed tree with canescent, densely stellate calyces, and that the name is properly applicable to the relatively localized species of the West Indies and southern Mexico which has glabrous or spar- ingly hirsute calyces and larger flowers, and which was described and current as C. gerascanthoides HBK. Ten years later, in his paper on Cordia § Gerascanthus, Chodat, 1. c., declared Urban’s interpretation of C. gerascanthus L. to be incorrect and used the name in the traditional sense, applying it to the widely distributed plant with stellate calyces. Further examination of this matter has recently been made to determine tthe correct specific name for use by 78 JOHNSTON Dr. W. M. Wheeler in his publications on myrmecophytes. For the convenience of others the results of this study are here put on record. Cordia gerascanthus L. is based upon the Jamaican plant which Patrick Browne, |. c., described and figured under the name “ Gera- scanthus.”’ Browne’s illustration, showing only the floral structures, portrays a corolla of large size which has broad short obtuse lobes with conspicuous pinnate veining, a broad saucer-shaped throat, a stocky weakly ribbed calyx, and deltoid calyx-lobes. These char- acters definitely associate Browne’s plant with C. gerascanthoides HBK. and prohibit the use of the Linnean name for the plant with stellate calyces. It is to be also noted that not only does Grisebach, l. c., cite Browne’s figure under “C. gerascanthoides HBK.,” but he gives C. gerascanthoides HBK. as “common in the lowlands and mountains” of Jamaica, and gives the plant with stellate calyces (under C. gerascanthus Jacq.) as “rare”? on that island. Browne’s plant was not rare, for he speaks of it as follows, “This tree grows in many parts of Jamaica, and is generally esteemed as one of the best timber woods in the island; it rises to considerable er eer ee especially in the low-lands, where it is most common, .... . x It is significant that concerning the Jamaican occurrence of the plant with stellate calyces, Urban, |. c. (under C. alliodora Cham.), com- ments parenthetically as follows, “fortasse a cl. Wilson introducta ex el. Stapf. in lit.” Since the identity of C. gerascanthoides HBK. and Gerascanthus Browne is certain from a study of Browne’s plate and description, and from distributional considerations, it is evident that Cordia gerascanthus 1. is, indeed, improperly applied to the widely distributed plant with stellate calyces. Among its close relatives in the West Indies and Central America, C. gerascanthus L. is readily recognized by its large flowers, saucer-shaped throat hirsute or gla- brescent stout weakly ribbed calyx-tube, and deltoid calyx-lobes. It is known only from Cuba!, Isle of Pines!, Jamaica!, southern Mexico!, and northern Central America. As Urban, Symb. Antil. iv. 516 (1910) and viii. 574 (1921), has pointed out, Cordia alliodora (R. & P.) Cham. is the correct name for the widely distributed plant with stellate calyces, or, in other words, for the one incorrectly current as “C. gerascanthus.” Cordia alliodora ranges from Mexico and the West Indies southward along the Andes to Bolivia. A number of critical species, doubtfully distinct from it, have been described from southern Brazil, adjacent Paraguay, and Argentina. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY New Series.—No. LXXIV. STUDIES IN THE BoraGinaceaE.—IV. THE NORTH AMERICAN SPECIES OF CRYPTANTHA. By Ivan Murray Jounston. PUBLISHED BY THE GRAY HERBARIUM OF HARVARD UNIVERSITY CAMBRIDGE, MASS., U.S. A. 1925 THE NORTH AMERICAN SPECIES OF CRYPTANTHA. THE genus Cryptantha is a group of annual herbs centering in western America and belonging to the tribe Eritrichicae of the family Boraginaceae. It has long been recognized as a genus of decided difficulty. This is due both to the variability of its species and to the minute size of the fruiting and floral structures which, as experi- ence has shown, are the only satisfactory basis for precise specific differentiation. In recent years the mass of collected material of this genus has become so large, and the failure of the provisional classification of the earlier authors so patent, that the need of a thorough restudy and reclassification of the group has been increas- ingly apparent. It is hoped that the present monograph of the North American species will partially meet this obvious need and to some extent bring order out of the chaos that has caused so many botanists to neglect this technical but highly interesting group of West American plants. Preliminary studies of the North American species of Cryptantha were begun by the writer at the University of California in 1920. Intermittently the work has been continued up until the present. During this time many species have been seen in the field and types and critical material have been studied in several of the major her- _baria of the United States. The present treatment was prepared at Harvard University and is based upon the material in the Gray Herbarium and the University of California Herbarium, and such other material as has been borrowed during the final critical study of the genus. The mass of material of Cryptantha from the herbarium of the University of California has been invaluable, since it is replete with authentic fragments, particularly of the species described by Greene, and since it consists in large part of the critical and extensive accumulations of Mrs. Katherine Brandegee, who was one of the keenest students of the group. The genus having its geographical cen- ter in California, the advantages of having available such an extensive and critically assembled representation from that region is apparent, especially so when it is realized that the collection was studied in conjunction with the large general collection, and very numerous types and historical specimens, contained in the Gray Herbarium. In combining the representations of the genus from the two herbaria mentioned over 1500 different specimens were made available for detailed study and comparison during the final review of the genus. 4 JOHNSTON The detailed investigation resulting in this paper was carried on at the Gray Herbarium under the direction of Professor B. L. Robinson who has unstintingly given me of his time and scholarly aid. It isa pleasure here to express my gratitude for his friendly interest and encouragement, and ready aid in matters of perplexing nomenclature and difficult classification. I am also indebted to Professor W. A. Setchell of the University of California, who, in continuation of his many favors, has made it possible for me to restudy in detail the Cryptantha material from the University of California Herbarium. For assistance in bibliographic matters relating to this as well as other papers, I wish to express my indebtedness to the late Mary A. Day, Librarian at the Gray Herbarium, and to her successor, Miss Ruth D. Sanderson. GENERAL DISCUSSION. History OF THE GENUS. The generic name Cryptantha seemingly first appeared in a seed- list af the Hamburg Botanical Garden published in 1833 by Leh- - mann,’ Del. Sem. Hort. Hamb. iv. The next appearance of the name was in a seed-catalogue from the gardens at St. Petersburg published in 1836 by Fischer & Meyer, Ind. Sem. Hort. Petrop. ii. 35. Two Chilean species, C’. glomerata and C. microcarpa, were newly de- scribed, but no generic diagnosis was given. The third appearance of the name was in 1837 when these Chilean species appeared under a formal generic diagnosis in George Don’s General System of Garden- ing and Botany, iv. 373. The effective publication of the generic name Cryptantha, therefore, dates from Don’s General System, since in that work was fulfilled for the first time the requirement of a generic description called for in Art. 38 of the International Rules of Botanical Nomenclature. The genus Krynitzkia was launched in 1841, being fully described in a seed-catalogue from the St. Petersburg fardens published by Fischer & Meyer, Ind. Sem. Hort. Petrop. vii. 52. The genus in- cluded a single Californian species, K. i ieie transferred from 1 blendtec not _ examined, being apparently unrepresented in North American lib ibr H. Barnhardt t has pointed out to me, however, no Cryp- Lé ; fii d. s Jahr 1833 as kh b briefly mann in the seed- it bable th I d 1833, it was merely as a nomen nudum. The earlies' t reference associating the 1 name eae with the Delectus Seminum of 1833 appears to have been aa of De Candolle, Prodr. x. 129 (1846). THE NORTH AMERICAN SPECIES OF CRYPTANTHA i 5 Echinospermum under which it had been described by the same authors in 1836. The next important event in the history of the group occurred in 1846 with the appearance of the tenth volume of De Candolle’s Prodromus. In this work Kryniizkia was kept up, although, except for a few poorly understood species given as doubtful members of Lithospermum or Myosotis, all species of Cryptaniha then known were transferred, along with various species of Plagiobothrys and Oreocarya, to Eritrichium § Cryplantha (containing only the two original species) and Eritrichium § Rutidocaryum. In 1871 Torrey’s genus Piptocalyx was published by Watson, Bot. King Exped. 240 (1871). It was based upon Lithospermum cireum- scissum, a plant of western United States described by Hooker & Arnott in 1840. : A few years later Gray, Proc. Am. Acad. x. 58-61 (1874), published the results of his first study of the eritrichioid borages of western United States. Following De Candolle he referred all the species of Cryptantha to Eritrichium, and cited as synonyms of the latter Pla- giobothrys, Krynitzkia and Piptocalyz. Gray’s presentation of the group in the Synoptical Flora, il. pt. 1, 193-197, in 1878 is essentially that of his paper in 1874. In the year 1876 Bentham & Hooker published that part of their Genera Plantarum, ii. 850-851, treating the Boraginaceae. These authors, accepting the work of De Candolle and Gray, added still more diverse elements to the already overburdened genus Eritrichium. The species of Cryptantha were placed under that genus and con- sidered generically indistinguishable from plants now classified under Oreocarya, Plagiobothrys, Amblynotus, Megastoma, Anoplocaryum, Trigonotis and Eritrichium. ’ The genus Eritrichium thus became so unwieldy and so obviously heterogeneous that its break-up was inevitable. The reaction began with Gray’s notable paper in 1885, Proc. Am. Acad. xx. 257-286, in which the species of Cryptantha and Piptocalyx were placed under Krynitzkia along with species now referred to Amblynotus, Oreocarya, Antiphytum and Plagiobothrys § Allocarya. This treatment was repeated in the supplement of the second edition of the Synoptical Flora, ii. pt. 1, 423-430 (1886), published shortly before his death in 1888. In a series of three papers published i in 1887 by Greene, Pittonia i. §-23, 55-60 and 107-120, the American representatives of De Can- dolle’s overburdened Eritrichium were segregated in detail. First the | 6 : JOHNSTON genus Allocarya was formed to cover the species Gray had treated as Krynitzkia § Myosotidea. Then Piptocalyx was resurrected, and two new genera, Eremocarya and Oreocarya, were founded, the first based upon the plant of southwestern United States described as Eritrichium micranthum by Torrey in 1859, and the second upon Eritrichium § Pseudokrynitzkia and part of Krynitzkia § Pterygium described by Gray in 1885. Amblynotus and Antiphytum, although not men- tioned, were apparently also considered distinct from Krynitzkia. Krynitzkia having been finally trimmed to evident homogeneity, Greene, commenting cynically on Gray’s reasons for accepting Kry- nitzkia, discarded the name and took up the earlier Cryptantha. In 1899 Piptocalyzx Torr. having been found to be a homonym of the valid name of an Australian monimiaceous shrub published in 1870, the substitute generic name Greencocharis was published by Giirke Harms. In 1906 Grant, Bull. So. Calif. Acad. Sci. v. 28, proposed the name Wheelerella as a substitute for Greeneocharis, since the latter name was said to have not met with Greene’s approval. In 1923 both Eremocarya and Piptocalyx were reduced to Cryptantha by Johnston, Contr. Gray Herb. n. s. Ixviii. 55-57. The genus Johnstonella Brand, Fedde Repert. 1925 (as learned from proof sheets) is based upon Eritrichium racemosum, a Cali- fornian plant described by Watson in 1882. In the present paper Cryptantha is taken as including Krynitzkia, Piptocalyx, Eremocarya and Johnstonella. Gross MorpuHo.ocy. Roots. All the North American species of Cryptantha are clearly annual. In most of the species the root is a slender herbaceous tap- root obviously of short duration. In C. racemosa and C. holoptera, however, the taproot frequently becomes lignified to such a degree that the species have been repeatedly described as perennial. Never- theless field observation has shown that these species are also annual. Professor E. C. Jeffrey has obligingly sectioned and examined material showing tl imum d-development in C. racemosa. No growth- rings were discernable in a woody root about 9 mm. thick or in a section of stem nearly 5 mm. thick. Since the species grows in a desert area which is subjected to a sharply defined rainy and dry season it is more than probable that the plant is indeed annual, since even in such circumstances it lacks annual-rings though these are produced in such associated desert shrubs as Ephedra, ete. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 7 A purple dye is secreted in the roots of some species, particularly so in C. micrantha. According to Norton, Rep. Mo. Bot. Gard. ix. 149 (1898), this is alkannin, a dye secreted by species belonging to various genera of the Boraginaceae. The presence of conspicuous amounts of alkannin in the roots of C. micrantha has been stressed as one reason for giving that species generic recognition (as Eremo- carya). However, very noticeable amounts of alkannin are frequently produced in the roots of C. Fendleri, C. Grayi, C. muricata, C. mari- tima, etc., although this fact seems to have been generally overlooked. In most species there is scarcely any evidence of the production of this ye. Srems. The stems are usually slender and herbaceous, not infre- quently somewhat fistulose. In C. racemosa the basal portion of the stem frequently becomes decidedly woody and almost 8 mm. thick. In this species, as well as in a number of others, the bark is frequently exfoliated in age. The common form of branching in the species of this genus is sparse, loose and ascending. Frequently, however, the branching becomes diffuse. Occasionally it is widely spreading, the branches becoming even subprostrate. Although the very lowermost branches are not uncommonly opposite, the middle and upper ones are more or less dichotomous. The dichotomy in Cryptantha varies much in the definiteness to which it is developed, although present in some degree in all the species. Strange to say, dichotomy has been considered a characteristic of C. circumscissa and C. micrantha and as characters of the segregate genera, Greeneocharis and Eremocarya. The dichotomy of these species differs from that of indubitable mem- bers of Cryptantha neither in degree nor in nature. The ascendingly branched habit, rather characteristic of the genus, is absent in the common forms of C. Fendleri and C. foliosa, in which the stem is stiffly erect and forms a conspicuous axis. Leaves. The leaves are narrow and elongate, becoming linear, lanceolate or oblanceolate, and having acute to obtuse or very rarely retuse apices. Though commonly sessile the lowermost leaves are sometimes more or less narrowed into a petiole. The texture of the foliage is usually firm. The several basal pairs of leaves are unmis- takably opposite with more or less short, sheathing, connate bases. Opposite lower leaves are produced by all the North American species of the genus, being obvious in the seedlings, although tending to be obscure in the mature plants. Despite this fact, Cryptantha has been repeatedly described as having consistently alternate leaves. Op- posite leaves frequently persist in fruiting plants of C. affinis, C. ros- Ss JOHNSTON tellata, etc. In some species, such as C. maritima and C. cireumscissa, the leaves, especially the basal portion, become tessellated through the silicification of the epidermal cells. Usually, however, the leaves are at most somewhat abundantly pustulate. Trichomes. The trichomes of Cryptantha are all simple, unicel- lular, and more or less siliceous. They differ in no striking way from the type of appendages occuring in most genera of the Boraginaceae, cf. Solereder, Syst. Anat. Dicot. ed. 2, i. 555-560 (1908). The sili- ceous hairs are either smooth and somewhat transparent, or are more or less roughened by encrustations and somewhat opaque. They are clear or more or less tawny in most species, but in C. flaccida, C. simulans, etc., they are noticeably pallid. Commonly the hairs are straight, but decidedly faleate and uncinate ones are developed on the calyx-lobes of C. flaccida and allies. The bristles vary notably in direction, length and rigidity, the common form being a stiff, long, slender one. Since this type of pubescence varies much in stiffness, two degrees of rigidity are distinguished under the names, “hispid” and “hirsute.” The very stiff, somewhat pungent extreme, exem- plified by the conspicuous hairs produced by typical C. intermedia, is termed “hirsute.” The less rigid, more slender. pubescence de- veloped on the stems of such species as C. Hendersoni is termed “hispid.” The trichomes are commonly spreading or somewhat ap- pressed. Not infrequently, however, in such species as C. flaccida, C. Clerelandi, C. dumetorum, ete., the hairs are short and very closely strigose. More or less stiffish villous hairs are frequently developed, particularly on the calyx-lobes. This is best shown in C. crinita and C. maritima, var. pilosa. Subvillous hairs are occasionally found on the stems of some species, notably C. pterocarya. The bristles on the calyx-lobes are frequently quite stout. Probably the most decided extreme of this development is to be found. in C.foliosa. Associated with the bristle-like trichomes on the stem and leaves are the pale blistery structures called pustules. These are composed of a circle of slightly elevated, silicified, opaque, tessellately arranged epidermal cells surrounding the base of the trichome. They show much variety in size and frequency, varying from total absence to decided a , and os to a diameter of 1-3 mm. Similar structures are known in many European borage genera, cf. Revedin, N. Giorn. Bot. Ital. ser. i ix. - 301-318 (1902). According to Solereder, l.c., the pustulate trichome-bases are cystolith-like in origin. : INFLORESCENCE. The flowers of Cryptantha are borne in two ranks in unilateral cymosely arranged spikes or racemes. Although funda- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 9 mentally the spikes or racemes are disposed in a rhipidial cyme, this fact is often obscured by the complete reduction of the internodes within the inflorescence and consequent crowding of the spikes or racemes into fascicles terminating the branches. In species such as C. micrantha, C. cireumscissa, C. albida, C. Grayi, C. micromeres, ete, the spikes are not at all fascicled, the cymose forking appearing to be indefinitely repeated and ending only at the death of the plant. With other species, of which C. affnis, C. ambigua, C. echinella, C. mariposae, C. patula, etc., are examples, the cymose forking is repeated only a very few times, perhaps only once. In all these cases the spikes or racemes are solitary or geminate, and not grouped into fascicles. The cymosely forking stem is terminated at any one time by a pair of spikes or racemes, and bears down its sides the solitary spikes which earlier were terminal. In such species as C. muricata, C. flaccida, C. intermedia, etc., by a complete suppression of the _ internodes separating the spikes, 3-6 of these have been crowded together to form a terminal fascicle. In C. micrantha scattered inter- nodes are suppressed, this being shown by the opposite or subopposite bracts sprinkled through the inflorescence. The racemes or spikes are loosely or somewhat densely flowered in age, but in C. glomeriflora, C. leiocarpa, C. Torreyana and C. simulans the flowers are frequently glomerate-congested. Most of the species have the spikes or racemes bractless or at most with 1 or 2 bracts oc- casionally subtending the lowermost flowers. In C. albida, C. mari- tima, C. minima, C. micrantha, C. circumscissa, etc., the spikes are leafy-bracted throughout or nearly so. Some forms exemplified by C. ambigua, C. mariposae, C. crassisepala, etc., have the peduncles of the spikes or racemes leafy, whereas others, such as C. intermedia, C. ‘pterocarya, C. microstachys, ete., have even these naked. The rhachis of the spike is terete in all species except C. dumetorum, in which it is decidedly compressed and rather fragile. Coroutia. The corolla is white, and subtubular or more com- monly short rotate-salverform. It may be very minute and incon- spicuous, or become conspicuous and as much as 7 mm. broad. The tube always about equals the calyx-lobes and bears below the middle 5 stamens, the filaments of which about equal the length of the short- oblong included anthers. The throat is poorly developed, producing at the base 5 small intruded hemispherical or subtrapeziform appen- dages that occasionally almost close it. The lobes are variously de- veloped, being either ovate-oblong or suborbicular, and either widely spreading or more or less strictly ascending. In the North American 10 JOHNSTON species there are no cleistogamous flowers such as those developed in several groups of Chilean species. ALYX. The calyx is usually much accrescent in fruit, C. micro meres being the species in which this is least so. Since the accrescent fruiting calyx is the most modified, in addition to being the most con- spicuous, it alone is considered. Commonly it is divided to near the base, although in species such as C. pusilla, C. sparsiflora and C. glomeriflora it is less deeply so. In C. circumscissa, however, the calyx is obviously united to near the middle and in addition is unique in the genus in being circumscissile just below the sinuses. The fruiting calyx commonly shows more or less evident tendency towards ir- regularity or asymmetry. The mature calyces of such species as C. albida, C. oxygona, C. pterocarya, C. holoptera, C. pusilla, C. mi- crantha, ete., are regular or subregular. On the other hand the fruit- ing calyces of C. dumetorum, C. flaccida, C. recurvata, C. echinosepala, ete., are obviously asymmetrical. Commonly the irregularity is manifest in the slightly greater length and more conspicuous pubes- cence of the abaxial calyx-lobe. Less commonly, however, as in C. recurvata and C. echinosepala, it is the axial calyx-lobe that is the longest and most hispid or hirsute. Because it commonly very closely invests the nutlets, the mature calyx frequently becomes somewhat asymmetrical in consequence of the irregularities in shape or abortion of the nutlets. Cryptantha dumetorum has the calyx axi- ally and downwardly gibbous, C. affinis has it compressed, C. re- curvata has it obviously curved, and such species as C. utahensis or C. flaccida have it oblique or distorted. The base of the mature calyx may be rounded, conical or more or less angulate, and may be regular or oblique or evidently asymmetrical. In texture it is usually some- what firmly herbaceous with the epidermis tending to become silicified. The calyces of C. circumscissa exhibit the extreme of this tendency towards silicification. The lobes of the mature calyces vary from narrowly ovate through lanceolate to linear. Though occasionally erect, they are commonly somewhat connivent above and have their tips spreading or even recurved. In pubescence the fruiting calyx shows various tendencies. The midrib, which is weakly developed in such species as C. pterocarya, C. pusilla, C. circumscissa, etc., and very strongly so in C. erassisepala, C. intermedia, ete., is commonly armed with evident bristles which vary in length, direction, slender- ness and rigidity. The fructiferous calyx is usually ascending, but may be strictly and closely appressed to the rhachis as in C. dumetorum, C. flaccida, C. microstachys, ete., or spreading or deflexed as in (. re- curvata or C. echinosepala. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 1] The pedicel of the completely developed calyx is commonly very short or almost undeveloped. Although pedicels are lacking or practically absent in most species, evident and well developed ones are produced by C. racemosa, C. holoptera and C. crinita. Since the pedicels are clearly developed only in these three species they alone, in the present monograph, as treated as having “racemes.” The remaining species, which have sessile calyces or very obscure pedicels, have been uniformly accredited with “spikes” although the unit of the inflorescence in some of the species might be more exactly de- scribed as “spicate-racemose”’ or “racemose-spicate.”” The pre- ponderance of species have deciduous calyces. As already mentioned C. cireumscissa has a circumscissile calyx. In this the lower cup-like portion is firmly attached and clearly persistent. Persistent calyces also occur in C. micrantha and C. dumetorum, and subpersistent ones are produced by C. albida and C. racemosa. Fruir.' The fruit of Cryptantha consists of 4 or fewer, elongate, ventrally grooved nutlets affixed to a usually elongate central gyno- base. Most of the species have 4 ovules, but in C. recurvata and C. maritima only 2 are normally present. Cryptantha maritima, how- ever, has an insular variety that is distinguished only by having 4 rather than 2 ovules. In shape the nutlets vary from broadly ovate or triangular-ovate to narrowly lanceolate. The back may be some- what flattened, but is commonly more or less convex, although in C. muricata it is obtusish with a suggestion of a medial dorsal ridge. The sides of the nutlets in most species are clearly rounded or some- what obtusely angled, but they may be quite acutely angled as in C. mohavensis, C. angustifolia, ete. In species like C. utahensis, C. costata, C. racemosa, etc., the margin is drawn out into a very narrow, thin, knife-like border. In other species the margin is greatly de- veloped, wing-like, and about the width of the body of the nutlet. Cryptantha holoptera and C. pterocarya produce nutlets of this sort. In C. pusilla and frequently in C. muricata, the edges of the nutlets are thickened to form a bead-like margin. The surface of the nutlet may be smooth and shiny as in C. Torreyana, C. leiocarpa, C. Fendleri, etc., or may be variously roughened. In species such as C. ambigua, C. simulans and C. Hendersoni the roughenings consist of small, low, rounded bosses, a condition here described as tuberculate. Other species may have the nutlets studded with conical warts and thus become muricate, or with rather elongate nipple-like warts rendering 1The nutlets as well as all other minute structures have been consistently studied iS oS = 12 JOHNSTON the surface papillate. Other species such as C. muricata, C. inter- media, C. barbigera, etc., have the surface of the nutlets verrucose, i. e. sprinkled with coarse simple warts. Species such as C. crassi- sepala, C. echinella and forms of C. Hendersoni have the nutlets covered with spinular papillae. On the roughened as well as the smooth nutlets the surface is frequently more or less covered with minute, -white discules. Such nutlets have been traditionally but not very precisely described as granulate. to be no prevailing tendency towards individualizing a particular nutlet. In this respect the specific group is unique. The natural THE NORTH AMERICAN SPECIES OF CRYPTANTHA 13 The ventral groove of the nutlet may be open or closed (not fused) and may be simple or forked below. Some species, such as C. Tor- reyana, have the groove almost invariably closed, others, such as C. Fendleri, have it characteristically open, while still other species, such as C. intermedia, may have it open or closed. In C. albida the groove is very broadly dilated and becomes excavated. A similar condition is present in the consimilar nutlets of C. crassisepala and C. minima. The groove of C. costata is open, but is extremely shallow. In C. flaccida the groove is not only closed, but frequently has one margin overlapping the other. The groove of C. leiocarpa is closed and not forked or is very obscurely so at the very base. In C. am- bigua, C. Torreyana, etc., it is very broadly forked below. In most species there is a small open areola formed at the forking. The groove on the more or less asymmetrical nutlets of C. affinis and C. glomeri- flora is evidently excentric and dilated below into an irregular tri- angular areola The gynobase varies from quadrangular-subulate to narrowly pyramidal or very shortly columnar. Usually it is subulate as in C. intermedia, C. letocarpa, C. Fendleri, C. pterocarya, etc. The nar- rowly pyramidal form is produced by C. albida, C. pusilla, ete. In species such as C. texana, C. glomeriflora, C. microstachys, the gynobase is reduced to an inconspicuous, exceptionally short column. It commonly reaches to about 24 the height of the nutlets, though fre- guently in such species as C. Grayi, C. costata, C. holoptera, etc., it reaches to the summit of the nutlets, or in species such as C. glomeri- flora and C. microstachys only to about 14 the height of the latter. Commonly the style is sharply differentiated from the gynobase. In C. micrantha, however, the style is not clearly set off and appears to be the subulate prolongation of the gynobase. In the length of the style and the height to which it reaches on the nutlets, the species noticeably vary. Generally the style reaches to 7/3~*/5 the height of the nutlets, but it may just reach the tips of the nutlets or even surpass them. In species such as C. glomeriflora and C. flaccida the style reaches less than !/3 the height of the nutlets. ABNORMALITIES. In Southern California and less commonly in the deserts of Nevada and Utah, the plants of Cryptantha frequently become fasciated. All or only some of the stems are affected. The abnormal stems are short and for the most part unevenly reddish- tinged, and are clothed with strictly ascending, scarcely reduced leaves. The spikes are usually undeveloped, or are partially developed and form a glomerate infertile mass above. The affected plant as a 14 JOHNSTON whole is very dense and broom-like. Although clearly abnormal the cause of this condition has not been ascertained. Cryptantha has been reported as host for Puccinia ryptanthes Diet. & Holw., P subnitens Diet. and. Synchitrium myosotidis Kihn., but none of these fungi causes such abnormal growth. A microscopic examination has failed to disclose any other fungi affecting the plant, nor any mites, aphids or similar parasites capable of profoundly disturbing it. The condition described has been noted in C. intermedia, C. barbigera, C. simulans, C. angustifolia, C. grgcilis, and C. pterocarya, geen it appears to be most common in the species first mentionec SysTEMATIC PosItION OF THE TRIBE ERITRICHIEAE. Cryptantha is obviously a member of the Eritrichieae and appears to have been derived from the closely related, and also West American genus, Oreocarya. It is believed that this genus was evolved from the Lithospermeae through some forms similar to the North American species of Antiphytum. These opinions assume the correctness of the arrangement of the tribes of the Boraginaceae given in a recent paper by Johnston, Contr. Gray Herb. n. s. Ixxiii. 42 (1924). Since the assumption there expressed, that the Cynoglosseae are derived from the Lithospermeae, is directly contrary to that accepted in stan- dard works, it seems well to state the reasons for this belief so that the direction of evolution may be understood and phylogenetic specu- lations regarding Cryptantha logically founded. Brand, Pflanzenr. iv. Fam. 252, pt. 1, 14 (1921), agreeing with previous authors, considers the Cynoglosseae among the Boraginoideae to be most nearly related to the Heliotropioideae, saying “ Der end- stindige Griffel, das Charakteristikum der Cordioideae, Ehretioideae und Heliotropioideae findet sich bei den Borraginoideae in 2 Gattungen, Trichodesma und Lacaitaea, wenigstens zur Bliitezeit. Diese beiden Gattungen miissen daher an der Spitze der Cynoglosseae stehen, . . . . However, Trichodesma and Lacaitaea clearly have pyramidal gyno- ane their yo teas oe seit are attached supramedially and have 1 elopments represent considerable departure fice the pr mses and very much greater departure than the common developments produced by the Lithospermeac. This may be appreciated after consideration of the following data. raginaceae appear to have sprung from ancestors with two, (at least) biovulate carpels. This is suggested by the occurrence in many of the shrubby genera of the Heliotropioideae and Ehretioideae THE NORTH AMERICAN SPECIES OF CRYPTANTHA 15 of fruit which is 2-celled, is more or less incompletely 4-celled, or has decidedly paired carpels, cf. Miers, Contr. Bot. ii. 190-261 (1869), and further suggested by stages in the development of the fruit in various members of the family, cf. Baill. Adansonia iii. 1-7, t. 1 (1862); Rosanoff, Jahrb. Wiss. Bot. v. 72-80, t. 5-6 (1866) and Payer, Organ. Vég. 546-549, t. 112 (1857). At an early stage in the develop- ment medial partitions form which divide each of the two carpels so that at maturity the fruit becomes virtually 4-carpellate. The primitive fruit was probably similar to that of the Hydro- phyllaceae, being 2- or imperfectly 4-celled, capsular, and terminated by a lobed style. The lobed style consistently occurs in those tribes of the Boraginaceae which are commonly considered primitive, i. e. the Heliotropioideac, Ehretioideae and Cordioideae. Within the Boraginoideae the style is lobed or bears geminate stigmas only in the Lithospermeae. Within in the subfamily it is borne on a flat recep- tacle or more commonly on an elevated gynobase. In the more primitive subfamilies the style is seated in the pericarp, usually at the tip of the more or less globular, 2—4-celled fruit. There is no thickened, persistent gynobasic column connecting the style directly with the receptacle, the style being seated in pericarpial tissue and falling away with some one of the carpels when the fruit breaks up. This condition prevails in the Heliotropioideae and Ehretiowdeae, and is completely and fundamentally different from that in the Cyno- glosseae and in Trichodesma or Lacaitaea in particular. The nutlets have resulted from a pinching in of the pericarpial walls to form lobes of the fruit each containing one carpel. The stages of this development may be appreciated by a comparative study of the fruit of Heliotropium or, better still, Coldenia. Coldenia canescens, T. & G. Pacif. R. R. Rep. ii. pt. 2, 169, t. 7 (1856), has an unlobed fruit bearing a decidedly terminal style. Coldenia Nuttallii, Torr. Bot. Wilkes Exped. 410, t. 12 (1874), or C. hirsutissima, T. & G. 1. c. 170, t. 9, has the lobing evident and the style attached to the peri- carp between and below the apices of the nutlets. In C. litoralis the lobing is almost complete and the style is affixed practically upon the receptacle. From a study of this, and similar series, it seems clear that the development of the nutlets has proceeded by the deepening downward from the apex, and inward from the sides, of the pinching in of the pericarp between the carpels. This finally results in an obviously basal attachment of the nutlets, and the gradual lowering of the style-base between the nutlets until it is at last directly and firmly affixed upon the receptacle. The gynobase appears to be sub- 16 JOHNSTON sequently developed, either by the thickening of the style-base, or by the pushing up of the central portion of the receptacle. It seems clear that the Cynoglosseae are not the primitive members of the subfamily Boraginoideae, for the nutlets are not attached basally, but apically or subapically, and, except in such anomalous genera as Harpagonella, Antiotrema and Bothriospermum in which the nutlets are completely inverted and attached by the (morphological) tip to a flattened gynobase, the gynobase is obviously developed. A study of the stages in the history of nutlet-development in the Helio- tropioideae and Ehretioideae is conclusive in showing that the base of the fruit-lobe, in these unspecialized groups, is rever free when the apex is not, although the contrary condition is of common.occurrence. The nutlets being basally attached and the gynobase commonly flat in the Lithospermeae it seems quite obvious that the group is indeed the most primitive one of the Boraginoideae and hence closest to the Heliotropioideae. Significant in this regard, is the fact that the heliotropioid stigma is suggested in certain species of Lithospermum. The stigma in this genus occasionally becomes somewhat lateral with the style-branches prolonged beyond them. Rarely the lobes become more or less fused and the stigmas, brought near one another, tend to cohere just as illustrated by Reichenbach, Icon. Fl. Germ. xviii. t. 113 (1858). The result is a stigma differing in no profound way from that characteristic of the Heliotropioideae. The primitive subfamilies being prevailingly woody it is also significant that within the Boraginoideae the most decided tendency towards woodiness is found in the Lithospermeae. Not only do the Cynoglosseae lack certain features suggesting close relationship in the Heliotropioideae, but they have developments which make such an affinity seem improbable. As previously mentioned, the style is gynobasic and though appearing to be terminal in 77ri- chodesma is not really so. A careful, examination of Trichodesma shows clearly that the style is not attached at the apex of the fruit and seated in pericarpial tissue, as in Heliotropium for example, but is definitely borne at the apex of a well developed gynobase. The nutlets are covered with highly specialized appendages which suggest nothing in the Heliotropioideae or Ehretioideae, although the develop- ment of these appendages can be traced back into the immediately related tribe Eritrichieae, from which the Cynoglosseae seem in fact to have evolved. The apical nutlet-attachment of the Cynoglosseae is wholly unlike any development in the less specialized subfamilies, but is obviously the termination of a strong tendency towards THE NORTH AMERICAN SPECIES OF CRYPTANTHA 17 elevation of the nutlet-attachment which is discernable in the Litho- spermeae and quite evident in the Eritrichieae. The Lithospermeae being thus considered the most primitive tribe of the Boraginoideae because of its basifixed unappendaged nutlets, non-gynobasic usually lobed style or double stigmas, unspecialized corollas, and frequent development of woody habit, it now becomes a relatively simple matter to place the tribe Eritrichieae. The Litho- spermeae appear to have evolved from some primitive member of the Heliotropioideae or specialized member of the Ehretioideae. The tribe appears to have given rise to two principal evolutionary lines. On one hand to the relatively unimportant line represented by the Anchuseae, in which the nutlet-attachment has tended to become con- spicuously margined and the attachment-surface on the nutlet tended to become elevated into a strophiolate plug. On the other hand it has evolved into a major line of development which ends in the Cynoglosseae, the most specialized group in the entire family. This latter evolutionary line is characterized by a tendency of the nutlet- attachment to move from the chalaza-end of the nutlet towards the radicle-end, i. e. from base to apex, and for the nutlets to vary from rounded and smoothish towards margined and variously roughened and appendaged. The Eritrichieae form a rather arbitrary group including the medium developments on the latter line. The tribe is usually regarded as including the genera in which the nutlet- attachment is typically lateral. The nutlets may be smooth or rough- ened or appendaged, margined or unmargined. The other characters indicative of the Fritrichicae have been already outlined in another paper, Contr. Gray Herb. n. s. lxxiii. 57 (1924), and need not be re- peated here. | GENERIC RELATIONS. As previously stated, Cryptantha is obviously a member of the tribe Eritrichieae and is apparently derived from the very closely related genus Oreocarya. Among their immediate relatives, Oreocarya and Cryptantha are together characterized by the possession of a medial ventral groove on the nutlets, this formed by the non-fusion of the pericarpial walls. This development appears to have been brought about by the gradual encroachment of the pericarp over the surface of a sharply cut triangular attachment-scar such as those present in the Lithospermeae. This encroachment gradually narrows the at- tachment-surface of the nutlet and forms a groove which is usually somewhat forked at the base. In Cryptantha and Oreocarya the groove 18 JOHNSTON is narrow, but not completely closed. In Amsinckia, Plagiobothrys, etc., the forked groove is entirely shut and its location is marked by a ridge of fused, pericarpial tissue which bears the caruncular scar, or small modified areola at the forking below the middle of the nutlet. It seems quite likely that Oreocarya has been derived from some form of Antiphytum, a genus of the Lithospermeae evidently derived from Lithospermum. Such a species as A. peninsulare not only has a habit suggestive of Or ocarya and related genera, but has nutlets, which, except in the nature of the attachment, are remarkably like those of some species of Oreocarya and Plagiobothrys. Its rather large at- tachment surface is lateral and submedial, but is otherwise like the attachment-surface of most Lithospermeae. The encroachment of the pericarp over the scar of A. peninsulare would result in a nutlet re- markably like that of Oreocarya or Plagiobothrys; the particular simil- arity depending on the extent to which the encroachment proceeded. It is to be noted, however, that Antiphylum has a style bearing geminate stigmas, whereas Oreocarya, Plagiobothrys, Cryptantha, etc., all have solitary stigmas. Inasmuch as the stigmas of Antiphytum are less obvious than are those of Lithospermum, its progenitor, it is not hard to suppose that this tendency was carried to an extreme and the geminate stigmas became coalescent in the progenitor of Oreocarya. Cryptantha is a specialized off-shoot from Oreocarya. The latter genus consists of rather coarse perennials, or rarely biennials, with persistent fructiferous calyces and homomorphous nutlets. In Cryptantha a successful, annual, herbaceous habit has been evolved, in addition to more or less deciduous fruiting calyces and somewhat heteromorphous nutlets. By taking on the annual habit, and develop- ing detachable somewhat bur-like fruiting calyces, Cryptantha has been able to reproduce, spread and evolve very rapidly, adapting itself to the variety of conditions in western America and greatly surpassing Oreocarya both in number and in the variability of its species. Oreocarya seems to have also given rise to Plagiobothrys, a group which seems to have evolved into Amsinckia, and into a line producing Microula, Craniospermum, etc. The principal derivative of Oreocarya, however, appears to be Hackelia, for through that genus Oreocarya appears to connect with Cynoglossum and the tribe Cynoglosseae. While the medial ventral groove on the nutlets of Oreocarya and Cryptantha characterize these genera among their immediate relatives, the development is by no means peculiar to them. It is present in Microcaryum, Amblynotus and Megastoma, and slightly developed in Oreogenia and Chionocharis. Microcaryum, Oreogenia and Chiono- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 19 charis appear to be Asiatic developments from the plexus containing Plagiobothrys. ‘The immediate relations of Amblynotus are wholly obscure. Megastoma, although remarkably simulating Cryptantha, cf. Bonnet & Barratte, Ill. Phaner. Tunis t. 11 (1895), and having been accepted as a close relative of the genus by all authors including the writer, Contr. Gray Herb. n. s. Ixxiii. 64 (1924), appears in fact to be of the Lithospermeae and to be most closely related to Echiochilon. The detailed study of specimens, recently at hand, shows the stigma to be decidedly geminate, and the yellowish unappendaged corollas to be slightly irregular and to bear the stamens at slightly unequal heights. Megastoma is immediately related to Echiochilon, agreeing withit in its irregular calyx and corolla, but differs from it in its broadly grooved nutlets, large gynobase, glabrous and much less irregular corolla and almost equally inserted stamens. Echiochilon is prac- tically an African genus, and the affinities just mentioned with Megastoma, a plant of the deserts of Algeria and Tunis, seem thor- oughly satisfactory on phytogeographical grounds, and vastly more so than the supposed relations between Megastoma and Cryptantha. GENERIC CONCEPT. As here taken the genus Cryptantha sarc Krynitzkia, Piptocalyz, Eremocarya and Johnstonella. Since two of these genera have achieved some currency in botanical ee it seems well to review the claims which these rejected genera have for recognition. Since the resurrection of Cryptantha by Greene in 1887, Krynitzkia has been almost universally regarded as a clear synonym of the earlier genus. Krynitzkia, in fact, appears to have been deliberately main- tained only by Katherine Brandegee and Prof. M. E. Jones. Mrs. Brandegee, I have been told, was not satisfied that the cleistogamous Chilean species, i. e. the true species of Cryptantha, were actually congeneric with the more abundant open-flowered species which form the bulk of the genus in South America and the entire representation of itin North America. Being in doubt as to the applicability of the earlier name, Cryptantha, she was content to follow Gray in accepting the name Krynitzkia, since there could be no uncertainty regarding the applicability of it to the North American plants. Similar reasons, as well as his confidence in Gray’s judgment, seem to have given Jones the basis for his persistent use of Krynztzkia in the broad inter- pretation of Gray. The two original species of Cryptantha and several other closely 20 JOHNSTON related ones depart from the common phases of the genus in no striking development other than the presence in the inflorescence of more or less numerous cleistogamous flowers. As these commonly occur variously mixed with open flowers, and since they are borne on plants which in gross habit and inflorescence, as well as in details of fruit and calyx, are quite like the completely normal-flowered plants, there seems to be no reason why we should differ from Reiche, Fl. Chile v. 217-237 (1910), who has studied the Chilean species, and consider the name Cryptantha applicable only to the few Chilean species bearing cleistogamous flowers in the inflorescence. Especially is this the case when it is remembered that another and more peculiar phase of cleistogamy, in such species as C’. phaceloides and C. linearis, goes completely unrecognized. the four genera reduced under Cryptantha, Piptocalyx is the most notable. Under the name Piptocalyx or Greencocharis this portion of the genus has gained rather wide generic recognition follow- ing its reinstatement by Greene in 1884. Previously, however, it had been usually treated as a section under Eritrichium and later under Krynitzkia. In 1921 Johnston gave it sectional rank under Cryp- tantha. The outstanding development of Piptocalyx is its persistent circumscissile calyx. In flowers, fruiting structures, as well as in- florescence and branching, it differs in no fundamental way from the Cryptantha species related to C. angustifolia. In fact, its gross habit is such that it was once described as a species of Cryptantha by A. Nelson. Greene, in resurrecting the genus, made much of the dicho- tomy of Piptocalyx, although the cymose branching is decidedly similar, if not exactly the same as that developed in C. Grayi, C. an- ia, etc. Rather than showing a profound difference, like the fruiting structures, the cymose branching suggests a very close af- finity between Piptocalyx and C. micrantha, C. Grayi and C. angusti- folia. The character of branching failing to separate Cryptantha and Piptocalyx the latter must, perforce, stand or fall according to the emphasis placed on its peculiar calycine developments. It is to be noted that the possible generic characters of Piptocalyx, the circum- scissile fission of the calyx and the persistence of the cupulate calyx- base, merely represent phases of a single aberrant structure. For this reason it has been considered unwise to separate the plant from its obviously close affinities in Cryptaniha. Furthermore, species with cireumscissile calyces and those without are universally admitted to the closely related genus Plagiobothrys, and in addition the degree to which the calyces are persistent in Cryptantha is variable, they being THE NORTH AMERICAN SPECIES OF CRYPTANTHA at readily deciduous, or persistent or subpersistent as in C. albida, C. racemosa, C. dumetorum, C. micrantha, ete Eremocarya, like Piptocalyx, has achieved wide recognition since its publication in 1887. Unfortunately it appears to lack characters of generic value. Greene, who launched the genus, laid much em- phasis upon its spikes, which were described by him as “ biserial and very dense, conspicuously leafy-bracted, and repeatedly forked.” Except for the bracts and the complete lack of reduced internodes, the inflorescence and branching of C. angustifolia and C. Grayi are exactly similar to those of Eremocarya. More or less similar biserial, leafy- bracted and repeatedly forked cymes are to be found in such species as C. circumscissa, C. albida, C. maritima, ete. It can be decisively said that Eremocarya can not be separated from Cryptantha by any character of inflorescence. The calyx of Eremocarya is said to differ from that of Cryptantha in its persistence. However, calyces as persistent as those of Eremocarya occur in C. dumetorum, a species no one has attempted to exclude from Cryptantha. The style in Eremocarya is scarcely distinguishable from the gynobase proper. This thickened style, appearing as a continu- ation of the gynobase, much surpasses the nutl d bears the stigma almost at the level of the tips of the calyx-lobes. Similarly elongate and persistent styles are developed in C. racemosa. The thickness of the style, therefore, appears to be the only peculiar character of Eremocarya. This is not considered of generic value. The genus Johnstonella is being proposed by Brand to include certain species thought to be generically ambiguous. Cryptantha racemosa, the type species, is said to have persistent calyces, thereby exhibiting a salient character of Oreocarya. In habit the plant is clearly a Cryptantha. It seems to be of particular note only because of its more or less suffrutescent character. In duration, however, it is unquestionably annual. Furthermore it can be definitely said to have deciduous calyces. There are at least twenty collections in the Gray Herbarium and University of California Herbarium which exhibit completely or partially denuded .raceme-rhachises. The calyces are no more persistent than are those of C. holoptera or C. al- bida, and not so decidedly so as are those of C. dumetorum. The genus is clearly synonym Expanding the limits of Cryptantha to include the four genera just discussed we dispose of several small or monotypic genera, and obtain a very naturally defined and readily recognized genus. This covers more or less stiffly haired, rather slender annual herbs which agree 22 JOHNSTON in having more or less distinctly biserial spikes or racemes of white flowers and later somewhat irregular deeply lobed hairy calyces, and which develop 1-4 somewhat heteromorphous nutlets that are at- tached laterally to an erect gynobase through a medial ventral roove. As a group it inhabits sunny open places, in North America ranging from southern Alaska to southern Mexico, although it is most common in western United States, and particularly so in California. It is to be confused only with Oreocarya and Plagiobothrys. From the former, its closest relation, it differs in its slender annual, rather than coarse biennial or perennial habit, and commonly deciduous, rather than invariably persistent calyces. From Plagiobothrys (in- cluding Allocarya and Sonnea) it differs in having the nutlets attached through a ventral groove, rather than through a caruncular scar or along the crest of the ventral keel. THE SPECIEs. ’ Fifty-seven species, two of which are new, are here recognized as occurring in North America. This is approximately half the total number of species, the genus having a similar development in Chile, Argentina and Peru. In North America most of the species occur in California, forty-five being known from that state alone. The genus, however, is not infrequent over most of western United States. One species, C. Torreyana, ranges northward to southern Alaska and another, C. albida, occurs as far south as southern Mexico. The most easterly ranging species are C. minima, which reaches central Nebraska, and C. texana, which reaches east-central Texas. Most of the species grow in warm, open, gravelly or sandy places, par- ticularly on the desert, others occur on dry, sunny, open flats in the pine-belt of the mountains and still others grow on open, grassy slopes or in clearings or burns in chaparral. The plants have no conspicuous place in the vegetational succession of the region, being secondary herbs commonly forming vernal or estival colonies in the earlier stages of the succession. describers of the oldest species made much of the shape and appendages of the corolla. With the increase of material the relative uniformity of the corolla-structures became apparent, and the at- tention, after having first been laid on the mere presence or absence of roughenings on the nutlets, was finally focused on the character of the nutlet-groove. Much of the confusion that has obscured the species of this genus is directly traceable to the persistent and ex- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 23 cessive emphasis which has long been laid on this character. The groove of the nutlet shows much more intraspecific variation than has been realized. While decidedly valuable in recognizing certain species, it is after all a character of only secondary importance, and, like every other character in this genus, is subject to decided and fre- quently erratic variations. It is obvious that the character is only to be used in conjunction with others. The most important specific characters are certain fruit-develop- ments. Among the most valuable of these are the extent and nature of the individualization of the axial or abaxial nutlet, the number, size, shape and surface of the nutlets, the nature of the nutlet-groove, the shape and height of the gynobase, and the height reached (in re- lation to the nutlet-length) by the style. The size of the corolla is frequently significant, as is also the size, shape, direction and depth of lobing of the fruiting calyx, and the shape, direction and pubescence of the mature calyx-lobes. The arrangement of the spikes or racemes, as well as the presence or absence of bracts, is important. So also is the character of the pubescence. The excessive variability of the characters of Cryptantha does not seem to have been fully appreciated. Particularly confusing in this genus is its propensity for unexpected, erratic variation in a single character or group of related characters. This is probably due to the fact that the species have a short life-cycle and are in active evolution. As a result the worker in this group is constantly confronted with much aberrant material. It is frequently necessary, therefore, to admit as atypical phases of a given species many specimens whose mass of characteristics indicate the accepted relation, even though the atypical character be a favorite one and perhaps that emphasized in the key. To attempt to name these atypical forms seems thoroughly unwise, since they are endless, and commonly represent odd plants or extremely localized phases, or, as seems likely, intraspecific hybrids. Some of the species admitted completely intergrade. In most cases these species occupy adjacent floral districts, and the inter- gradation is confined to a definite region lying between them, being thus in all probability of hybrid origin. In any case it seems best to maintain certain intergrading groups as specific, since to insist upon absolute non-intergradation as a criterion of specific difference would result in extensive reductions and give a condition as unsatisfactory as would extensive segregation. The groups treated as species, how- ever, can be readily recognized after a little study and appear to be eminently natural, having a characteristic gross aspect and credible ranges. 24 JOHNSTON SYSTEMATIC ACCOUNT. GENERIC CHARACTERS. Cryptantha Lehm. Calyx lobed to below middle or more com- monly almost to base, accrescent in fruit, usually deciduous; lobes linear to lanceolate or rarely lance-ovate, erect or connivent above in fruit, often slightly unequal; pedicels erect to widely spreading, com- monly very poorly developed. Corolla white, minute to evident, regular, glabrous; tube cylindrical, equalling calyx-lobes or surpassed by them; throat saucer-shaped or somewhat funnel-form, more or less closed below with 5 semicircular or subtrapeziform intruded ap- pendages; lobes 5, imbricate, orbicular to obovate-oblong, spreading or occasionally somewhat erect. Stamens 5, included; filaments short, equally inserted below the middle of the corolla-tube; anthers short- oblong, about as long as the filaments. Ovules 4 or rarely 2, amphi- tropous. Nutlets 1+, straight, usually vertical, ovate to lanceolate, affixed laterally through an elongate medial ventral groove to a pyramidal or subulate erect gynobase, unmargined or with more or less well developed marginal wing, smooth or somewhat warted or spiculiferous, neither rugose nor keeled, usually deciduous, tending to be heteromorphous. Seeds ascending, vertical, chalaza-end lower- most. Cotyledons flat, broad, undivided. Style included in the corolla-tube at anthesis, surpassing the tips of the mature nutlets or greatly surpassed by them, bearing a distinctly solitary simple ter- minal stigma.—West American annual herbs with stiffish pubescence. Leaves linear to lanceolate or spatulate, elongate, several lower pairs always opposite, the upper ones alternate. Flowers in bracted or naked fasciculate or cymosely disposed spikes or racemes. BIBLIOGRAPHY OF THE GROUP AND ITS CHIEF SUBDIVISIONS. Cryptantha Lehm. Del. Sem. Hort. Hamb. iv. (1833); F. & M. Ind. Sem. Hort. Petrop. ii. 35 (1836); Ann. Sci. Nat. ser. 2, V. 207 (1836); Linnaea xi., Lit. 103 (1837); G. Don, Gen. Syst. iv. 373 (1837); Greene, Pittonia i. 110-120 (1887); Baill. Hist. Pl. x. 373 (1891); Coville, Contr. U. S. Nat. Herb. iv. 164-166 (1893); Greene, Bot. San - Francisco Bay Reg. 261-262 (1894); Giirke in E. & P. Nat. PAlanzenf. iv. Abt. 3a, 110 (1893); Rydb. [Fl. Mont.] Mem. N. Y. Bot. Garden i. 330-332 (1900); Howell, Fl. N. W. Amer. 486-489 (1901); Jepson, Fl. W. Mid. Calif. ed 1, 444-446 (1901); ibid. ed. 2, 346-348 (1911); Piper, [Fl. Wash.] Contr. U. S. Nat. Herb. xi. 483-485 (1906); Rydb. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 25 FI. Colo. 289 (1906); Nels. New Man. Bot. Cent. Rocky Mts. 415- 416 (1909); Abrams, FI. Los Angeles ed. 1, 333-335 (1904); ibid. ed. 2, 306-308 (1917); Hall & Hall, Fl. Yosemite 205-206 (1912); Frye & Rigg, Northwest Fl. 327-328 (1912); Piper & Beattie, Fl. S. E. Wash. 209-210 (1914); Wooton & Standley, [Fl. N. Mex.] Contr. U.S. Nat. Herb. xix. 546-547 (1915); Henry, Fl. So. Brit. Columbia 255 (1915); Rydb. Fl. Rocky Mts. 725-729 (1917); Millsp. & Nutt. [Fl. S. Cata- lina Isl.] Pub. Field Mus. Nat. Hist., Bot. v. 232-234 (1923); Davidson & Moxley, Fl. So. Calif. 305-306 (1923); Johnston, Contr. Gray Herb. n. s. Ixx. 45 (1924); Johnston, Proc. Calif. Acad. Sci. ser. 4, xii. 1143-1148 (1924). Eritrichium § Cryptantha A. DC. Prodr. x. 129 (1846). Cryptantha § Eucryptantha Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a. 110 (1893). Lappula § Cryptantha Post & Kuntze, Lexicon 316 (1904). Krynitzkia F. & M. Ind. Sem. Hort. Petrop. vii. 52 (1841); Ann. Sci. Nat. ser. 2, xvi. 55 (1841); Linnaea xv., Lit. 116 (1841); Endl. Gen. Pl., Suppl. 2, 59 (1842); Meisn. Gen. ii. 366 (1842?); DC. Prodr. x. 134 (1846); Gray, Proc. Am. Acad. xx. 264-277 (1885) and Synop. Fl. N. Am. ii. pt. 1, Suppl. 423-429 (1886) in part; Greene, Bull. Calif. Acad. Sei. i. 203-208 (1885); Coulter, Man. Bot. Rocky Mts. 260-261 (1885) in part; Coulter, [Fl. W. Texas] Contr. U. S. Nat. Herb. ii. 286-287 (1892). Eritrichium § Krynitzkia Gray, Proc. Am. Acad. x. 58-61 (1874) and Synop. Fl. N. Am. ii. pt. 1, 193-197 (1878) in part. Krynitzkia § Eukrynitzkia Gray, Proc. Am. Acad. xx. 267-275 (1885) and Synop. Fl. N. Am. ii. pt. 1, Suppl. 424-428 (1886). Eritrichium § Rutidocaryum A. DC. Prodr. x. 130-132 (1846) in part. Eritrichium in part, Benth. & Hook. Gen. PI. ii. 850 (1876); Wats. Bot. Calif. i. 527-529 (1876); Hemsl. Biol. Cent.-Amer. ii. 378 (1882); Macoun, Cat. Canad. PI. i. 337 (1883). Piptocalyx Torr. in Wats. Bot. King Exped. 240 (1871); Torr. Bot. Wilkes Exped. 413, t. 12 (1874); Greene, Pittonia i. 59 (1887); Baill. Hist. Pl. x. 373 (1891); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 109 (1893); Howell, Fl. N. W. Amer. 483 (1901); Piper, [Fl. Wash.] Contr. U. S. Nat. Herb. xi. 481 (1906); Abrams, Fl. Los Angeles ed. 1, 331 (1904); ibid. ed. 2, 305 (1917); Frye & Rigg, Northwest Fl. 326 (1912); Henry, Fl. So. Brit. Columbia 255 (1915); Davidson & Moxley, FI. So. Calif. 305 (1923). Eritrichium § Piptocalyx Gray, Proc. Am. Acad. x. 58 (1874). 26 JOHNSTON Krynitzkia § Piptocalyx Gray, Proc. Am. Acad. xx. 275 (1885). Lappula § Piptocalyx Post & Kuntze, Lexicon 316 (1904). Cryptantha § Piptocalyx Johnston, Contr. Gray Herb. n. s. Ixviii. 55 (1923). Krynitzkia § Pterygium Gray, Proc. Am. Acad. xx. 276 (1885) and Synop. Fl. N. Am. ii. pt. 1, Suppl. 428-429 (1886) in part. Cryptantha § Pterygium Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 110 (1893). Lappula § Pterygium Post & Kuntze, Lexicon 316 (1904). Eremocarya Greene, Pittonia’ i. 58 (1887); Baill. Hist. Pl. x. 373 (1891); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 109 (1893), Howell, Fl. N. W. Amer. 483 (1901); Abrams, Fl. Los Angeles ed. 1, 331 (1904); ibid. ed. 2, 304 (1917); Frye & Rigg, Northwest Fl. 326 (1912); Wooton & Standley, [Fl. N. Mex.] Contr. U. S. Nat. Herb. xix. 544 (1915); Rydb. Fl. Rocky Mts. 719 (1917); Davidson & Moxley, FI. So. Calif. 306 (1923); Johnston, Contr. Gray Herb. n. s. Ixviii. 56-57 (1923). Lappula § Eremocaryum Post & Kuntze, Lexicon 316 (1904). Greeneocharis Giirke & Harms in E. & P. Nat. Pflanzenf., Gesamt- reg. 462 (1899); Rydb. Fl. Rocky Mts. 718 (1917). W heelerella Grant, Bull. So. Calif. Acad. Sci. v. 28 (1906). Johnstonella Brand in Fedde, Repert. in press. ARTIFICIAL Ky To SPECIES. = with margins decidedly winged or —S cels usually evident, “resvauped r, 1-4 mm. | utl RON a 1. C. holoptera. Nutlets decidedly 1 i uieneccchous ieee 2. C. racemosa. Pedicels obscure or none, less than 2 ial. Fruiting calyx ca. 2 mm. long; nutlets 0.6-0.7 mm. long.. ~ C. angelica Fruiting calyx 2.5-3 mm. long; nutlets ca. 1.7 mm. long. 4 . C. inaequalta. Nutlets pear or if slightly heteromorphous t Nutlets et he paseues bent, margin thickish; gynobase NINE gig is es. cs ke kk 5. C. pusilla. Nutlets dlls, _—e ap thin; gynobase subulate. Nutlets ughened, decidedly plano- tent in spovteeies Pics (the back rounded and We TOO AOE ss, ee a 6. C. costata. aeeeoeet = tuberculate or reesei not noticeably plano- ion. Nutlets shia or shaely 2; calyx obliquely conical Re RBG ee i 17. C. utahensis. Nutlets 4, —* symmetrical. — es conspicuous; mature calyx viously waar Gan bee 18. C. oxygona. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 27 . Corolla os seas calyx eet al s broad:.as lone tise. C. pterocarya. —— with margins ath Ls raided never with a ing or cng edge Nutlets all smooth. Hairs on sae uncinate or decidedly arcu St nia? eg oe height of ss os ‘this with an Segal e AULT S Fae Pewee earns 53. C. rostellata Style peachiter Tes than 1/2 height of nutlet; groove nutlet clo: Nutlets nearl Sith rostrate; hairs on calyx usually Be ee as OO a a 54. C. flaccida. Nutlets decidedly co cr i acute; hairs on calyx Parga and less pale... ......2...5..5. 55. C. sparsiflora. Hairs oi aly Nutlet with Besson groo Nutlets 1; flowers a cllaey t biserial. .... br. t, apa eng Nutlets 4; 4; flowers in "beaial sake “ clo, affinis. Nutlet with a mneelly placed groove Style reaching cha -3/4 Pr of oe ntlets oblon: = forl Stems she 8 ’ nutlets io aidal e 16. C. maritima. Stems green; nutlets 1-4, bain! when solita segues less than 1 mm. broad; diffuse prostrate ee ee ae . C. Brandegei. oC aeeg rey ary Ss gear . Abramsit. Spikes snaked 3 with a few bracts at base. Nutlets peoasiiy ovate. Cc ee 4-7 mm. broad; ere mly ternate.................-- C. Hendersoni. Corolla inconspicuous, 1-2 mm. broad; ae usually so solita ary or geminate. Ret Sey ected above the leafy mass of plant, well cette . Torreyana. Spikes usually solitary, not sharply differenti- ated from the leafy peduncular stems. Nutlets a Montana and shor Nutlets oblong-0 hover Style reaching Mgt height ot seer these 1-4. C densely appressed hispid- villous rior. Calyx sparsely fnehisut; coastal and evid ently spreadi C3 28 JOHNSTON 7 Hairs on upper part of calyx-lobes spreading or ascending. Style reaching to 2/3-3/4 height of nutlets. 50. C. Clevelandi. Style reaching to 1/4-1/2 height of ats 8. C. microstachys. Style almost reaching the nutlet-tips or pi oe them ian of nutlets acute, at least above the Corolla conspicuous; style Mg ke surpassing 2-2 nutlets, these .42. C. mohavensis. Corolla i Eeainies Fae reachitig ‘to the nutlet-tips or a trifle surpassed by them; nutiets 1.2-2 mm. long. ............. 43. C. Watsoni. Margin of nutlets rounded or oove of nutlet opened balan into a tri- ith a definite central axis; not Californ Sage 45. C. Fendleri. reer of nutlet closed throughout; pla arly sahege pe d; = rnian. . 47. "C. hispidissima. N — all rough or atleast oo . them Nutlets decidedly per ature pee sinoacloa seal somata to the flattened ey ome decidedly gibbous on axial side, persistent. 4. C. dumetorum. — Rigen mewhat spreading, nor at all ee hi Reo jal gine spreading or reflexed, most aE. e on axial side..... cette: woes 13. _C. echinosepala. Odd res abaxial, surpassed by style. Pap bracteate throughout; seed a 7. C. micrantha. ikes naked or ate 80; calyx deciduo Pedicels slender ME IONE oss 2. C. racemosa. Pedicels stout a obscure, les than 1 mm. long. Nutlets 0.6-0.7 mm. long............... 3. C. angelica. Nutlets 1-1.7 m - ia eine 1.3-1. 7 mm. gies calyx 2-3 mm. ee Wiens cad ewe cere see Bh C. inaequata. Nutlets ca. 1 mm. long; calyx 3-4 m A res pe ee eee a 3. "C. angustifolia. Odd — axial; style surpassed or occasionally o the nutlet-ti Nutlets small 0.7-0.9 mm. eon triangular-ov a 5. C. micromeres. Nutlets an 1-2.3 mm. long, ovate tolanceolae Odd nutlet smooth and shiny............ . C. maritima. Odd eer tuberculate or papillate. PR TORO es eee 23. C. minima. Spikes naked. Odd anutle oe ee calyx-lobes ckened....... 24. €. crassise pala. y Odd ‘mr cuesiet more - less granulate; calyx rately thickened. eagoar ovate, smoothish or sparsely reulate, odd one ca. 1.9 mm. i Per ee oe Oe 20. C. Pattersoni. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 29 oe lanceolate or narrowly ig see a ——— odd one 2-2.6 aa Sait aes 2 Oe Wo aie 22. C. Kelseyana. “Oak Gireumacinsile £00. ee 10. C. circumscissa. ni ghou Groove of nutlet eoaais ee to form a bas angular areola occupyin; of the ventral side of a gynobase na Hetty pyramidal. 11. C. albida. — of nutlet narrow, ssatesly dilated below aohiise wubMhecu. 5555 err ok 7 C. micrantha. Spikes get if at all bracted. Nu gp s bent, lucid; gynobase narrowly pyra- Bh te ee a pusilla. Nutlets = straight, usually dull; gynobase subu- Nutlets small, 0.5-0.7 mm. — oe sk 9. C. Grayi. Nutlets larger, 1.5-3 se Fag Nut — a te, "i ith a suggestion medial diriel ridge; plant dull dark Se Ray fore Re eae A tas - muricata. Nedets: ee ate to lance-ovate; plants canescent. Sots SONAR re oes as 26. C. patula. Spikes praca EN ternate or re, geminate or quinate.......... C. intermedia. eae ger fs p Race by or about reaching to an f the pan Core olla a conspicuo broad. Calyx evi ens Ore esteeliats, conspicuously lonz- t hispid or hirsute............... 34. C. crinita. ase on a Nutlets usually geneon ovate-triangular, uminate; calyx -lobes shorter than autlet, very iocaely connivent....... 35. C. excavata. i and closely connivent over ete AL ey | + Anan Lae ing 1/3-1/2 eal eg of sere ee a 25. C. decipiens. Axial nutlet developing; gynobase about — = motets 17. C. utahensis. Normally all 4 ts developing; gynobase elongate, 2/ 3 a3) r ‘height of Satie. Plant low, spreading. and widely branched; SUIMER SONUNEY ee en mariposae. Plant erect one pene aciy branched; spikes eminate or ternate. Hairs on calyx subinflated, extremely coarse; stem usually stiffly erect and forming a conspicuous axis, branches 30 JOHNSTON usually widely divaricate........... 27. C. foltosa. Hairs on calyx slender, not notably coarse; stem irregularly branche ad ee ungent-hirsute; nutlets ob- Lie oaks C. intermedia. (Gist Beaty i hispid nutlets ovate, ubercul RUPOWUAUE 5s i, os es 6. C. Hendersoni. Corolla inconspicuous, 0.5-2 mm. broad. Ovules 2, nutlet and calyx bent........... 12. C. recurvata. forsick 4; nutlets and calyx straight. nee small, ca. 1.5 mm. long; spikes IAS DPRCION oo fae es vas as 37. C. Traskae. Nutlets larger, 2-3 mm. long, spikes very spar- gly if at all bracteate. Nutlets usually see Pe gynobase 1/4-1/2 height of nutlet. utiet. persistent, axial... ..;.. 23.4.0. 2t.-C; terana, Nutlet i deciduous, abaxial. ...25. C. decipiens. Nutlets Nutle decwdedl ly ovate Plant closely pik, ong pallid, usually 2-3 pikes commonly gemina ate or ternat i Foren ee eae es ee 39. C. simulans. Plant gape tag: -hispid, usually 1-1.5 dm. tall; s a 8b usually solitary or rarely geminate. Nutlets with tg rounded tubercula- MM eso a ee .ambigua. Nutlets with elongate papillae or WIS ie pet xa Oud 41. C. echinella. Nutlets more or less oar Stems i Rig HiMsie so 29. C. barbigera. Stems strig Nutlets verrucose or verrucose-muri- ee ee oe at 30. C. nevadensis. Nutlets ailiciarcee on 31. C. scoparia. DESCRIPTION AND CLASSIFICATION OF THE SPECIES. Ser. I. ANGUSTIFOLIAE. Nutlets 4, muricate or tuberculate, dark with pale roughenings, triangular-ovate or triangular-oblong, with sides acute or knife-like or definitely winged, homomorphous or in most species decidedly hetermorphous with odd nutlet abaxial larger and sometimes slightly less roughened than the others; style definitely surpassing the nutlets. Pedicels slender, 1-4 mm. long; Yon. annuals. rphous, broadly winged................. 1. C. holoptera. Nutlets hetero us, narrowly winged............... 2. C. racemo. cho Pedicels sto winged stout and very short, le sn aa 1 mm. long. Nutlets with a knife-like Nutlets heterom: nei tog THE NORTH AMERICAN SPECIES OF CRYPTANTHA 31 Fruiting calyces ca. 2 mm. long; nutlets 0.6-0.7 mm. BOW 5 a oe i ve eggs Oe ee ee ae ye 3. C. angelica. Fruiting calyces 2.5-3 mm. long; nutlets ca. 1.7 mm. sa . C. inaequata. Nutlets homomorphou ee icin Lahepeulie. bent; gynobase — Moises e te tee eet eae . C. pusilla. Y SUeptt sea wep GS es Re Gr ELT yee ok ae 6. C. costata. spikes bracteate arcuate 5 WEV10 Chi kc ds 7. C. micrantha. Spikes bractless or practically so; style slender. Nufleta heteromorphots, 0.75555. . occa S.-C. a nen Nutlets homomorphous............. era y Vet ee oN ete 1. C. holoptera (Gray) Macbr. Coarse erect annual, iat or rarely somewhat suffruticose towards the base, 1-6 dm. tall; branches rather numerous and ascending, hirsute as well as strigose, usually drying brownish; leaves oblanceolate to lance-linear, 3-6 cm. long, 3-8(-12) mm. broad, acute or obtuse, coarsely pustulate and hirsute beneath but less so above; racemes geminate, naked or sparsely bracteate, usually ca. 5 cm. but becoming 10 cm. long; corolla inconspicuous, tube shorter than calyx, the ovate lobes as- cending and less than 1 mm. long; fruiting calyx oblong-ovate, 2.5- 3.5 mm. long, subsymmetrical, rather persistent; pedicels ascending _ or decurved, 0.7-1.5 mm. long; mature calyx-lobes lanceolate, some- what connivent, midrib thickened and hirsute, margins strigose; nutlets 4, homomorphous, 1.5-2.5 mm. long, body oblong-ovate or triangular-ovate, dark with pale tuberculations, margin narrowly to quite broadly winged, groove open or closed above but dilated into an areola below; gynobase slender, about equalling the nutlets; style clearly surpassing the nutlets but shorter than calyx-lobes——Contr. H i Deserts of California and western Arizona. Not common. Arizona: Ehrenberg, 1876, Palmer (G, TYPE); near Peach Springs, Grand Canyon, 1884, Lemmon (UC). Cattrornta: sandy places near Palm Springs, 120 m. alt., cer 2066b and 2070 (G); Marshall Canyon, 16 km. west of Conthells; 90 m. alt., Hall 5786 and 5788 (UC); oe Davy 8007 (UC); rocky talus slopes at foot of The Needles, Grinnell (UC); ‘“‘The Caves,” sink of the Mohave River, Parish 9855 (UC). This species was founded by Gray on two collections cited thus: “S. Utah, Capt. F. M. Bishop; Ehrenberg, Arizona, Palmer.’’ The material collected by Bishop is an immature scrappy specimen of C. pterocarya. Palmer’s Ehrenberg collection consists of two mounted oe JOHNSTON sheets with fruit. As would be suspected almost the whole of Gray’s original description is based upon the Arizonian plant. Subsequently Gray excluded Bishop’s collection and cited only the Ehrenberg material as representative of his species. For these reasons Palmer’s Ehrenberg collection is taken as the type of C. holoptera. C. racemosa (Wats.) Greene. Long-lived annual often de- cidedly suffruticose towards the base, 1-10 dm. tall; stems single with numerous ascending branches or many and diffusely branched, younger parts green, inconspicuously strigose and commonly hirsute, older parts woody and becoming brown from the falling away of the pale bark; leaves oblanceolate, acute, hirsute, pustulate, the early ones 3-6 cm. long and 6-12 mm. broad, the later and more abundant 1.5-4 cm. long and 2.5 mm. broad; racemes apparently forked and paniculately disposed, inconspicuously and irregularly bracted, 3-15 em. long; corolla very inconspicuous, limb ca. 1 mm. broad; fruiting calyces oblong-ovate, ascending, 2-4 mm. long, slightly asymmetrical, tardily deciduous, inconspicuously biserial; pedicels usually well de- veloped, 1-4 mm. long, slender, frequently nodding; mature calyx- lobes lance-linear, somewhat strigose, hirsute along the thickened midrib; nutlets 4, heteromorphous, triangular-ovate, the acute tips slightly out-curved, groove open or closed above but below broadening cut into a shallow broadly triangular areola; odd nutlet next the ab- axial calyx-lobe, 1-2 mm. long, subpersistent, finely muricate or tuberculate or both, light or dark; consimilar nutlets 0.8-1.5 mm. long, acute, tending to be very narrowly winged, dark with pallid tuberculations; gynobase subulate, 34 length of odd nutlet and about equalling consimilar nutlets; style much surpassing the nutlets, almost as long as the mature calyx-lobes.—Pittonia i. 115 (1887); Johnston, Proc. Calif. Acad Sci. ser. 4, xii. 1147 (1924). Eritrichium racemosum Wats in Gray, Proc. Am. Acad. xvii. 226 (1882). Kry- nitzkia racemosa Greene, Bull. Calif. Acad. Sci. i. 208 (1885). John- stonella racemosa Brand, Fedde Repert. in press. K. ramosis- sima Gray, 1. c. xx. 277 (1885). C. suffruticosa Piper, Proc. Biol. Soc. Wash. xxxii. 42 (1919). C. racemosa, var. lignosa Johnston, Univ. Calif. Pub. Bot. vii. 445 (1922). J. racemosa, var. lignosa _ Brand, Fedde Repert. in press. thern Nevada and western Arizona, and southwestward to middle Lower California and Carmen Island; usually growing in rocky places. Nevapa: semishaded places, Las Vegas Mts., Goodding 2381 (G, UC); Petrified Forest Canyon west of Logan, 540 m. alt., Heller 10446 (G) and 10447 _ THE NORTH AMERICAN SPECIES OF CRYPTANTHA 33 G, UC); rocks, Ash Meadows, Purpus 6024 (UC). segue Grand tice re (eb, Gray (G); Diamond Creek Canyon, Wilson 170 (UC). CALrFoRNIA onand about soca: Silver Canyon, White Mts. east if Laws, Heller & 8209 (G): Beano Canyon, Hall & Chandler 7034 (UC. TYPE of C. racemosa, var. lignosa); Surprise Canyon, Parish Bete (UC); Morongo Wash, 900 m. alt. Parish 3337 (UC); among rocks on canyon-side, oi Canyon, 300 m. alt., meg 3 1002 (G); s sandy bet pte Palm Sprin 120 m. alt., Spencer 9 (G); among rocks, Cathe ral Canyon near Palm Spri rings, 150 m. alt., plea 2078 (G); Borregos amit 1894, Brandegee (UC): ; San Felipe Creek, chepensege 2704 (G); Split Mt., 1905, Brandegee (UC); canyon near Mesquite tati ish 7 tion, Par 75 Se ae a m; UC, ISOTYPE) ; desert sand, Colo- rado , san 54 (G). Lower Cauirornia: Cantillas Mts., 1883, Orc ult (Ge Cantillas Mts.. “884, Orcutt (G, oe tGh Julio Canyon 1889, gee (UC); San Se astian, 1884, Bra ); San eguis, 889, Brandegee (UC); Santa Maria, 1889, Brandegee UC); Cajon de Santa e ); 88 Maria, 1889, Brandegee (UC); talus, Ports. Refugio, Angel de la Guarda Island, Johnston 3374 (G); aey slope, San Estaban Island, Johnston 3175 (G); cobble “tone fees: South San Lorenzo sages Johnston 4192 (G); 188 Se ; foot of gypsum one San Marcos Island, Se Ba 3621 (G); Carmen A very distinctive ee to be confused only with C. holoptera, which has homomorphous, more broadly winged nutlets. The very diffusely branched form of C. racemosa, which is most common in the Mohave Desert, has been described as C. racemosa, var. lignosa. It may be worthy of recognition. 3. ©. angelica Johnston. Much branched finely strigose herb 15-25 em. high; stems spreading, rebranched, brown and glabrous below, strigose and canescent above; leaves linear, 8-24 mm. long, 1-2 mm. broad, loosely strigose, densely and minutely pustulate, ascending; spikes solitary, crowded, slender, naked, 4—9 cm. long, very floriferous; corolla minute, ca. 1 mm. long, ca. 0.5 mm. broad; fruiting calyces oblong-ovate, 1.5-2 mm. long, ascending, subsessile, somewhat asymmetrical; mature calyx-lobes linear, erect, ribbed, sparsely hispid, abaxial one slightly the longest and most hispid; nutlets 4, hetermomorphous, triangular-ovate, dark with small pale tuberculations, margins sharp and with a very narrow knife-edge, back convex, groove closed above but dilated below into a shallowly triangular areola; odd nutlet next the abaxial calyx-lobe, ca. 0.7°mm. long, somewhat persistent; consimilar nutlets ca. 0.6 mm. long; gynobase narrow, almost as long as the consimilar nutlets; style very much surpassing the nutlets.—Proc. Calif. Acad. Sci. ser. 4, xii. 1143 (1 aes only from Angel de la Guarda Island in the Gulf of Califor- nia. Lower Cauirornts: on silty flats, Angel de la Guarda Island opposite Pond Island, pe aia 4221 (G, UC, IsoTyPEs). 34 JOHNSTON A well marked species, probably most related to C. racemosa, from which it differs in such developments as smaller and sessile calyces, much smaller nutlets, and narrower leaves. To judge from the de- scription of C. inaequata it differs from that species in its denser spikes, and much smaller calyces and nutlets. Except for the very narrowly winged margin, the nutlets of C. angelica are very suggestive of those of C. Grayz. 4, C, inaequata Johnston. Loosely and sparingly branched herb 3-4 dm. high; stems erect or ascending, hispid and strigose or hirsute towards the base; leaves oblanceolate to linear, 2-4 cm. long, acute, not numerous, more or less hispid, pustulate especially underneath; spikes geminate or solitary, 4-12 cm. long, at times sparsely bracted below; corolla inconspicuous, tube shorter than calyx; fruiting calyx ovate-oblong, 2.5-3 mm. long, ascending, pedicels less than 0.5 mm. long; mature calyx-lobes lanceolate, midrib thickened and hirsute, axial lobe most pubescent and thickened; nutlets 4, heteromorphous, triangular-ovate, dark with small pale tuberculations, margins de- cidedly acute, groove closed above but below gradually dilating into a shallow triangular areola; odd nutlet ca. 1.7 mm. long, somewhat persistent, slightly lighter than the others, next the abaxial calyx- lobe; consimilar nutlets ca. 1.3 mm. long; gynobase equalling the con- similar nutlets but surpassed by odd nutlet; style much surpass the nutlets.—Univ. Calif. Pub. Bot. vii. 444 (1922). Joiciouals inaequata Brand, Fedde Repert. in press. Southeastern California. CALIFORNIA: among rocks, Pleasant Canyon, Panamint Mts., 600 m. alt., Hall & Choniiee 6925 (UC, TrPE); Baxter, Parish 9859 (UC). The status of this species is problematic, since the collections upon which it was based have not neen available to me for several years. The description given above is adapted from a preliminary diagnosis made in 1922. It is possible that the plant is only a form of C. race- mosa, although it has been noted as differing in its subsessile calyces and less extended duration. 5. C. pusilla (T. & G.) Greene. Low plant usually with very numerous prostrate-ascending stems; these very slender, usually strictly branched, 3-15 cm. long, canescent, appressed villous-hirsute; leaves crowded near base of plant but distant above, spathulate- linear to linear, 1-3 cm. long, 1-2 mm. wide, somewhat pustulate and hispid below but less so above; spikes solitary or geminate, oe cm. long, naked or with a few minute b orter than calyx, 1.5-2 mm. long, lobes slightly spreading hae ca. 0.3 THE NORTH AMERICAN SPECIES OF CRYPTANTHA 35 mm. long; fruiting calyces 2-2.5 mm. long, broadly ovate, sym- metrical, sessile or subsessile, early deciduous; mature calyx-lobes lance-ovate or oblong-lanceolate, hirsute, midrib only slightly thick- ened; nutlets 4, homomorphous, lucid, ca. 0.8-1.2 mm. long, tri- angular-ovate, bent, tan-colored with low light-colored tuberculations, margin angled and beveled, groove opened or closed above but ex- panding below into a shallow open triangular areola; gynobase nar- rowly pyramidal, about equalling the nutlets; style much surpassing the nutlets, ca. 14 length of gynobase.—Pittonia i. 115 (1887). Eri- trichium pusillum T. & G. Pacif. R. R. Rep. ii. pt. 2, 171 (1856). Krynitzkia pusilla Gray, Proc. Am. Acad. xx. 274 (1885). Southern Arizona and New Mexico southward to Durango. Arizona: near Nogales, 1902, Orcutt (UC); Douglas, 1907, Goodding 2264 (UC). New Mexico: without locality, Wright 1571 (G, NY). Texas: Fort near Chihuahua, Pringle 184 (G, UC); vicinity of Chihua ua, 1300 m. alt., Palmer 65 (G). Duranco: Tepehuanes, Palmer 28 (G, UC); Durango, Palmer 139 (G, UC) and 227 in pt. (G). A very distinct species readily recognized by its beveled, bent, lustrous nutlets that just equal the somewhat basally constricted gynobase. The type is given as coming from “ Rio Pecos to Llano Estacado, ete. in sandy soil, March.” From a study of the journal in Pope’s Report, the locality given is found to be within about 100 kilometers traveled between March 24th and 26th 1854. This would make the type locality lie somewhere near the juncture of Ward, Crane and Ector counties, Texas. 6. C. costata Brandg. Coarse stiff few-branched herb 1-2 dm. high; stems (especially younger parts) canescent, densely villous- strigose and usually somewhat hirsute as well; leaves lanceolate to linear, 1-3 cm. long, 2-4 mm. wide, broadest near base, apex acute, above villous-strigose and sparsely hispid, beneath more hispid and frequently also pustulate; spikes rigid, 2-5 cm. long, solitary or geminate, sparsely leafy-bracted; corolla inconspicuous, ca. 2 mm. long, tube shorter than calyx, lobes broad and ascending; fruiting calyces ovate-oblong, 4-6 mm. long, subsymmetrical, spreading, de- ciduous, subsessile; mature calyx-lobes linear-lanceolate, somewhat connivent above with tips slightly spreading, midrib thickened hir- sute hs ee 1 1 or ae aS pl with the nutlet next the abaxial calyx-lobe slightly the largest, ca. 1.8 mm. long, triangular oblong-ovate, back strongly convex, incon- spicuously rugulose or faintly verrucose, face noticeably flat or even 36 JOHNSTON slightly convex, margins sharp and narrowly winged; groove very shallow, closed above but dilating below and merging into the deltoid shallow areola; gynobase subulate, equalling the nutlets; style not sharply differentiated from the gynobase, much surpassing the nut- lets—Bot. Gaz. xxvii. 453 (1899). C. seorsa Macbr. Contr. Gray Herb. n. s. xlviii. 46 (1916). Deserts of Southern California. a cer (G); in sand, Palm Canyon, 180 m. alt., Spencer 1627 (G); margin of wash, Borregos Spring, 1889, K. Brandegee UC) ; Borregos Spring, 1895, Brandegee TYPE . er ers (G) A very distinct species readily recognized by its peculiar nutlets: and by its rather coarse, stiff habit and very canescent herbage. 7. C. micrantha (Torr.) Johnston. Slender strigose ascendingly branched dichotomous herb 5-15 em. high, drying brownish; root and lower parts of stem stained with dye; leaves oblong-oblanceolate, 3-7 mm. long, 0.8-1.4 mm. broad, canescent-strigose and occasionally short villous-hirsute, rounded at apex, uppermost scarcely reduced and extending through the inflorescence; spikes very numerous, soli- tary or geminate, dense, strongly unilateral, leafy-bracted throughout, 1+ em. long; corolla inconspicuous or medium-sized, limb 0.5-2.5 mm. broad, faucal appendages poorly developed; fruiting calyx ovate-oblong, 1.8-2.5 mm. long, slightly asymmetrical, decidedly bi- seriate, base broadly conical; pedicels 0.5-0.8 mm. long; mature ealyx-lobes oblong-lanceolate, broad, erect, hirsute, Ghic not par- ticularly thickened; nutlets 4, homomorphous or somewhat hetero- morphous, 1-1.3 mm. long, tlamsbeous or brown, smooth or tuber- culate, abaxial one usually a trifle the largest and most persistent; groove extending full length of nutlet, narrow, scarcely broadened at base; gynobase subulate, nearly as long as the calyx, much surpassing the nutlets and bearing at its summit the sessile stigma.—Contr. Gray Herb. n. s. Ixviii. 56 (1923 Var. . Corolla Gacaiavicaos, 0.5-1.2 mm. broad.— Eritrichium micranthum Torr. Bot. Mex. Bound. 141 (1859). Kry- nitzkia micrantha Gray, Proc. Am. Acad. xx. 275 (1885). Eremocarya micrantha Greene, Pittonia i. 59 (1887). C. micrantha Johnston, |. c. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 37 Eremocarya muricata Rydb. Bull. Torr. Bot. Cl. xxxvi. 677 (1909); Macbr. Proc. Am. Acad. li. 545 (1916). Southeastern Oregon to Utah, southward to northern Lower California and Arizona, and eastward along the Mexican boundary to western Texas. OREGON: without Piet 1898, Cusick 2020a (G). Utan: Stansbury Island, oa m. alt., Watson 856 (G); valley of the Virgin near St. George, Parry 164 (G, sory TYPE of "E. muricatum). Nevapa: Logan, Kennedy 1832 (G); - sag ‘tandy ravines, Moapa, Goodding 2203 (G); deep sand, Beaver Dam Wash, Goodding 2144 (G). CALIFORNIA: sandy place, Mohave Desert, m. alt., Spencer 436 (G); Barstow, Jones 106 (G); Mohave Desert, 1880, Lemmon (G); Acton, Elmer 3682 ies Angeles, 1880, Nevin (G); sand in Arroyo ; } prings, 678 m. alt., Spencer 857 (G). pti ie a — Sebastian, 1880, vabeateg ae Pit C); eithout locality, a Orcutt (G). ARI- NA: near T se ee me 1112 (G); W raphy tage 371 (G); near Camp penta 1881, le (G); Prescott, Rusby 745 7 New Mexico: without locality, oe "Wright 1565 (G). Trex ion ee Rio Grande, El Paso, Thurber 181 (G); El Paso, Jones 3703 (G ). Var. lepida (Gray) Johnston. Corolla medium-sized, 1.0-1.5 mm. broad; plants usually coarser than in var. genuina.—Contr. Gray Herb. n. s. Ixviii. 57 (1923). LEritrichium micranthum, var. lepidum Gray, Synop. Fl. N. Am. ii. pt. 1, 193 (1886). Krynitzkia micrantha, var. lepida Gray, Proc. Am. Acad. xx. 275 (1885). Eremocarya lepida Greene, Pittonia i. 59 (1887). Eremocarya micrantha, var. lepida Macbr. Proc. Am. Acad. li. 545 (1916). Southern California, most common in warm montane valleys. & z 5) ro > & z, 5 g — nutlets of C. micrantha are 1. variable. Some plants have all the nutlets smooth, others have all of them tuberculate, while still others have the abaxial nutlet of each calyx tuberculate and the remaining ones smooth. Eremocarya muricata is described as dif- fering from C. micrantha in having tuberculate nutlets, but in the isotype of that species contained in the Gray Herbarium I find that 38 JOHNSTON although most of the calyces have roughened nutlets a few of them (and these always the oldest) have decidedly smooth nutlets. In plants characteristically smooth-fruited it is quite common to find that the oldest calyces contain three smooth nutlets and a tuberculate one. Not only is there a decided tendency towards heteromorphism in markings and roughenings, but to a slight degree also in size and firmness of attachment. The abaxial nutlet is commonly a trifle larger than the other nutlets and is somewhat more firmly attached to the gynobase. None of the nutlet-variations can be geographically correlated. Despite its variable fruit the species is readily recognized because of its densely bracteate inflorescence, deeply dye-stained root and long-protruded gynobase. 8. C. angustifolia (Torr.) Greene. Diffusely branched from the base, 5-20(-45) cm. tall; stems canescent, villous-hirsute, commonly somewhat strigose-villous, lowermost branches decumbent or loosely ascending; leaves linear, 1.5-4 cm. long, 14 mm. wide, spreading, not crowded below, hispid or strigose, somewhat pustulate especially underneath; spikes geminate, usually ca. 5 cm. long, rather dense, commonly naked; corolla usually inconspicuous, tube 1-2 mm. long, limb 1-2.5 mm. broad; fruiting calyces ovate-oblong, 3-4 mm. long, stiffly ascending, strongly biseriate, slightly asymmetrical; pedicels less than 0.5 mm. long; mature calyx-lobes lance-linear, rigid, slightly connivent, midrib thickened and hirsute, margin somewhat villous- ciliate, abaxial lobe longest and most hirsute; nutlets normally 4, heteromorphous, ovate-oblong, brown or plumbeous with pale tuber- culations or rarely murications, back convex, face flattish, margin somewhat angular; odd nutlet next the abaxial calyx-lobe, a trifle larger and more persistent than the similarly colored and shaped con- similar nutlets which are ca. 1 mm. long, groove usually narrowly open above but broadening at the base; gynobase columnar, equalled by consimilar nutlets but shorter than odd nutlet ; style usually sur- passing even the odd nutlet.—Pittonia i. 112 (1887). Eritrichium angustifolium Torr. Pacif. R. R. Rep. v. 363 (1857). Krynitzkia angustifolia Gray, Proc. Am. Acad. xx. 272 (1885). Southern California and southern Utah, southward to Lower Cali- fornia, Sonora and western Texas. Cauirornia: Death V: Tov ; Orcutt (UC); Barstow, 1915. @ ce oo OO) eure pled a, 960 m. alt., Spencer 2091 (G); Needles, 1884, Jones (G); Riverside Mt., 1910, Grinnell (UC); Palm Springs, 1913, Eastwood (G); in sandy places, Palm ‘ Spencer 843, 849, 853, 855, 1526a, 2066a, 207 1b in pt. and 2118 (G); sand dunes, Old Beach, near Holtville, Parish 8124 (UC); Twentynine Palms, THE NORTH AMERICAN SPECIES OF CRYPTANTHA 39 _—_ ——— (UC); desert — Indio, 30 m. alt., Spencer 1517 (G); andy Shaver’s Well near Mecca, 59 m. below sea-level, Munz & Kec 63 Nay, satis. Mecca, Sune 1512 and 1785 (G ye Cameron Lake, eal (UG); Yaqui Wells, Eastwood 2632 {S5 bottom lands near Colo- rado River, 750 m. per Hall 5922 (UC); in sandy places. Colorado Desert, Spencer 195, 196, 197 oe 201 (G); without locality, a 500 (G). Lower ALIFORNIA: stony ridges, Los Ange s Bay, Pa 06 (G); San ae ee 241 (G, UC); Angel de os Guarda Island, Johnston 4227 (G). NEV } a, Gooding 2181 (G, UC); O vison 450 alt pe sag 10430 (@); pees isd osa in moist place n site stati ller 10976 (G sandy wash, Meadow Valley Wash, Bashing 2169 (G): ‘Moday Valley, Ken- nedy & Gooding 26 Seu Arizona: Fort Yuma, Thomas (NY. Type); Fort uma, DuBarry (NY, cotType); Gila fetter Thurber 690 (G); Tueson, 1907, Loyd (G); oan 1884, Parish (G); Tue 1894, Toumey (UC); campus of University of Arizona, Thornber 407 sa “516 (UC); without locality, 1876, Palmer (G); without locality, 1881, Pringle (G). Sonora: Torres, 1902, Purpus (UC); dry ie wears of Guadaloupe, Smith (NY); Las Durasnillas a Brandegee (UC); Guaymas, Palmer 169 (G, UC). New Mexico: of Organ Mts., 1905, Wooton (UC). Texas: El Paso, 1884, Jones XG). Probably the most common Cryptantha in the lower deserts of California. Itis readily recognized by its characteristic ashy herbage, dense spikes, and heteromorphous dark nutlets which are covered with small light colored low tubercles. It is a well marked species, having its closest relation in C. Grayi. 9. C. Grayi (Vasey & Rose) Macbr. A small slender herb 5-15 (-18) em. high; stems usually several, strict or spreading, appressed or spreading villous-hispid; leaves quite numerous, linear, 1-4 cm. long, 1-2 mm. wide, densely pustulate-setose beneath but much less so above, basal ones somewhat aggregated, upper ones reduced; spikes usually geminate, naked, densely flowered, 1-4 cm. long; corolla minute to medium-sized, tube shorter than calyx, limb 0.5-3 mm. broad; fruiting calyx oblong-ovate, 1-2 mm. long, sessile or sub- sessile, decidedly biseriate; mature calyx-lobes lance-linear, rigid, slightly connivent, midrib short-hirsute, margins somewhat villous- hispid, axial lobe most pubescent; nutlets 4, homomorphous, very small, 0.5-0.7 mm. long, triangular-ovate, dark colored, usually bearing light-colored low tuberculations, edges angled or rounded, areola shallow and deltoid with the groove above it commonly closed; gynobase subulate, equalling or a little shorter than the nutlets; style much surpassing the nutlets.—Contr. Gray Herb. n. s. xlviii. 43 (1916). Var. genuina. Corolla conspicuous, 2-3 mm. broad; nutlets tuber- culate.—Krynitzkia Grayi Vasey & Rose, Proc. U. S. Nat. Mus. xi. 536 (1888). C. Grayi Macbr. |. ec. Middle Lower California. 40 JOHNSTON ER CALIFORNIA: Lagoon Head, Palmer 801 (G, ye rae = 1- malli, Wine urpus (UC); Magdalena Island, Orcutt 27, 52 and 80 (G); Magdalena Bay, 1889, Brandegee (UC); Santa Margarita Island, 1889, Brandegee (UC). Var. nesiotica Johnston. Corolla inconspicuous, 0.5-1 mm. broad; nutlets very sparsely or not at all tuberculate——Proc. Calif. Acad. Sci. ser. 4, xii. 1146 (1924). Dunes on islands in southern part of Gulf of California. OWER CALIFORNIA: very common on dunes, San Francisco Island, Johnston soar (G, UC, Itsoryps). Var. cryptochaeta (Macbr.) Johnston. Corolla inconspicuous, 0.5-1 mm. broad; nutlets abundantly tuberculate—Proc. Calif. Acad. Sci. ser. 4, xii. 1145 (1924). OC. micromeres, var. cryptochaeta Macbr. ace Gray Herb. n. s. xlviii. 46 (1916). C. filiformifolia Macbr. I. c Extreme Satie Lower California and Sonora. Lower Cairornia: Cape San Lucas, Xantus 76 (G); sandy clearings, La Paz, bigger ct 3071 (G); San José del Cabo, Anthony a {> TYPE of C. micromeres, Va cryptochaeta). Sonora: Alamos, Palmer 397 (G, TYPE of C; ‘iliformifolia. UC, Isotype); Site ed ci 308 (G). : Um ptantha Grayt i is very closely related to C. angustifolia, and from ly slender forms of the latter it is distinguishable as by its decidedly houiesorphces nutlets. Usually, however, it is readily recognized by its very slender habit, and by its herbage which usually dries a darker color than that of C. angustifolia. Ser. II. CIRCUMSCISSAE. Nutlets 4, smoothish or incon- spicuously muricate, dark, triangular-ovate, with acute sides, homo- morphous or subheteromorphous with the abaxial nutlet just appreci- ably the largest; style equalling or barely surpassing the nutlets; calyx circumscissile at matu is 10. C. circumscissa (H. & A.) Johnston. Low herb 2-10 cm. high, trimly erect or much branched and forming hemispherical masses; stems few to numerous, more or less branched above, strigose or hirsute, the outermost somewhat decumbent; leaves oblanceolate, 3-15 mm. long, 1-2 mm. broad, obtusish, surface siliceous especially toward the pale base, strigose or hirsute, obscurely pustulate, upper ones scarcely reduced and extending through the inflorescence as foliaceous bracts; flowers axillary, the racemose arrangement obscure; corolla more or less inconspicuous, 1—2(-3) mm. broad; fruiting calyx 2.5-4 mm. long, oblong-ovate, united to near the middle, at maturity upper half falling away by a circumscission just below the sinuses; THE NORTH AMERICAN SPECIES OF CRYPTANTHA 41 basal persistent part of calyx-tube siliceous, ee ae gaieasd hirsute; mature calyx-lobes firm-herbaceous cely ribbed, more or less hirsute; edits pit ca. 0.5 mm. ghey nutlets 4, homomorphous or with abaxial one barely the largest, smooth or obscurely muriculate, triangular-ovate or oblong-lanceolate, 1.2-1.7 mm. long, back flattened especially towards apex, margins angled, groove closed and forked below; gynobase about 24 height of nutlets, pyramidal-oblong; style eiunlling or barely exceeded by nutlets.—Contr. Gray Herb. n. s. Ixviii. 55 (1923). Var. genuina. Stems with short appressed hairs. —Lithospermum circumscissum H. & A. Bot. Beechey 370 (1840). Piptocalyx circum- scissus Torr. in Wats. Bot. King Exped. 240 (1871); Torr. Bot. Wilkes Exped. 414, t. 12 (1874). Eritrichium cireumscissum Gray, Proc. Am. Acad. x. 58 (1874). Krynitzkia circumscissa Gray, Proc. Am. Acad. xx. 275 (1885). Wheelerella cireumscissa Grant, Bull. So. Calif. Acad. Sci. v. 28 (1906). Greeneocharis circumscissa Rydb. Bull. Torr. Bot. Cl. xxxvi. 677 (1909). C. depressa Nels. Bot. Gaz. Xxxiv. 29 (1902). Southern British Columbia and Idaho, southward to Arizona and northern Lower California. WASHINGTON: Junction of Coal and Crab creeks, 730 m. alt., Sandberg & Leiberg 228 (G); in dried up pools, Pasco, Piper 2966 (G); a 1883, Brandegee (G); plains, Narene’s Ferry, Yakima River, Suksd (G). OREGON: open sandy places, Baad. elson 859 (G); dry sandy slope g Des Chutes River, 8 km. below Bend, Peck 9709 (G); plains between Pineville and Bear Buttes, 1110 m. alt., L vont ad (G); , Burns, Peck 6108 (G) AHO: Boise, 1881, ); dry soil, 11.2 km. west of St hony, / ll & Wilcox 873 (G, Pons of en ; dry Il bottom-land, Falk e, Mac 29 (G); Snake Country, Burke (G); sandy slopes, Emmett, 660 m. alt., Macbride 786 Lg ag innings Soma Island, 1290 m. alt., Watson 847 (G); valley of the Virg ar St. Geo (G); St. Georg e, 1880, Jones (UC); without eatite. 1873, Bakes iG). 4 NeEv- VADA: Pocamat Lake, K. get was (G, UC); Carson City, 1500 m. alt., Watson 848 (G); Humboldt Plains, 1872, Gray (G ii Carson City, Anderson 163 (G); Lake W: orrey G); ern Rix a $ 8, 8235 (G); sandy places, Erskin Creek, Purpus 5304 (G) ; Acton, Elmer 3705 (G); hills bordering Mohave Desert, 1 1882, Pringle (G); sand, Mohave, Heller 7764 (G); = rt sand near Victorville, 600 m. alt., 1918, Spencer (G); desert r oi O§ : as) go Sod Fy a”) ag P Pe Op = 2 iy: ne a3 = A % S (G); ae agian Jacumba, Abrams 3659 (G); Cott sagas Springs, Jaeger 1846 (G). Lower Catirornta: without locality, 1884, Oreutt (G). 42 JOHNSTON Var. hispida (Macbr.), comb. nov. Stems clothed with spreading bristles; plant usually coarser than in var. genuina.—Greeneocharis ctreumscissa, var. hispida Macbr. Proc. Am. Acad. li. 546 (1916). Krynizkia dichotoma Greene, Bull. Calif. Acad. Sci. i. 206 (1885). Piptocalyx dichotomus Greene, Pittonia i. 60 (1887). Wheelerella dichotoma Grant, Bull. So. Calif. Acad. Sci. v. 28 (1906). Greeneocharis dichotoma Macbr. |. ce. C. dichotoma Johnston, Contr. Gray Herb. n. s. Ixviii. 55 (1923). Western Nevada and adjacent California. Infrequently collected. Nevapa: between Boca and Verdi, 1884, Curran (G, isotypE of K. dicho- toma); Carson City, Anderson (G). i Creek, 2400 m. alt., Davidson 2700 (G); trail to Mt. Whitney, 3000 m. alt., Culberton 4240 (G, tTyPE of G. circumscissa, var. hispida); sandy plains, Erskin Creek, Purpus 5304 in pt. (G). A detailed study of C. circumscissa and C. dichotoma has failed to disclose any fundamental character capable of separating the two. The nutlet-characters used by Gray and Greene are not distinctive, since among plants of indubitable C. circumscissa the nutlets not only vary from triangular-ovate to oblong-lanceolate, but also from smooth and shiny to very finely granulate and minutely muriculate. The size of the nutlets does not separate natural groups. The only tangible character separating C. cirewmscissa and C. dichotoma seems to be one of pubescence. Since all the material with spreading bristles comes from a definite region along the east base of the Sierra Nevada I believe that this hirsute form is best treated as a geographical va- riety. Consequently Macbride’s varietal name is taken up. It is to be noted that annotations in the Gray Herbarium indicate that even Gray inclined towards treating C. dichotoma as a mere variety. Ser. III. ALBIDAE. Nutlets 4, coarsely tuberculate, dark, tri- angular-ovate, thickish, very broadly obtuse or convex on the sides, homomorphous but with the abaxial nutlet subpersistent; style much surpassing the nutlets. ascen herbaceous or fruticulose plant 1-3(-4) dm. high; stems solitary or usually several, simple or commonly paniculately branched, more or less villous-strigose as well as hirsute ; early leaves in evanes- cent basal rosettes, spathulate, becoming 5 cm. long, 1.5-4 mm. wide; cauline leaves numerous and somewhat crowded, reduced above but extending as linear bracts through the inflorescence, 3 em. or less long, hirsute, usually minutely pustulate; spikes 1-7 cm. long, solitary, THE NORTH AMERICAN SPECIES OF CRYPTANTHA 438 numerous, loosely bracteate; corolla inconspicuous, ca. 2.5 mm. long, tube about equalling calyx; fruiting ast Seward ovate, 2-3 mm long, only slightly accrescent, becoming remo t iehaeen calyx-lobes oblong-ovate to LE seine conniv ent, midrib slightly thickened and short-hirsute, margin appressed-hispid; pedicels ca. 0.5 mm. long; nutlets 4, homomorphous, triangular ovate, 1-1.5 mm. long, tan or brownish, usually sparsely granulate, low-tuberculate, thickish, one next abaxial calyx-lobe subpersistent, margins rounded, ventral side occupied by a very large open triangular areola which appears to be excavated in very mature nutlets by the breaking away of the attachment-scar from the nutlet-walls; gynobase narrowly pyramidal, about equalling the nutlets; style much surpassing nutlets. —Contr. Gray Herb. n. s. Ixviii. 53 (1923). Myosotis albida HBK. Nov. Gen. et Sp. iii. 91 (Aug. 1818). Lithospermum ramosum Lehm. Asperif. ii. 328 (Nov. or Dec. 1818). Eritrichium ramosum A. DC. Prodr. x. 132 (1846). Krynitzkia ramosa Gray, Proc. Am. Acad. xx. 274 (1885). C. ramosa Greene, Pittonia i. 115 (1887). E. hispidum Buckley, Proc. Acad. Philad. 1861, pg. 462 (1861). K. mexicana ler Zoe v. 182 (1904); Macbr. Contr. Gray Herb. n. s. xlviii. 49 (1916 Wester Texas to eastern Arizona in the vicinity of the international boundary, thence south to Durango 2 Puebla. Texas: Davis Mts., Tracy & Earle 1 fee Maxon peeing. hye 46 (G); valley of Limpia, 1858, Hayes G): Pecos, 1858, Hayes (G); without locality, 1874, Buckley (G, ‘pia of TYPE of E. “his pidum); co oe, 5% J iw) authentic fragment of : Ex1co: sandy 8 ca, 3 rina. 6648 eS UC); San Tien Teotihuacan, District of Tezcoco, Seler 5304 (G). LA: Mt. Orizaba, 3000 m. alt., Seaton 173 (G). The nutlets of this species are very thick in relation to their breadth. This condition, as well as the very large open (in age excavated) tri- angular areola, makes the species readily recognizable. It is practic- ally confined to eastern Mexico where in Puebla it sets the southern- most outpost for the genus in North America. Not only is it the 44 JOHNSTON most southerly ranging of the North American species, but it is the only North American species which appears to have its immediate relationships in the continent to the south. Cryptantha albida is very closely related to C. argentinica Brand of northern Argentina, if indeed it is not the same species. Ser. IV. MARITIMAE. Nutlets 1-4, tuberculate or muricate, usually dark with pale roughenings, lanceolate to triangular-ovate, with rounded or obtuse sides, decidedly heteromorphous with odd nutlet (sometimes alone developing) axial and larger as well as oc- casionally less roughened than others; style surpassed by the nutlets or reaching their tips or rarely surpassing them. ; 4, straight; style reaching to about the ps OF Ce wie 13. C. echinosepala. OE TOI se 14. C. dumetorum. Calyx ascending or spreading, not at all gibbous; odd nutlet rather smooth. Nutlets triangular-ovate, 0.7-0.9 mm. long; mature calyx subglobose, minute, with lobes scarcely surpassing ew a 15. C. micromeres. 16. C. maritima. 12. C. recurvata Coy. Ascendingly branched rather slender herb 1-3 dm. high; root frequently dye-stained; stems usually strigose, rarely appressed-hispid; leaves remote, oblanceolate or linear-ob- lanceolate, 1-2(-3.2) mm. long, 2-4(-5) mm. broad, rounded or obtuse, appressed-hispid, densely and minutely pustulate; spikes naked, slender, loose, 2-10 cm. long, solitary or geminate; corolla incon- spicuous, subtubular, ca. 2mm. long, not exserted, lobes short; fruiting calyces slender, subdistichously biseriate, very asymmetrical, char- acteristically bent and recurved, 3-4 mm. long, tardily deciduous, sessile; mature calyx-lobes linear, midrib somewhat thickened and hirsute, rarely merely strigose, axial lobe longest thickest and most hirsute; ovules 2; nutlet 1, subpersistent, oblong-lanceolate, inwardly curved, brownish, dull, granulate-muriculate, next the axial calyx- lobe, edges obtusish; groove somewhat oblique, narrow or closed, opening into a small basal areola; gynobase slender, ca. 14 length of THE NORTH AMERICAN SPECIES OF CRYPTANTHA 45 the matured nutlet, slightly exceeded by aborted one; style commonly much surpassed by nutlet—Contr. U. S. Nat. Herb. iv. 165, t. 16 (1893), --- Eastern Oregon to Utah and the Inyo Region of California. REGON: Alword Desert, 1600 m. alt., Leiberg 2425 (G, UC). Nevapa: in open sand along water- “Pipe. Candelaria, 1950 m. ae Shockley 260 (G); sandy soil, Palette) Range, 1800-2100 m. alt., Purpus 5806 (UC). Urvan: 1891; Jones (UC). CALIFORNIA: Surprise ‘Canyon, Panamint Mts., 8 alti Coville & Funston 713 (G, UC, 1soryPEs); Silver Canyon in White M east of Laws, Heller 8221 (G, UG); oe Canyon, 1913, K. Brandegee a UC); Fish Lake Valley, 1897, Purpus. One of the most distinct and interesting species in the genus. It may be readily recognized by its biovulate fruit and spreading, re- curved fruiting calyces. 13. ©. echinosepala Macbr. Loosely branched herb 1-2 dm. high; stems usually reddish, strigose as well as somewhat hirsute; leaves 24.5 cm. long, 2-5 mm. wide, linear to linear-lanceolate, below pustulate and short villous-hirsute, above subglabrate or minutely strigose; spikes geminate or solitary, usually 2-5 cm. long, naked or rarely with a single bract; corolla inconspicuous, tube shorter than calyx, limb ca. 1 mm. broad; fruiting calyx ovate, 2-3 mm. long, sessile or subsessile, spreading or even refiexed, asymmetrical; mature calyx-lobes very unequal, lance-linear, strigose and usually hirsute along the thickened midrib, axial lobe the longest and most hirsute; nutlets heteromorphous, usually 4, pallid, groove narrow and scarcely dilated below; odd nutlet minutely muriculate-tuberculate, ovate, acute, 1.2-1.5 mm. long, subpersistent, next the axial calyx-lobe, margin angulate; consimilar nutlets ca. 1 mm. long, minutely tuber- culate, lance-ovate; gynobase evidently shorter than consimilar nutlets, oblong; style longer than consimilar nutlets, equalling or just surpassed by the odd nutlet.—Contr. Gray Herb. n. s. lvi. 57 (1918); Johnston, Proc. Calif. Acad. Sci. ser. 4, xii. 1147 (1924). ar Lower California. WER CaLirorNiA: Santa Agueda, Palmer 242 (G); Magdalena ag F i a (UC), Bryant (UC), ro (UC); Magdalena Island, Orcutt (G, TyPE); La Paz, Palmer 26 (G). At once distinguishable from nearly all other species of the genus by having its fruiting calyces deflexed and most heavily hirsute on the axial side. It is probably most related to C. angustifolia. 14. ©. dumetorum Greene. Laxly branched closely strigose herb; stems at first erect but later commonly much elongated and sprawling 46 JOHNSTON or scrambling among bushes; leaves lanceolate, thickish, 2-4 cm. long, 2-4(-8) mm. aes sparsely appressed hirsute-villous, closely pustulate below and finely or geminate, usually remotely flowered, 5-10 cm. long, ‘occasionally with foliaceous bracts towards base, rhachis brittle and tortuously flattened; corolla inconspicuous, ca. 1 mm. broad; fruiting calyx closely appressed to rhachis, 2-3 mm. long, very asymmetrical, not at all deciduous, base very oblique and downwardly gibbous on axial side; mature calyx-lobes connivent and reaching about equal height; 3 abaxial lobes lanceolate, somewhat strigose, with the thickened midribs deflexed-hirsute; 2 axial lobes partly united, hirsute only on outer margins; nutlets 4, heteromor- phous, granulate and muriculate; odd nutlet persistent, axil, broadly lanceolate, 2-3 mm. long, base much developed and distorting the calyx, groove open and broad; consimilar nutlets 1.5-2 mm. long, deciduous, lanceolate, groove closed or very narrow; gynobase narrow, shorter than consimilar nutlets; style shortly surpassed by nutlets or reaching to their tips.—Pittonia i. 112 (1887). Krynitzkia dumetorum Greene in Gray, Proc. Am. Acad. xx. 272 (1885). Deserts from western Nevada to Southern California. Nevapa: Muddy Valley, 510 m. = Kennedy & Goodding7?4 (UC). Cauir, orNiA: half climbing among bushes at Tehachapi Pass, 1884, Curran (G- TYPE); Kramer, Parish 9810 UC); Kramer, K. Brandegee (G, UC); Barstow, . Brandegee 158 (G, UC); Lancaster, K. Brandegee (UC); Whitewater, 1909, 300 m. alt., 1903, Jones (UC); Sthout locality, Lemmon (UC). An anomalous species very peculiar in habit and in calyx and nutlet developments. Although the tips of the 4 nutlets in each calyx are of equal height, their bases are decidedly not so. The base of the axial one is more developed than that of the others causing the axial side of the calyx to be gibbously distended downwards along the pedicel and thereby making the base of the calyx conspicuously oblique. In habit the species is unique in the genus. It commonly grows about bushes and scrambles up through them, often reaching a meter in height. So sinuous and brittle are the stems that it is commonly quite impossible to disentangle from the supporting bush a perfect specimen of this plant. The pedicels are unquestionably persistent, being as decidedly so as are those of C. micrantha. The type of the species is given as having come from Tehachapi Pass. An annotation in the University of California Herbarium gives the type-locality as: Tehachapi Pass, probably at the “ Bend,” between Mohave and Te- hachapi Station. 15. C, micromeres (Gray) Greene. Slender usually erect-growing THE NORTH AMERICAN SPECIES OF CRYPTANTHA 47 herb 1-5 dm. tall; stems dull dark-green, short-hirsute throughout; leaves linear to oblong-linear, somewhat hirsute on both surfaces and usually somewhat pustulate beneath, 1.5-4.5 em. long; spikes com- monly ternate, very slender, naked, 2-8 cm. long; corolla incon- spicuous, subtubular, ca. 0.5 mm. broad; fruiting calyces very small, 1-2 mm. long, subglobose, fulvous, sessile or nearly so, early deciduous; mature calyx-lobes ovate-lanceolate, decidedly connivent, scarcely surpassing the fruit, margins ciliate, midrib slightly thickened, hirsute and frequently uncinate; nutlets 4, heteromorphous, triangular-ovate, 0.7-0.9 mm. long, margin subangulate; odd nutlet slightly the largest, smooth or sparsely papillate, next the axial calyx-lobe; consimilar nutlets papillate, groove open, gradually dilated into a small open areola; gynobase equalled by consimilar nutlets but somewhat sur- passed by odd one; style short, equalling or bearly surpassing the odd nutlet.—Pittonia i. 113 (1887). LEritrichium micromeres Gray, Proc. Am. Acad. xix. 90 (1883). Krynitzkia micromeres Gray, Proc. Am. Acad. xx. 274 (1885). Central California to northwestern Lower California. Ca.iFrorniA: Ione, 1886, K. er (UC); near pa on Hill, = n (G); Point Loma, 1906, K. Brandegee (UC). Lower CALIFORNIA! near Ensenada, 1882, Jones (UC). This Species develops the smallest flowering and fruiting parts known in the genus. Its occurrence is sporadic, and in the south at least somewhat determined by the presence of chaparral-burns. C. maritima Greene. Ascending loosely branched herb be- coming 1-3 dm. tall; stems commonly reddish, strigose or frequently hirsute; leaves linear to lanceolate, acutish, usually somewhat con- tracted at the base, 1-3.5 cm. long, 1-4 mm. wide, commonly hirsute, coarsely pustulate; spikes solitary or geminate, 1-6 cm. long, usually more or less crowded and frequently glomerate, irregularly leafy- bracted throughout; corolla inconspicuous, tubular, 1.4-2 mm. long, 0.5-1 mm. broad; fruiting calyx 1.8-3(-3.5) mm. long, ovate-oblong, stiffly ascending, slightly asymmetrical, tardily deciduous, subsessile; , 48 JOHNSTON mature calyx-lobes lance-linear, stiff, connivent, midrib of three abaxial lobes thickened and hirsute, margins appressed short hispid- villous or loosely villous; ovules 2 or 4; nutlets 1-4, heteromorphous; odd nutlet frequently alone developing, smooth, shiny, brownish, oblong-lanceolate, 1-2 mm. long, firmly affixed, next the axial calyx- lobe, groove narrow or closed throughout or opening below into a small areola; consimilar nutlets grayish, minutely tuberculate, readily deciduous, slightly smaller but otherwise like odd nutlet; gynobase subulate, 14-24 length of nutlet; style about equalling the height of consimilar nutlets.—Pittonia i. 117 (1887). Var. genuina. Ovules 2; nutlets 1 or 2; calyx not conspicuously pilose—Krynitzkia maritima Greene, Bull. Calif. Acad. Sci. i. 204 (Aug. 1885). C. maritima Greene, Pittonia |. c.; Munz & Johnston, Bull. Torr. Bot. Cl. xlix. 38 (1922). K. ramosissima of Greene, Bull. Calif. Acad. Sci. i. 203 (Aug. 1885); not K. ramosissima Gray (Jan. 1885). C. ramosissima of most recent authors Nevada to Lower California. ells, 8 m. alt., Spencer 1520 (G); sandy places, Schaffer’s Canyon near a 48 m. below sea-level, Spencer 2069 (G) and Munz & Keck 4755 (G); : o Dese 1890, Wright (UC); Point Loma, 1884 Ay hae G, UC); windswept arroyos, San Nicolas Island, ge 56 and 57 (G):8 ; Santa Barbara Island, Trask (UC); Catalina Island, 1884, L yon (G); Catalina Island, 1890, Brandegee (UC). OWER CALIFORNIA: nort i Lower Califo ornia, Orcutt 2257 (UC); Lagoon Head, —— Sb BAG: Sant Guadalupe Island, Palmer 879 (G), Anthony 241 (G, UC), P 7 (G), Gr 1885 (G, c. —— >. = ? be phate Anthony in i806 (UC). Pa ge in a (UC n Benito I Palmer 912 so), Pond 21 (G), ries 276 (UC) Brondege in 1897 7 (U0) Natividad Var. cedrosensis 7 Si comb. nov. Ovules 4; nutlets 1-4; otherwise as in preceeding variety.—Krynitzkia scdvomeusts Greene, Bull. Calif. Acad. Sci. i. 204 (1885). C. cedrosensis Greene Pittonia it; (1887); Brandg. Bot. Gaz. xxvii. 454 (1899). Endemic to Cedros Island. R Catrornta: Cedros Island, Palmer 691 (G), Greene in 1885 (G, ue. mee Veatch (G), Anthony 289 (G, UC). ai ar. pilosa Johnston. Ovules 2; nutlets.1 or 2; calyx conspicu- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 49 ously clothed with long white villous pubescence.—Univ. Calif. Pub. Bot. vii. 445 (1922) Nevada to Lower California, occurring with var. genuina but less common, ADA: Logan, 1909, Kennedy (UC). Ca trrorniaA: Death Mont Coville & Pnston a is Furnace Creek, Pa ica Sag ree 10050 (UC); Calico, Parish 9798 (UC); Inyo County, Rizford (UC); Kane Springs, Ord Mts Hall & Chandiot 6807 (UC); Needles, 1884, Lemmon (VO); Signal Mt., 1901, Brandegee (UC). Arizona: Tempe, 1892, ‘Gan ong & Blaschka (G). LOWER CALIFORNIA: stony ridges, Los Angeles Bay, Palmer 541 (UC, TyPE; G, Iso- TYPE); Santa Agueda, Palmer § 2142 (UC). A well defined and common species in he Californian deserts and on the coastal islands. The usual form is readily recognized by being biovulate and having reddish strigose stems. It has generally gone as C. ramosissima but that name being based on a renaming of C. racemosa properly falls into synonymy. Ser. V. PTEROCARYAE. Nutlets 1-4, tuberculate or papillate or verrucose, lanceolate, with knife-like or conspicuously winged sides, homomorphous or frequently heteromorphous and with the odd nutlet (sometimes alone developing) axial and usually lacking a thin margin; style reaching to tip of nutlet-body or definitely beyond. ope — or rarely 2, morn with a knife-like a quely cont axa? Shee. 3 ae. . C. utahensis. Nutlets 4; ca ays 6 symmetr rolla conspicuous ; mate calyx obviously longer than road; nutlets always homiomorphnous: 55.25... tes. 18. C. oxygona. gees inconspicuous ; mature calyx nearly as broad as long; nutlet frequently wingless..............-... 19. C. pterocarya. 17. C. utahensis (Gray) Greene. Erectly branched herb 1-3 dm. high; stems closely strigose or more or less appressed short-hirsute; leaves not numerous, strongly reduced above, linear to oblance- linear, 1-5(-7) em long, 1-4 mm. wide, rounded at apex, commonly pustulate and short-hirsute especially beneath; spikes geminate or sometimes solitary, commonly 0.8-2.5(-5) cm. long, dense, naked; corolla evident, 2-3 mm. broad; fruiting calyces ovate or ovate- oblong, 2-3(—4) mm. long, quite asymmetrical, subsessile by a broadly conic oblique base, spreading or somewhat recurved, deciduous, usually densely appressed-hirsute and notably silky; mature calyx- lobes lanceolate, strongly connivent, midrib thick and usually brown- ish and infrequently bearing spreading or recurved hairs; ovules 4; nutlets 1 or rarely 2, next the abaxial calyx-lobe, 1.7-2.5 mm. long, 1-1.5 mm. broad, pale, broadly lanceolate, granulate, muricate- 50 JOHNSTON papillate or rarely spinulose, back low-convex or flat, margins sharp- angled or with a very narrow knife-like margin; groove open, narrow, opening into a small areola below; gynobase subulate, ca. 24 height of nutlet, not markedly differentiated from style; style usually a trifle shorter than the nutlet.—Pittonia i. 120 (1887). Krynitzkia utahensis Gray, Synop. Fl. N. Am. ii. pt. 1, Suppl. 427 (1886). Eritrichium holopterum, var. submolle Gray, Proc. Am. Acad. xiii. 374 (1878). C. submollis Cov. Contr. U. S. Nat. Herb. iv. 166 (1893). Southern Utah and western Arizona and westward into the deserts of California. Uran: volcanic rocks and ashes, Diamond Valley, Goodding 828 (G, UC); St. George, Palmer 352 (G, Type). ARIZONA: Yucca, Jones (G); Yucca, Jones 3910 (UC). Nevapa: Candelaria, Shockley 347 (G) and 650 (UC); foot of cliffs, Meadow Valley Wash, Goodding 2165 in pt. (G); rocky slopes, Mes- quite Well, Goodding 2252 (G); Rhyolite, 1080 m. alt., Shockley 69 (UC); Gold Mt., Purpus 5986 (UC). Catirornia: Inyo County, 1891, Brandegee (UC); Surprise Canyon, Panamint Mts., m. alt., Coville & Funston 714 (G); without locality, Purpus 5433 (G); Providence Mts., Munz, Johnston & Harwood 4241 (UC); Daggett, 1914, K. Brandegee (UC): sandy places near Barstow, Spencer 2082 and 2093 (G); in sandy places, Palm Springs, Spencer 1526, 2073, 2074, 2075 and 2076 (G); in sandy places, Mission Canyon, 180 m. alt., Spencer 1782 (G); Colorado Desert, 1889, Orcutt (UC). A very neat species most readily distinguished by the appressed silky indument on the calyx. It seems to be an ally of C. pterocarya. Its rough nutlets at once distinguish it from C. gracilis and C. Watsoni with which it has been frequently confused. 18. C. oxygona (Gray) Greene. Sparsely branched herb 1-4 dm. tall; stems usually solitary with several well-developed ascending branches from near base, appressed villous-hispid or strigose, often sparsely hispid, leaves linear or lance-linear, 1-4(—-6) em. long, 1-2(-3) mm. broad, strigose or short-hispid, ascending, obtusish, densely and inconspicuously pustulate, upper ones evidently reduced; spikes geminate or ternate, usually short and dense, 1-3(-6) cm. long, naked; corolla conspicuous, limb 4-7 mm. broad; fruiting calyces ovate or oblong-ovate, ascending, 2.5-4 mm. long, deciduous, obscurely bi- serial, symmetrical, base rounded, pedicel ca. 0.5 mm. long; mature calyx-lo Ss lanceolate, somewhat connivent above, margin more or less silky-strigose, midrib slightly thickened and frequently sparsely hirsute; nutlets 4, homomorphous; body of nutlets oblong-ovate, 2 or rarely 3 mm. long, only slightly shorter than the calyx-lobes, muricate or tuberculate, back low convex; margin of nutlet narrowly winged or knife-like; groove closed or rarely open, broadly forked below where always opened to form a triangular areola; gynobase columnar- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 51 subulate, about 24 height of nutlets; style clearly surpassing the nut- lets.—Pittonia i. 120 (1887). Eritrichium oxygonum Gray, Proc. Am. Acad. xix. 89 (1883). Krynitakia oxygona Gray, Proc. Am. Acad. xx. 277 (1885). Borders of the San Joaquin Valley and southward along the desert margins to the northern end of the Colorado Desert. Cauirornta: Alcalde, 1892, Brandegee (UC); Estrella, 1897, Jared (UC); open places in rich ground, McKittrick, Heller 7789 (G, UC); northern slope of Tehachapi, 1905, K. Brandegee (UC); Tehachapi, K. Brandegee (G, U hills bordering the Mohave Desert, 1882, Pringle (G, TyPE); mountain slopes, ino Co., 1200 m. , Spencer 415 (G); ie locality, McLean ong : 9 (UC); near Minerets, Madera Co., 1899. Congdon (UC); hillsides, Erski Creek, Purpus 5369 (G, UC); Deep Spring, Purpus 58. (UC); Coyote Canyon, 150 m. alt., Hall 2849 (UC); between San Jacinto and ts at Van Deventer Ranch, 1350 m. alt., ‘Hall 1161 (UC). Nevapa: Palmetto Range, Purpus 5897 in part (UC). : This species, although having the gross habit of C. muricata, is clearly related to C. pterocarya and particularly to the variety cy- cloptera. In fact, decisive characters for separating the two species appear to be lacking. As a general rule, however, C. oxygona differs in having conspicuous corollas, narrower usually more silky calyces, and brownish nutlets rarely if ever with scolloped or lobed wings. It is maintained as a species largely because of its natural range, which, generally speaking, is apart from that of C. pterocarya. The most satisfactory character for separating C. orygona from C. pterocarya seems to be corolla-size. It is significant, however, that a specimen (Purpus 5715a) from the South Fork of the Kern River, occurring within the range of C. orygona and having the characteristic habit, calyx and nutlets of that species, nevertheless has minute corollas. For practical purposes this specimen has been refered to C. pterocarya. . pterocarya (Torr.) Greene. Erect ascendingly branched herb 1-5 dm. high, finely strigose or short-hirsute; leaves broadly linear or the reduced, upper ones somewhat lanceolate, 1-2.5(-4) em. long, 1-3(-5) mm. broad, obtuse, strigose or hispid, coarsely pustulate below but usually finely so above, spikes geminate or rarely ternate or solitary, naked or inconspicuously bracted below, 2-6(-12) em. long, b ing loosely fl d; lla i pi ,0.5-1(-2) mm. broad; fruiting calyces becoming notably accrescent, usually broadly ovate, (2-)3-5 mm. long and usually about 34 as broad, tardily deciduous, symmetrical, base obtuse or rounded, pedicels 0.5-1 mm. long; mature calyx-lobes ovate to lanceolate, connivent, only a little sur- passing the nutlets, margins more or less tawny appressed-hispid, midrib slightly thickened and weakly and sparsely hispid; nutlets 4, 52 JOHNSTON homomorphous and all winged, or heteromorphous with axial nutlet wingless; body of nutlet oblong-lanceolate or lanceolate, 2—2.5(-3) mm. long, muricate or verrucose; wing-margin of nutlet broad or narrow, entire or crenate or lobed, extending completely around the _ nutlet or only down the sides; groove open or closed (even in the same plant) and dilated below into an open excavated areola; gynobase slender, about 24 height of nutlets; style subulate, slightly surpassing or somewhat surpassed by the wing-margin of = nutlets but always exceeding the body proper.—Pittonia 1. 120 (18 Var. genuina. Nutlets heteromorphous, ax sie one win Eritrichium pterocaryum Torr. Bot. Mex. Bound. 142 on, Bee Wilkes Exped. 415, t. 13 (1873). Krynitzkia pterocarya Gray, Proc. Am. Acad. xx. 276 (1885). C. pterocarya Greene, |. c. E. ptero- caryum, var. pectinatum Gray, Proc. Am. Acad. x. 61 (1874). K. pterocarya, var. pectinata Gray, Proc. Am. Acad. xx. 276 (1885). C. pectocarya Frye & Rigg, Northwest FI]. 328 (1912). Eastern Washington and southern Idaho and southward to Southern California and Utah. Sporadic in Arizona. WASHINGTON: junction of Crab and Wilson creeks, Sandberg &Leiberg 260 (G, UC); in sand, Pasco, Piper 2961 (G); Walla Walla, Brandegee 995 (G UC) and 994 (UC); rocky bank of Columbia River near Columbus ’ Suskdor’ 9 (UC); without) locality, Vasey 421 (G). Orecon: clay bank, Mathew aw, 750 m. alt., Leiberg 2041 (G, UC); near Lexington, 400 m. alt., Leiberg ‘A : G); Stansbury Island, Wats 59 (G). Nevapa: Peavine Hills, 1895, Hillman UC); Pyram mid Lake, aaa 998 (UC); Teakké ee Lake, Kennedy 1345 A ©); Lawton’s Springs, ee B sagore (UC); Reno, 1890, Hillman (UC); Re 1884, Curran (UC); R 1885, K. Brandege e (UC); ghoak Carson *Gity, 1446 m. alt., ape toe 975 “G. UC): 4 Carson CE Wataon 859 (G); Carson pend son 165 (G); Candelaria, Shockley 282 (UC); ravine amo g hills n ina, at 8365 (G); on scoria on mesa west of Goldfield, Heller 10971 ; oodding 2. Barnwell, K. ndegee ; Le , Paris : — > het ge 10138 Sha, = Johnston 6494 (UC); sa: nd near Bars gs Ay apr x Var. cycloptera (Gri) ae ‘Nutlets homomorphous, all — Gray Herb. n. s. xviii. 44 (1916). Krynitzkia THE NORTH AMERICAN SPECIES OF CRYPTANTHA 53 cycloptera Greene, Bull. Calif. Acad. Sci. i. 207 (1884). C. eycloptera Greene, Pittonia i. 120 (1887). Southern California to western Texas and sporadic in southern Nevada and Utah and in eastern Colorado. Catirornia: Inyo, 1891, Brandegee (UC); Surprise Canyon, Panamint Mts., 800 m. alt., Coville & Funston 720 (G); Ord Mts., Hall & Chandler 6807.4 UO); rocky places, Snow Creek near Palm Springs, 150m alt., Spencer 2065a (G); desert sand, Palm Springs, eee on age cer 846 and 847 (G); rocky places, Cathedral Canyon nea 120 m. alt., ae 2079 (G); Coyote Canyon, 150 m. alt., Hall 28 Pe (UC CS aent caked Mounta n Smart 132 (G); hills near Tucson, 1884, Pringle (G, isotyPE); Tucson, 1894, Toumey (UC); Rio Cienega, Greene 1111 (G): Nogales, 1892, Br wie ray (U ©); Lowell, Parish 167 (G); Fort bad inge Palmer 346 (G). Cotorapo: Grand Junction, 1892, Eastwood (G). gs Fee rocky hillsides, south end of Black Range, 1380 m. alt. Meee 1573 (G). Texas: El Paso, Jones 3753 (UC); Fronteras, Wright 1570 This is one of the most s Accesting species in the genus. Its con- spicuously winged nutlets and broad large fruiting calyces are very distinctive. Although for the most part readily determined, it has some forms that are very puzzling. The outstanding variation is the well understood one regarding nutlet-form. As a general rule, the northern material has heteromorphous nutlets, whereas the southern has them homomorphous. Actual intergrades connecting the two forms are rare. In these the axial nutlet is frequently smallest and is more narrowly and less completely winged than the others. The northern plant with heteromorphous nutlets has been _commonly taken as the typicalform. Th in the Mexican Boundary Report. The f few notes there given ‘clearly apply to the southern homomorphous form. From internal evidence, however, it is very clear that the Mexican Boundary Report was written sub-. sequently to the Botany of the Wilkes Expedition. Hence the much later publication in the Wilkes Report is primarily considered in typifying the species, especially since in that work the plant was illustrated and fully described. The material from Walla Walla, Washington, collected by Pickering & Brackenridge and illustrated by Torrey is taken as the type. . This material has hetero- morphous nutlets. The southern plant with homomorphous nutlets was named C. cycloptera by Greene and was subsequently reduced to a variety of C. pterocarya by Macbride. It is possible that the 54 JOHNSTON varietal name pectinata should be taken up in place of var. cycloptera. Gray originally published it as Eritrichium pterocaryum, var. pec- tinatum, basing it upon material collected by Parry (numbers 168 and 169) in the Virgin River Valley near St. George, Utah. Un- fortunately, however, Gray hastily mounted collections by Greene and by Palmer on the type sheet of the var. peclinatum and it is now quite impossible to decide just which are the original Parry collections. Since both var. genuina and var. cycloptera are represented on the mixed sheet it seems best to drop the varietal name pectinatum as a nomen confusum, particularly so since the lobing of the nutlet-wing seems to be too hopelessly variable and unimportant to justify nomen- clatorial recognition. The most puzzling forms of C. pterocarya come from southern Ne- vada and Southern California. Purpus has collected material in the Gold Mountains of Nevada which have very small, scarcely accrescent calyces and a habit suggesting that of C. utahensis or C. gracilis. The nutlets, though small, are quite like those of typical C. pterocarya and it seems best to refer the specimens to that group. A specimen collected by Munz (number 5746) above Cactus Flats in the San Bernardino Mts. of California, although clearly related to C. pterocarya, may represent an unnamed species. The calyces are rather small and quite hirsute, although with the characteristic broad form of C. pterocarya. The nutlets are heteromorphous, but instead of having the odd nutlet wingless, it is winged and the consimilar nutlets are wingless and suggest those of C. utahensis. More material of this peculiar variation is greatly desired. I doubtfully refer to the var. cycloptera a collection made by Purpus (number 5714a) in the South Fork of the Kern River. The specimen has the habit of C. holoptera, in fact appears to differ from thoroughly typical members of that species only in having inconspicuous corollas. Since flower-size seems to be the only character capable of separating . C. holoptera and C. pterocarya in a manner that is at all practical and satisfying, I am arbitrarily referring Purpus’s collection to C. pterocarya although it is realized that in range and habit it unmistakably suggests C. holoptera. THE NORTH AMERICAN SPECIES OD CRYPTANTHA 55 Nutleta smooth or nearly: 90.605 sesh Sees ae i eee 20. C. Pattersoni. Nutlets coarsely granulate or tuberculate. et 1, axial one alone developing; ascendingly branched, 24d. talks 6 20 eee 21. C. texana. usua Nutlets 4; spreading, usually 1-1.5 dm. Consimilar nutlets lance-ovate, ac cantinte. 1.8-2 small and suprabasal; eveRibes = only ately: thickenedinics Seay eee ree dee be Kelseyana. Contailan: nutlets ovate or oblong, acute, ca long; areola large and ventral; calyx-lobes PBN se ously ‘hicks ned, har ee Drnciante: gaeoualowhe odd nutlet papillate- er ee Cai eee page ies Kee 23. C. minima. Spikes ‘naked; odd nutlet finely granulate and — TS ek a A oe ea eae . C. crassisepala. 20. C. Pattersoni (Gray) Greene. Loosely branched hirsute herb 1-1.5 dm. high; stems usually several, ascending, branched, hirsute and usually somewhat strigose; leaves oblanceolate 1-3 cm. long, 2-4 mm. wide, rather firm, obtuse, hirsute, more or less pustulate, upper ones little reduced; spikes solitary or geminate, naked, 2-5(—7) cm. long; corolla inconspicuous, 1—1.5 mm. broad; fruiting calyx oblong-ovate, 4-5 mm. long, spreading, slightly asymmetrical, evi- dently biserial, lowermost becoming mm. distant; pedicels ca. 0.5 mm. long; mature calyx-lobes linear-lanceolate, tips more or less connivent, midrib thickened and hirsute, margins appressed hispid; nutlets 4, heteromorphous; odd nutlet next axial calyx-lobe, slightly the largest, ca. 1.9 mm. long, ovate, acute, smooth or obscurely ru- gulose or sparsely tuberculate, somewhat persistent, standing off slightly from the gynobase; consimilar nutlets oblong-ovate, ca. 1.6 mm. long, deciduous, closely appressed to gynobase, smooth, back convex, sides rounded or obtuse, groove opened or closed and abruptly broadening below into a small triangular areola; gynobase narrow, reaching to ca. 24 height of consimilar nutlets; style exceeded by odd nutlet, equalling or a little shorter than consimilar nutlets.—Pittonia i. 120 (1887). Krynitzkia Pattersoni Gray, Proc. Am. Acad. xx. 268 (1885 ae of Colorado and Wyoming. Rarely collected. lahat a a soil, 32 km. east of Point of Rocks, Payson 2518 gone des, Junction Butte, Nelson 5887 (G). CoLorapo: emling, et ay ney (G); Sulphur Springs, Osterhout 35596 (G); without locality, 1875, Patterson (G, TYPE); without locality, 1877, Hooker & Gray. This species is related to C. Kelseyana and C. ambigua and seems to intergrade with both. In habit it is quite like C. ambigua, and with immature material alone available it is scarcely possible to distinguish 56 JOHNSTON it from that species. The most troublesome intergrades come from northwestern Wyoming. Certai1 material from this region has the nutlet next the axillary calyx-lobe slightly more persistent and more erect than the others, and hence suggesting a relationship with C. Pat- tersoni. Despite these suggestive developments the specimens are referred to C. ambigua, since for all practical purposes the nutlets are similar in size and shape, and usually indistinguishable in markings and hence quite like those of C. ambigua. The type of C. Pattersoni probably came from the Rocky Mountains near Golden, Colorado. The species appears to occur at altitudes somewhat higher than those affected by C. Kelseyana and C. minima. 21. C. texana (A. DC.) Greene. Ascendingly branched hirsute herb 1.5-4 dm. high; stems usually several, with few short branches; leaves oblanceolate 2-5(-7) cm. long, 2-5(-8) mm. wide, obtuse or rounded, not particularly firm, uppermost little reduced; spikes soli- tary or occasionally geminate or ternate, naked or at times sparsely bracted at base, 4-7(-11) cm. long, not sharply differentiated from leafy mass of plant; corolla inconspicuous, ca. 1 mm. broad; fruiting calyces ovate-oblong, 4-5 mm. long, somewhat asymmetrical, widely spreading or subdeflexed, becoming remote; pedicels short but evident, 0.5-1 mm. long; mature calyx-lobes lance-linear, more or less connivent with the herbaceous tips spreading, midrib indurated and strongly hirsute, margin sparsely short-hispid; ovules 4, only the one next the axial calyx-lobe maturing; nutlet solitary, persistent, broadly ovate, acute, ca. 2 mm. long, pale, densely and evenly coarse-granulate, back convex, sides rounded; gynobase short-oblong, 44-14 as long as the style; style reaching to above the middle of the nutlet.—Pittonia i. 112 (1887). FEritrichium texanum A. DC. Prodr. x. 130 (1846). Myosotis texanae Hook. Kew Jour. Bot. iii. 295 (1851). Krynitzkia terana Gray, Proc. Am. Acad. xx. 268 (1885). Central and western Texas. eExAs: Llano County, 1885, Reverchon (G); dry ground, ge Hall 469 (G); val G); valleys among Thulis beyond Limpia Pass, Wright 488 (G); without y, Drummond (G, UC, TSOTYPES); without sacnliey. Weight (G). Peasy venta to ] t quite distinct, since it differs in its more restricted easterly range, looser and taller habit of growth, and coarsely granulate solitary nutlets. Although C. texana has four ovules only one nutlet develops, that being the one next the axial — and clearly the homologue of the odd nutlet in C. crassi- The type of C. texana was collected by Drummond probably se ae in the — of Austin. - THE NORTH AMERICAN SPECIES OF CRYPTANTHA 57 22. C. Kelseyana Greene. Spreading or ascending hirsute herb 5-25 cm. high; stems one to several, hirsute and also hispid-strigose; leaves 1.5-3(-4) em. long, 2-4 mm. wide, rounded or obtuse, thickish, linear-oblanceolate, hirsute, pustulate, the upper ones searcely re- duced; spikes usually solitary, 4-9 cm. long, naked or with a few bracts near base; corolla i inconspicuous, 1-2 mm. broad; fruiting calyx 4—6 mm. long, ovate-oblong, spreading, somewhat asymmetrical, loose or dense; pedicels short but definite, ca. 0.8 mm. long; mature calyx- lobes linear, slightly connivent above, midrib thickened and hirsute, margins inconspicuously villous-strigose; nutlets 4, rep sia odd nutlet next the axillary calyx-lobe, broadly lance-ovate, 2-2.6 m long, smoothish or granulate or granulate-muriculate or rarely some- what tuberculate, standing off slightly from the gynobase; consimilar nutlets lance- or oblong-ovate, 1.8—-2.3 mm. long, coarsely tuberculate and commonly granulate, darker than odd nutlet, sides rounded; groove narrow or closed, near base abruptly dilated to form a small triangular areola; gynobase subulate, a little longer than style, 14- 24 height of consimilar nutlets; style surpassed by odd nutlet and just surpassing or even exceeded by consimilar ones. siege li. 232 (1892); Macbr. Contr. Gray Herb. n. s. xlviii. 49 (191 Saskatchewan and Montana southward Reh iss Ww yoming to northern Colorado and Utah SASKATCHEWAN: Medicine Hat, Macoun 5803 in pt. (G); ehorrtes stn 1858, Bourgeau (G). Montana: scratch gravel near lean randegee 26 (UC); Boakiew Desert, 1898, Brandegee 36 (UC); Elligton. “Ae eg Greene (G. IsoTyPE); Northern Boundary, French Creek to Roe 1874, Coues (G). Wyomrine: Teton National Forest, 1897, Reendives we snow-drift beds, Powder River, Nelson 9377.4 and 9415 (G); : anc e Plains, Laramie, N. Nelson 8897 (G, UC); Pole Creek, Nelson 1335 (G, UC); Laramie Hills, Nelso 412 (G); Sherman, 1893, Greene (UC); Gorfield Peak, Nelson 672 in pt. (UC); sandy roadside 5 km. north of Saratoga, Payson 2535 ee Dea deep sand, Oden, Goodding 1176 (G). C : Kremling, erhout 3464 (G); mountain side near Georgetown, 1885, Patterson (G, tic): - Castle Rock in foothills near Golden, Patterson 1 111 (G); west of Craig, Osterhout 6188 (G); Rocky Mts., lat. 40°-41°, Vasey 434 (G). This plant has been often confused with C. crassisepala although quite distinct in range and in structures. It is a larger more loosely branched plant with more elongate less indurated calyx-lobes, nar- rower ventrally less excavated nutlets, and a range separated from true C. crassisepala by half the width of Colorado and Utah. Al- though frequently growing with C. minima it is at once separable by its ebracteate inflorescence and very different nutlets. 58 JOHNSTON 23. C. minima Rydb. Erect or widely spreading hirsute herb 1—1.5(—2) dm. high; stems usually numerous, branched, finely strigose and coarsely hirsute; leaves oblanceolate, thickish, 1-3 cm. long, 2-4 (-5) mm. broad, obtuse or rounded, hispid or hirsute and usually pustulate, upper ones little reduced and continuing through inflo- rescence as evident foliaceous bracts; spikes solitary or rarely geminate, bracted, 2-8(—15) cm. long, frequently springing from even the lowest axils; corolla inconspicuous, 1—-1.5 mm. broad; fruiting calyces oblong- ovate, 5~7(-9) mm. long, spreading, asymmetrical; pedicels short but definite, 0.5-1.2 mm. long; mature calyx-lobes linear-lanceolate, con- . nivent above, midrib hirsute and strongly indurated and thickened, margins sparsely appressed-hispid; nutlets 4, heteromorphous; odd nutlet persistent, next the axial calyx-lobe, 2-3 mm. long, brownish, very finely and closely papillate-granulate, ovate, acute; consimilar nutlets 1.2-1.5 mm. long, ovate, thickish, strongly tuberculate, not granulate; groove broadly dilated at least to beyond middle, commonly excavated, not forked; | Synobase oblong, slightly more than 4 length of consimilar dently surpassed by odd nutlet, equalling or surpassing consimilar natlets: —Bull. Torr. Bot. Cl. xxviii. 31 (1901). Plains east of continental divide from Saskatchewan southward to northern Texas; occuring west of the mountains only in southwestern Colorado. SASKATCHAWAN: Medicine Hat, Macoun 5803 in pt. (G). MoNnrTANA: Great, en 1887, Anderson (UC). Wromine: hig agg nag a elson 304 (G . Cotorapo: Castle Rock near Golden tterson 111 in ae oF. _ Golden, Greene 301 (G); Ft. Collins, pty Crciiaa (UC); lat. 39°-41°, Hall & Harbour 433 and 434 (G); 10 km. north of Pueblo, Johns- ton & Hediccct 509 (G); Canyon City, Brandegee ); Cuchara River, abov eta, 2100 m. alt., g & Vreeland 5697 (NY, TYPE); g adobe flats and egteegrr Paradox, 1620 m. alt., W (G); dry hill- s, Naturi t., Payson 296 (G) without locality, 1842, Fre- nae H aa: Bad Lands, 1881, H r BRASKA Ke 1890, Bates (G); Callaway, 1901, Bates (G); Ha , 1890, Bates (G); top of butte, Ft. Robinson, 1200 m. alt., Be ). Kansas: Ellis, 1876, Watson (G); plains, Ellis, st wep 346 (G). OKLAHOMA: low waste place, Was , Stevens 598 (G ); sandy places, Alva, Stevens 245, 616, 664 ka. and 30. G)iw . v 2” EN a A LTS lor 2120 (G); Big Wichita, 1880, Ball (G). Although a very readily recognized species, C. minima has not been distinguished from its southerly ranging relative, C. crassisepala. Ors plasthe minima has thick, hardened calyx-lobes and a gross habit very similar to that possessed by C. crassiscpala, but differs very sharply in ia definitely bracted inflorescences and in the fact that THE NORTH AMERICAN SPECIES OF CRYPTANTHA 59 the odd nutlet is simply papillate-granulate rather than covered with spiculate murications. The ranges of C. minima and C. crassisepala overlap only in southwestern Colorado. In the northern part of its range C’. minima grows with C. Kelseyana, and, strange to say, has been frequently confused with it, although C. minima can at once be separated by its stockier nutlets with larger areolae, and, of course, by its bracteate inflorescence. The name C. minima was originally proposed to cover certain material thought to be related to C. pusilla. An examination of the type, however, shows clearly that it is a very reduced form of the species here treated, and that though inappropriate, it is the name to be taken up for the bracteate material from the Great Plains until now passing as C. crassisepala. 24. C. crassisepala (T. & G.) Greene. Erect or widely spreading herb 5-15 em. high; stems commonly numerous, loosely ascending, branched, hirsute or rarely hispid; leaves oblanceolate, 2—3(-6) cm. long, 3-4(-6) mm. wide, rounded or obtuse, thickish, hirsute, pustu- late, the upper scarcely reduced; spikes solitary or rarely geminate, naked or few-bracted below, 5-8(-15) em. long, frequently produced from the lowest axils; corolla inconspicuous, 1-1.5 mm. wide; fruiting calyces 6-7(-10) mm. long, oblong-ovate, somewhat asymmetrical, becoming distant below; mature calyx-lobes linear-lanceolate, con- nivent above, midrib very hirsute and strongly thickened and indur- ated, margins inconspicuously short-hispid; pedicels short but definite, 0.5-1.2 mm. long; nutlets 4 (J or 2 rarely aborted), decidedly hetero- - morphous; odd nutlet next the axial calyx-lobe, persistent, 2-2.5(-3) mm. long, brownish, ovate, acute, finely granulate and spinular- muricate; consimilar nutlets readily deciduous, 1.2-1.5(-2) mm. long, oblong-ovate, thickish, coarsely tuberculate, very obscurely if at all granulate, groove usually dilated and commonly excavated to form an areola occupying much of the ventral face of the nutlet; gynobase nar- rowly oblong, usually about 24 height of consimilar nutlet's; style - equalling or a trifle exceeding the consimilar nutlets, surpassed by odd nutlet.—Pittonia i. 112 (1887). Eritrichium crassisepalum T. & G. Pacif. R. R. Rep. ii. pt. 4, 171 (1855). Krynitzkia crassisepala Gray, Proc. Am. Acad. xx. 268 (1885). (. dicarpa Nels, Proc. Biol. Soc. Wash. xvi. 30 (1903). Southwestern Colorado and southern Utah southward to western Texas, Arizona and adjacent Mexico. Cotorapo: Deer Run, Gunnison Watershed, 1380 m. alt., Baker 75 (G); Grand Junction, 1892, Eastwood (G, UC); dry sandy flat, Grand Junction, 60 JOHNSTON bride & Payson 689 (G). UrTan: acand 1890, Jones {Ge yO. ree 1895, Stokes (UC); Green River, 1890, Jones (UC). Wingah, Mathews 40 (G); Gila River bottom satin Cliff, 1350 m Bes Meicalfs 52 (G, UC); Santa Fe, Fendler 640 (G); Hueco Mts., Thurber 61 (G); A Albu- querque, Jones 3709 ; mesa west of renal Spree 8, t., 1905, Wooton (G, UC); Silver City, Eastwood 8401 (G). ARIZONA: Pi Lowell, 1883, Pringle (UC); Chino Valley, Twomey 240a (UC); Clifton, atidagn 401a (UC); Aztec Ruins, Little Colorado, 1905, cee ven Verde Mesa, roa 141 (G); Douglas, 1200 m. alt., Goodding 2273 (G, UC); Tucson, 1907, iG) Bern ardino, 1602, Orcutt (UC); a locality 1882, Pringle 6). vite? vicinity of permanent camp o , April 6-7, 1856, Pope d. (G, wee abundant on iy hi ‘lls Sa re of El Paso, Hover 102 (G, NY); western Texas, March 9th, Wright (G). CutHuaHva: rere Pa soil, * oes 1905, Purpus (UC); ye Diego, 1800 m. alt., Hartman Cryptantha crassisepala is a very interesting and readily recognized species most related to C. minima. Although it usually produces four nutlets some forms of it regularly mature only two or three. It is one of these forms, indistinguishable in other respects, that was made the type of C. dicarpa. The consimilar nutlets of C. crassisepala and C. minima are rather thick for their breadth. In this respect, as well as in general contour, they strikingly simulate the nutlets of C. albida, a species not closely related. The only material suggestive of a con- dition intermediate between C. minima and C. crassisepala is East- wood’s collection from Grand Junction. In that collection nutlets typical of C. minima are associated with the naked —— and habit of C. crassisepala. Ser. VII. BARBIGERAE. Nutlets 1-4, verrucose or muricate, lanceolate to ovate-lanceolate, dorsall¥ convex, laterally rounded or obtuse, homomorphous with the abaxial one always developing; style reaching only to 14 height of nutlets or in varying degrees longer, sometimes even somewhat surpassing them. Nutlets normally solitary or rarely 2; style reaching 1/3-1/2 Dre Ol Be 25. C. decumens. Nutlets preety * 4; style reaching 2/3 height of nutlets or be- 01 Spike solitary; style evidently surpassing nutlets........... 26. C. patula. Spike 6 ens nate or ternate; style almost aieays shorter than nu le Hairs arming calyx-lobes very coarse, subobese; plan neal orgs with an erect central axis; insular endemic... "37 C. foliosa., Hairs calyx-lobes slender, not conspicuously thick wisimge ter be irregularly branched. Corolla’ uous; hirsute — almost confined to eee tema ae ry ae Spans oh ta - 28. C. intermedia. oro inconspicuous; plants of the reat Basin and Californian deserts Plant spreading hirsute... .. ta a, veo, 29. C. barbigera. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 61 Plant with appressed hairs Nutlets verrucose or verrucose-muriculate. .30. C. nevadensis Nutlets spinular-mltrieate, oo. era 8 31. C. scoparia. 25. ©. decipiens (Jones) Heller. Loosely branched herb 1-4 dm. high, slender, strigose and frequently short-hispid; leaves rather few, linear, obtuse, 1-3 cm. long, 1-3(—4) mm. broad, strigose and some- times hispid, minutely pustulate; spikes geminate or rarely ternate or solitary, slender, becoming loosely flowered or congested, 4-10 cm. long, naked; corolla inconspicuous to conspicuous, 0.8-3.5 mm. broad; fruiting calyces ovate to ovate-oblong, strictly ascending, asym- metrical, 2.5-7(-9) mm. long, deciduous, sessile; mature calyx-lobes lance-linear, decidedly connivent above with the tips frequently spreading or even recurving, midrib thickened and usually evidently hirsute, margins strigose, abaxial lobe evidently the longest and most hirsute; ovules 4; nutlets 1 or rarely 2, next the abaxial calyx-lobe, ovate-lanceolate or occasionally narrowly ovate, 1.5-2.4 mm. long, usually granulate or muriculate-granulate, muricate-papillate or low- tuberculate, usually brownish, back convex, sides rounded, groove open or closed but always dilated below to form a definite areola; gynobase short, 14-1 height of nutlet; style much surpassed by nutlet, 16-% the height of nutlet. —Muhlenbergia viii. 48 (1912). Var. genuina. Corolla inconspicuous, less than 1 mm. broad.— Knees decipiens Jones, Contr. W. Bot. xii. 6 (1910). C. decipiens Heller, |. ¢ Southern Nevada, western Arizona and Southern California. Nevapa: Logan, Kennedy 1838 (G, eee Rhyolite, 1072 m. alt., Heller 9632 (G). Arizona: Hot Springs, ee mey (UC). CALIFORNIA: Kern- ville, Brandegee (UC); candy argo nea Palm sg “S ings, 120 m. alt., Spencer 2072 (G); desert rt sand, mouth o Piew: ae cee 210 m. alt. , Spencer 1522 (G); Whitewater, 300 m. alt., Jones (UC); without eal, Palmer 150 (G). Var. corollata, var. nov., a varietate genuina differt corolla con- spicua 2-3.5 mm. lata. Extreme western margin of the Mohave Desert and the adjacent coastal slopes. Caurrornia: Fort <> Xantus 85 (G); Sespe Creek near Ten Sycamore Flat, Abrams & McGregor 173 SE Haghs Matilija Canyon, Ojai Soe ong: 1896, Hubby 20 @ towards Meootkills , Ojai Valley, 1896, Hubby 21 (G, TYPE); Roble Canyon, San Rafael Mts., 1020 m. alt., Hall 7408 (G, UC); Salita | Inez Mts., 1888, Brandegee (UC); Huron, Fresno Co., Brandegee (UC); without locality, Brandegee (UC). This species is probably most related to C. intermedia and C. neva- densis and has been somewhat confused with them. It differs, how- 62 JOHNSTON ever, in having a very short style and gynobase and normally but one or rarely two nutlets. 26. C. patula Greene. Sparsely and loosely branched herb 5-15(-30) em. high; stems strigose and sparsely short-hispid; leaves linear or lance-linear, 1-5 cm. long, 1-3 mm. broad, acutish, appressed- hispid, minutely pustulate; spikes solitary, with a few leafy bracts towards the base; corolla medium-sized, tube ca. 2 mm. long, limb 1.5-3 mm. broad, lobes short-oblong, ascending, ca. 1 mm. long, throat funnelform, appendages hemispherical; fruiting calyces ovate- oblong, 5-6 mm. long, obscurely biserial, subsessile, asymmetrical, base broadly conical or rounded; mature calyx-lobes lance-linear, connivent above with the herbaceous tips spreading, margins shortly white-villous, midrib thickened and densely tawny-hispid, abaxial lobe evidently the longest; nutlets 4, homomorphous, ca. 1.9 mm. long, oblong-ovate, acute, finely tessellate-granulate, tuberculate or muricate, base somewhat truncate, back convex, groove closed or nearly so and divaricately forked at base, at times open at the fork to form a small triangular areola; gynobase subulate, almost as long as the nutlets; style evidently surpassing nutlets.—Pittonia i. 265 (March 1889). C. Pondii Greene, 1. ¢. 291 (April 1889). Middle western Lower California and adjacent islands. Lower Cauirornia: San Bartolomé, 1889, Pond (G, 1soTyPE of C. Pondii); San Benito Island, 1897, Brandegee (G, UG). Although probably most related to C. intermedia this species is readily separated by its solitary spikes and extreme southern range. Cryptantha Pondii is clearly a synonym. Although Greene described it as having “smooth and shiny” nutlets and ternate or quadrinate spikes the isotype sent Gray has granulate and tuberculate nutlets and solitary spikes as described above. 27. C. foliosa Greene. A stiffly erect divaricately branched herb 6-20 cm. high; stems solitary, straight, usually forming a conspicuous central axis, short-hispid below but b ing hat strigose above, branches well developed, widely spreading; leaves lanceolate to broadly linear, obtuse or rarely acutish, 2-6 cm. long, 2-5(-7) mm. broad, appressed-hispid, abundantly and minutely pustulate; spikes dense, 1-4 em. long, geminate or ternate, naked; corolla evident, limb 2-3 mm. broad; fruiting calyces ovate-oblong, stiff, 5-7 mm. long, strongly a ee ris, 1 . 1 x lew biseriate, subsessile, ye = ghiy conical; ca lobes lance-linear, towards the tips herbaceous and somewhat spread- ing, towards base indurated, thickened midrib armed with short ay THE NORTH AMERICAN SPECIES OF CRYPTANTHA 63 excessively coarse almost inflated pungent tawny hairs, margin stri- gose; nutlets 4, homomorphous, narrowly ovate, acute, ca. 1.5 mm. long, brownish and somewhat mottled, finely tessellate-granulate, tuberculate or muricate, back convex, edges obtusely angled, base rounded, groove narrowly dilated towards base where divaricately forked and closed; gynobase narrow, ca. 0.8 mm. long, ca. 24 height of nutlets; style reaching tip of nutlets——Pittonia i. 113 (1887). Krynitzkia foliosa Greene, Bull. Calif. Acad. Sci. i. 205 (1885). Endemic to Guadalupe Island, off the west coast of Lower Cali- fornia. Lower Cauirornia: Guadalupe Island, Palmer 68, 842 and 877 (G), Anthony 238 (G, UC), Greene in 1885 (G, UC, IsOTYPES), Townsend (UC), randegee in 1897 (UC). Readily recognized because of its peculiar habit of branching, con- gested spikes, coarsely armed calyx-lobes and small nutlets. 28. C. intermedia (Gray) Greene. Erectly branched commonly stiff and very hirsute herb 1.5-5 dm. high; stems several or solitary, erect, commonly hirsute but frequently more or less strigose; leaves lanceolate to linear or rarely somewhat oblanceolate, acute to obtuse, 2-6(-7.5) em. long, 1-5(-7) mm. broad, hirsute or strigose, usually inconspicuously pustulate; spikes naked, geminate to quinate but commonly ternate, 5-15 cm. long, usually stiff; corolla conspicuous, 2-8 but commonly about 5 mm. broad; fruiting calyces ovate-oblong, 2-7 but commonly 4-6 mm. long, ascending or strict, deciduous, slightly asymmetrical, lowermost not conspicuously biserial, pedicels ca. 0.5 mm. long; mature calyx-lobes lance-linear, connivent above with tips usually spreading or recurving, margin appressed-hispid or short- villous, midrib thickened and pungently hirsute, abaxial lobe longest and most hirsute; nutlets commonly 4, homomorphous, lance-ovate, ca. 2 (1.5-2.3) mm. long, more or less coarsely and decidedly tuber- culate or verrucose, frequently somewhat granulate, grayish or brown- ish, margins slightly angled, back convex, groove narrow or closed but gradually dilated towards base into a small triangular areola; gyno- base 24-34 height of nutlets, narrow; style usually about reaching the nutlet-tips or rarely slightly surpassing or surpassed by them.— Pittonia i. 114 (1887). LEritrichium intermedium Gray, Proc. Am. Acad. xvii, 225 (1882). Krynitzkia intermedia Gray, Proc. Am. Acad. xx. 273 (1885); Synop. Fl. N. Am. ii. pt. 1, Suppl. 426 (1886). C. quentinensis Macbr. Contr. Gray Herb. n. s. lvi. 58 (1918). C. bar- bigera, var. F. ergusonae Macbr. |. ec. 59. C. intermedia, var. Johnstonii rie. 59. Mach 64 JOHNSTON Coastal drainage from northern California to northern Lower Cali- fornia, and infrequent along the desert borders. British Cotumstia: probably introduced in ballast, vicinity of Victoria, Macoun 672 (G). Cattrornta: Scott River Valley, 1899, Gilbert (UC); Cas- tilla, Hastwood 1366 (G); 7883 in River near fish hatchery, Heller 13021 ; ne ne m yon nea + & Brown 5510 (G); 16k ) Alder i Shee Hele 11461 “GO: Cae Co., 1884, Curran (G); Stony Creek, Colusa Co., an 43 (G); north side of Marysville Butte, Heller 11368 (G);8 akley, 1900, The Postmaster (UC); table ‘isl S oo oO pe] omy OS rs © S oO po ber | io?) a oe ao ic] a SE Ss is So el: ean en — ti" 7 7 50) | ee | et ° Ei, ° e: oe is°) oO Pe ~} 8 Ss 7 Qr SS e Ra , pan ck 88 ; Santa Barbara, Elmer 3866 (G); Ojai Valley, 1896, Hubby 22 (G); Saugus, 1889, Brandegee oe ie sith 1901, Davy (UC); San Pomado Wash, 1913, Eastwood (G); San Fernando Mts ” 1883, Nevin (G, UC); San Fernando, Nevin (UC); Bin Angeles, Nevin (U C); Los Angeles, Gambel (G); Los Angeles, 1880 to 1882, Nevin (G, including Typ of C. intermedia); sandy creek margin, Jerdugo Can nyon, Macbride & Payson 748 (G); wey hee 1901, Setchell (UC); Santa Monica Experiment Station, Barber 48 (UC); T. grat Canyon, 1916, Crawford & Hiatt 257 (G): Playa del Rey, Abrams 2512 (G, UC); San Pedro, 1889, iret (UC); dry hillsides back od Laguna Beach, 150 m. alt., Johnston 1934 (G); Santa Catalina Island, 1890, Brandegee (UC); Santa Catalina jaeed yaw 242 (UC); Santa Clemente Teles 1880, Nevin & Lyon ston n Anto yon, John 2046, 2047, — and 2049 (G); ; near Upland, Parish 11154 (G, UC), Johnston 957, 1959 and 1960 (G); east of Victorville, Spencer 838 (G); rocky hillside, torville 30D m. al 2087 2063 (G); Sno Caheann Station, A iidins ots (G): : Paks Springs, Spencer 850, 851, 852 and 863 (G); Palm Springs, Ferguson 42 (G, TPE of C. barbigera, var. Fergusonae) ; issi pe pencer : . 920, ‘921 aha "99765 (G); near Campo, 1885, bead 1278 tte near Campo, Abrams 3556 (G). Lower Canirorn1a’ Todo THE NORTH AMERICAN SPECIES OF CRYPTANTHA 65 pegs Island, Anthony 213 (G, UC) and 204 (G); Ensenada, 1889, K. Brande- ee (UC); Burro Canyon, 1895, Brandegee (UC); Cari so Creek, 1893, cen (uc ); San Quentin Bay, Palmer 608 08 (UC) and 695 (G. TYPE of C’. quentinensis; UC, 1sorypr); without locality 1883, Orcutt (G). A variable and perplexing species the sharp delimitation of which seems quite impossible. At the northern extreme of its range it passes into C. Henderson?. But outside the zone of intergradation in the northernmost counties of California, C. intermedia is readily distinguished from the more northerly ranging C. Hendersoni by having narrower, lance-ovate, strongly verrucose nutlets and de- cidedly pungent hairs on stem and calyx. In the southern part of its range C. intermedia passes even more completely into C. barbigera and C. nevadensis. The two latter species inhabit the desert and probably represent modifications of the C. intermedia-stock as adap- tations to that extreme environment. Material transitional between C. intermedia, C. barbigera and C. nevadensis comes from the western margin of the Californian deserts or from the hot interior coastal valleys more or less connected with the desert proper. Cryptantha barbigera and C. nevadensis differ from C. intermedia primarily in their minute corollas. Cryptantha barbigera appears to be a derivative of C. intermedia with bristly stems and evidently villous as well as hirsute calyces. On the other hand, C. nevadensis has tended to develop a wiry habit, strigose pubescence and very elongate nutlets. Inter- gradation among them being beyond question, C. Hendersoni, C. intermedia and C. nevadensis, with much justification, might be treated as mere varieties of C. barbigera. It seems best to adhere to the traditional classification, however, and retain the four groups as specific, since such an extended concept as their union would be unwieldy. As various other species in this genus are more or less connected by- local intergradation an unfortunate precedent might be set if, in the present instance, drastic reductions were made because of the occurrence of local transition. ies varies considerably in pubescence. Though the plant is sometimes closely short-strigose, the common form is densely pungent- hirsute throughout. In some forms the calyx is appressed-hispid, in others hispid-villous or even somewhat silky, though the midrib of the calyx-lobe is practically always more or less pungent-hirsute. Cryptantha quentinensis is a form which is not hirsute and has a rather silky calyx. It is clearly a phase of C. intermedia and is not a re- lative of C’. oxygona as its author suggests. Cryptantha barbigera, var. Fergusonae is a form of C. intermedia with rather long-villous calyx- lobes. The type is one of the suite of puzzling plants from the desert 66 JOHNSTON border which connect C. barbigera and C. intermedia. Cryptantha in- termedia, var. Johnstonii differs from C. barbigera, var. Fergusonae only by not being conspicuously villous on the calyx. It appears to be one of the very large-flowered, coarse phases of C. intermedia. Material of C. intermedia from the eastern margin of the Colorado Desert not infrequently has the style definitely surpassing the nutlet-tips whereas that from other regions seems uniformly to have the style not reach- ing beyond the nutlets. 29. C. barbigera (Gray) Greene. Erectly branched hirsute herb 1-4 dm. high; stems solitary or several, very bristly and sparsely if at all strigose; leaves oblong to lance-linear, obtuse, 1-5(-7) em. long, 3-7(-13) mm. broad, hirsute, inconspicuously pustulate; spikes geminate or rarely solitary or ternate. naked, becoming as much as 15 cm. long; corolla inconspicuous, limb 1-2 mm. broad; fruiting calyx 5-10 mm. long, ovate-oblong or oblong-lanceolate, ascending, asym- metrical, deciduous; pedicels 0.3-0.7 mm. long, villous; mature calyx- lobes lanceolate to linear-lanceolate, connivent above with tips re- curved, margin conspicuously long white-villous, midrib thickened and hirsute, abaxial lobe slightly the longest; nutlets 1-4, homomor- phous, lance-ovate, 1.5-2.5 mm. long, strongly verrucose, usually brownish, back convex, edges obscurely angled or rounded, groove opened or closed but towards base gradually dilated to form a tri- angular areola; gynobase narrow, 24-34 height of nutlets; style reach- ing to or slightly beyond the nutlet-tips.—Pittonia i. 114 (1887). Eritrichium barbigerum Gray, Synop. Fl. N. Am. ii. pt. 1, 194 (1878). Krynitzkia barbigera Gray, Proc. Am. Acad. xx. 273 (1885). K. mixta Jones, Contr. W. Bot. xiii. 6 ( 1910). Southern California and southern Utah, to Arizona and northern Lower California. : Uran: deep sand, Diamond Valley, Goodding 889 (G, UC); St. George, 900 m. alt., Jones 5106 (UC, tsotyrE of K. mixta); Utah, Parry 171 (G, TYPE of E. } ); — re mn : a): . alt., Hall 8940 . Ben Lomond, Santa Cruz Co., 1890, Brandegee (UC); Soledad, 1882, Fence (G); Santa Cruz, Jones 2810 (G, TyPE of K. Jonesii) ; above Alma Soda Springs near ‘‘ French Settlement,’’ Santa Clara Co., Heller 7491 (G ee): ae Cruz gee }> dry sandy wba San ‘Bernardino, 360 m ae . Johnston 1886 (G); og ong San ardino, $67 929 G); dry mes esas, San —- 0 alley, 300 m : CaALIFO ay 1885, ene (UC); near U. 8S. eee er, 1884, Orcutt 1022 (G); near janat 1885, Orcutt 1381 (G); Topa, 1884, Orcutt 1127 (G); + iedeok ites. Bene e (UC). Var. denticulata (Greene), comb. nov. Corolla inconspicuous, 1-2 mm. broad; plant with a few well developed loose branches; spikes not numerous; nutlets usually larger than in var. Jonesit.—Krynitzkia denticulata Greene, Bull. Calif. Acad. Sci. i. 205 (1885). C. denticu- lata Greene, Pittonia i. 114 (1887). C. Gace Nels. & Kenn. Proc. Biol. Soc. Wash. xix. 156 (1906). Western Nevada and adjacent California, southward to western Arizona and the mountains of Southern California 72 JOHNSTON Nevapa: Dog Valley Road, 1895, Hillman (UC); Clear Creek Canyon, -2615 m. alt., Baker 1381 (G , UC); Reno, 1884, Curran (G); near Reno, 1893, Hillman (UC); along railroad above Tetgehtiri's: 1894, Hillman (UC), Verdi, 1904, Kennedy 952 (UC, 1soTyPE of C. densiflora); Hunter Creek, Washoe Co., 1800 m. alt., Kennedy 1917 (G); western Nevada, 1884, Curran (G, UC, tsoTyPpErs of K. “denticu wlata). ARIZONA: mesas near r Camp owell, 1881, Pringle; Skull Valley, 1290 m. alt., 1903, Jones 7028 ied Yucca; 1884, Jones 26 (G). CALIFORNIA: Trucke ee, 1887, Sonne (UC); Donner Lake, 1888, Curran (UC); Summit Station, Placer Co., 2083 m. alt., Heller 12886 (G); Mammoth, 1913, K. Brandegee (UC); open pine woo ar Mt. Wilson, Abrams 2: 00 m. alt reek, 2250 m. alt., Johnston 2077 (G, UC); sunny rocky summit » Telegraph Peak, 2700 m. alt., Johnston 1542 (G); open Sites jo Or ieee a Fk. San Antonio Canyon, 2400 m. alt., Johnston 1598 (G, yo side, San Antonio Canyon, 1110m. “ Johnston 1950 (a. dite eck Vahey, Hall 12 (UC). A species readily recognized by its ovate-triangular nutlets, tawny pubescence, and characteristic yellowish-green herbage. It breaks _up into three intergrading varieties. The var. genuina, characterized by its large corollas, occurs in western California. It varies notice- ably in the size of the nutlets and fruiting calyces. The type of C. muricata is one of the coarse, large-flowered forms and was probably collected by Douglas somewhere between Monterey and Santa Barbara. The other varieties are small-flowered. The var. Jonesii commonly assumes a peculiar erect habit by which it can be distin- guished at a glance from all other forms in the genus. The stems are usually single or several and fastigiate, and are erect, producing in the upper half or two-thirds abundant very short floriferous branchlets. Rarely the plant becomes diffusely branched. The distribution of var. Jonesii is peculiar. It occurs in the Coast Ranges from Santa Cruz to Glenn County, and in the middle Sierra Nevada. Jumping _ the several hundred kilometers occupied by the var. genwina it re- appears in the vicinity of San Bernardino and is frequent from there south into Lower California. The var. denticulata is perhaps un- worthy of recognition since it is frequently distinguishable from the var. Jonesii only with difficulty. It differs chiefly in habit, being sparsely and loosely long-branched. It grows in western Nevada in the general region of Reno, i in California about Truckee, in western Arizona and in the pine belt of Southern California. The material from Nevada usually has coarsely granulate nutlets and perhaps te leaps rats the southern material that I have associated with Ser. Ix. AMBIGUAE. Nutlets 1-4, smooth to papillate or tuberculate, ovate, dorsally low-convex, laterally rounded or obtuse THE NORTH AMERICAN SPECIES OF CRYPTANTHA 73 or occasionally acutish, homomorphous, no particular nutlet always developing; style reaching to ?/3~*/3 height of nutlets. Plant low, stiff, spreading, widely branched; spikes solitary..33. C. mariposae. Plant erectly brane ed, less rigid; spikes ternate or geminate. DES. 6 eee OCS Ok VRC nae a ai are hee 5. C. excavata, Nutlets usually 4, erect, evidently shorter than the ecidedly connivent calyx-lobes............. 36. C. Hendersoni. Corolla inconspicuous. Nutlets minute, less than 1.5 mm. long; spikes bracted CURIONI oo oe Be ee 37. C. Traskae. Nutlets usually 2-2.5 mm. long; spikes naked or bracted only at base. Nutlets smooth, rarely finely granulate; spikes usually inate; plant commonly 2-4 dm. tall......... . C. Torreyana. ge Nutlets somewhat tuberculate, densely granulate or granu- ate-muricate as well as more coarsely roughened. Plant closely strigose, pale, usually 2-3 dm. tall; spikes commonly geminate or ternate.............. . C. simulans. Plant spreading-hispid, usually 1-1.5 dm. tall; spikes usually solitary or rarely geminate. s ro bercule: 40. C. ambigua. Nutlets with elongate papillae or spicules.......41. C. echinella. 33. ©. mariposae, sp. nov., humilis saepe basaliter ramosa 5-15 em. alta; ramis paucis laxe ascendentibus breviter strigosis; foliis paucis firmis oblanceolatis vel oblongo-oblanceolatis 0.8-1.8 mm. longis 2-5 mm. latis obtusis integerrimis adpresse breviter hispidis inconspicue pustulatis, superioribus paullo reductis, inferioribus op- positis subpersistentibus; spicis solitariis vel geminatis 3-10 cm. longis rigidis ebracteatis vel basem versus pauce folioso-bracteatis; floribus obscure biseriatis, inferioribus 5-10 em. distantibus; calycibus fruc- tiferis oblongo-ovatis 5-7 mm. longi dentil bsessilibus; lobis calycis lineari-lanceolatis in costa infra medium cum setis flavescenti- bus horridis et in marginibus adpresse hispidis, supra medium breviter hirsutis et saepe recurvatis; corolla conspicua 3-6 mm. lata; nuculis homomorphis 4 (rare 1-2 abortis) ovatis acuminatis 2-2.5 mm. longis saepe granulatis basi truncatis margine rotundis vel obtusis ventre ?/s-'/2 longitudinis ad gynobasem quadrangulari-columnarem ca. 1 mm. longam affixis; sulcis basem versus paullo gradatim ampliatis ad imam basem divaricato-furcatis et areolam triangularem profundam formantibus; stylo ca. 0.8 mm. longo quam nuculae evidenter breviori. Known only from Mariposa County, California. 74 JOHNSTON Cauirornia: Mariposa, April 27, 1898, Congdon (UC); Mariposa, cae 1903, Congdon (UC); Mariposa County, May 2, 1890, Congdon C59 (TYP Gray Herb.). A peculiar plant with small, broad, thickish leaves and low, loosely branched, coarse, rigid stems. It has passed as C. ambigua and C barbigera but-is distinct from both, differing in its large corollas, low stiff habit, strigose stems and in the shape and attachment af the nutlets. In having basally truncate and acuminate nutlets it some- what suggests C. muricata, var. denticulata, but is very different in its low habit, strigose pubescence and short style and gynobase. The young spikes are not so tawny as are those of C. muricata. 34. C. crinita Greene. Erectly branched herb 2-3 dm. high; stems appressed and spreading-hispid, branches numerous; leaves oblance- olate to oblance-linear, 2-4 em. long, 2-4 mm. broad, obtuse, hirsute, evidently pustulate; racemes ternate or geminate, naked, 3-6 cm long; corolla conspicuous, 3-5 mm. broad; fruiting calyces 5—6.5 mm. long, spreading, ovate-oblong, obscurely biserial, deciduous, divided, conspicuously villous, hairs very long and white, oldest calyces 5-10 mm. distant; pedicels well developed, 2-3 mm. long, long-villous; calyx-lobes linear, erect, slightly unequal, without a thickened midrib and pungent bristles; ovules 4; nutlet 1, next the axial calyx-lobe, ca. 3 mm. long, 1.3 mm. broad, erect, dull, brownish, ovate-lanceolate, densely muriculate-granulate, frequently coarsely tuberculate es- pecially above the middle, apex attenuate, base obtusish, back convex, margin rounded or obtuse, groove opened towards base to form a small deep triangular areola; gynobase elongate, about half length of nutlet; style reaching to about 34 height of nutlet.—Erythea iii. 66 (1895). Known only from Shasta County, California. CaALiFornia: bed of ae aga Ra at Leightone, 1900, Baker (G, UC); Cow Creek, 1894, Baker & Nutting (UC); Stillwater, Nutting (UC); Shasta County, 1894, Baker & Nutting MGC, “Dup. of t : Distinguished from all other members of the genus by its unusually long white hairs on the calyx and on the well developed ees eed distinct from, but probably most related to, C. Henderson . C. excavata Brandg. Loosely and ascendingly caked herb ps du: tall; stems short hispid-villous and usually appressedly so; leaves not numerous, 1.5-3 em. long, 1-2 mm. broad, linear or spathu- late-linear, obtuse, appressed-hispid, minutely pustulate, upper ones eens reduced; spikes ternate or geminate, 3-8 cm. long, naked; conspicuous, ca. 4 mm. broad; fruiting calyx, broadly ovate, pe 5 mm. long and oo as wide, asymmetrical, sessile, spreading, THE NORTH AMERICAN SPECIES OF CRYPTANTHA 75 base rounded or very broadly conic, becoming remote and obscurely biserial; mature calyx-lobes lanceolate, subequal, erect or loosely con- - nivent, frequently one displaced by the tip of the decumbent nutlet which they barely exceed in length, margin short hispid-villous, midrib scarcely thickened and somewhat short-hirsute ; ovules 4; nutlet 1 and usually horizontal or infrequently 2-3 and then erect, ovate-triangular, recurved-acuminate, ca. 2.5 mm. long, densely granulate or granulate-muriculate, commonly sparsely tuberculate, base decidedly truncate, sides obtuse or rounded, back convex, groove dilated towards. base and forming a broad deep narrowly triangular suprabasal areola; gynobase narrow, ca. 1 mm. long, about 44-4 length of nutlet; style reaching to 24-34 height of nutlets.—Bot. Gaz. XXVil. 452 (1899). Inner North Coast Ranges of California; rare. CALIFORNIA: occasional on shelving slopes of sand, Cache Creek, Yolo Co., Baker 2886 (G, UC); Lake County, April, K. Brandegee (UC). The solitary, decumbent, triangular nutlet and short calyx are distinctive of this species. It appears to be related to C. Hendersoni on one hand, and apparently to C. decipiens on the other. No par- ticular nutlet seems to develop, even in the same spike, for adjacent calyces may each develop its nutlet in a very different position re- lative to the spike-axis. 36. ©. Hendersoni (Nels.) Piper. Ascendingly branched hispid herb 1.5-5 dm. tall; stems single or numerous, sparsely and loosely branched above, hispid; leaves oblanceolate or linear, 2-5(-—7) em. long, 2-5 mm. broad, acute or obtuse, appressed-hispid, lower ones somewhat persistent and pustulate, upper ones reduced; spikes usu- ally ternate, rarely geminate or quadrinate, naked or at times bracted toward very base, 2-8 or rarely even 20 cm. long; corolla conspicuous, tube about equalling calyx, limb 4-7 mm. broad; fruiting calyx ovate- oblong or narrowly ovate, 3-6 mm. long, ascending, slightly asym- metrical, lowermost becoming obscurely biserial and distant, base rounded or broadly conic; pedicels ca. 0.5 mm. long; lobes lance- linear or linear, somewhat connivent above with herbaceous tips usually hat spreading, margins densely appressed villous-hispid, midrib obscurely thickened and hispid; nutlets 4 or by abortion rarely fewer, broadly ovate or very rarely lance-ovate, 2-2.8(-3) mm. long, smooth or more or less coarsely granulate, frequently coarsely tuber- culate and at times finely papillate-muricate, back low convex, sides rounded or rarely obtuse, base rounded or somewhat truncate; groove 76 JOHNSTON closed or very narrow, broadly forked below; gynobase narrow, ca. 1.3 mm. long, becoming 14-24 as high as nutlets; style reaching to about */; height of nutlets or barely exceeding them.—Piper ex J. C. Nelson, Torreya xx. 44 (1920). seek i Hendersoni A. Nelson, a ( s (1920). C. monosperma Greene, Pini v. 53 vote C. incana Greene, Leaflets i. 79 (1904). C. grisea Greene, I. C. trifurca Eastw. Bull. Torr. Bot. Cl. xxxii. 203 (1905). C. srusiietowe Rydb. Bull. Torr. Bot. Cl. xxxvi. 679 (1909). C. Torreyana, var. grandiflora Nels & Macbr. Bot. Gaz. Ixi. 43 (1916). A. dichotoma Brand in Fedde, Repett. xviii. 313 (1922). C. scabrella Piper, Proc. Biol. Soc. Wash. xxxvii. 95 (1924). Western Idaho to southwestern British Columbia and south to the Sierra Nevada and northern parts of California. Ipano: sandy hillsides in open places or amongst trees, Juliaetta, Hender- son 4815 (G); hills about ae Henderson 2811 (G); Mey n, 1895, Piper (G); about Lewiston, 450-600 m. alt., Heller 2998 (UC - Valle ey of Clearwater River, Sa idbrg: ‘M acDougal Vet Heller 10 (G, ISOTYPE 4 C. grandi- flora); Clearwater, Spalding (G); abunda vice over wooded an paren tS hills, Palouse Country, Henderson 2811 ee t. (G); lava soil slope, Big Willo, 900 m. alt., Macbride 109 (G, UC) : ie (G). a 44 (G);' 18 894, Van Renselaer (G); Steilacoom, Piper 372 (G); , Cape Horn, Paper 5018 River, western Klickitat cote Suksd of C. sperma). Orgcon: along Hood River, Haller fe 12 (Ue): L ‘pale Rock, 390 m. alt., Leiberg 116 (G, UC); rocky bank, 1.6 km. north of Tonquin, Nelson Falls, Marion Co., Nelson 4869 (G); 3.2 km. east of Curry-Douglas county- line on trail from Marial to West Fork, Peck 3995 (G); am soil vy sie Marial, Nelson 1414 (G); rocky woods, Elk Rock, Nelson 12 = (G); dr a ac near mouth of Mule Creek, Curry Co., Nelson 1369 (G); ttech pe, The Heads, Port Oxford, Peck 8473 (G); Grant Pass, Piper 3105 and 5106 (G); Grant Pass, sig 043 (G, 200 al Le m. a : 1882, Cleveland (UC); southeast side of Snow Mt. above Bonnie View Helier 13236 Gy, Rush Cesk T Trinity Co., Yates 402 (UC); Minmathon. opeland 3550 50 (G, ISOTYPE of C. Foes, dry hills near Yre ka, Butler 733 THE NORTH AMERICAN SPECIES OF CRYPTANTHA 77 (UC) and 1416 (G, UC); Edgewood, 1887, Curran (UC); Siskiyou County, 1890, Edmonds (UC); along streams in me adows, Parker Cr Se! Warner Mts., 16 alt., 1910, Taylor & Bryant (UC); Goose Lake Valley, 1884, Austin (UC); Milford, 1892, Brandegee (UC); Mormon B , Congdon 50 (G); Salm reek, Tula 2100 m. alt., Hall & Babcock 5132 (UC); Nine Mile Creek, Tulare Co., Culberton 4537 (G, isotype of C. incana); Redrock Meadows, Tulare Co., 2610 m. alt., Hall 8394 (UC). Nervapa: Reno, 1898, Purpus (UC); Dog Valley Road, 1895, Hillman (UC). In northwestern United States, where this plant appears to be rather common, it is almost the only species of Cryptantha with con- spicuous corollas. In the past it has been much confused with C. am- bigua and C. Torreyana, although in fact it seems to be most closely related to C. intermedia. In Idaho it tends to intergrade with C. Torreyana, but over most of its range it is readily separated by its conspicuous corollas, commonly ternate spikes and usually tuber- culate nutlets. Occasional plants are hard to separate from C. am- bigua although C. Hendersoni is for the most part pretty easily recog- nized by its large-flowered, ternate spikes projected above the leafy mass of the plant. Crypiantha intermedia completely intergrades with C. Hendersoni in northern California, although the overwhelm- ing mass of material of C. Hendersoni is readily distinguished from its relative by its less stiff and scarcely pungent pubescence, and broader tuberculate nutlets. It has been found expedient to admit consider- able range of nutlet-variation in C. Hendersoni. Among plants quite similar in gross habit, and commonly from within a small natural region, the nutlets frequently (from plant to plant) vary from entirely smooth and shiny, to simply granulate or decidedly tuberculate. Roughly it can be said, however, that more of the northern than southern material is smooth-fruited. Cryptantha incana and C. gran- diflora are names applied to the phase with smooth and shiny nutlets. The other names cited apply to forms with roughened fruit. Cryp- tantha scabrella was based on a phase of C. Hendersoni from southern Oregon in which the nutlets are papillate-muricate. Cryptantha monosperma has similar though less abundant and less well develo papillae, and appears to be a form in which only 1 or 2 nutlets develop. The aborted nutlets appear to be the abaxial ones. 37. C. Traskae, sp. nov., pumila sparse laxeque ramosa 8-10 cm. alta; caulibus gracilibus strigosis ca. 1 mm. crassis; foliis paucis linear- ibus 1-2 em. longis 1-1.5 mm. latis acutis strigosis rare hispidis in- conspicue pustulatis; spicis solitariis vel geminatis 1-5 cm. longis cum bracteis linearibus 2-5 mm. longis numerosis ornatis; floribus obscure biseriatis maturitate 5-10 mm. separatis; corolla inconspicua ca. 1.5 78 JOHNSTON mm. lata; calycibus fructiferis ovatis subsessilibus deciduis 2-3 mm. longis; lobis calycis maturi lanceolatis saepe acutis in costa cum setis flavescentibus brevibus horridis et in marginibus adpresse hispidis; nuculis 4 homomorphis ovatis vel anguste ovatis vix 1.5 mm. longis minutissime granulatis apicem versus plus minusve_ tuberculatis dorso convexis margine obtusis ventre 34 longitudinis ad gynobasim angustam ca. 0.9 mm. longam adfixis; sulcis clausis basi in areolam minutam deltoideam dilatatis; stylo nuculas vix superante. Known only from San Nicolas Island off the Californian coast. Ca trornia: one locality, bare windswept cliffs, San Nicolas Island, April, 1901, Trask = TYPE); infrequent on bare windswept heights, April 1897, Trask 57 (UC). This endemic of San Nicolas Island was reported by Eastwood, Proc. Calif. Acad. Sci. ser. 3, i. 109 (1898), as C. Torreyana. It is evidently quite distinct from that species in its low habit, pubescence, bracteate inflorescence and small tuberculate nutlets. In habit it somewhat suggests C. leiocarpa. Its nutlets are about the same size and shape as those of that species, but differ in being tuberculate and in having a broadly forked groove and a small areola. 38. C. Torreyana (Gray) Greene. Commonly scantily and more or less strictly branched hispid herb, 14 dm. tall; stems solitary or frequently several, usually inconspicuously short-strigose as well as hispid; leaves oblanceolate.or linear, strict or ascending, 2-5(—7) em. long, 3-6(-8) mm. wide, obtuse or rounded, hispid, inconspicuously pustulate if at all so; spikes usually geminate, naked, 4-8(-15) cm. long, more or less projected from the leafy mass of the plant, very elongate and loosely flowered or congested and glomerate; corolla inconspicuous, ca. 1 mm. broad; fruiting calyces oblong-ovate or ovate-lanceolate, 2-7 mm. long, ascending, asymmetrical, base rounded or broadly conic, pedicels ca. 0.5 mm. long; mature calyx- lobes lanceolate to lance-linear, connivent above with tips usually spreading, midrib slightly thickened and hispid-hirsute, margins hispid-strigose; nutlets 4 (occasionally 1 or more aborted), usually broadly ovate, 1.5-2.2(-2.5) mm. long, 0.8-1.3 mm. broad, smooth and polished, usually mottled, rarely finely granulate, back very low- convex, sides rounded or obtuse, groove broadly forked below and closed throughout; gynobase about 14 height of nutlets, ca. 1 mm. tall; style reaching to ie a of nutlets or rarely even to their tips.— Pittonia i. 118 (188 ar. genuina. ue ruiting calyx 3.5-8 mm. long; style clearly sur- passed by nutlets; plants usually over 2 dm. tall, not conspicuously THE NORTH AMERICAN SPECIES OF CRYPTANTHA 79 hispid.—K rynitzkia Torreyana Gray, Proc. Am. Acad. xx. 271 (1885). C. Torreyana Greene, |. c. K. T, orreyana, var. calycosa Gray, 1. c. Torreyana, var. ealycosa Greene, |. c. 119. C. calycosa Rydb. Mien: N. Y. Bot. Gard. i. 331 (1900). C. affinis, var. flexuosa Nels. Bot. Gaz. xxx. 195 (1900). C. flexuosa Nels. New Man. Bot. Cen. — Mt. 416 (1909). Extreme western Wyoming and northern Utah, westward to Washington and ass and northward through British Columbia to southern Alas ALASKA: Skagway, Macoun 78736 and Eastwood 802 (G). Britisn Cotum- BIA: Ly tton, 1885, ey her: dry ground, Spences Bridge, 1889, Macoun (G); Cascade, Macoun 6658 0 (G).. Wyomine: — slopes, Nez Perces Creek, Nelson 6224 (G); ie sagebrush, 32 km. west of Par Piney, Payson 2628 G); on cobblestone flat, Jackson Lake, pe 6546 (G, UC, tsotyres of C affini » var. fleruosa); Teton Forest icin og 1897, Bra legee iC); abandoned field, Alpine, Payson 3392 (G): dry hillside, hills f Afton, 1980 t., Payson & Armstrong 332 TAH: waste y groun hae Goodding 1068 (G, UC); dry plains, Red Butte Canyon, hillsi . alt. try, son 2811 in pt. (G); Snake Country, Burke (G); pes, loose soil, Silver City, 2100 m. alt., Macbride 373 (G). WASHINGTON: on prangle, Piper 4140 (G); Clark Springs, Kreager 93 (G); cheney. Pucker 129 127 127 (G); err , Piper ; a. ry hillsides al lue one Falcon Valley, Suksdorf 78: 9 (UC): ; Faloots Valley, Suk Suksdorf 593 (G); : = matains Klickitat County, 1886, Suksdorf (G); Yakima Region, 1882, Brandegee (UC) c : ON: und, Rock Creek, 10 km . west of Haines, Peck 3609 (G); a8 woods on West — Peck : alls road, of Kes oth Fel Pek 7 (G). ak Pass, Piper (G) VADA ee hillsides, Bieroth’s Ranch, McDo: Creek, 1950 m. alt., Ne & tapascieg 215 Hum — Mts., 2100 m Watson 858 (G, ar Carson City, 1866, A on 12 (G); Spru mont, J (UC); Tnakine Road, Lake Tahoe. 1890 m m. alt., Kennedy 1428 (UC); Peavine , 1895, Hillman (UC); foothills near rson’s Ranch, Truc Valley, 1894, Hillman (UC). Ca irornra: Hupa Indian Reservation, 150 m. alt., Chandler 1303 (G, UC); Supply Creek, Hupa Valley, Davy lale 5730 (UC); Kneeland Prairie, 750 m. alt., Tracy 3071 (UC); High Prairie on Bald Mt., 1650 m Bo oe Tracy 4560 (UC); apse ‘ago m. Po “s one, _ (UC); open ‘lacie n brus’ abe slopes, 1650 m. Tracy 3944 (UC); y hillside at Alder Point on Eel River, 150 m. alt., gai 1879 (GC), Eel 80 JOHNSTON River, 1893, cna otf (UC); Ukiah sepred 3916 (UC); gravelly slope about 16 km. e of Alder Springs, Heller 1 0 (UC); gravelly places, Alder Sorties 1260 ok: Heller 12759 (G); ae near trees, betw on ue e am ; UC i shady ground, Lake age — 42 (G); Napa Valley, Torrey 330 (G); hy places near stream, 13 above Ruth, Tr ai ay Co., Tracy 4301 (UC); reka, Heller 7995 (G, UC); ey y land nea: r Yreka, Butler 762 and 1310 (UC); " ; y, 1892, avera av (UC); Yosemite Valley, Torrey 337 (G); Yosemite Valley, Bolander 6283 (UC): Yosemite Valley, Hall 9168 (UC); Mariposa Big Tree Grove, 1893, pee wood (UC); Mt. Bullion, 1889, Congdon (G); poeiees. et, a (G, ; : 0 n, Mono Co. , Var. ae var. nov., varietatem genuinam simulans differt stylo nuculis subaequilongo vel rare eis longiore. Vicinity of Pameiaics California, in the southern part of the North Coast Range CALIFORNIA: bushy eon near Callen, 120 m. alt., Tracy 1865 (UC); hills east of Calistoga, m. alt., Tracy 2094 (trpx, U. C. ‘Herb. no. 175,784); : Rosa Creek Ah Baker 620 (UC); ? ‘ne Honda, 1890, Brandegee Var. pumila (Heller), comb. nov. Fruiting calyx 2-3.5 mm. long; style shorter than nutlets; plant usually rather conspicuously hispid- hirsute, usually under 2 dm. tall.—C. pumila Heller, Mubl. ii. 242 (1906). Middle Coast Ranges of California from Marin to Santa Clara County. CaLIroRNia: Mt. Reaver og a 8403 (G, IsoTyPE of C. pumila); M amalpais, 1893, ines ged (UC); in brus h at Sugar Loaf Peak, Berkeley Hills, 300 m. alt., Trac 2060 (G, UC); in _— Strawberry Canyon, 150 m. alt., "Tron racy 793 and 2054 (UC); wooded slope THE NORTH AMERICAN SPECIES OF CRYPTANTHA Sl Smith Creek ry foot of Mt. Hamilton, Heller 8588 (G); foothills west of Los Gatos, Heller 7458 (G, UC); a, brian Searsville, Bolander 39 (G); with- out locality, anes & Harford 770 (G). This species is at once the most widely distributed and the most northerly ranging of all the North American species. It has been generally accepted and well understood, being readily recognized by its erect loosely branched rather slender habit and broadly ovate smooth nutlets. Although it may intergrade with C. Hendersoni, possible intergrades are few and the ranges of the two species are such as to suggest specific difference. Cryptantha Torreyana differs from C. Hendersoni in having much smaller corollas, prevailingly geminate rather than ternate spikes, and always smooth rather than smooth or tuberculate nutlets. Except in California and Alaska C. Torreyana occurs in the dryish interior east of the high mounains, whereas C’. Hendersoni, while occurring east of the mountains, is very common in the moister valleys of western Oregon and Washington. Piper, Contr. U. S. Nat. Herb. xi. 484 (1906), has given the type locality of C. Torreyana as “Grassy hills near San Luis Rey, Cali- fornia,” which is a locality quite beyond the known range of the “species and in a region in which it is not to be expected. I have been unable to locate in the Gray Herbarium a collection of this species made by Torrey at the locality mentioned. There appear to have been only two of Torrey’s collections of this species available to Gray. Gray based his species on a number of specimens and had no “type” in mind. If a type must be selected it seems better to choose either Torrey 330 from Napa Valley, or preferably Torrey 337 from Yosemite Valley since the latter is more characteristic of the species. Most recent botanists, following Gray, have maintained the variety calycosa, a variation distinguished by its glomerate spikes and usually elongate calyx-lobes. This variation occurs throughout the range of the species, and although an obvious sort of variety seems to be a trivial one better treated as a mere forma or phase of the species worthy of no particular nomenclatorial attention. In California the typical form of the species pushes southward along the length of the Sierra Nevada, the most southerly undoubted station being in Tulare County. In the herbarium of the University of Cali- fornia there is a specimen of C. Torreyana labeled as having been collected by G. F. Reinhardt in the San Jacinto Valley of Riverside County in Southern California. The data accompanying this speci- men I doubt, since no other of the numerous collectors visiting the San Jacinto Valley has detected the species, and since the hot plains 82 JOHNSTON and buttes about San Jacinto seem scarcely a likely locality for the southern outpost of a species which over most of its range in Cali- fornia is confined to the mountains and chiefly to the pine belt. In the Coast Ranges of California the species occurs in its typical form as far south as Lake and Napa counties. Just south of these counties it is replaced by var. calistogae and var. pumila. Over its extensive range C. Torreyana is very constant in having styles evi- dently shorter than the nutlets. This condition is departed from in the available material from the vicinity of Calistoga, Sonoma County, California; the styles in this material clearly reaching to or slightly beyond the nutlet-tips. To this variation, because of its geographical correlation, the varietal name calistogae has been applied. The var. pumila is of different character. In the Middle Coast Range the plants representing C. Torreyana are more hispid and smaller in all parts. This plant has been confused with C. leiocarpa but is quite distinct, for, like the typical C. Torreyana, it differs in its broadly ovate nutlets, shorter style, and ebracteate spikes. Macbride, Contr. Gray Herb. n. s. xlviii. 43 (1916), confused the plant with C. hirsutis- sima, a close relative of C. leiocarpa. From C. hirsutissima the variety differs in its broad nutlets, somewhat shorter style, and different habit. Although the plant was described as a distinct species by Heller it is at most a weak en variety of C. Torreyana and perhaps not worthy of recognition at a . Simulans Greene. Erect strigose pallid herb with few strictly ecending branches, 1.5-3(-4.5) dm. high; scantily and loosely strigose or below rarely shortly and loosely appressed-hispid; leaves not numerous, oblanceolate or oblance-linear, 2-5(-7.5) cm. long, 2-5(—7) mm. wide, strigose, pustulate (especially the lower ones), frequently extending into the lowermost part of the spikes; cotyledons and early leaves frequently persistent at anthesis; spikes solitary or frequently geminate or ternate, slender, usually elongate and sparsely flowered but at times glomerate; corolla inconspicuous, ca. 2 mm. broad; fruiting calyx 3-8 mm. long, oblong-ovate, slightly asymmet- rical, strict or ascending, obscurely biserial, base rounded or broadly conic, pedicels ca. 0.5 mm. long; mature calyx-lobes lance-linear, con- nivent above with the green tips spreading, midrib slightly thickened and shortly arcuate-hirsute, margins white villous-hispid; nutlets 4, homomorphous, broadly ovate, 2-2.5 mm. long, 1-1.5 mm. broad, densely granulate or granulate-muriculate, sparsely broad-tuberculate, back low-convex, margins rounded, groove spr forked below and usually closed throughout; gynobase ca. 1-1.5 mm. high; style reach- ing to about 3/,—*/; height of nutlets.—Pittonia v. 54 (1902). THE NORTH AMERICAN SPECIES OF CRYPTANTHA 83 In the pine forests from southern Oregon to Southern California, and local in northern Idaho and middle southern Washington. an dry pine woods, Mt. Moscow, Henderson 281114 (G). WASHINGTON: Ris oods, Faleon Valley, Suks deck, 46 a nd 595 (G), 181 (UC). OrEgGon: Kiamath Valley, 1260 m. alt., 1864, Cronkhite (G); summit of Cascade Mts., along Ashland-Klamath Falls road, Peck etl @); pi locality, 1883, Howell (UC). Nevapa: Incline Road, Lake Tahoe, m. alt., Kennedy 1446 (UC); Kings Cany y 1700-2000 ag It., Baker 1194 (G, C); Dog Valley Road, 1895, Hillman (UC); log railroad north of Verdi, 1590 It 0873 (G, UC). IFORNIA: in conife rests r isson, Heller 8035 (G); G Valley, Eastwood 765 ; Milford, 1892, Brand tc): Prattville, Heller & Kennedy 8 (G); Prattville, 1892, Brandegee ( rican River at Strawbe 16 F 8 ); Truckee, Heller 7060 (G, UC) er County, 1892, Carpenter (UC); ( ; Plac Jackson, 1892, Hansen (UC); beneath pines in Pig ae of Hog Ranch, Yose- mite Park, Hall & Babcock 3318 poe 3444 (UC); Hog Ranch above Hetch- etchy, Congdon 51 (G); —— Valley, 1200- 1350 m alt. ., Abrams 4379 n n Mt., Mari Ss UC); i m. alt., - alt. aoe rie! : YA cari Peak, 1894, Brandegee (UC); oe aga California, 1880, t 11 (G); sags mage a Parry & Lemmon 279% (G); This readily ecseiued species, so characteristic of the Yellow Pine forests of California, has passed as C. ambigua, although very different from the low hispid plant of the Northwest properly bearing that name. The pallid strigose pubescence and the broad granulate and tuberculate nutlets readily distinguish it from C. Torreyana with which it grows and somewhat approaches in habit. 40. C. ambigua (Gray) Greene. Ascending hirsute herb 1-2.5 dm. tall; stems usually loosely branched from the base, hirsute and some- what short-strigose; leaves narrowly lanceolate to linear, 2-3(-5) em. long, 1-4(-5) mm. broad, obtuse or subacute, usually somewhat ap- pressed hispid-hirsute, commonly inconspicuously pustulate; spikes usually solitary, 5-15 cm. long, naked or with the lowermost flowers bracted, commonly not projected clear of the leafy mass of the plant and usually not sharply differentiated from the leafy peduncular 84 JOHNSTON branches; corolla inconspicuous, 1-2 mm. broad; fruiting calyces ovate-oblong or oblong, 4-7 mm. long, slightly asymmetrical, spread- ing, crowded or distant and obscurely biserial, base rounded or broadly conic; pedicels 0.5-0.9 mm. long; mature calyx-lobes linear or lance-linear: usually more or less connivent above, midrib slightly thickened and tawny hirsute, margins evidently short strigose-villous, nutlets 4, broadly ovate, 1.6-2 mm. long, granulate and coarsely tuberculate or very rarely tending to be smooth especially towards base, back low convex, sides obtuse and rounded, groove closed or rarely somewhat dilated but always broadly forked at base; gynobase narrow, 1-1.2 mm. long, 24 height of nutlets; style reaching +/5-°/5 height of nutlets Beet 1. 113 (1887). Eritrichium muriculatns var. ambiguum Gray, Synop. Fl. N. Am. ii. pt. 1, 194 (1878). Kry- nitzkia ambigua Gray, Proc. Am. Acad. xx. 273 (1885). Eritrichium muriculatum of Torr. Bot. Wilkes Exped. 416, t. 13 (1874). C. poly- carpa Greene, ee 114 (1887). C. multicaulis Nels. Bot. Gaz. xxx. 194 (19 Southern oe to southwestern Montana and thence south- ward to northern Colorado, extreme western Nevada and northeastern California nish B asin, en sey 4886 a $37 (G); a ercte fos fale to Bidets Valley, Watson 286a (G, UC); Bridger Mts g & Bessey 9 MING; on dry loose soil of a e, Snake River, Nels (G, ISOTYPE 0 icaulis) ; : River near Joust ction Butte, Nelson 57614 a (G); under granite cliff, Cow Creek, bany Co. » Nelson 8888 (G); dry loose soil, Centennial, Nelson 8731 (G). :'m DAH Sar: *s 1 3305 sagebrush covered hillside, C orral, 1650 mm. alt., Macbride & Payson 2936 (Gy fs diel owt held, Martin, 1800 m. alt., Macbride & Payson 3044 (G); hills, Castle Rock, 1500 m. alt., ‘Macbride & Payson 2861 (G); (G); loose dry soil, Ho on & M. 1773 (G); dry soil, St nthony, Merri ilcox pt.(G). Urtan: Thistle, 1590 It., Jo 537 N: upper end of Star Valley near Dee , Heller 9074 (G); Frank 1 . alt., He 4 (G); Galena Creek, m » Kennedy 1293 ); Lake Washoe, Torrey 335 (G); Palisade, 1903, Stokes econ Ca, oe of — - n, © eo was on, Suksdorf 594 : ckitat, Howe aon ; Walla Walla n, 1883, Brandegee (UC); Washington Terri ran 193 ( N Sisters, hag? 3607 (Gel Ae ee "Nelson 8 G). Oregon: UC); ngage along Des Chutes Rive 8 km. below Bend, Peck orb if 498 8 Sw e Valley, gate 370 (G). Casreeaa: ortola, K. Brandegee (UC); lake icmelgr ae Co., Austin & Bruce 2267 (UC); ice pond below ike 1887, Sonne (UC, ISOTYPE of C. Sepiepay: THE NORTH AMERICAN SPECIES OF CRYPTANTHA 85 In the past this species has been greatly confused and the name has been used in a variety of applications. The plant treated here is that figured and described by Torrey in the Botany of the Wilkes Ex- pedition, 416, t. 13 (1874), since Gray appears to have based his Eritrichium muriculatum, var. ambiguum largely upon Torrey’s plate and description. Torrey gives his plant as from Nisqually, but since it represents a species characteristic of the dry interior of Washington it seems likely that, as with other material collected by the Wilkes Expedition, the data had become confused and the plant was probably collected in the Walla Walla Region of southeastern Washington, cf. Piper, Contr. U. S. Nat. Herb. xi. 15 (1906). Although usually distinguishable by its habit, C. ambigua ap- proaches C. Kelseyana and C. Pattersoni very closely in Idaho and Wyoming, and is occasionally distinguishable from them only with difficulty. Doubtful intermediate plants with homomorphous or subhomomorphous nutlets I have referred to C. ambigua even though some so referred have one nutlet subpersistent and occasionally some- . what less tuberculate than the others. Some plants from Idaho referred to C. ambigua have nutlets smooth quite like those of C. Tor- reyana. These plants are so referred because their habit is that of C. ambigua. Cryptantha Hendersoni usually agrees with C. ambigua in the size, shape and roughenings of the nutlets, but has a very dif- ferent habit and inflorescence. . C. echinella Greene. Usually low and loosely branched hispid herb 5-20 or rarely 40 cm. tall; stems short-hispid, sparsely branched; leaves oblanceolate or oblance-linear, 1-2.5(-4.5) cm. long, 14(-6) mm. broad, obtusish, appressed-hispid, minutely pustulate, not numerous; spikes solitary or at times geminate, 1-5 cm. long, slender, commonly leafy-bracted towards base; corolla inconspicuous, 1-1.8 mm. broad; fruiting calyx oblong-ovate, 5-6 mm. long, deciduous, spreading, obscurely biseriate; pedicels 0.1-0.5 mm. long; mature calyx-lobes linear-lanceolate, connivent above and usually recurved, midrib slightly thickened and pale-tawny hirsute, margins appressed short-hispid; nutlets 4, homomorphous, broadly ovate, 2-2.2 mm. long, more or less finely granulate, conspicuously and narrowly papil- late, back convex, margin rounded, groove very narrow or closed and widely forked at base; gynobase about 24 height of nutlets; style shortly but definitely surpassed by tips of nutlets.—Pittonia i. 115 (1887). aie Sierra Nevada to the mountains of Southern California and the Charleston Mts. of southern Nevada. 86 JOHNSTON Nervapa: Densmore Camp, Hunte —— Canyon, 1800 m. alt., Kennedy 1606 in eu rine ); Peterson’s Ranch near Reno, 1894, Hillman (UC); Peavine Foothills, 1895, Hillman (UC); Charleston Mts., Purpus 6077 (UC). Cati- FORNIA: Mt. St : anford, 2640 m. alt., 1886, ne (UC, 1soTyPE); Castle Peak near highest point, Heller 7079 (G); Tahoe, 1901, Boring (UC); Lake Tahoe R 1901, Setchell & Dobie (UC); Luthers Pass, 2340 m. alt., Abrams 4759 (G); Yosemite Bea ie 1200-1350 m. alt., Abram ms 4379 in pt. ne dry e ati Brandegee (UC); ee Camp above Bishop, 1913, K. Brandegee (G, UC); jasenill. Mt. Pinos, 2490 m. alt., Hall 6523 (UC); rocky ground etn pines, ae ; Kelly’s Cabin, Ontario Peak, 2460 m. alt., Johnston 1620 (G, UC); dr ry — east of Ontario Peak, 2520 m. alt., Munz 6076 (UC); Coldwater Fork of Lyt Creek, 1725 m. alt., Johnston 2057 in pt. (G); Mare Flat, 2400 m. alt. phe is ford 934 (G); Little Greet Valley, 2160 m. alt., Hall 24 (UC). This characteristic species has been greatly misunderstood, and repeatedly confused with C. ambigua and C. intermedia. It grows in dry sunny clearings in the Yellow Pine belt of the California moun- tains usually in the company of C. simulans, C. affinis and C. Tor- reyana. Ser. X. MOHAVENSES. Nutlets 4, smooth, oblong-ovate or - lanceolate-ovate or lanceolate, clearly angled at the sides, decidedly homomorphous; style usually equalling height of nutlets or shorter than latter. * ——. conspicuous; style clearly surpassing the nutlets, binge PAs MOM cine Sra as Gea nie ce eb ere Clery mee a mohavensis. cc : vs gg” aie — or a trifle mas is oon: em 1. oo mm. long.. "43. C. Watsont. 42. C. mohavensis Greene. Ascendingly branched herb 1-4 dm. tall; stems usually freely branched, short-hispid to hispid strigose; leaves linear or lance-linear, 1-4 cm. long, 1-3 mm. broad, appressed- hispid or strigose, minutely and densely pustulate, obtusish, upper ones reduced; spikes ternate or geminate, usually crowded, 2-6 cm. long, naked; corolla conspicuous, 4—-7 mm. broad; fruiting calyces oblong-ovate, 3-5 mm. long, ascending, becoming obscurely biserial, symmetrical, base rounded, deciduous, pedicels ca. 0.5 mm. long; mature calyx-lobes lanceolate, connivent above, midrib somewhat thickened and frequently sparsely hirsute, margins usually more or less silky strigose; nutlets 4, homomorphous, smooth and shiny, rarely obscurely granulate, oblong-ovate or lance-ovate, 2-2.5 mm. long, back low-convex or flattish, margins definitely angled especially towards the apex, groove closed above but forked below and opened at the fork to form a small triangular areola; gynobase columnar- sobuiate, about 34 height of nutlets; style clearly surpassing tips of THE NORTH AMERICAN SPECIES OF CRYPTANTHA 87 nutlets.—Pittonia i. 120 (1887). Krynitzkia mohavensis Greene, Bull. Calif. Acad. Sci. i. 207 (1885). Southern Sierra Nevada of California, best known from the vicinity of peneee Mountains. and Cameron, 1905, K. vets (UC); Monte” Deshet, 1884, Curran (G, ISOTYPE). In habit quite similar to C. oxygona, and like that species much suggesting C. muricata in gross aspect. Although having smooth, wingless nutlets it seems very closely related to C. oxygona. 43. C. Watsoni (Gray) Greene. Slender strictly branched hispid herb 1-3 dm. high; stems solitary, sparsely to loosely branched, spreading short-hispid; leaves linear to oblanceolate, 1—4(—5) cm. long, 1-4(-5) mm. wide, obtuse or rounded, ascending, hispid and rarely pustulate; spikes solitary or geminate, 1—4(—6) em. long, occasionally leafy-bracted below; corolla inconspicuous, ca. 1 mm. broad; fruiting calyx ovate or oblong-ovate, 2-3.5(-4) mm. long, subsessile, rounded at base, early deciduous, oldest ones becoming distant; mature calyx- lobes lanceolate, tips usually connivent, midrib hispid and scarcely thickened, margins appressed short-hispid; nutlets 4, Speayoer, or practically so, lanceolate, 2 mm. long, ca. 0.8 mm. broad, smooth, shiny or at times dulled by pe ee granulations, beck nearly flat, margins definitely angled, groove closed or nearly so and forked at base; gynobase subulate, ca. 24 height of nutlets; style equalling nutlets or a trifle surpassed by them.—Pittonia i. 120 (1887). Kry- nitzkia Watsoni Gray, Proc. Am. Acad. xx. 271 (1885). C. vinctens Nels. & Macbr. Bot. Gaz. lxii. 143 (1916). Eastern Washington to western Montana, southward to Nevada and northern Colorado. NTANA: Canyon Ferry, 1898, Brandegee 30 (UC). WyomIna: steep aa 2! tone River near Junction Butte, Ne Seca n S761 It., m bi 75 (UC) AHO: loos $ near road, Ch 620 m. alt., 2 ide & eg oa 22 (G); lopes, Mackay, 1750 m. alt., Nelso Mac vstate sandy slopes ew 660 t., Mac 2G, UG), * m. alt., Watson 858 (G, Type of K. Watsoni) Ravan ve i est base of Mt. Grant old- field, Heller 10970 in pt. (G, UC); Tonopah, 1800 m. alt., Shockley 81 (UC). WASHINGTON: junction of Crab and Wilson creeks, 390 m. alt., Sandberg & 88 JOHNSTON Leiberg 249 (G, UC). Oregon: dry ground, Narrows, Peck 3587 (G); Juniper Springs, 1350 m. alt., Leiberg 2271 (G, UC); rocky slopes, Mathew near Harper’s Ranch, 1100 m. alt., Leiberg 2235 (G, TyPx of C. vinctens; UC, ISOTYPE) ; clay banks, Mathew Butte, 750 m. alt., Leiberg 2041 (G). An interesting species characterized by its four, lanceolate, angled nutlets and well developed style. Although it has been confused with C. gracilis it is really quite distinct from that plant in the angling and number of nutlets, length of style, and shape and pubescence of calyx. Cryptantha vinctens is a peculiar form of this species having somewhat appressed-pubescent and inconspicuously hispid calyces. The segre- gate is not separated geographically, and appears to be merely an extreme form whose characters of pubescence are rather completely obliterated by transitional forms clearly referable to C. Watsoni. Ser. XI. GRACILES. Nutlets 1 or rarely 2, smooth, lanceolate, laterally rounded or obtuse, subhomomorphous, axial one always de- veloping and in general slightly larger than the second nutlet when that develops; style reaching to 24-34 height of nutlet. 44. C. gracilis Osterh. Slender erectly branched herb 1-2 dm. high; stems usually solitary, sparsely branched, densely spreading short-hispid; leaves not numerous, linear to narrowly oblanceolate, 1-3 em. long, 1-3 mm. broad, obtuse or rounded, ascendingly short- hispid, usually minutely pustulate, upper leaves reduced; spikes solitary or geminate, usually dense, 1-2 cm. long, naked; corolla in- conspicuous, limb 0.6-1 mm. broad ; fruiting calyx ovate, divaricate, 2-2.8 mm. long, promptly deciduous, base decidedly conical, sessile; mature calyx-lobes lanceolate, rather densely appressed tawny hispid- villous, tips erect, midrib slightly thickened and inconspicuously short-hispid; nutlets 1 or rarely 2-3 and then more or less unequally developed, lanceolate, 1.5-2 mm. long, ca. 0.8-1 mm. broad, smooth and shiny, acute, back nearly flat, sides rounded at least towards apex, groove usually opened to above middle and scarcely forked ow; gynobase ca. 14 height of nutlet; style reaching to 24-34 height of nutlet—Bull. Torr. Bot. Cl. xxx. 236 (1903). C. Hill- gracilis, var. Hillmanii Munz & Johnston, Bull. Torr. Bot. Cl. xlix. 39 (1922). : Southern Idaho and eastern Colorado to northern Arizona and southeastern California. Ipano: plains of the Snake River, Palmer 72 (G, : wi locality, Henderson 2561 (G). _Conorapo: Ww es ce (G: dry mesa among junipers, Nucla, 1800 m. alt., Payson 395 (G). Uvan: THE NORTH AMERICAN SPECIES OF CRYPTANTHA 89 to Range, Purpus (UC); near Reno, 1894, Hillman (UC); Cande- laria, Shockley Mesia west of Goldfield, 1860 It., Heller 10970 (G). vA: Ash Rusby 747 (UC); Grand Canyon, RIZON: For Dougal 184 (G, UC); G Grand Canyon, 1907, Setchell (UC). Catirornta: near Bonanza King Mine, Providence Mts., Munz, Johnston & Harwood 4222 (UC); Barnwell, K. Brandegee (UC); Silver Canyon near Laws, K. Brandegee (UC). This is a very distinct species and is scarcely to be confused with any other once its characters are understood. The calyx is notably conical at the base, densely covered with short appressed hairs, and apparently lacking pungent hairs on the lobes. One nutlet commonly develops; this is obscurely if at ell angled on the edges; and evidently surpasses the style. Occasionally two or even three nutlets are matured, and then they appear to be unequal in length and develop- ment. The normal and fully developed nutlet is always near the axial calyx-lobe. Ser. XII. RAMULOSISSIMAE. Nutlets 4, smooth, lanceolate, laterally rounded or obscurely angled, practically homorphous but with the axial one minutely and obscurely though definitely larger than the others and always present when for any reason less than the normal number develop; style reaching the tips of the nutlets or barely surpassed by them. 45. C. Fendleri (Gray) Greene. Herb 1-5 dm. high, usually with a definite straight erect axis (commonly simple below but producing _numerous ascending laterals above), more or less densely hispid and frequently appressedly so; leaves narrowly oblanceolate, acute, 2-5 cm. long, 2-4 mm. broad, appressed-hispid, frequently pustulate beneath; spikes solitary or geminate, 2-12 cm. long, sparsely if at all bracteate, loosely flowered; corolla inconspicuous, ca. 1 mm. broad; fruiting calyces ovate-oblong, 4-5(-7) mm. long, ascending, slightly asymmetrical, obscurely biserial; pedicels ca. 0.5 mm. long; mature calyx-lobes linear to lance-linear, usually loosely connivent with the tips somewhat spreading, midrib thickened and hirsute, margins strigose; nutlets homomorphous, 4 (exceptionally with 1-3 aborted, but then the axial nutlet always present), smooth, somewhat shiny, lanceolate, acuminate, 1.5-2 mm. long, back convex, sides rounded or obscurely obtuse, groove closed or nearly so but at base opening into a definite deltoid areola; gynobase subulate, twice length of style, at least 24 height of nutlets; style equalling or barely surpassing the nutlets.—Pittonia i. 120 (1887). Krynitzkia Fendleri Gray, Proc. Am. Acad. xx. 268 (1885). Eritrichium hispidum, var. leiocarpum 90 JOHNSTON Kuntze, Rev. Gen. ii. 437 (1891). C. ramulosissima Nels. Erythea vii. 68 (1899). C. wyomingensis Gandoger, Bull. Soc. Bot. France Ixv. 62 (1918). Southern Alberta and Saskatchewan “ eastern Nebraska, northern New Mexico and Arizona; also in eastern Washington and western Nevada. ALBERTA: Hand Hills, Macoun 165 See SASKATCHEWAN: Crane Lake, Moe 5804 (G). Wromtna: sandy dry bottom lands, Dunn’s Ranch, Nelson 7611 (); sandy plains, Laramie, Nelson oB8e ater 7670 (G); Sores. he elson 6275 (G, , ISOTYPES of C. ramulosissima); Pine Bluff, N 7 3510 Gas L : i 152 Baker 780 ( EW Mexico: between Santa Fe and Canoneto, 2190 n alt., Heller $7838 @. without locality, 1847, Fendler (G, TPE of K. Fendleri) TAH: n er RIZONA: Cosnino, Rusby 750 (G); openings in yel e forest, east base of San Francisco Mts., 2190 m. alt., 1920, Hall @ OTe cate 1894, Toumey (UC); San Francisco es 1884, mon (UC);s dunes, Navaho Reserva- tion, Vorhies 124 (G, UC). WasninaTon: ata and, Pasco, pipe 2951 yy without locality, Henderson 2562 (G). Nevapa: Palisade, Brandegee G). Cryptantha Fendleri is the most easterly ranging of the smooth- fruited series of Cryptantha. In its common typical form the erect habit and paniculate branching are very characteristic, and as a general thing the species has been well understood. A diffusely branched form from southern Wyoming has been described as C. ramulosissima. In floral and fruiting structures this form is quite ee from typical C. Fendleri. Similar material has been ted in eastern Washington. The range of C. Fendleri seems siecitaent discontinuous. Its roots are frequently charged with a purple dye. The three extra-Californian species with narrow, smooth and more or less shiny nutlets may be distinguished as follows,— Calyx broad] s eben cia as sist wan ab gr oe height ee C. gracilis. Calyx ormigast at base, hispid or hirsute, inconspicuously along Pans Mee pallets 4; style about a = [eaves oe pil a a ht ree A a C. Watson. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 91 Nutlets with margins rounded or somewhat obtusish, groove opened at base to form an areola; leaves acute; plant usually with a stiffly erect straight axis............... . Fendleri. Ser. XIII. LEIOCARPAE. Nutlets 1-4, smooth, ovate or somewhat lanceolate, laterally rounded or obtuse, homomorphous with the abaxial one always developing; style reaching to 14 height of nutlets or barely surpassing them; calyx-lobes hirsute with straight hairs. Style creche to the tips of the nutlets or barely surpassing them. Nutlets ovate with a subsimple groove; spikes decidedly racteatescs 25. d ics eee ee eee . 46. C. leiocarpa. Nutlets ovate-oblong, usually with a forked groove; spikes Ee ET Os ee eee 47. C. hispidissima. Style reaching to 1/4-1/2 height of nutlets........... 48. C. microstachys. Style reaching to 2/3-3/4 height of nutlets. irs cn upper part of calyx-lobes conspicuously retrorse. 49. C. nemaclada. Hairs on calyx spreading or ascending. : Soikes ankeds. 502245. s eee see 50. C. Clevelandi. eR os ee et 51. C. Brandegeit. Corolla 1.5-2 mm. broad; sparsely branched erect ; plaiiia. 625 os ee hee wes ree ee 52. C. Abramsit. 46. C. leiocarpa (F. & M.) Greene. Laxly branched usually de- cumbent or’ prostrate herb; branches usually long and numerous, his- pid-strigose or frequently loosely appressed-hispid, becoming 1+ dm. long; leaves oblance-linear to oblanceolate, strigose-hispid and often sparsely hispid, occasionally pustulate, obtuse or rounded or rarely emarginate, 1—-2.5(-4) em. long, 1-4(-9) mm. broad; spikes solitary or geminate or rarely ternate, becoming 6 em. long but usually shorter, conspicuously leafy-bracted, not at all sharply differentiated from the mass of the plant; corolla usually inconspicuous, limb 1-2.5(-3.5) mm. broad; fruiting calyces ovate to oblong-ovate, 2-3 mm. long, usually ascending, tardily deciduous, subsessile, usually crowded or becoming loose below, subsymmetrical; mature calyx-lobes lance-linear, loosely connivent above, midrib somewhat thickened and usually decidedly tawny-hirsute, margins strigose; nutlets 4 or very rarely fewer by abortion, oblong-ovate to ovate, 1.6-2 mm. long, smooth, polished or minutely granulate and dull, back convex, margins obtuse, face con- vexo-obtuse, groove closed and very shortly forked at base if at all; gynobase subulate, 24-34 height of nutlets; style equalling or slightly surpassing the tips of the nutlets.—Pittonia i. 117 (1887). Echino- 92 JOHNSTON spermum leiocarpum F. & M. Ind. Sem. Hort. Petrop. ii. 36 (1835). Krynitzkia leiocarpa F. & M. 1. ¢. vii. 52 (1841). Eritrichium leio- carpum Wats. Bot. King Exped. 244 (1871). Along the ocean beach from southern (Curry County) Oregon to (Santa Barbara County) Southern California. Orecon: beach, Gold Beach, Peck 8691 (G); beach near Harbor, Peck 8755 (G); he reeas Howell 222 (G). CALIFORNIA: ‘sand dunes of ocean beach, Humboldt Bay, Tracy 2457 (G, UC); Bodega Point, Eastwood (G); Bodega ithe Heller 5615 (a): Point Reyes, 1886, Curran (UC); San Francisco, = Davy (UC); — Gate Park, San Francisco, er Jones (UC); Lake ae ncisco, 1 ; UC); Casmaila Station, K. Brandegee (UC); sporti ag wena K. Brandegee (UC); Surf, 1909, K. Brandegee (Gy: eel 1909, K. Brandegee a, b, c (UC). This is a very well marked coastal species, but has been greatly misinterpreted in the past and at various times made to include nearly all the smooth-fruited species of the genus. Study of material in the Gray Herbarium which was raised from authentic seeds re- ceived from St. Petersburg, Hamburg and Geneva, shows clearly that the name should be applied to the sea-shore plant of middle and north- ern California which has bracteate spikes, long styles, and small ovate smooth nutlets with a simple or barely forked groove. Occasionally it grows with C. hispidissima and has been confused with that species, although it is readily separable from it by its bracteate spikes, shorter ovate nutlets, simple or barely forked groove, and commonly smaller corollas. The leaves are usually 1-3 mm. broad, but in some pe- culiar forms from Surf, Santa Barbara County, California, which ap- parently grew with the common form, the leaves are broadly oblong, retuse | nearly 10 mm. broad. 47. C. hispidissima Greene. Erect and ascendingly branched or loosely cnehed and somewhat decumbent, 1.5-5 dm. high; stems hirsute or somewhat appressed-hispid; leaves oblance-linear to linear- lanceolate, ascending, 1.5-5.5 em. long, 1.5-4 mm. wide, obtuse or acute, appressed or frequently spreading-hispid, occasionally hirsute- ciliate; spikes ternate or geminate or rarely quadrinate, dense or re- motely flowered, bractless or occasionally with 1-2 bracts near base, 1-8(-15) em. long; corolla more or less conspicuous or occasionally in- conspicuous, 1-5 mm. broad; fruiting calyx ovate-oblong, usually , 2-5 mm. long, slightly asymmetrical, base broad, sessile; mature calyx lobes lance-linear or almost lanceolate, connivent above with the tips somewhat spreading, margins strigose, midrib thickened and decidedly hirsute; nutlets 4, eproks cant ovate-lanceolate, * THE NORTH AMERICAN SPECIES OF CRYPTANTHA 93 1.5-2 mm. long, smooth or very finely granulate, usually shiny, back convex, sides obtusish, groove simple or forked at very base; gynobase elongata, ca. 24 height of nutlets; style reaching to nutlet-tips or de- finitely surpassing them.—Pittonia i. 118 (1887). C. letocarpa, var. ) hispidissima Macbr. Contr. Gray Herb. n. s. xlviii. 43 (1916). West of and in the Coast Ranges of California, from San Francisco to Point Conception and doubtfully to near Los Angeles. Point, Monterey, Eastwood 86 (G); Seaside, Monterey, Eastwood 161 (G); Monterey, 1900, Eastwood (G); s 4559 (G): Casmaila Station, K. Brandegee (UC); Surf, 1909, z Brandegee d (UC); Redondo, 1903, Grant 5500 (UC). . Obviously related to C. Clevelandi, var. florosa and perhaps not to be kept specifically distinct, although differing in general range and length of style. 48. C. microstachys Greene. Erect slender herb 1-5 dm. high; stems commonly with numerous ascending simple or rebranched laterals, hirsute or rarely more or less strigose; leaves linear, obtuse or rounded, 1-6 mm. long, 1.5-5(-8) mm. broad, broadly sessile or the lower ones with a contracted base, hispid or hirsute, rarely somewhat strigose or pustulate; spikes slender, solitary or geminate, naked, 2-8 cm. long, frequently somewhat crowded towards the end of the stem and apparently paniculate; corolla inconspicuous, usually 0.4-1 mm. or rarely even 2.5 mm. broad; fruiting calyces ovate or oblong-ovate, 1.5-3 mm. but commonly 1.5-2 mm. long, sessile, strict or ascending, asymmetrical with the abaxial lobe the longest and most hirsute; mature calyx-lobes linear or linear-lanceolate, connivent above with the tips somewhat spreading, usually united below to form a short siliceous tube, midrib slightly thickened and divaricately short- hirsute, margins ciliate; ovules 4; nutlets 1 or rarely 2, next the ab- axial calyx-lobe, acute-ovate to lanceolate, 1.5 mm. long, smooth and shiny, back and sides rounded, groove closed and simple or forked at -very base; gynobase very short, 14 or less height of nutlet; style about as long as gynobase and commonly attaining about )2 height of nut- let or when two nutlets develop reaching to beyond their middle.— Pittonia i. 116 (1887). Krynitzkia microstachys Greene in Gray, Proc. Am. Acad. xx. 269 (1885). 94 JOHNSTON oast Ranges of middle California and the coastal drainage of Southern California. CALIFORNIA: oa slope aon of Alder ay Glenn Co., teas 11444 (@); Sone | Count ies Cur woes Wildcat Canyon, Davy 7 046 (UC); =s species is very closely related to C. Clevelandi from which it differs in its coarser habit, shorter style, and usually fewer nutlets. The forms most suggestive of that species come from Southern Cali- fornia and have more or less closely appressed pubescence and calyces becoming 2-3 mm. long. These forms are few, however, most speci- mens being conspicuously bristly and having calyces only 1.5-2 mm. long. 49. C. nemaclada Greene. Slender much-branched erect herb 1-3 dm. tall, minutely and sparsely strigose, finely hispid; leaves linear, rather few, 1-3 mm. long, 1-2 mm. broad, obtuse, somewhat appressed, finely hispid, very minutely pustulate; spikes solitary or geminate, slender, naked, becoming loosely flowered, 2-9 em. long; corolla in- conspicuous, less than 1 mm. broad; fruiting calyx oblong-ovate, strictly ascending, 2-4 mm. long, deciduous, obscurely biserial, sub- sessile by an obliquely conic base; mature calyx-lobes linear, con- nivent above with the slender tips spreading, midrib thickened and more or less abundantly hirsute but near the tip retrorsely setulose, margins sparsely strigose; ovules 4; nutlets 1-4, lanceolate to ovate-. lanceolate, smooth, 1.7-2 mm. long, back convex, sides obtuse, the one next the abaxial calyx-lobe always developing, groove opened or closed but usually open at the broad forking; gynobase slender, about % the length of th g to about 34 the height of the es i. 118 (188 a). Diggs Ranges of California from Tehachapi to Colusa County; THE NORTH AMERICAN SPECIES OF CRYPTANTHA 95 Ca.irorNiIA: Keene Station, Tehachapi, 1905, K. Brandegee (UC); Paso Robles, K. Brandegee (UC); Alcalde, 1892, Brandegee (UC); Colusa County, 1884, Curran (UC, part of TYPE). The plant is obviously related to C. Clevelandi and C. hispidissima, and further study may justify the reduction of it to one of the species named. Its outstanding feature is the possession of small stout retrorse hairs on the upper part of the calyx-lobes. The type came from Colusa County in the North Coast Range. 50. C. Clevelandi Greene. Usually erect freely branched herb 1-5 dm. tall; stems with long branches, strigose or hirsute; leaves linear to linear-lanceolate, usually acutish, 1-4(-6) em. long, 1-4 mm_ broad, hirsute or occasionally strigose, spreading; spikes solitary or geminate or ternate, 4-10 cm. long, naked, slender, usually remotely flowered; corolla inconspicuous or conspicuous, 1-5 mm. broad; fruiting calyces ovate-oblong or nearly ovate, 2-5 mm. long, strict, asymmetrical, deciduous, subsessile by an obliquely conical base; mature calyx- lobes linear or lance-linear, usually decidedly connivent with tips somewhat spreading, midrib thickened and decidedly hirsute, margins densely hispid-ciliate; nutlets 1-4, ovate-oblong to broadly lanceolate, 1.5-2(-2.5) mm. long, smooth, usually very finely granulate, bac convex, sides obtuse or rounded, axial nutlet always developing; groove closed, broadly forked at base or rarely with a small areola; gynobase elongate, 14-24 as high as the nutlets; style reaching to 24- 4/5 height of nutlets.—Pittonia i. 117 (1887). ar. genuina. Corolla inconspicuous, ca. 1 mm. broad; leaves 1-2(-3) mm. broad; spikes solitary or geminate, not sharply dif- ferentiated from leafy mass of plant; nutlets 1-2.—C. Clevelandi Greene, l. c. West of the mountains from the vicinity of Los Angeles southward to northern Lower California. Var. florosa, var. nov., a varietate genuina differt corolla conspicua 2-5 mm. lata, foliis 1-4 mm. latis, spicis saepe ternatis et supra folia projectis, nuculis 1-4.—C. Rattani Greene, Pittonia i. 160 (1887). In and west of the Coast Ranges of California from San Luis Obispo to San Diego, and infrequent northward to Lake County. 96 JOHNSTON m Summers (UC); Luis Obispo, 1905, Roadhouse ee i C); San Luis Obispo, 1911, Condit (UC); — Station near San Luis 0, K Es gee (UC); steep hills near Lompoc, Suksdorf 220 and 7757 G): Ga viota, K. Brandegee (UC); near Frazier ede Mine, Mt. Pinos, Abrams & Malreger 211 (G); 1S 22 Brandegee (UC); grassy hillsides, — | anyon, Puente Hills, Johnston 1935 (G); es ree ostir Laguna, Munz 2214 (G); roadside, Linda Vista, Macbride & Payson 797 (G, TyPE); =H Dea 1906, K. Brandegee (UC). ‘This species mre almost all the material from South California passing as C. leiocarpa. It is related on one hand to C. hispidissima and on the other to C. microstachys, apparently intergrading with both. In having a very short style, elongate nutlets and bractless spikes, it differs from C. leiocarpa. Doubtfully included in the species are gine collections from the North Coast Ranges. 1. C. Brandegei Johnston. Much branched decumbent or pros- pe herb; stems numerous, slender, hispid-strigose, 14 dm. long; leaves oblong-lanceolate or lanceolate, spreading, 5-15 mm. long, 2-3(—4) mm. broad, obtuse or acute, hispid-strigose and often sparsely hispid, minutely pustulate; spikes solitary or occasionally geminate, —§ cm. long, dense or loose and elongate, slender, more or less evi- dently leafy-bracted; corolla inconspicuous, less than 1 mm. broad; fruiting calyx ovate-oblong, 2-4 mm. long, strictly ascending, asym- metrical, becoming obscurely biserial, sessile; mature calyx-lobes lance-linear, usually connivent above with the tips spreading, midrib thickened and hirsute, margins strigose; nutlets 1-4, ovate-lanceolate, 1.5-2 mm. long, smooth or very minutely granulate, usually shiny, back cance) sides rounded, face flattened, groove closed with a well developed basal fork; gynobase subulate, about 14-24 height of nutlets; style reaching to 7/3-*/; height of nutlets—Contr. Gray Herb. n. s. Ixviii. 53 (1923) _ Known only from Santa Rosa off the coast of California. Catirornta: Santa Rosa Island, 1888, Brandegee (G, TvpE; UC, 1soTyPE)- In habit much suggesting C. leiocarpa but differing in its more aS es nutlets, widely forked groove, short style and more southern insular range. It is much more closely related to, and perhaps it is THE NORTH AMERICAN SPECIES OF CRYPTANTHA 97 only a phase of, C. Clevelandi with which it agrees in nutlets although differing in its longer style and bracteate inflorescence. Cryptantha Clevelandi, var. florosa occurs on Santa Cruz and San Miguel islands which are adjacent to Santa Rosa Island. C. Abramsii Johnston. Erectly and sparsely branched herb 1-3 dm. high; stem slender, finely strigose; leaves linear, rather nu- merous, 1-3 em. long, 1-1.5 mm. broad, acutish, hispid-strigose, usually sparsely hirsute-ciliate towards the base; spikes solitary or rarely geminate, 2-10 cm. long, evidently leafy-bracted; corolla evident, 1.5-2.5 mm. broad; fruiting calyx ovate, strict, 2.5-4 mm long, asymmetrical, becoming distant and obscurely biserial, sessile by an oblique broadly conic base; mature calyx-lobes lance-linear, loosely connivent, midrib somewhat thickened and on abaxial lobe sparsely and very shortly hirsute, margins strigose; nutlets 1+, lanceolate, acuminate, ca. 2 mm. long, smooth, shiny, back convex, sides obscurely obtuse, groove closed and broadly forked at very base of nutlet; gynobase narrow, about 24 height of nutlets; sty le reaching to 3/4- 4), height of nutlets—Contr. Gray Herb. n. s. Ixviii. 52 (1923). Southern California, known only from the type locality. Catrrornia: San Pedro Hills near Malaga Cove, Abrams 3139 (G, TYPE; UC, tsoTyPeE). Although suggesting C. leiocarpa in its bracteate inflorescence, this species differs in its shorter style and elongate nutlets with a broad decidedly basal forking of the groove. It is most related to C. Cleve- landi, var. florosa and possibly is only a bracteate phase of it. Ser. XIV. FLACCIDAE. Nutlet 1, smooth, somewhat ovate, laterally rounded or obscurely angled, always abaxial; style reaching to 14-24 height of nutlet; calyx-lobes armed with pallid encrusted arcuate or uncinate hairs. Styl hine to 1/2-2/3 height of nutlet; nutlet with an open dati : oathee : phiekat at fee as aes ed eae 53. C. rostellata. Pegi reaching to ‘en ae 1/2 height of nutlet; nutlet with a closed groove; more slender Aged 3. auger any terete oe : hairs on calyx usually de- cid Gherinted S00 pale: << <0 ios eee 54. C. flaccida. edly e — evidently com japeaeons acute; hairs on ealyx rather ooth and less pale 55. C. sparsiflora. Dea a a ea ee ee Re eee ee Ne 53. C. rostellata Greene. Stems isla usually somewhat red- dish, 1-2 dm. high, with few ascending branches above, strigose, canescent; leaves few, notably persistent ieee catpamnees thickish, firm, oblanceolate, 10-15 mm. long, 2-3 mm . broad, ascending; spikes 98 JOHNSTON solitary or geminate, stiffish, naked, 2-4 cm. long; corolla incon- spicuous, 0.5-1 mm. broad; fruiting calyces 3-4 mm. long, oblong- ovate, spreading or ascending (not strict), coarse, rather few, sub- persistent, subsessile by a very asymmetrical base; mature calyx- lobes lanceolate, margin sparsely ciliate or strigose, midrib on all lobes armed with stout encrusted uncinate or arcuate hairs; ovules 4, the one next the abaxial calyx-lobe alone developing; nutlets 1, smooth, compressed, ovate-lanceolate to lanceolate, 2-3 mm. long, back con- vex, sides rounded, base truncate; groove closed above but dilated below into a definite areola; gynobase very short and stout; style reaching up to 14-14 height of nutlet.—Pittonia i. 116 (1887). Kry- nitzkia rostellata Greene, Bull. Calif. Acad. Sci. i. 203 (1885). K. Suks- dorfit Greenm. Bot. Gaz. xl. 146 (1905). C. Suksdorfit Piper, Contr. U.S. Nat. Herb. xi. 484 (1906). Southern Washington (Klickitat County) southward through eastern Oregon to the Sacramento Valley of California. WASHINGTON: on dry hills near Rockland, Suksdorf 1495 (G, T Suksdorfii; UC, isotype). Orgcon: on dry hills near Dalles City, Peace 2346 (G, UC); camp by ay ae Crook Co., 1170 m. alt., Leiberg 282 (G, UC). Cattrornra: Hornbroo , 1889, Howell U C); Hornbrook, Howe U 1386 (UC); Leesburg, 1884, ae (UC, ‘ banks an d Springs Lake, Baker 468 (G, UC); Mt. Hamilton, Elmer 5038 (UC); between Brick Yard and Licks Observatory, 1110 m. alt., Pendleton 907 (UC); along Mt. H scatter road, 22.5 km. from San Jose, Heller 7437 - UC); Coyote, Chandle ); Lewis Creek, 1893, Eastwood (UC); Jolon, Brandegee (UC); without locality, 1899, Jarea 1 (G); Lancaster, K. Bra (UC); Saugus, ee bere: a8 (UC); Los Angeles, 1884, [Nevin?] 57g (G); grassy pee Re RSS - alt., Johnston 1941 (G); dry ridges near Cuyamaca ke, Airatis 3824 (G); Colorado Desert, 1905, Brandegee (UC); without locality, Blake (G). One of the best known species in the genus, readily recognized by its pallid strigose pubescence, stout arcuate bristles on the calyx- lobes, very short style and solitary terete ovate and rostrate nutlet. The type was collected by Douglas probably in eastern Oregon or Washington 55. C. sparsifiora Greene. Very slender, sparingly and ascendingly branched sparsely strigose herb 1-3 dm. tall; cotyledons somewhat persistent, ovate to orbicular, 2.5 mm. broad, contracted to a petiole 2 mm. long; leaves few, narrowly linear, 1-3 cm. long, ca. 1 mm. broad, strigose, herbaceous, noticeably opposite below; spikes geminate or solitary, 2-6 cm. long, slender, not stiff, with 1-2 bracts near very base; corolla inconspicuous, less than 1 mm. broad; fruiting calyces 2-3 mm. long, ovate or oblong-ovate, ascending, few, early deciduous, subsessile by a very asymmetrical broadly conic base; mature calyx- lobes Bike fansohis: united for 44-14 length of calyx, loosely con- nivent, margins sparsely ciliate, midrib slightly thickened, armed with _ short stout smoothish uncinate hairs; ovules 4, the one next the ab- axial calyx-lobe alone developing; nutlet 1, ovate, acute (scarcely if at all acuminate), decidedly compressed, smooth or finely granulate, 2 mm. long, equalled or somewhat t surpassed by calyx-lobes, base somewhat truncate, back convex, margins angled, groove closed and broadly forking near the base; gynobase low; style attaining 14-4 height of nutlet—Pittonia i. 116 (1887). Krynitzkia sparsiflora Greene, Bull. Calif. Aead. Sei. i. 203 (1885). ing the Sacramento and San Joaquin valleys of alterna. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 101 ar tnagg lo a rat 3 County, 1884, Curran (G, UC); Coburn Mills, 1892, Brandegee (UC rmon Bar, Mariposa Co., 1889, Congdon (G); Whitlocks, 1897, se et ian, Havilah, 1891, Brandegee (UC). Although having broad nutlets suggestive of C. Torreyana, this species is clearly related to C. flaccida as shown by its somewhat encrusted strigose pubescence, single abaxial nutlet, and falcate or uncinate bristles on the calyx-lobe. The plant is characterized by its broad nutlets, very slender habit, and short uncinate bristles on the calyx. The type of the species was collected by Curran somewhere in northern California, probably in Lake or Colusa counties. Ser. XV. AFFINES. Nutlets 1 or 4, = ovate, laterally rounded, asymmetrical with an excentric groove and an irregular areola, when single always abaxial; style Laces ey. 24 oie of nutlets. N Hee 4; ae reaching at least to 2/3 height of nutlets; veuped Nutlet t solitary’ te phe, te ua ee lowers ary; plant “spreading 646055 ee . C. glomerifilora. 56. C, affinis (Gray) Greene. Usually sparsely branched herb 1—2(-4) dm. high; branches commonly few and ascending but plant occasionally much branched from the base, hispid or short-hirsute throughout; leaves narrowly to broadly oblanceolate, 14(—5) em. long, 2.5-6(-8) mm. broad, few, short-hirsute, usually minutely pustulate, obtuse or rounded at tip, lowest pair clearly opposite; spikes geminate or solitary, usually 2-8 but becoming 15 mm. long, slender, remotely flowered, commonly with a very few large leafy bracts below; corolla inconspicuous, 1-2 mm. long, limb ca. 1.5 mm. broad; fruiting calyx 2.5-4 mm. long, usually about as broad as long, laterally compressed, ascending; pedicels 0.5-1 mm. long; mature calyx-lobes lanceolate, somewhat connivent, not greatly surpassing the nutlets, midrib weakly thickened and on the abaxial lobe sparsely hirsute, margins appressed-hispid; nutlets 4, homomorphous, smooth or very finely granulate, shiny, brownish to greenish, frequently mottled, 1.8-2.5 mm. long, ovate, obliquely compressed, back low- convex, margins rounded; groove evidently excentric, closed, simple or shortly and unequally forked at the base; gynobase short, stout, ca. 14 height of nutlets; style evidently surpassed by nutlets or rarely equalling them.—Pittonia i. 119 (1887); Macbr. Contr. Gray Herb. n. s. xlviii. 46 (1916). Krynitzkia affinis Gray, Proc. Am. Acad. xx 270 (1885). C. geminata Greene, |. c. C. confusa Rydb. Bull. Torr. Bot. Cl. xxxvi. 679 (1909). C. affinis. 102 JOHNSTON Washington and western Montana to Southern California, northern Nevada and southern Wvoming. Wasuineton: Falcon Valley, puketor 455 (G); open pve wood, Faleo Mey Suksdorf 179 and 3207 (UC); valley of Swauk Riv , Sharples 198 and 197 (G); Kamiach Butte, Piper 3092 (G); dry rocky hillsides, Blue Mts., 1897, “flract (G): dry low ground northwest of Chenowith, — dorf 2768 : G 286 (G); Payette Lake, 1899. (UC) ONTANA: Priests, 1894, Brandegee 9 per Madison Canyon, Yel Seiane Par alt., Rydberg & Bessey 4884 ( OTY ore Sa) Tau: City Creek G3 onfu : anyon near Salt Lake City, Jones 269 (G) and 107 (UC). Nevapa: log railroad north of Verdi, ae ~ alt., He _ 10873 in ’ Sg CALIFORNIA: near Shasta Springs, Heller 2 (G); moist open plac ar Durney’s Mill, north base of Mt. Eddy Hello. 13887 (G,UC); Metealf’s ny cotter tate rtheast base of Mt. y,1170m vat. Heller 1 ale i); — Bis gy Eastwood 737 (G); Susanville, 1891, Brandeg ee (UC);m ws 3 km. south of Red Clover Valley, Heller & Kennedy 9 (G); Piaitie 1892, Eranieaes (UC); Liga pe Heller & Rasiedy Fe (eae "edge of forest about Chico Meadows, 1200 m alt., Heller 11489 (G, UC); Sutton House, Butte Co., Curran (UC); * ave yon 1c ‘ Truckee, Sonne (UC); Truckee, 1885, gf Area (UC); lower end of Donner Lake, Heller O48 G); ane Mt., Lake CoC C. F. L. (UC); near summit of Mt. Sanhedrin, , Heller 5882 (G); near Cape Horn, K. Brandegee (UG); Cathedral reat a Me Tallac, bong m. alt., Smiley 219 (G); Sly Park in sierran foothills of El] Dorado Co., 1020 m. alt.. = 11276 (G, UC); Lake Valley, 1908, K. Brand ieee Awe Silver Lake, 1 882, Hansen (UC); Bear Valley, Calaveras Co., m. alt. Hansen 516 {G: UC); Yosemite Valley, Abrams 4379 (G); Pian Creek, Yosemite Park, 2190 m. alt., Hall 9172 (G); eres Falls, Yosemite Park, Hall 8925 (UC); summit of Mt. Surprise, es oe (UC); Nerth Fork of San Joaquin, Madera Co., 1895, Hall & Babcock 5632 (UC); Ellis Meadows, 1914, K. Brandegee (G); Sequoia National Forest, Davidson 1820 (G); Sequoia Mills, 1892, Eastwood (UC); General Grant Big Tree Grove, 1892, Brandegee ryih ee pe of . ° ea brea Flats, San Bernardino Mts., 2400 m rawford 934 in pt. Middle Peak, rg a Mts., Abrams 3856 Gy: prey California. 1888" Palmer 183 (G). Especially characteristic of this plant are its obliquely compressed nutlets and the resulting excentrie position of the groove. The nut- lets appear to be compressed by a force perpendicular to the axial and abaxial faces of the stout gynobase. Greene, I. c., has remarked on this development as exemplified in the type of the synonymous U. aera _ 57, ©. glomerifiora Greene. Small very slender herb 1-S(-10) em. high; stems simple or with more or less numerous ascending re- THE NORTH AMERICAN SPECIES OF CRYPTANTHA 103 branched basal branches, finely and very Appressed|y short-hispid or Lally or lanceolate, 5-10 mm. ‘long, 0.8- 1.3 mm. broad, appressed short- hispid, minutely pustulate; flowers axillary, frequently borne along short branchlets and glomerate, distributed over nearly all parts of the plant; corolla inconspicuous, tubular and just surpassing the calyx, 1.3-1.8 mm. long, ca. 1 mm. broad, lobes broadly orbicular and erect, appendages trapeziform; fruiting calyx obovate, 2-2.5 mm. long, ascending, asymmetrical, united for 14-4 its length, base conical and somewhat siliceous, sessile, deciduous, not at all biserial; mature calyx-lobes lanceolate, erect or slightly connivent, midrib thickened and hirsute especially on abaxial lobe, margin short-hispid or hispid-strigose; ovules 4; nutlet 1, ovate, acute, 1.5-2 mm. long, next the abaxial calyx-lobe, asymmetrical, persistent, smooth and shiny, usually greenish, just surpassing or about equalling the axial calyx-lobes, back and sides rounded, face somewhat flattened; groove slightly off-center, closed except just above very base of nutlet where opening abruptly into an irregular roughened areola; gynobase very reduced, ca. 1/10 height of nutlet; style very short, not much surpass- ing the aborted nutlets, surpassed by the mature nutlet by over 1 mm. —Pittonia i. 116 (1887). Central Sierra Nevada of California. Rare. 5 eel RNIA: borders of ice-ponds below Truckee, 1887, Sonne (UC, sil ; dry gravelly pine forest, Tuolumne Meadows, 2600 m. alt., War stone () @ Volesas Creek, Upper Kern River, 2400 m. alt., Hall & Dabeork A very rare and distinct species with a habit recalling Plagiobothrys § Allocarya. It is very well marked in having asymmetrical solitary nutlets that about equal the calyx-lobes, and in having decidedly axil- lary glomerate flowers with no suggestion of a biserial arrangement. The nutlets are suprabasal in attachment. EXxcLUDED OR UNIDENTIFIED SPECIES. CrYPTANTHA BArTOLOMAEI Greene, Pittonia ii. 232 (1892). “ Aspect, pubescence and inflorescence of C. utahensis, but the minute (14 line long) ovate-lanceolate nutlets (4 and consimilar) perfectly smooth and lucid, and without margin; the ventral groove shortly bifurcate at base, but closed throughout. Bay of San Bartolomé, Lower California, Lieut. Pond, 1889. A connecting link between the orygona and leiocarpa groups in the genus.”—The identity of this species is wholly obscure. 104 JOHNSTON CRYPTANTHA FALLAX Greene, Pittonia v. 54 (1902). “With much the aspect of a slender C. muriculata, less than a foot high, the branches with scattered oblong-linear obtuse foliage and terminating in 3 divergent short and rather dense spikes; stem sparsely villous-hirsute, the foliage as sparingly somewhat strigose, the calyxes with a dense pubescence under the rather few very hispid spreading hairs: calyx- segments lanceolate, short, the tips not attenuate; corolla minute; nutlets ovate-lanceolate from a truncate base, traversed ventrally by a closed scar, this shortly forked at base, the whole surface grayish, smooth and shining, the back but slightly convex and distinctly sharp- edged. A very remarkable species in the almost wing-margined character of the smooth nutlets; for the plant has the aspect of that group of species whose nutlets are obtuse all around, and muricate. The only specimen known was collected by myself in the mountains above Tehachapi, California, 22 June, 1889, and was mixed with my duplicates of C. muriculata until now.”—Perhaps conspecific with C’. mohavensis. ; Cryptantha heliotropoides Logs. in Fedde, Repert. xii. 243 (1913) = ANTIPHYTUM HELIOTROPIOIDES A. DC. Prodr. x. 122 (1 1846). Cryptantha T orreyt Rydb. Mem. N. Y. Bot. Gard. i. 331 (1900).— Although taken up in place of C. Torreyana (Gray) Greene, this name is based on Eritrichium Torreyi Gray and so falls into the synonymy of PLacgiopoturys Torreyi Gray, Proc. Am. Acad. xx. 284 (1885). Kryniizkia californica Gray, Proc. Am. Acad. xx. 266 (1885) = PLAGIOBOTHRYS TRACHYCARPUS (Gray) Johnston, Contr. Gray Herb. n. s. Ixviii. 78 (1923). Krynitzkia californica, var. subglochidiata Gray, Proc. Am. Acad. xx. 266 (1885) = PLacioporurys scopuLorum (Greene) Johnston, Contr. Gray Herb. n. s. Ixviii. 79 (1923). Krynitzkia Chorisiana Gray, Proc. Am. Acad. xx. 267 (1885) = PiaciopoTurys Cuoristanus (Cham.) Johnston, Contr. Gray Herb. n. s. Ixvili. 77 (1923). Krynitzkia Cooperi Gray, Proc. Am. Acad. xx. 267 (1885) = PiaciosoTurys Parisau Johnston, Contr. Gray Herb. n. s. Ixviii. 78 (1923). Krynitzkia depressa Jones, Contr. W. Bot. xiii. 5 (1910) = ORE- -OCARYA DEPRESSA (Jones) Macbr. Contr. Gray Herb. n. s. xlviii. 32 (1916). Krynitzkia echinoides Jones, Proc. Calif. Acad. Sci. ser. 2, v. 709 (1895) = OrEocaRYA ECHINOWES (Jones) Macbr., i. e. plant treated _ by Maebride, Contr. Gray Herb. n. s. xIviii. 36 (1916), as O. ful- -- vocanescens. THE NORTH AMERICAN SPECIES OF CRYPTANTHA 105 Krynitzkia floribunda Gray, Proc. Am. Acad. xx. 265 (1885) = ANTIPHYTUM FLORIBUNDUM (Torr.) Gray, Proc. Am. Acad. x. 55 (1875). Krynitzkia fulvocanescens Gray, Proc. Am. Acad. xx. 280 (1885) = OREOCARYA FULVOCANESCENS (Gray) Greene, i. e. 0. echinoides of Macbride, Contr. Gray Herb. n. s. xlviii. 31 (1916) Krynitzkia fulvocanescens, var. idahoensis Jones, Contr. W. Bot. xill. 6 (1910) = OrEocaRYA sp. Krynitzkia glomerata Gray, Proc. Am. Acad. xx. 279 (1885) = OREOCARYA GLOMERATA (Pursh) Greene, Pittonia i. 58 (1887). Krynitzkia glomerata, var. acuta Jones, Zoe ii. 250 (1891) = OreocaryA WETHERILLU Eastw. Bull. Torr. Bot. Cl. xxx. 242 (1903). Krynitzkia glomerata, var. virginensis Jones, Contr. W. Bot. xiii. 5 (1910) = OrREOCARYA VIRGINENSIS (Jones) Macbr. Proc. Am. Acad. li. 547 (1916). Krynitzkia heliotropioides Gray, Proc. Am. Acad. xx. 265 (1885) = ANTIPHYTUM HELIOTROPIOIDES A. DC. Prodr. x. 122 (1846). Krynitzkia Jamesii Gray, Proc. Am. Acad. xx. 278 (1885) = OREOCARYA SUFFRUTICOSA (Torr.) Greenin, Pittonia i. 57 (1887). Krynitzkia Kingii Wats. acc. Hillman, Nevada Agric. Exper. Sta., Bull. xxiv. 71 (1895) = Puaciopoturys Kine (Wats.) Gray, Proc. Am. Acad. xx. 281 (1885). Krynitzkia leucophaea Gray, Proc. Am. Acad. xx. 280 ae OREOCARYA LEUCOPHAEA (Dougl.) Greene, Pittonia i. 58 (188 Krynilzkia leucophaea, var. alata Jones, Proc. Calif. Acad. aH ser. 2, v. 710 (1895) = OrEOCARYA CONFERTIFLORA Greene, Pittonia iii. 112 (1896). Krynitzkia lithocarya Greene in Gray, Proc. Am. Acad. xx. 265 (1885) = PLAGIOBOTHRYS LITHOCARYUS (Greene) Johnston, Contr. Gray Herb. n. s. xviii. 76 (1923). Krynitzkia mensana Jones, Contr. W. Bot. xiii. 4 (1910) = OrEo- CARYA EULOoPHUS Rydb. Bull. Torr. Bot. Cl. xxxi. 637 (1905). Krynitzkia mollis Gray, Proc. Am. Acad. xx. 267 (1885) = Puact- OBOTHRYS MOLLIS (Gray) Johnston, Contr. Gray Herb. n. s. Ixviii. 74 (1923). < Krynitzkia multicaulis, var. abortiva Jones, Contr. W. Bot. xiii. 5 (1910) = OREOCARYA SUFFRUTICOSA, var. ABORTIVA (Jones) Macbr. Proc. Am. Acad. li. 547 (1916). a Krynitzkia multicaulis, var. setosa Jones, Contr. W. Bot. xiii. 4 (1910) = OrEocarya s Krynitzkia oblata jae Contr. W. Bot. xiii. 4 (1910) = OrEocarya OBLATA (Jones) Macbr. Proc. Am. Acad. li. 548 (1916). 106 JOHNSTON Krynitzkia Palmeri Gray, Proc. Am. Acad. xx. 278 (1885) = OreocaryA Patmert (Gray) Greene, Pittonia i. 57 (1887). Krynitzkia Parryi Gray, Proc. Am. Acad. xx. 265 (1885) = AN- TIPHYTUM Parry! Wats. Proc. Am. Acad. xviii. 122 (1883). Krynitzkia peninsularis Rose, Contr. U.S. Nat. Herb. i. 85 (1890) = ANTIPHYTUM FENINSULARE (Rose) Johnston, Contr. Gray Herb. n. s. Ixviii. 51 (1923) Krynitzkia plebeia Gray, Proc. Am. Acad. xx. 266 (1885) = PLAGIOBOTHRYS PLEBEJUS (Cham.) Johnston, Contr. Gray Herb. n. s. Ixviii. 77 (1923) Krynitzkia pustulata Blankinship, Mont. Agric. Coll. Sci. Studies, Bot. i. 96 (1905) = OrEocARYA AFFINIS Greene, Pittonia iii. 110 (1896). Krynitzkia Scoulert Gray, Proc. Am. Acad. xx. 267 (1885) = Praciosoturys Scoutert (H. & A.) Johnston, Contr. Gray Herb. n.s Ixviii. 75 (1923). Krynitzkia sericea Gray, Proc. Am. Acad. xx. ae (1885) = OREOCARYA SERICEA (Gray) Greene, Pittonia i. 58 (18 Krynitzkia sericea, var. fulvocanescens Jones, Proc. aah oad Sci. ser. 2, v. 710 (1895) = OREOCARYA FULVOCANESCENS (Gray) Greene, Pittonia i. 58 (1887). Krynitzkia setosissima Gray, Proc. Am. Acad. xx. 276 fase = OREOCARYA SETOSISSIMA (Gray) Greene, Pittonia i. 58 (188 Krynitzkia trachycarpa Gray, Proc. Am. Acad. xx. 266 ace PLAGIOBOTHRYS TRACHYCARPUS (Gray) Johnston, Contr. Gray Herb. n. s. Ixviii. 78 (1923). Krynitzkia virgata Gray, Proc. Am. Acad. xx. 279 (1885) = OREOCARYA VIRGATA (Porter) Greene, Pittonia i. 58 (1887). Pretrocatyx Moore Oliver in Benth. Fl. Austral. v. 292 (1870).— A shrubby Australian plant belonging to the Monimiaceae, cf. Perkins & Gilg, Pflanzenr. iv. Fam. 101, 22-23 (1901). INDEX TO NUMBERED EXSICCATAE. The figures enclosed in pcan refer to the numbering given to the species in this paper Abrams, L. R. 2512 intermedia (28); 2555 muricata v. genuin (32a); 2578 simulans (39); 2590 muricata v. denticulata (32¢) ; 2904 290 : a _ microstachys 4 18 muric v. nesii (2b); 3556 ialennaits (28): 3594 micranth av (7b); 3659 circu ssa v. genuina (10a); 3824 flaccida ( (54): 3856 afinis ~ 4379 inell s (41, 4471 nia xin 7 39); - 4560 ‘cecude (54); 4759 echinella rams, L. R., & McGregor, E. A. 94 muricata Vv. genuina (32a); 173 decipiens v. corollata (25b); 211 Clevela ndi v. florosa (50a); 343 mu av. genuina (32a): 474 pohaveas (42). ogg bee . 18 micrantha v. genui- (/a). Andacs rson, C. L Torreyana v. genuina (38 a); 163 circumscissa v. genuina (10a); 165 pterocarya v. genuina (19a). A. W. 204 ee media (28): 238 foliosa “27; 2 276 maritima v. 347 Grayi v _ eryptochaeta (9e). Applegate, E. ei 70 ambigua (40). Austin, R. M 624 os (54). Austin, R. M., & Bruce, C. 2267 "ambigua (40). C. F. 75 crassisepala (24); 468 flaccida (54); 780 Fe ndleri ; a ca Vv. 2810 microstachys (48); D886 cavata (35); 2966 Chevilatadi: . florosa (50b); 3003 flaccida (54); 4137, 4744, 4769, 4773 intermedia (28); 4779 4780 microstachys (4 Baker, S. 22 Clevelandi v. florosa Ob); 620 Torreyana v salinteins (388b Barber, 48 intermedia (28); 116 Clevelandi Vv. ag Ob). Bessey, Bolander, H. N. orga (38¢); 3 16 Brandeg E. 6 Kelseyana (23)- 30 P Wetec aby. 36 Kelsey- ana ( randegee, genuina (30a yn 14 Brandegee, T. 8. 406 minima (23); 995 pterocarya v. genuina (19a); 16 gg oa tee 1646 muri- cata v. Jones Brewer, W. H. 232 muricata v. genuina Sg 278 Clevelandi v. florosa (50b); 1129 flaccida (54); 6284 echinella (41). Butler, G. D. 733 iy ages (36); 734 flaccida (54); 7 157 nevadensis v. 158 dumetcrum Orr yana Vv. genuina (38a); 1416 Heidences (36). Chandler, H. P. 937 flaccida (54); 1302 Hendersoni (36); 1303 Torrey- ana Vv. genuina (38a); 1304 flaccida GM; 5355 muricata v. Jonesii (3 b). Clark, J. A. 165 Torreyana v. 108 don, J. W. i 41 muri a: — (32 b); 4 a s (15 4, 45 flaccida (54); 50 He pderaon! rreyana v. genuina (38a); 3550 Hendersoni ( Coulter, T. 500 5 oo 2a Eo Sea, ere eee ee Var. glauca. c, Boielets 4.5-7 mm. long: panicle soenly contracted, occasionally diffuse, during anthesis........-..----- Var. littoralis. Var. Genuina Gren. & Godr. Fl. de a iii. 507 (1855). Aig cespitosa L. Sp. Pl. i. 64. (1753). A. altissima Moench, Meth. 1 7 . A, cespitosa genuina Reichenb. Ic. Fl. Germ. i. t. X (1834). A. cespitosa b) pe ne a (Moench) Aschers. Fl. Brand. i. 833 (1864).—Newfoundland; Montana to southern, British Columbia, margin of Junction Pond, Whitbourne, Fernald & Long, no. 26,282; gravelly Muehers along Harry’ s River, Fernald & Wiegand, no. 2589. Montana: Bozeman, Rydberg, no. 2219; wet meadow, Sheep Creek, Rydberg, no. 3301. WyoMInG: Laramie, Merrill, no. 25. Kennedy, no. 8790; Mt. Eddy, Copeland, no. 3798. OREGON: wet soil in shade of Coniferae, Big Meadows, Des Chutes River, Leztberg, no. 523; eee Prairie, Griffith & Hunter, no. 76; ditch, Salem, Nelson, 1354. British CoLuMBia: summit of Selkirk Mts., J. SP bth’ Chilliwack Valley, J. M. Macoun, no. 26,084. : Var. ae hry (Thuill.) Neos Pl. Eu. i. 56 (1890). Aira parviflora Thuill. Fl. Par. ed. 2 1799). 2A. cespitosa Q. vires- Wimm. & i Fl. Sil. i. 60 (1827).—Locally in New England where introduced from Europe. Marne: damp hollow, w an ro vigghe along shore, Islesboro, Woodward, Biss Isl sell & Fe no. MASSACHUSETTS: edge of field, Washington, July 17, 1919, iene ConnecticuT: open swamp, very abun dant, Franklin, July 8, 1906, Woodward; introduced in grassland, Southington, June 21, 1903, Bissell; spreading in oe shaded dooryard, South Windsor, Weatherby, no . 4945. 154 Rhodora [AuGUST Var. GLauca (Hartm.) Lindm. fil. Svensk Fanerogamfl. 81 (1918). Aira ambigua Michx. Fl. Bor.-Am. i. 61 (1803). D. glawea Hartm. Handb. Skand. Fl. 448 (1820). A. aristulata Torr. Fl. N. Mid. U. S. i. 132 (1824). A. cespitosa 8. glauca Hartm. |. c. ed. 2: 25 (1832). A.c pains var. minor Lange, Fl. Dan. xvii. fase. L 4, t. MMDCCCC- XLV, fig. 1 (188 Rien: lake-shores and damp, chiefly calcareous, soil, Newfoundland to Yukon, south to northern New sal Pennsylvania, Ohio, Illinois, Arizona, and California; Faroe Islan Seandinavia. The following, selected from a large series of Pe ns, are representative. NEWFOUNDLAND: Roc larsh, Flower Coxe. Fernald, g & Dunbar, no. 26,279. QUEBEC: “circa lacus Mistassins et juxta amnes in lacum s. Joannis defluentes, “ Ab Arra cespitosa foliis tantum differt etiamque dimidio minor est” Ile-aux-Couleuvres, Lac St-Jean, Victorin, no. iat 258; Hoaia vente River, August, 1904, Collins, Fernald & Pease; St. Lawrence River below Quebec, July 9, 1905, Churchill. New ‘ Brunswick: Wood- stock, Fernald & Long, n no. 12,658; Westfield, Fassett, no. 2183. ova Sco ag? Lisi ieemyes Grand Lake, F srnald, Bartram & Long, no. 23,285; r Lake, Fernald, Bissell, Pease, Long & Linder, no. 19,946; sealers Bissell & Pesie; no. 19 944. Maine: Fort Fairfield, F ernald, no. 183; Milford, Fernald, no. 12, pet Winn, Fernald g, no. 12,657" Pembroke, Fe rnald, no. ; West Dresden, Fassett, no. 1045. New HAMPSHIRE: Lake oe bagog, 1, Pease, no. 10,515; Dalton, Pies no. 17,387; Summer’s Falls, June 12, 1897, Walaa. VERMONT: Lyndon, June, 1871, Congdon; Burlington, Jane 18, 1877, Pringle; Brattleboro, July 8, 1895, Grout. SSACHUSETTS: Merrimack R., Lowell, July 20, i Swan. Connnericvr: ae te R., East Haddam, Weatherby, no. 4295. New Y er Lake, Spencer, Metcalf, no. 5698; * Conduncl Wiegand, no. 3526: Dae, — no. 6521. New Jersey: White Pond, Sussex Co., Mackenzie, no. 4648. PENNSYLVANIA: ae Grove, Heller, no. 4819. Chinas: Belleville, Macoun, 2239; ingham, June 17, 1892, Morton. Mucnican: Isle nt neg Cooper, no. 164. Wisconsin: Jackson Harbor, June 21, 1896, Schue Minnesota: Thompson, Carlton Co., Sandberg, no. 395. "NaNtTons Lake Winnipeg Valley, 1857, Bourgeau. { peace Jaw, — no. 13,080. Atperta: Red Deer Valley near S ipanale Moodie, 1042; near Banff, July 4, 1891, Macoun. MonrTANA: Forks of ‘the Madison, R ydberg & Bessey, no. 3588. Ipauo: Victor, Merrill & Wilcox, no. 185. Wyomine: Laramie River, Albany Co., Nelson, no. 430; Waahvietor Ranch, Merrill & Wilcox, no. 78. CoLo- RADO: Gunnison, Baker, nos. 530, 553; Trout Creek Pass, July 18, AD Trees, Clavortia Ca. “Hillebrand. no. 2246; near 1926] Fernald,—Botanizing in Newfoundland 155 Donner Lake, Torrey, no. 558. Orgecon: Wallowa Mts., Cusick, no. 3126; Lower Albina, Portland, Sheldon, no. 11,125. WaASHING- TON: Calispel Valley, Kreager, no. 328. YUKON: Carcross, East- wood, no. 697e. The identity of our plant with the Scandinavian var. glauca is definitely indicated by beautiful material collected in Torne Lapmark by Professor Alm. This is clearly matched by many American speci- mens, especially such small plants as F ernald, Long & Dunbar, no. 26,279 from Newfoundland; Fernald & Collins material from Gaspé Co., Quebec; Eggleston, no. 1 from Summer’s Falls, New Hampshire; Sandberg, no. 395 from Minnesota; Tweedy, no. 616 from Yellow- stone Park; Var. tirroratis (Reut.) Richter, Pl. Eu. i. 56 (1890). Aira cespitosa, var. @. littoralis Reuter, Cat. Pl. Vase. Genéve. 116 (1832). A. littoralis (Reut.) Godet, Fl. Jura, 803 (1852). cespitosa, V alpina Vasey in Beal, Grasses N. A. ii. 368 (1896). alpicola Ste. Anne de Beaupré, Macoun, no. 69,230. ALBERTA: mountains north of Cavell Creek, J. M. Macoun, no. 98,349; south of Wilcox ass, Brown, no. 1394; Lake Louise, Brown, no. 549. IpaHo: Soda (To be continued.) 1926] Fernald,—Botanizing in Newfoundland 161 TWO SUMMERS OF BOTANIZING IN NEWFOUNDLAND. M. L. Fernald. (Continued from p. 155.) DESCHAMPSIA ATROPURPUREA (Wahlenb.) Scheele. Wet quartzite rocks and gravel by brooks, gulches of Bard Harbor Hill, Highlands of St. John, the first stations in Newfoundland: upper Deer Pond Brook, Fernald, & Long, no. 27,462; Southwest Gulch, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 27,463. p- AGROSTIS CANINA L. Moss er silicious rocks along rill, slope of South Hill, St. John’ s Fernald, Long & Dunbar, no. 26, 257; margin of cold brook i in spruce thicket, Bay Bulls, no. 26,258; gravelly — of Goose Pond, Whitbourne, no. 26,259; dry gneiss crests an ges and peaty barrens, Port aux Basques, nos. 6, 260, 26,262. ‘Older collections indicate that the species is in of southern SS See p digenous across the breadth p. 50, 56, 81, 85. A Sceribn. Dry carci limestone barrens, St. John’s A. PALUDO Island, Ferciald; Wiegand, Long, Gilbert & Hotchkiss, no. "97, 479, the t Newfoundland station. Described from Blanc Sablon, Labrador but represented in the Gray Herbarium er —— Co., Quebec (St. ian AY wie Trin. Springy swales the Rock Marsh, — Cove, Fernald, m as far west as Ouapita- no. 90,123 as A. borealis }: and limy bog-barrens near Long & Dunbar, nos. ro 253, 26,280, Fern ald, P e & Long, no. 27, 481, Fernald, un dripping quartzite cliffs and ledges, upper Deer Pond pack. igh. 162 Rhodora [SEPTEMBER lands of St. John, Fernald & Long, bs 27,482; the only stations nown in eastern America. See pp. -CALAMAGROSTIS PicKERINGII Gray. "Wet boggy tundra in central valley of Quirpon Island, Fernald & ra no. 27,484, the first from north of Notre Dame Bay at Bonne C. PICKERINGII, var. DEBILIS (feet, Fernald & Wiegand. Boggy barrens (“mesh”’) ances of Ship Cove, Fernald, Wiegand & Long, no. 27,485, the first from north of Notre Dame Bay and Bonne Ba AmMopPHILA BREVILIGULATA Fern. Dunes and sandy beach, Bard Harbor, St. John Bey: Gilbert & cee no. 27,504; : the jal stations known north of Port-a-Port. See p . AEQUALIS, var. NATANS (Wahlenb.) ai ; " Ruopora, XxVil 198 (1925). Pools in tundra, Boat Harbor, Fernald, Wiegand & Long, no. 27,505. See p. 107. Known — the arctic and sub- arctic regions only along the Straits of Belle MUHLENBERGIA RACEMOSA (Michx.) BSP. ‘Bo oggy meadow b-rd- ering pond, Bard Harbor, St. John Bay, abet no. 27,509, the first from north of Bay of Islands. See p. 127. Oryzopsis ASPERIFOLIA Michx. Knoll in spruce thicket, St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 27,510, the first from north of Bay of Islands. Mitium errusum L. Abundant in glades and gulches of Bard Harbor Hill, Highlands of St. John, the only Newfoundland region known for it except shores of Ingornachoix Bay. Se Pp. gee HrerocuLok oporata (L.) Wahlenb. Abundant in swales and tundra along the Straits. The previous Newfoundland collections belong to the seashore var. FRAGRANS (Willd.) Richter EriopHorum Scueucuzert Hoppe. Wet swales and crag of ponds, heen Cove, Fernald, Long & Dunbar, no. 26,334, d Pease & Long, nos. - on 27,444, the first authentic ica in Newfoundland. See p E. Cuamissonis C oA. tes: ., Var. AQUATILE (Norman) Fernald, Ruopora, xxvii. 207 (1925). Shallow pool at base of Cape Dégrat, Quirpon Island, Fernald & Long, no. 27,545, the only known American station. See p. 120. ry opacum (Bjérnstr.) Fernald. Abundant in limy bog-barrens d tundra along the Straits of Belle Isle. p. 53, or CALLITRIX Cham. See Fernald, Raopora, xxvii. 205 (1925). Wet peaty limestone barrens and tundra along the Straits of Belle Isle, the only region known for the species except St. Lawrence Bay on the western side of f Bering Straits. See pp. 98, 105. 1926] Fernald,—Botanizing in Newfoundland 163 Ryncnospora atpa.(L.) Vahl. Boggy meadow bordering pond, Bard Harbor, St. John Bay, Hotchkiss, no. 27,570, the first from north of Bonne Bay. See p. 127. Carex capirata L. Peaty margins of pools in limestone barrens back of Big Brook, Fernald & Long, no. 27,575; peaty margins of pools in limestone barrens on the Highlands northeast of Big Brook, Pease & Griscom, no. 27,576; dryish tundra, Schooner (Brandy ) Island, Pease & Long, no. 27, 577; the first stations in Newlodhilland. See pp. 98, 99. C. incurva Lightf. A characteristic turf-plant on damp limestone along the Straits, eastward to Eddies Cove, southwestward to Dead- man’s Cove. See. pp. 61, 108. CHORDORHIZA Ehrh.; * var. SPHAGNOPHILA Laestad. Springy swales and quaking bogs is the limestone barrens, Flower Cove, Fernald, Long & Dunbar, nos. 26,382, 26,383; swales on limestone barrens, Sandy (Poverty) Cove, Fernald ilbert, no. 27,588; pee swales, Eddies Cove, Fernald, Wiegand & Long, no. 27,587; the first American records for the variety. c STERILIS Willd. ramet y in peat overlying limestone, north to the Straits. See p C. Dewryana Sian. Peaty borders of spruce thickets near Mistaken Cove, Fernald & Griscom, no. 27,601; glades and openings in spruce thickets east of Big Brook, Fernald, Wiegan d & Hotchkiss, no. 27,602; meadow below limestone escarpment, western face of Bard Harbor Hill, Highlands of St. John, Fernald & Long, no. 27,603; the first authentic collections from Newfoundland. See p. 100. C. TRIsPperMa Dew., var. Bruincsu Knight. Wet ema Trepassey, Fernald, Long & Dunbar, no. 26,376. See p The ae — Newfoundland collection is from the East A sate of the C. prpartira All. C. Lachenalii Schkuhr. See Mackenzie, Bull. Torr. Bot. Cl. 1. 348 (1923). Brookside on slaty hills east of Little Quirpon, Fernald & Gilbert, no. 27,622; wet mossy quartzite rocks along Mans Humbug Brook, Highlands of St. icine a me & Long, no. 27,617; new to Newfoundland. See pp C. GuarEosa Wahlenb. Tuirfy slopes of Site hills, Little ~~ Fernald & Long, no. 27,618, new to Newfoundland. See p ir. True C. brun 1S var. SPHAEROSTACHYA (Tuckerm.) Kiikenthal. It abounds in thickets and woodlands and extends well to the south in the eastern 164 Rhodora [SEPTEMBER stachya ee) Dew. Am. Journ. Sci. xlix. 44, t. Ee, fig. 110 1845). C. witilis b. sylvatica Meinsh. Fl. Ingr. 402 (1878). C. canescens, var. vulgaris Bailey, Bot. Gaz. xiii. 86 (1888). C. brun- nescens gracilior Britton in Britt. & Brown, Ill. Fl. i. 351 (1896). C, RUPEsTRIs All. wig on Sratne, western Newfoundland, south bes Bay St. George. See C. NOVAE-ANGLIAE Schwein. ate and silicious rocky slope of Joan Plains Hill, Bay Bulls, Poteait Lon aan Dunbar, no. 26,417, the first from the 5 pee eae See ee: C. UMBELLATA Schku Barren silicious oon of Joan Plains - Hill, Bay Bulls, Fernald, ge & Dunbar, no. 26,414; seen but not collected ~ Trepassey : the first statins on the ‘Avalon Peninsula. See p C. 6 NCULATA Muhl. Woods ae thickets on Bard Harbor Hill, Highlands of St. John and on St. John’s Island, Fernald, Wieg- and, Long, Gilbert & Hotchkiss, nos. 27,647, 27,648; extensions north- s. C. conctnna R. Br. Shaded limestone ledges and escarpments, northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert Hotchkiss, no. 27,649, the only Ne wfoundland station except Mackenzie & Griscom’s on Cape St. George. See pp. 101, 102. ACIALIS Mackenzie. Dominant on limestone gravel in west- ern Mdoumiant from Quirpon to Bay St. George. See pp. 62, 79, 103, — C. EBuRNEA Boott. Shaded limestone ledges and escarpments, oben half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 27,659; mossy coniferous woods, lower south- western slope of Bard Harbor Hill, Highlands of St. John, Fernald, Wiegand, wie Gilbert & Hotchkiss, no. 27,660; extensions north from ae of Islands. » «. Lor All. Wet depressions and margins of rills and_ pools mig stow ne barrens, generally distributed from Pistolet Bay to Ingomacis eth , ~ — known region for the species in America. . 76, 91, LIVIDA (Wahtenb,) Willd. See Fernald, Ruopora, xxviii. 8 (1926). Turfy limestone barrens, northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no- Sa ,672, the only known station outside northern Europe and Alberta. of oe var. RUFINAEFORMIS Fernald, |. ¢. (1926). Wet muddy hollow in in seal gravel-barren, Four-Mile Cove, Fernald, Wiegand hoe no. 27,637 (TYPE of the var.). See p. 108. CAREX q! n.hybr. (C. atratiformis x. Haller), laxe iat rae pais 2-6 a altis rigidulis superne acute angulatis seabrisque basi vaginis aphyllis atropurpureis deinde fibrillosis obtectis; foliis culmo valde brevioribus 2-4 mm. latis laete viridibus; 1926] Fernald,—Botanizing in Newfoundland 165 spicis 2-4 ellipsoideis 0.7-1.3 cm. longis 4—6 mm. crassis, lateralibus breviter pedunculatis subapproximatis vel imo remotiusculis adscen dentibus, terminalibus majoribus gynaecandris; squamis late ovatis albo-hyalinis; perigyniis vacuis ovatis planis 3 mm. longis dense papillosis enerviis stramineis apice purpurascentibus.—NEWFOUND- N ndant as scattered clumps on the turfy and rocky western slope of Cans Dégrat, Quirpon Island, August 7, 1925, Fernald & Long, no. 27,696 (ryPE in Gray Herb.) Although abundant over a large area on the eastern half of Quirpon Island (See p. 121), Carex quirponensis is here described as a hybrid because in all the scores of clumps we examined no good achenes could be found and because it is an exact blend of C. atratiformis Britton and C. Halleri Gunn. (C. alpina Sw.), both of which are dominant and very fertile species from Cape Onion to Cape Bauld and Cape Dégrat. A very similar hybrid of C. atrata and C. Halleri (X C. Candriani pecsnataee is known in Europe. C. Hattert Gunn. Dominant in turf or on crests f slate or trap lis, panic end of the Straits Coast: Little Quirpon, Fe ; 7,697; Quirpon Island, Wiegand, Gilbert & ‘Hotehkiss, no. 27, 698: Mais Bay, Fernald, Wiegand, Long, Gilbert & Hotch- kiss, no. 27, 701; Cape Onion, Fernald, no. 27,700; Anse aux Sauvages, Fernald, Ws iegand & Long, no. 27,699; new to Newfoundland. See pp. 121, C. vie C. A. Meyer. Springy and seepy slopes in peaty barrens among the gneiss hills, Port aux Basques, Fe , Long & quartzite gravel or in turf, summit-barrens of Bard Harbor Hill, Highlands of St. John, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 27,702, Ses mica Gilbert & Hotchkiss, no. 27,703, Fernald & Long, no. 27,704; mossy brooksides and turfy slopes, Sacred Island, ~ Wiegand, Gilbert & Hotchkiss, no. 27,707, Fernald & Long, no. 27,708; turfy and rocky slopes of Cape Dégrat, Quirpon Island, Fernald d& Long, no. va 705. See pp. 56, 110, 121. C. niga Good. Al though common on the bare crests of south- western Newfoundland C. rigida seems to be rare north of Bonne Bay, a single station: turfy and mossy quartzite rocks along Mans eae Brook, Highlands a St: Soha, Fernald & Long, no. 27,711. 56. Pp C. LenticuLarts Michx., var. ALBI-MONTANA Dewey. Wet quartzite rocks and gravel along brook, Southwest Gulch, Highlands of St. John, Fernald, Wiegand, — Gilbert & Hotchkiss, no. 27,719, new to Newfoundland. p- (C. satina Wahlenb. The aan typical form of the species, with short scales is common on saline shores along the Straits. 166 Rhodora [SEPTEMBER C. SALINA, var. PSEUDOFILIPENDULA (Blytt) Kiikenth. As noted on p. ere this variety occurs on the Labrador side of the Straits and on the Gaspé coast. A gp ie from the boggy barren south of Ship Cove “(R ernald, Wiegand & Long, no. 27,732) is an exact match for Scandinavian material cited by Kiikenthal as C. aquatilis X C. salina, var. pseudofilipendula, so that the presence in the region of var. peeudofilipendula may be inferred. C. Hostrana DC., var. LAURENTIANA Fern. & Wieg. Glades in spruce ane Brig ‘Bay, Fernald, Long 2 a no. 26,445, an Soh acca been from Bay of Islands. See 1 Retz., var. SUBGLOBOSA (Melick, Richter. In clay mixed as Tideatene gravel, St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 97, 733. This and material from Pointe scamege (Fernald & Wiegand, no. 2912) is all we know from America. p. 119. C. LEPrpocARPA Tausch. Wet a soil on the West Coast, northward to the Straits. See p. 118. MICROGLOCHIN Wahlen Dominant on peaty limestone barrens and in =~ tundra, from Pistolet Bay to St. John Bay. See pp. 53, 61, 118 VESICARIA L., var. Raeana (Boott) Fern. Boggy barren ben mesh’ ”), south of Ship Cove, Fernald, Wiegand & Long, no. 27,758, the only preg station yet known except on the lower Humber. See p. 1 JUNCUS EFFUSUS <, var. PyLart (Laharpe) Fern. & Wieg. Bushy swale on flat north of Does Hill, St. John Bay, Fernald & Wiegand, no. 27,764, the first station north of Bay of Islands. See p. 127. J. BULBOSUS L. Cold brooks, serine ally margins of not sa “9 i en os Long & Dunbar, nos. 26,482, 26,483, and Bay sula. gos pp. 50 ), 81, 83. J. ALPINUS Vill., var. UNI-BICEPS Laestad. Wet depressions and borders of joes in limestone barrens and tundra , Cook Point, Fe st & Hotchkiss, no. 27,779; Burnt Cape, Fernald & Long, 27,777; St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotch- kiss, nos. 27,774, 27,776; the first records from America. J. ALBESCENS (Lange) Fernald. Dominant on wet calcareous barrens from Quirpon to Bay St. George. See pp. 62, 118. U: au. ward: , Flower Cove, Fernald, Long & py no. 26 ,498; pools a tundra, central valley of Quirpon Island, Fernald & Long, no. 27,792; the first stations north of Bonne Ba L. ously known only locally from the East Coast. See p. 110. L. campestris (L.) DC. 1926] Fernald,—Botanizing in Newfoundland 167 Dry spruce thickets, Curling, Fernald, Long & Dunbar, no. 26,499. All previous stations are on the East Coast. L. CAMPESTRIS, var. CONGESTA (Thuill.) Meyer. Wet spruce thickets among the gneiss hills back of Port aux Basques, Fernald, Long Dunbar, no. 26,500, the first from eastern America. 56 ALLIuM ScHOENOPRASUM L., var. laurentianum, n. var., a var. sibirico recedit floribus minoribus (8-10 mm. longis) intense coloratis; segmentis ovatis vel ovato-lanceolatis.—Near the Gulf of St. Law- rence, western Newfoundland, and locally across the continent. NEWFOUNDLAND: along Eddies Cove Brook, July 24, 1925, Griscom, no. 27,822; muddy depressions in gravelly limestone barrens one mile back of Savage Cove, July 14, 1925, Fernald & Long, no. 27,821; wet places, Flower Cove, July 23, 1921, M. E. Priest, no. T5; damp crevices and gravel of limestone, Rock Marsh, Flower Cove, July 30, 1924, Fernald, Long & Dunbar, no. 26,512; in clay mixed with lime- stone gravel, barrens, St. John’s Island, July 31, 1925, Fernald, Wieg- and, Long, Gilbert & Hotchkiss, no. 27,824 (type in Gray Herb.); turfy calcareous western slope under summit of Bard Harbor Hill, July 28, 1925, Wiegand, Gilbert & Hotchkiss, no. 27,823; dry lime- stone cliffs and talus, western face of Doctor Hill, August 24, 1925, Fernald & Long, no. 27,825; damp talus of limestone sea-cliffs, Pointe Riche, August 4, 1910, Fernald & Wiegand, no. 3069. QUEBEC: Restigouche River, Metapedia, July 23, 1904, A. S. Pease, no. 4448. EW Brunswick: in grass field, Casey’s Cape, Kent Co., July 9, 1914, F. 7. Hubbard. New Yor: bank of Black River, Watertown, July 3, 1857, Wm. Boott. Montana: Belton, Glacier National Park, September 17, 1921, S. D. McKeloy. Orecon: Wallawalla, Tolmie. Wasutncton: Marcus, August 18, 1902, Kreager, no. 459. In typical var. sibiricwm (A. sibiricum L.) the perianths are 10-14 mm. long, usually with more attenuate segments and commonly paler (though sometimes intensely colored); and in this variety the pedicels are commonly rather elongated so that the well developed flowering umbels are 3.5-5 cm. in diameter. In var. laurentianum the short perianths are intensely colored, the segments commonly less attenuate than in var. sibiricum and the pedicels mostly short, the well developed flowering umbels being usually 2.3-3 (rarely —3.3) cm. in diameter. Some specimens from eastern Asia (Ochotsk Sea, Small; Amur, Mazximowicz) apparently belong to var. laurentianum and in the note on p. 62 they were mistaken for true var. sibiricum. SrREPTOPUS OREOPOLUS Fernald, Ruopora, viii. 70 (1906). As noted on pp. 104, 116 and 124, this characteristic plant, heretofore wn only from the mountains of Gaspé and Matane Cos., Quebec, occurs about Pistolet Bay and abounds in the gulches of Bard Harbor 168 Rhodora [SEPTEMBER Hill. In the latter area it fruits heavily, its berries, previously unknown, being slightly mange as in S. amplezifolius (L.) DC. but of a Hoa cherry-red color as in S. roseus Michx. Ir1Is SETOSA Pallas, var. Hee ENSIS is Foster, forma pallidifiora, nit. Fe abeeentne sepalis plus minusve caeruleo-tinctis.— NEWFOUNDLAND: turf ore near Nameless Point, Flower Cove, August 2, 1925, Fernald & Long, no. 27,843 A peculiarly attractive color-form of the usually deep blue-violet northern Iris; the petals and stigmas are white and the large sepals also white except for a faint bluish his CYPRIPEDIUM PARVIFLORUM Salisb., var. planipetalum, n. var., humile, 0.8-2.3 dm. altum; foliis 2—4 ellipticis vel ovatis 3-9 cm. longis 1.7—4.5 cm. latis; sepalis viridibus vel stramineis plus minusve purpureotinctis, mpage 2 ovato 2.2-4 em. longo 0.8-2.2 cm. lato apice acuto vix acuminato basi rotundato, inferioribus connatis minoribus; petalis neues vel ge i oblongo-lanceolatis planis 2.5-4 cm. longis 0.4-1 . latis; lal ello dorso-ventraliter of Burnt Cape, Pistolet Bay, ok Be ye 1925, Fernald, Wiegand, Pease, g, Griscom, Gilbert & Hot chee. no. 27,855; southern half of Burnt aS. sere 3, 1925, Fernald & Long, ni 27,857; peaty slo f lim rrens on the Highlands coe of Big Brook, Straits of Belle ae, July 16, 1925, Fernald, Wiegand & Hotchkiss, no. 27,854; peaty and bushy areas on limestone barrens one mile bac of Savage Cove, Straits of Belle Isle, July 14, 1925, Fernald & Lon no. 27,851 (type in Gray Herb.), 27,852 ‘(aberrant individuals); mossy spruce thickets bordering limestone barrens near Ice oint, Sa Ba in Bay, July 14, 1925, Wiegand, Gilbert & Hotchkiss, no. 7,853; turfy limestone barron, St. John’s Island, St. John Bay, aly 31, 1925, Fernald et al. no. 27,856; Bay of Islands, July 5, 1906, CW: "Townsend; turfy sl seas of the marble region between Mt. L. 8. Sanford; in humus us or turf on the limestone tableland, alt. 200- m., Table Mountain, Port-d-Port Bay, July 16 & 17, 1914, ties & St. John, no. 10,815. ar. planipetalum is distinguished from both C. parviflorum and Seapine, var. pubescens (Willd.) Knight by its short and com- paratively broad, flat, usually purplish petals, by the relatively shorter and broader upper sepal with less acuminate or elongate tip and with rounded rather than abruptly narrowed or subcuneate bane, and by the cordate ar, no. 26,573; Bard Harbor, Fernald & Long, no. 28,000; Bard Harbor Hill, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 27,970; St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,001; Pointe Riche, Fernald & Wiegand, nos. 3202, 3205; Ingornachoix Bay, Fernald & Wiegand, nos. 3206, 3208, 3215-3217, 3299: Cape St. George, Mackenzie riscom, no. 11,043. QUEBEC: Fireok: St. John, no. 90,845; Caribou Island, Martin, no. 4 (cited by Rydberg as 8S. glauca); Archipel Ouapitagone, St. John, no. 90,842; Archipel de Mingan, St. John, nos. 90,835-90,840, 90,846, Victorni & Rolland, nos. 18,910, 18,911, 18,913, 18,916, 18,922-18,927, 18,929-18,931; Baie Sainte-Claire, Anticosti, Victorin, nos. 4349-4351; Table-top Mountains, Fernald Fe 25,685; Pease Basin, between Mts. Logan ‘and Pembroke, Fe Griscom & Mac kenzie, nos. "95, 673, 25,674, Pease & Smith, nos. 25,677-25,680; pass between Mts. Logan and Fortin, Fernald & Pease, no. 25, 008. Var. intonsa, n. var., frutex depressus vel erectus ad 1.7 m. altus; ramulis villosis; foliis maturis lanceolatis vel phy vicar vel ellipticis vel anguste ovatis basi apiceque acutis vel obtusis dense villosis 2-4.5 em. longis.—Lasprapor: 20 miles north of Nakvak Spee 28, 1908, H. S. Forbes, no. 100; Makkovik, Augus ge Stecker, no. 3; Turner’s Head, Hamilton Inlet, ah: 6, 1892, 1F oe no. 36 (the first-cited specimen be) . labradorica Rydb. and paca by at base contradicting Rydberg’ s description,—see discussion under var. typica); wet moss along spring-brooks on calcareous sandstone escarpment, Blanc Sablon, September 4, 1925, Fernald & Long, no. 28,031. NEWFOUNDLAND: spruce thickets at base of Yankee Point, July 12, 1925, Fernald & Griscom, no. 28,025; rocky meadows and 186 Rhodora [OcToBER brook-bottoms, upper Deer Pond Brook, Highlands of St. John, August 20, 1925, Fernald & Long, no. 28,030 (ryPE in Gray Herb.). QUEBEC: dry granite rock-slides at 950-1200 m. altitude, *outherh slope of Mt. McNab, Tabletop Mts., August 1, 1923, F ernald, Dodge Smith, no. 25,681. As already noted, the Turner’s Head specimen which Britton has designated as the “type” of Salix labradorica Rydb. belongs with S. cordifolia, var. intonsa. In its pubescence alone does it match Rydberg’s description and since both his key and his description emphasize the broadly ovate leaves of S. labradorica, it seems wholly unsafe to take up for a variety of S. cordifolia the misbegotten name S. labradorica. Similarly, I am quite unable to separate S. cordifolia, var. intonsa from some specimens from Greenland which Schneider places in his species, S. anamesa. But one has only to read the origi- nal discussion of S. anamesa Schneider, Bot. Gaz. Ixvi. 348 (1918) to see that it is scarcely separable from S. cordifolia “from which it chiefly differs by the presence of stomata in the upper leaf surface.” Schneider looks upon S. Waghornei Rydb. as a hybrid of S. cordifolia and S. anglorum, because it resembles the former but has stomata in the upper surface of the leaves (as in S. anglorum). This interpre- tation may be correct but not all shrubs of S. cordifolia with villous leaves and stomata in the upper leaf-surface can be regarded as hybrids of S. anglorum. For instance, Fernald & Long’s no. 28,031 from Blanc Sablon, Labrador has such stomata, yet Schneider cor- rectly states that the southern limit of S. anglorwm in Labrador is in latitude 55°—fully 250 miles north of Blanc Sablon. That the pres- ence of stomata is a specific character separating the Greenland shrub from the American, I am quite unwillung to admit; and Schnei- der’s memoranda on the Greenland sheets in the Gray Herbarium are specially significant: “cf. S. anamesa Schn.”; “S. anglorum Cham. f. satis sericeo-pilosa vel S. anamesae forma”; “f. incerta, prob. ad S. anamesam referenda”; “ab S. anglorum typica certe distinguere non possum (an S. anamesam accedens)”’; “S. anamesa Schn. vide- tur”; “f. mihi incerta, prob. ad S. anamesam Schn. referenda”; “5. anamesa m.”; “forma incerta. S. cordifolia v. atra vel S. anamesa Schn. viletice ” Until its own author can recognize S. anamesa its claims to specific rank are not likely to appeal to others. __ Var. Macountt (Rydb.) Schneider, Bot. Gaz. Ixvi. 347 (1918). S. Macounii Rydb. Bull. N. Y. Bot. Gard. i. 269 (1899). S. Ryd- = Heller, Cat. N. Am. Pl. ed. 2:4 (1900). . vaceiniformis Rydb. 1926] Fernald,—Botanizing in Newfoundland 187 in Britt. Man. Fl. N. St. and Can. 319 (1901).—Depressed or low shrub: branchlets villous: mature leaves elliptic, see F or narrowly obovate, acute or acutish, mostly 1-2.5 cm. long, 0.5-1.5 em. broad, villous or glabrate. —The fo ee are Baer oa Lisaaioe. Hopedale, Sornborger, n NEw : Quirpon Harbor, Huntsman; Schooner ra " Brandy) Seek Pistoe Bay, Pease Long, no. 27,997; Cape Norman, Wiegand, Griscom & Hotchkiss, no. 27,987, 27,988; Boat Harbor, Straits of Belle hyd Fernald, Wiegand & Long, n 0. 245 998; Big Brook, Straits of Belle Isle, Pease & Griscom, no. 27,985; Wiegand, Gilbert & Hotchkiss, nos. 27,994, 27,995; Sandy (or Poverty) Cove, Fernald, Long & Dunbar, no. 26,575; Savage neha riscom, no. 27,992; Yankee Point, W7 iegand & Hotchk no. 27,984; Flower Cove, ‘Mary E. eg? no. Cl, Fernald, Die é Dunbar, no. 26,571, Fernald, Grisco é Gilbert, no. 27,980; Ice Point, Wiegand, ‘Gilbert & Hotchkiss, no. 0. 27 ,993; Brig Bay, Fernald, Long Wiegand, nos. 3203, 3207, 3218, 3221; ee Bay, Fernald & Wiegand, nos. 3229, 3230, Kimball, nos. 136, 137 | Rasaleli Island, Notre Q i Martin e : : 18, 932; Ellis Bay, Anticosti, John Macoun, no. 18,830 (TYPE); Femald ass, between Mts. Mattaouisse and Fortin, Fernald, Griscom Mackenzie, Pease & Smith, no. 25,670; Pease Basin, between Mts. Logan and Pembroke, Fernald, Griscom & M ackenzie, nos. 25,671, 25,672, “35, 675. e la, n. var., var. Macounii similis, oli orbicularibus — peetes oblongis apice pobalatie 0.7-2 mm. latis—-NEWFOUND- D: turfy limestone barrens, Capstan Point, fine Cove, July 27, 1924, Fernald, Long & Dunbar, no. 26,570; limy barrens southeast and south of Powe Cove, July 10, 1925, Wiegand, Pease, Long & Hotchkiss, nos. 27,982, 27,991; limestone pega ent sea-level, Pointe Riche, August 4, 1910, Fernald & Wiegand, no. 3204 (TYPE in Gray Herb.). QuEBEc: sprawling on ledge, Ile arene Archipel Ouapitagone, — 15, 1915, St. John, no. 90,841. Var rutex depressus vel erectus ad 1 m. altus; ramulis glabris jucidis gracilibus; foliis obovatis 2.5-9 cm. longis 2-5 em. latis glabris submembranaceis.—LaBrapor: wet land, Red Bay, August 7, 1894, Waghorne, no. 33 (Arnold Aboretum) ; springy banks and damp places, Forteau, July 30, 1910, Fernald & Wiegand. nos. 3213, 3214. NEWFOUNDLAND: rich thickets on lower slopes of Ha-Ha Mountain, July 17, 1925, Fernald, Wiegand, eras Gilbert & Hotchkiss, nos. 27,967, 27,968; shelves, crests and talus of diorite cliffs, Ha-Ha Mountain, August 5, 1925, Fernald & good nos. 27,971 (ryPE in Gray Herb.), 27,972; wet slaty cliffs, John s Ladder, western face of Doctor Hill, ‘August t 24, 1925, Fernald "¢ ony no. 27,979. QueEBEc: schistose talus and wet shelves at base (alt. 400- 600 m.) of Big Chimney, Mt. Mattaouisse, July 10, 1923, Dodge, Griscom & Pease, no. 25,659. 188 Rhodora [OcTOBER The Waghorne number from Red Bay was cited by Rydberg under S. Waghornet Rydb. Bull. N. Y. Bot. Gard. i. 271 (1899), but it is quite different from the type, a poor fragment, scarcely identifiable, collected by Dr. Bryant in 1860. Rydberg’s description calls for a shrub with the leaves “somewhat hairy when young, but the long white hairs . . . appressed and parallel to the midrib” and the type-specimen is too near S. Me eg var. callicarpaea. SaLrx pedunculata, n. sp., frutex 0.5-1.5 m. altus erectus; ramulis novellis glabris fuscis deinde Gad: nitidulis; gemmis bene evolutis nondum visis; foliis immaturis membranaceis glabris ellipticis oblongis vel anguste obovatis 2-5.5 cm. longis 1-3 cm. latis subacuminatis vel plus minusve acutis, basi angustatis petiolatis petiolis 3-8 mm. longis margine subintegris vel undulato-crenatis supra viridibus sublucidisque subtus glaucis; stipulis lanceolatis 2-3 mm. longis glanduloso-dentatis deciduis; amentis foemineis coetaneis er pated latis sub anthesi nondum visis adultioribus 4-8 em. longis 1.1-1.4 ¢ crassis densifloris; pedunculis 1-2.5 cm. longis foliis 2-3 cae munitis pedunculo rhachique griseo-pilosis; bracteis anguste ovato- oblongis subacutis nigrescentibus 3. mm. longis longe pilosis; capsulis lanceolato-ovoideis longe rostratis 6-8 mm . longis breviter griseo-pilosis pilis nitidulis; stylis distinctis 1.2-1.5 mm. longis supra subclavatis stigmatibus oblon ngis bifidis adscendentibus stylo duplo brevioribus basi vix distinctis; pedicels 1.5-2 mm. longis glandulam 0, se duplo superantibus. EWFOUNDLAND: brooksides and -margins one mile back of Savage Cove, Straits of Belle Isle, hilly 14, 1925, Fernald & Long, no. 28,035. See p p. 9. In some characters Salix pedunculata suggests S. discolor Muhl., S. planifolia Pursh and S. arctophila Cockerell. As an erect coarse shrub with subentire to undulate glabrous leaves glaucous beneath it suggests the two former and its very long fruiting aments are suggestive of those of S. discolor. It is at once distinguished from both S. discolor and S. planifolia by its late flowering, the aments terminating definite and elongate leafy-bracted peduncles. The bracts (or scales) of the aments are more elongate and acute than in S. discolor, the style longer, and the stigmas shorter than in that species, and the capsules are shorter than in well-developed S. dis- color. Besides by its long-peduncled instead of sessile aments S. pedunculata differs from S. planifolia in its very long fruiting pea the longest mature aments of S. planifolia rarely exceeding 4 cm.; longer pedicels and capsules and the much shorter pubescence of ss In its large peduncled fruiting aments, its black scales and its 1926] Fernald,—Botanizing in Newfoundland 189 glabrous foliage, herbarium-specimens of S. pedunculata superficially . suggest S. arctophila, but the latter species belongs in a very different section, with prostrate and repent habit, sessile or only short-pediceled capsules, short round-tipped bracts, etc. The only other near relatives of S. pedunculata besides S. discolor and S. planifolia are S. phylicifolia L. of Europe, with which S. planifolia was formerly confused, and S. pennata Ball, of Washington and Oregon. From the former it is distinguished by its long peduncles and much larger fruiting aments, its longer bracts (or scales) its less deeply cleft stigmas and its longer style; from the latter, as described and illustrated by Ball, by its peduncled aments much longer in fruit, long pedicels and griseous instead of white pubescence of the capsule. At the type-station S. pedunculata formed an extensive thicket at the border of a pond in the limestone barrens back of Savage Cove. It was observed only at this station but further exploration of the little known barrens along the Straits of Belle Isle will doubtless show that it, like most other species endemic in the region, extends over a considerable area. SaLIx amoena, n. sp., frutex 1-2 m. altus erectus; ramulis novellis longe rostratis 6-8 mm. longis breviter griseo-pilosis pilis nitidulis; stylis distinctis 1 mm. longis stigmatibus oblongis bifidis subadscend- is In habit Salix amoena is very similar to S. pedunculata (described in this paper) but it is at once distinguished by the shorter yellow bracts of the aments, the more slender capsule, shorter style, shorter pedicels and relatively longer glands. It also simulates S. Bebbiana, var. perrostrata (Rydb.) Schneider but is quickly distinguished by the short pedicels and definite slender styles. From all forms of the protean S. cordifolia Pursh it stands out by its long and in maturity 190 Rhodora [OcTOBER flexuous or drooping aments, its distinctly undulate or crenate leaves (as in S. discolor and S. Bebbiana) and its slender pedicels. In a springy meadow at the northern base of Ha-Ha Mountain it formed a thicket of considerable extent but staminate shrubs, if they there occur, were too far advanced for us to secure good male aments. S. MyRTILLIFoLIA Anderss. Cold mossy brooksides and_ pond- margins one mile back of Savage Cove, Ferna Long, no. 28,004; swale by Mile Brook, west of Big Brook, Long & Gilbert, no. 28,00; the only Newfoundland records from north of Ingornachoix Bay. p. 95. S. MYRTILLIFOLIA var. BRACHYPODA Fernald. Mossy brooksides and damp turfy slopes, Sacred Island, Fernald & Long, no. 28 meadow below limestone escarpment, western face of Bard Harbor Hill and dryish limestone talus, western face of Doctor Hill, Highlands of St. John, Fernald & Long, nos. 28,006, 28,010; the first stations except on Table Mt., Port-a-Port Bay. See p.122. This pe prove to be the problematical S. latiuscula Anderss. See p. 126 S. caLcicoLa Fern. fieg. Dominant on wet or dry limestone barrens from Pistolet Bay to Ingornachoix Bay. See p. 62. Very variable, the mature foliage varying from aiotiedensantat to sub- a irl and from 0.8—5.6 em. long. S. cryproponta Fern. Cold brooksides and pond-margins one mile hoe of Savage Cove, Fernald & Long, no. 28,026; cold springy glade, — of southern half of Burnt Cape, F ernald & Long, no. 28,028; the first stations except the original, on the East Branch of the Humber. See pp. 95, 119. ANA Sargent, var. PERROSTRATA et) Schneider. Cove Brook, Griscom, no. 98,038 ; Savage Cove, Fernald & ting, no. 28,033; the Rocky Mt. extreme, new to Newfoundland. See p. 95. S. PLANIFOLIA Pursh. Schneidert excludes this species from the flora of poe oeniiens It abounds in northwestern Newfoundland from Quirpon to St. John Bay. S. PELLITA die Pehle below John Kanes’s Ladder, ——. face of Doctor Hill and meadow below calcareous sandstone ments, western face of Bard Harbor Hill, Highlands of St. Toh Fernald & Long, nos. salt 28,058; the only Newfoundland records from north of ie Hum Berowa Puma L., oe renitnwts’ var., frutex depressus; ramulis villosis; foliis eidisesits fructiferorum plerumque reniformibus vel oblatis basi cordatis margine grosse dentatis.— QueEBEC: gneiss Journ. Arn. Arb. iff, 89 (1921). 1926] Fernald,—Botanizing in Newfoundland 191 ledges, Mutton Bay, Saguenay County, September 7, 1925, Fernald Long, no. 28,089 (TYPE in Gray Herb.); Bonne Espérance, August 27, 1882, J. A. Allen, no. 70 (in part). In typical Betula pumila, which abounds in Newfoundland, southern Labrador and much of eastern Canada, the principal leaves of the fertile branches are obovate and narrowed to the base. At Mutton Bay, var. renifolia was a dominant shrub, covering the ledges over a considerable area. No typical B. pumila was observed there, although our visit was brief—during the short stop of the steamer. URTICA GRACILIS AND SOME RELATED NorTH AMERICAN SPECIES.— There are three indigenous Nettles in the region from New England to Newfoundland and Labrador. One of them has regularly passed as Urtica gracilis Ait., another as U. Lyallii Wats., and the third has been doubtfully referred sometimes to one, sometimes to the other of those species. In attempting to clear the identity of the third plant it has become obvious that we have always misidentified U. gracilis and that the wide-ranging boreal species (extending from Newfoundland to British Columbia, south to northern and western New England, northern New York, New Mexico and Oregon) which we have called U. Lyallii is not satisfactorily identifiable with the Lyall plant which Watson had before him. And not merely the type of U. Lyallii Wats. has been misunderstood, but the identity of his other species, U. Breweri of California, has been wholly mis- interpreted. There are few good diagnostic characters to be found in the flowers and fruits of our perennial nettles, all of which were treated by Weddell as varieties of the Old World U. dioica L. U. dioica is, as its name implies, dioecious; the native American plants strongly tend to be monoecious. In the latter series wholly pistillate individuals frequently occur but in the whole range of native American plants studied I have found no specimens which have only staminate inflorescences; various mixtures occur but all American plants ap- parently have some of the terminal racemes pistillate. On this account and since the American plants have clearly defined and wholly characteristic and consistent ranges, I am treating them as species, although it is not possible to assign them diagnostic fruit-characters. A large proportion of specimens are so poorly collected (for obvious reasons) that the best diagnostic character, the mature stipule, is 1 Weddell, Mon. Urt. 78 (1856) and in DC. Prodr. xvi.! 50-53 (1869). 7 192 Rhodora [OcToBER not well displayed. The texture and shape of the mature stipules, a character already used by Rydberg in his Flora of the Rocky Moun- tains, is probably as important a character as any. In the eastern and transcontinental plants the stipules are scarious to herbaceous in texture and green to at most pale-brown; in the three species confined to the Pacific slope, U. holosericea Nutt., U. Brewert Wats. (? U. californica Greene) and U. Lyallii Wats., they are subcoriaceous to coriaceous and soon become deep-brown to castaneous. As a result of the present study the following memoranda on the principal perennial species of temperate North America are reco : Urtica Gracitis Ait. Hort. Kew. iii. 341 (1789). As usually interpreted, U. gracilis is the tall species of the eastern United States with the stems only slightly, if at all bristly, except at base, but cinereous-pilose above; the primary leaves ordinarily oblong-lanceo- late or narrowly ovate-lanceolate, with 20-35 teeth on each margin— those subtending the lowest inflorescences with 19-38 (av. 25); the petiole 1-2 as long as the blade; the lower leaf-surfaces ashy- puberulent and the stipules densely cinereous-puberulent. This plant is of wide range in the United States, westward across the Mississippi Basin; and, so far as shown in the Gray Herbarium, its northeastern limit extends from the warm southern side of the Gulf of St. Lawrence (Nova Scotia and Prince Edward Island—lat. 46°-47°) across southern Aroostook County, Maine (lat. 46° 40’) to the region slightly north of Montreal (lat. 46° 20’) and the counties of Ontario bordering Lake Ontario (lat. 44° ; en Aiton’s original description is studied some points are apparent which at once indicate that we have been in error in applying to the plant just defined the name U. gracilis. Aiton’s description was brief but clear: “6. U. foliis oppositis ovato-lanceolatis nudiusculis, caule petio- lisque hispidis, racemis geminis. Slender-stalk’d Nettle. Nat. of Hudson’s Bay. Introd. 1782, by the Hudson’s Bay Company.” The description alone might be given little weight, were it not for the type region, Hudson Bay, hundreds of miles to the north of the known northern limit of the tall species with puberulent closely toothed leaves and cinereous stipules. But in view of the occurrence in northwestern Newfoundland and on the Labrador Peninsula, thence westward to British Columbia, and south to the len 1926] Fernald,—Botanizing in Newfoundland 193 Islands, northern and western New England, northern New York, New Mexico and Oregon, of a plant with ovate-lanceolate usually glabrous slender-petioled leaves with the upper half of the stem and the petioles often bristly, it seems clear that it is this more northern species which should bear the name U. gracilis. The more boreal plant has the stems glabrous or merely setulose above; the leaves lanceolate to ovate, glabrous or only sparingly pilose beneath; their margins more coarsely toothed than in the more southern species— each side of the leaves subtending the lowest inflorescences having 13-23 (av. 17) teeth; the petiole slender and elongate, 1.5-5.5 (av. 3) em. long, 4-3 as long as the blade; and the very thin stipules often nearly glabrous. This is the plant which in eastern America has passed as U. Lyallii Wats., but the Lyall plant, as already noted, is a species of the north Pacific slope with firm and finally castaneous stipules. With U. gracilis now identified as the northern transcontinental long-petioled species, the ‘“Slender-stalk’d Nettle” of Aiton which has erroneously passed as U. Lyallii, it is necessary to find the correct name for the more 2S plant with cinereous puberulence which has passed as U. gracilis. This seems to be, without question, U. procera Muhl. in Willd. Sp. Pl. iv.! 353 (1805). The third indigenous species of northeastern America is with us a plant of the coast, chiefly at the upper borders of beaches, occurring from southeastern Labrador to Maine but reappearing in the Rocky Mountains, from Alberta to New Mexico. It is a lower but coarser plant than either Urtica gracilis or U. procera, its leaves with glabrous to pilose-hirsute lower surfaces, coarse teeth and short, stout petioles; the leaves subtending the lowest inflorescences having 11-25 (av. 17) pairs of teeth and petioles 0.7-2.5 (av. 1.3) em. long, only -} as long as the leaf-blade. Its staminate inflorescences are short and thick and its stipules comparatively long. This short-petioled species was described from the Rocky Mountains as U. viridis Rydb. Bull. Torr. Bot. Cl. xxxix. 305 (1912). Since these three species have all passed as Urtica gracilis it is well to reenumerate their diagnostic characters and to cite some characteristic specimens. U. cracrus Ait. Hort. Kew. iii. 341 (1789). U. cardiophylla Rydb. Bull. Torr. Bot. Cl. xxiv. 191 (1897). U. _ of eastern Am. authors, not Wats.—Slender, 0.3-1 m. high: stem glabrous above or somewhat setulose and sparingly qillee: jae lanceolate 194 Rhodora [OcTOBER to ovate, rounded to cordate at base, slender-petioled, glabrous on both surfaces or sparingly pilose beneath, coarsely serrate; those subtending the lowest inflorescences 5-15 (av. 9) em. long, with cl weste dag he a to northern Maine, western New. England and secthert New York; Alberta and British Columbia, south to New Mexico and Oregon. The following are characteristic. NEW- FOUNDLAND: rich thickets and glades near timber-line, Bard Harbor Hill, “eager of St. = Fernald, Wiegand, Long, Gilbert John ie Fernald & Wiegand, no. 28,104. QurBec: Seven Islands, Saguenay Co., C. B. Robinson, no. 946; thicket, Dartmouth River, Gaspé Co., Collins, Fernald & Pease, no. 5503; limestone detritus in thickets, Les Murailles, Percé, Fernald & Collins, no. 1003; cal- careous sea-cliffs and rock-slides by the Gulf of St. Lavette’: alight west of Marten River, Gaspé Co., Fernald & Peas 25,029; thickets and wet rock-talus, Nettle Gully, alt. about 400 m., ake base of Mt. Collins, Fernal, Pease, no. 25,038 (locality given on August 9, 1892, G. G. Kennedy. Maapaten Istanps: cool wooded bank, Grindstone Island, Fernald, Long & St. John, no. 7321; open spruce woods, Basin sland, Fernald, Bartram, Long & St. John, no. 7319; sandy bank, Amherst Island, St. John, no. 1850. PRINCE Epwarp Istanp: by brook in woods, Harmony, Fernald & St. John, no. 11,033. Marve: Fort Kent, 1881, Kate Furbish; alluvial island, Seven Islands, St. John River, St. John, no. 2278. VERMONT: Burlington, August 16, 1901, L. R. Jones (probably introduced) ; hid ancs etd May 25, 1908, E. F. Williams (probably introduced). NECTICUT: Aeon Stafford, August 27, 1903, Bissell, Graves nshably introduced). w York: — Potsdam, Phelps, no. 385 (probably segs aN Montana: banks of Missouri River, alt. 3600 ft., Scribner, no. =o Bald Ree ice alt. 7000 ft., Watson, no. 357 (distributed as U. Breweri). WyYomING: Centennial Valley, A. Nelson, nos. 1273, 1862 Ge U. besen Cotorapo: Bob Creek, western La Plata Mts., Baker, Earle & Tracy, no. 282; near Pagosa Peak, alt. 9000 ft., Baker, no. 281; North Elk Calving. Rio Blanco Co., July 26, 1902, Sturgis. NEw ‘Mexico: bottom of Moro River, Fendler, no. 821; White Mts., alt. 7000 ft., Wooton, no. 305; Windsor Creek, alt. 8400 ft., Pecos River National Park, Standley, no. eee Ipano: river-bank, St. Anthony, ee rrill & Wilcox, no. 853. ORE mountain streams and bogs, al 4000-6000 ft., oe an 2A Wasnincton: Columbia River, Klikitat Co., Suksdorf, no. 58; near 1926] Fernald,—Botanizing in Newfoundland 195 Ellensburgh, Brandegee, no. 1075; low meadows, Pullman, July 10, 1895, Hardwick, Piper; along streams in woods, Blue Mts., July 17, 1896, Piper; near streams, Waitsburg, Horner, no. B443; Montesano, Heller, no. 3920; among bushes along streams, Mt. Paddo, Suksdorf, no. 6354. British CoLumpra: Upper Alloknejik Lake, July, 1882, McKay; thickets, Beaver Creek, Selkirk Range, July 13, 1885, Macoun; deep thicket at Nelson, Shaw, no. 667; thickets, Yoho Valley, July 28, 1916, Hunnewell. prRocERA Muhl. in Willd. Sp. Pl. iv. 353 (1805). U. gracilis of most Am. auth., not Ait. . dioica, t. procera (Muhl.) Weddell, Mon. Urt. 78 (1856). U. longifolia Raf. acc. to Weddell, 1. c. (1856). —Slender and tall, up to 1 m. or more high: stem only slightly if at all bristly but cinereous-pilose or -puberulent toward the summit: leaves oblong-lanceolate to narrowly ovate-lanceolate, usually ci- nereous-puberulent beneath, slender-petioled, rather closely serrate; those subtending the lowest inflorescences 6.5-18 (av. 11) em. long, with 19-38 (av. 25) pairs of teeth and with petioles 1.5-8.5 (av. 4.5) ville, Fernald & Long, no. 23,786. Maine: moist roadside, Lime- stone, Cushman, no. 2265; river-thicket, Masardis, September 8, 1897, Fernald; Fort Fairfield, July 19, 1902, Williams, Collins & Fernald; thicket, Bradley, July 17, 1890, Fernald; Buckfield, Septem- ber, 1889, Parlin; moist roadside-thickets, Monhegan Island, Jenney, Churchill & Hill, no. 3231; river-shore, Topsham, August 29, 1912, ate Furbish; roadside clearing, Baldwin, F ernald, Long & Norton, no. 13,511; Kittery Point, August 27, 1895, E. F. Williams. NEw HampsuirE: wet roadside, Colebrook, Pease, no. 10,904; shaded street, Jefferson, Pease, no. 16,704; open roadside, Hillsborough, farm-yard, Concord, September 27, 1896, Williams; near outlet of Spot Pond, Middlesex Fells, August 19, 1920, Kidder; Needham, 196 Rhodora [OcToBER August 1, 1886, 7. O. Fuller; North Scituate, August 15, 1901, Ken- nedy; thickets along roadside, Barnstable, Fernald & Long, no. 16,724; open grassy roadsides, West Chesterfield, Robeson: no. 493; low ground, Monterey, July 7, 1920, Hoffmann; alluvium, Lanes- boro, July 20, 1915, Churchill. Ruope Istanp: Rumford, September 6, 1903, Williams; Wickford, June 17, 1908, Kennedy. CoNnNECTICUT: sandy ‘hank, Thompson, June 10, 1922, Bill & Grigg; roadside, Steg fe Bissell, no. 533; dry thicket, North Guilford, August 19, , G. H. Bartlett; roadside waste, Newton, September 26, 1901, eae New York: western New York, Sartwell. PENNSYLVANIA: . shore of Delaware River, Morrisville, Bartram, no. 1298. KENTUCKY: Lexington, rt. ONTARIO: woods, Hay sland, Gananoque, July 20, 1908, ic al Plevna, August 5, 1902, Fowler; Belleville, Macoun, no. 1584. MicuiGan: under scattened hardwood; Turin, Marquette Co., July 30, 1901, Barlow; Grand Rapids, September 29, 1860, Wm. Boott. Tuuinots: Mahomet, Gleason, no. 75; Bloomington, = e g oe is oma oO ot °o BPE 6. = or ° =a © oS ct io oo wm 9 mM = ag (=) ba | or petioled, glabrous on both surfaces or pilose-hirsute beneath, ype serrate; those subtending the lowest inflorescences 3.5—10 v. 7) only 3-6 mm. thick. eaten of Belle Isle, ai Laie along he coast to Lincoln Co., Maine; Alberta t — Dakota and New Mexico. The following are characteristic. se ADOR: springy banks and damp hillsides, Forteau, Fernald «& reach no. 3279; sand-bank bordering stim Forteau, Long, no. 28,101; abundant by streams on the calcareous ee terraces, Blanc Sablon, Fernald & Wiegand, no. 3280. WFOUNDLAND: turfy barrens an slopes, Sacred Island, Wiegand, Gilbert & Hotchkiss, no. 28,100; rich thickets on lower slope of Ha-Ha Mountain, Fernald, Wi iegand, Gilbert & Hotchkiss, no. 28,097; limestone ‘shingle and gravel along shore, Flower Cove, Fernald, Long & Dunbar, no. 26,604; damp thickets, Bear Cove, Wiegand & Pease, no. 28,098; turfy limestone shore, St. Barbe, Fernald, Long & Dunbar, no. 26, 605; sandy shore -back of Cow Head, Fernald be Wiegand, no. 3278; woods, i _ Harbor, Waghorne, no. oor-yard, near Frene hman’s Cov Mackenzie & seem no. oe QUEBEC: aay shore, Piuchishaiee 1926] Fernald,—Botanizing in Newfoundland 197 St. John, no. 90,396; sur le rivage caleaire, Ile Herbée, Mingan, ictorin & Rolland, no. 18,570; autour des habitations, Pointe-aux- Esquimaux, Victorin & Rolland, no. 18,497; abondant sur les rivages, Ilets de la Baie A Jean, Victorin & Rolland, no. 18,408; Baie Sainte- Claire, Anticosti, Victorin, no. 4197; sur les rivages calcaires, Baie du Renard, Anticosti, Victorin, Rolland & Louis, no. 22,053. Macgpa- LEN IsLANnpDs: Ile du HAvre-au-Ber, Victorin & Rolland, no. 9627. Ew Brunswick: clearings, Bathurst, Blake, no. 5434. MAINE: Machias Seal Island, July 27, 1902, Mrs. A. H. Norton; waste ground, Seal Harbor, August 30, 1889, Redfield; beach, Great Cranberry Isle, August 20, 1888, Rand; south and east shores of Little Cranberry Isle, August 6, 1889, Redfield; sandy ridge back of beach, Flye’s Island, Brooklin, Hill, no. 586; barrier beach, Atlantic, Swans Island, Hill, no. 1552; old cellar-hole, Outer Heron Island, Boothbay, Fassett, no . Norrn Daxora: damp canyons, Deadwood, Carr, no. 14 Atperta: Rocky Mountains, 1858, Bourgeau (cited by Rydberg); Red Deer River, Macoun, nos. 200, 1584; thickets, Red Deer Valley, vicinity of Rosedale, Moodie, no. 1094; Montana: Emigrant Gulch, Rydberg & Bessey, no. 3935 (TYPE no.). Wyomine: along the creek, creek bottoms, Halleck Cafion, A. Nelson, no. 7444 (cited by Rydberg); Merican Mines, A. Nelson, no. . Cotorapo: Gunnison, Baker, no. 601; low wet bottoms, Mancos, Baker, Earle & Tracy, no. 40 New Mexico: Parry, 1867, no. 196. Urtica viridis is the North American counterpart of the Asiatic U. angustifolia Fisch. In its short-petioled narrow leaves with coarse teeth and often quite glabrous surfaces it exactly matches sheets of the latter from Siberia, Mandchuria and Afghanistan, but all the Asiatic specimens agree with the descriptions in being strictly dioeci- ous, like the Old World U. dioica, while U. viridis is nearly always definitely monoecious. It is probable that some references to U. angustifolia in America belong to U. viridis. As stated in the introductory paragraph Watson’s species, Urtica Lyallii and U. Breweri have both been misinterpreted. U. Lyallii Wats. Proc. Am. Acad. x. 348 (1875) was primarily “A very large leaved species collected by Dr. Lyall in the Cascade Mts. in lat. 49°” but “more densely hispid” specimens were also cited from Marin Co., California, collected by Bolander and Kellogg (“H. N. BOLANDER, KELLOCK, M. D. & CO. ” on the label) and a slender plant collected by Lyall on Vancouver Island was cited as “probably the same.” The first-cited plant, which was col- lected by Lyall and which formed the chief basis of the species must be taken as type. This and the Lyall material from Vancouver Island are fortunately conspecific and represent the glabrous or but slightly 198 Rhodora [OcToBER pubescent species of Alaska, British Columbia, Washington and Oregon, with slender petioles, ovate, often cordate thin leaves and firm brownish, often finally castaneous stipules. The plant from Marin County is a specimen of the much coarser and more southern U. Breweri Wats. (? U. californica Greene). U. Brewert Wats. 1. e. (1875) was based primarily on Brewer’s no. 95 from Los Angeles, but specimens collected by Bigelow in western Texas and by Wolf in southern Colorado were also cited. The Brewer plant from Los Angeles, the type of the species, is not at all like the Bigelow and Wolf specimens. The two latter are very characteristic specimens of the annual species, U. gracilenta Greene. In fact, Watson himself, after originally calling these U. Breweri, penciled on the sheet con- taining them: “Probably distinct species. Lyvs. large, thin, coarsely toothed, few-nerved, long-petioled. Stupiles herbaceous, linear-lanceo- late. Calyx villose.” Nevertheless, on a sheet of the original material of U. gracilenta sent by Greene he penciled “ = U. Brewert, Watson.” The type of U. Breweri (Brewer’s no. 95 from Los Angeles) consists of two fragments picked obviously late in the season from a large almost woody stem. There are no primary leaves, only those of secondary shoots, but their characteristic pubescence and their firm brown stipules show them to belong apparently to the very characteristic coarse species of the coast of California farther north, afterward described as U. californica Greene, Pittonia, i. 281 (1889). The Marin County plant included by Watson under U. Lyallii also belongs here (a good match for authentic U. californica). Dr. I. M. Johnston informs me that the uncharacteristic fragments which formed the type of U. Breweri have not been matched by recent collections from Los Angeles and that the Brewer material might have been collected nearer San Francisco and erroneously labeled. The widely distributed plant which generally passes as Urtica Brew- eri is quite distinct and no material of it was included by Watson in his original account. As usually incorrectly interpreted, U.- Breweri is identified as a tall cinereous-puberulent plant with the coarsely toothed narrow leaves velvety-puberulent beneath, and with the lanceolate puberulent stipules drab or greenish, occurring from Idaho and Washington southward through California, Arizona - and New Mexico into Mexico. The plant is the western represent- ative of U. procera but it seems to be a reasonably distinct species. Its correct name seems to be 1926) Fernald,—Botanizing in Newfoundland 199 U. serra Blume, Mus. Bot. Lugd.-Bat. ii. 140 (Nov., 1855). U. dioica, B. angustifolia Schlecht: “pani vil. 141 (1832), not U. angustifolia Fisch. (181 9). U-. aquatica Li ebm. K. Dansk. Vidensk. Selsk. Skrift. ser. 5, ii. 291 (1851), not Moon (1824). U. mexicana Blume, |. c. (1855), not Liebm. (1851). U. dioica, ~ occidentalis Watson, Bot. U. S. Geol. Expl. Fortieth Par. 321 (1871). Blume’s species, U. serra and U. mexicana, and the other species (including Lythraceae) published at the same time are usually cited as dating from 1852. The latter date appeared on No. 1 of vol. ii. of the Museum Botanicum, but in Bot. Zeit. xiv. 186 (1856) Miquel showed that Nos. 1-8 of vol. ii. were not put on sale until 1856; and he later indicated (p. 540) that the four numbers, 9-12, were published respectively in November and December, 1855, and January and February, 1856. CoMANDRA RicHaRpsIANA Fernald. Dominant on turfy or sed knolls in the limestone barrens from Pistolet Bay to St. John See p. 91. Previously known in Newfoundland only from Masia & Griscom’s collection from the Bay of Islands. Oxyria pDicyNna (L.) Hill. Wet quartzite rocks and gravel along brook, Southwest Gulch, Highlands of St. John, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,115, the first station in Newfound- land since La Pylaie’s discovery of it on Grois Island. See p. 124. UMEX OCCIDENTALIS Wats. The only indigenous dock of north- SALICORNIA EUROPAEA L., var. PROSTRATA (Pall.) "Fern. Damp depressions in sand and gravel back of barrier beach, Argentia, Fernald, Long & Dunbar, no. 26,644; the first in Seog ogies except from Bay St. George and Port-a-Port Bay. See p SuaepA Ricum Fernald. Damp depressions in sand a ey gravel back of barrier beach, sir geste F ease Long & Dunbar, no. 26,645, the first fruiting material from Newfoundland. Records of young plants from Bay St. George and Notre Dame Bay probably belong ere. See ARENARIA DAWSONENSIS Britton (A. litorea Fernald). pres- — sions of limestone barrens along the Straits, widely distributed. but rather local; previously — from north of Bay of Islands and Notre Dame Bay. 62. A. CYLINDROCARPA Fernald. Widely distributed in clay and gravel of the limestone barrens from Pistolet Bay to Port-a-Port ay. STELLARIA rioripa Fischer. Springy swale and turfy upper border of strand, Anse aux Sauvages, Pistolet Bay, Fernald, Wiegand ng, nos. 28,196, 28,197; a Kamtchatkan species collected for the first time in America in 1923 on Tabletop Mts., Gaspé. See p. 123. 200 - Rhodora [OcroBER CrRASTIUM ALPINUM L. The typical form of the species was see only once: mossy brooksides and damp turfy slopes, Sacred Tail Fernald & Long, no. 28,208, probably new to Newfoundland. C. ALPINUM, var. GLANDULIFERUM Koch. Dry limestone barrens, northern half of Burnt Cape, F se Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 28,204, new to Newfoundland. C. aLPINUM, var. LANATUM why Hegetschw. Common on turfy or gravelly slopes or on cliffs from Quirpon westward to Four- ae ele ; ui Ostenf. Limestone ledges on the — side of Schooner (randy) “land. Pistolet t Bay, Pease & Long, no. 28,212. See p. An arctic Eurasian species heretofor > eee de America only ia Cape Nome, Alaska een, the riateridl of the latter col- Tabletop Mts., Gaspé, south of northern Labrador. See Only a single over-ripe specimen was noticed, snatched at the last moment after the order to quit had been given. It was on gravel at restate of the Gulch and had presumably washed down from gher u C. viscosum L. Turfy roadside, bese Fernald, Long & Dunbar, no. 26,678, an obvious adventive, new to Newfoundlan RANUNCULUS HYPERBOREUS Rottb. Brooksides and springy cal- careous meadows, Savage Cove, Fernald, Long & Dunbar, no. 26, 686, Fern = Griscom, no. 28, abt; dried-out Ria pools i in limestone ao) sana Fea of pool in limestone barrens, Capstan Point, F lower & beaver pond near the Yellow Marsh, back of Bard Harbor, St. John Bay, Gilbert & Hotchkiss, no. 28,253. See pp. 61, 127. New to Newfoundland, the old —— having been based on either R. Cymbalaria Pursh 0 r R. hederaceus L. R. PEDATIFIDUS 7. E. Sm., var. Lerocarpus (Trautv.) Fern. (A. afinis R. Br.). Turfy or gravelly shelves, crests and talus of diorite cliffs, Ha-Ha Mountain, Pease & Griscom, no. 28,258, Fernald & Long, no. 28,259; “ibe and turfy trap cliffs and talus, Anse aux Sauvages, Pistolet Fernald, Wiegand & Long, no. 28, 260; the first stations south oe the Torngat region of pean See pp. 104, 123, also pl. 153, | R. PENSYLVANICUS a f. Boggy meadows and clearings back of Bard Harbor, St. John Bay, very scarce, Wiegand, Gilbert & Hotch- kiss, ts 28,266, the first authentic station except from the lower RR. Macounn Britton. Brookside in bushy swale on flat north oo of Doctor Hill, —— Bay, no. 28,267, the first from north of Hawke . p. 127 1926] Fernald,—Botanizing in Newfoundland 201 ANEMONE MULTIFIDA Poir. var. HUDSONIANA DC. Dry limestone cliffs and talus, western face of Doctor Hill, Highlands of St. John, Fernald & Long, no. 28,278, our only station except on the lower Exploits and Acti Dame Bay. See p. 126. BARBAREA ORTHOCERAS Ledeb. Peaty borders of spruce thickets, Savage Cove, Fernald & Griscom, no. 28,317; brooksides and slaty hills, Little Quirpon, Wiegand, Gilbert & Hotchkiss, no. 28, 318, Fernald & Gilbert, no. 28,319, Wiegand & eg — 28,320; our rst stations outside the Exploits Valley. See p. CARDAMINE PRATENSIS L., var. ANGUSTIFOLIA Hoo ve Dominant in peat and damp turf on the limestone barrens from Pistolet Bay westward to Deadman Cove, the first area known south of northern Labrador. See p. 91. LesQuERELLA ARCTICA (Wormskj.) Wats. Dry gravelly limestone barrens, Burnt Cape to: Big Brook; St. John’s Island; previously own from Table Mt., Port-a-Port Bay aa Cape St. Geor ge. pp. 100, 103, 118 L. arctica, var. Purshii Wats., seems to have no characters of value. RABA NIVALIS Lilj. Gravelly shelves, crests and talus of Ha-Ha Mountain, Pease & Griscom, no. : crevices of trap cliffs, Anse aux Sauvages, Pistolet Bay, Fernald, Wiegan d & Long, no. 28,347; open peaty and gravelly spots on crests of trap cliffs, Cape Onion, Fernald & Long, no. 28,346; dry slaty crests of hills, Little Quirpon, Fernald & Long, no. 28,343; rocky crests of Cape Dégrat, Quirpon a x ernald & Long, no. 28,344; new to Newfoundland. See pp. 104, 23. as incana L. Typical D. incana with glabrous fruit is rare on the Straits: gravelly ia shore, Cape Norman, Wiega Griscom & Hotchkiss, no. 28,348. D. INCANA, var. CONFUSA (Ehrh.) Poir. Abundant on peaty and turfy calcareous slopes and shores from Pistolet Bay to Cow Head. See pp. 60, 90. D. MEGaAsPERMA Fern. ~ ee : ~~ and bed limestone John; only stations north of i. of Islands. See p. 117. D. pycnosperma Fern. & Knowlt. Dry limestone cliffs and talus, western face of Doctor Hill, os of St. John, Fernald & Long, no. 28,369; heretofore known only from the type region, Percé, Gaspé Co. , Quebec. See. p. 125. D. nirta L. Dominant on calcareous areas, northwestern New- foundland. See p. 90. D. RUPESTRIS R ‘Br. Dominant on calcareous areas, northwestern Newfoundland. See pp. 60, 90. 202 Rhodora [OcroBER ArABIS DrumMonDI Gray. Meadow below limestone pie: : western face of Bard Harbor Hill, Fernald & Long, no. 28,422; only known Newfoundland station except along lower Exploits. “i p. 124. ; UMMONDI, var. CONNEXA (Greene) Fernald. Mossy and turfy trap cliffs and talus, Anse aux Sauvages, Pistolet Bay, Fernald, Wregan no. 28,421; dryish limestone talus, western face of Doctor Hill, F ernald. & Long, no. 28,420; new to Newfoundland. e p. 123. Braya Longii, n. sp., planta aha ci humilis plerumque multiceps; radice crassa longe descendenti; ¢ bus aigiabegarsey: ener pur- purascentibus 1-10 cm. longis rare fructiferis usque ad 15 em. longis nudis vel sub racemo monophyllis, pilis simplicibus ef a inaequilongis griseis sparse obsitis; foliis rosulatis carnulosis lineari- oblanceolatis obtusis 1-4 em. longis 1-3 mm. latis integris glabris vel basi dilatato apiceque sparse ciliatis; racemo sub ant esi dense corymboso, fructifero plus minusve elongato 1-7 em. longo 3-20- floro, fructibus contiguis vel infimis 1-2 aosiansibas: pedicellis crassis ~4.5 mm. longis fructiferis arcte adscendentibus; sepalis deciduis pler- umque viridibus rare purpurascentibus glabris 2-3 mm. longis oblongis waet late obovatis albis tandem basin versus violaceis; staminibus mm. longis; antheris 0.4-0.6 mm. longis; pistillo lanceolato ice ovario ovulis 10-16, stylo crasso nee stigmate depresso style latiore bilobato; siliquis lanceolato-subulatis 4-9 mm. longis mm. latis stylo crasso 0.4—0.8 mm. longo coronas valvis glabris, septo albo uninervo vel enervo; seminibus oblongo-ovatis brunneis 1.2-1.5 mm. longis 0.5-0.8 mm. Jatis.—Shores of is Straits of Belle Isle, northwestern sine ne turfy or peaty pockets in lime- stone ledges, Sandy (or Poverty) Cove, August 1, Fernald, Long Dunbar, no. 26,723 (distributed as B. purpurascens); among loose ee aah limestone barrens, Sandy Cove, sae ‘Te, "1925, Fernald & 28,423; gravelly and peaty limestone barrens, Sandy ene ay 25, 1925, Fernald, Long & Gilbert, no. 28,424 (TYPE, in Gray Herb.); dry gravelly ‘inestone barren, Savage Point, July 13, 1925, Fernald, Wiegand, Pease, Long, Gilbert & Hotchkiss, no. 28,425; dry gravelly and turfy ai estone ine Savage Point, August 29, 1925, Fernald & Long, no. 28,426; crevices of limestone barrens, Yankee Point, August 16, "1925, Fernald, Wiegand & Long, no. 28,427. A beautiful species, originally detected by my companion on many hard trips, Bayard Long, to whose keenness in the field we owe many of the most interesting discoveries of the recent Newfoundland trips. B. Longii superficially resembles B. purpurascens (R. Br.) Bunge of the Arctic, with which the first collection was placed. It differs, however, in its narrower basal leaves; sparser but more often forking 1926] Fernald,—Botanizing in Newfoundland 203 pubescence of the stems; longer and broader petals; glabrous (not pilose) ovary and silique, the latter lance-subulate rather than ellipsoid or oblong-ovoid; and fewer (in B. purpurascens 16-20) and slightly larger seeds. In having seeds fewer than in B. purpurascens, B. Longii comes closer to B. rosea (Turcz.) Bunge of Siberia; but that species, of which three authentic collections are before me, has broader leaves; the stems very densely pilose with long hairs; the purple sepals 1.8—2.2 mm. long and densely pilose on the back; the narrowly obovate or spatulate-oblanceolate petals becoming wholly roseate and only 2.5-3.5 mm. long; the mature fruiting raceme but slightly elongated; the mature pedicels slender and elongating some- times to 8 mm.; and the capsule oblong-ellipsoid and at most 6 mm. long. See pp. 77, 94. Braya americana (Hook.) n. comb. B. alpina, 8. americana Hook. Fl. Bor.-Am. i. 65 (1830). Planta perennis p plerumque multicaulis; caulibus adscendentibus 1-10 cm. longis teretibus simplicibus pur- purascentibus, pilis furcatis brevibus inaequilongis griseis obsitis; foliis rosulatis carnulosis lineari-oblanceolatis 0.6-3 cm. longis 0.5-2 racemo sub anthesi capitato-corymboso, fructifero densiuscule 3-20-floro, floribus infimis 1-3 distantibus plerumque in axillis foliorum superiorum; pedicellis 1-4 mm. longis arcte adscenti- us; sepalis subpersistentibus plerumque Mesautecaihce 1.5-2.5 mm. longis spathulato-oblanceolatis vel anguste obovatis albidis deinde roseis apice rotundatis; staminibus 1.5-3.5 mm. longis, antheris 0.3-0.5 mm. longis; pistillo lanceolato, ovario ovulis 10-16 pilis plerumque bifurcatis hirto, stylo gracili ovario valde angustiori, stigmate depresso-capitato stylo valde latiori manifeste bilobato; siliquis lanceolato-subulatis 4-7 mm. longis 1. mm. latis stylo gracili 0.5-1 mm. longo coronatis, valvis hirtellis, septo manifeste uninervo; seminibus anguste ovoideis 1-1. . longis 0.5-0.7 mm. latis brunneis.—Alberta and Newfoundland. The following have been examined. ALBERTA: Cataract Pass below Summit, August 4, wFou open soil on limestone barrens near Ice Point, St. Barbe Bay, July 14, 1925, Wiegand, Gilbert & Hotchkiss, no. 28,429; gravelly limestone barrens 0 one mile back of Savage Cove, Straits of Belle Isle, July a 1925, Fernald & Long, no. 28,428, July 23, Fernald, Pease & Long, n 28,438; gravelly limestone barre rren, Four-mile ina Straits of Belle Isle, July 20, 1925, Fernald, Wiegand & Long, 28,437; moist gravel of limestone barrens on the Highlands porticast of Big Brook, Straits of Belle Isle, July 15, 1925, Pease & Griscom, no. 28,430; gravelly and peaty limestone barrens back of Big Brook, July 15, 204 Rhodora [OcToBER 1925, Fernald & Long, no. 28,431; swale near mouth of brook, Watts Big ht, Straits of Belle Isle, July 19, 1925, Pease, Griscom, Gilbert Hotchkiss, no. 28,436; sandy and clay y spots in limestone gravel- barrens, Boat Harbor. Straits of Belle Isle, Fernald, Wiegand & Long, no. 28,435; dry rocky and gravelly limestone barrens and boggy depressions in limestone barrens, Cape Norman, July 18, 1925, Wiegand, Griscom & oo nos. 28,433, 28,434, August 13, 1925, Wiegand & Long, no. 28,440; dry limestone barrens, ge: Cape, Pistolet Bay, July 17, 1925, Fernald, W tegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 28, 432, August 5, 1925, F ernald é ‘aa no. 28,439 Although I have not seen the Drummond material upon which Hooker founded Braya alpina, 8. americana, there is scarcely any ques- tion that the plant of northern Newfoundland with small pinkish flowers and hirtellous siliques is identical with the Drummond plant. After detailed study I placed the Newfoundland material close to B. alpina Sternb. & Hoppe of the European alps. From that species it differs in its narrower leaves, mostly shorter pedicels and sepals, much narrower petals with rounded rather than emarginate apex, slender and columnar rather than tumid style, broader and sharply defined stigma, and shorter lance-subulate rather than linear-cylindric siliques. The only material at hand from the region whence m- mond collected the type of B. alpina, B. americana is Stewardson Brown, no. 1457. That material is so close to the Newfoundland plant that at least until one knows more about it, it would be quite unwise to attempt to separate them. Brown’s material is in anthesis, Drummond’s was in fruit; but the one point emphasized by Hooker, “stylo longiusculo gracili’”’, is clearly shown in it. O. E. Schulz! places B. alpina, 8. americana with doubt under the Siberian B. siliquosa Bunge; but, as he clearly states, Bunge’s species has linear siliques up to 1.5 em. long. These are well shown by Trautvetter,? who also illustrates the leaves as broadly oblanceolate and definitely toothed. The American plant with its more slender and strictly entire leaves and its very short siliques is certainly not B. siliquosa. See pp. 96, 97, 98, 106. ? Braya Ricnwarpsonn (Rydb.) Fern. Sandy and ayes spots in limestone gravel-barrens, Boat Harbor, Fernald, Wieg ’ no. 28,441, very scarce, only two sterile plants found, rsa apparently the Canadian Mt. s See p. 107. 10. ie Schule fi Engice. Pflanzenr. iv. 231 (1924). ? Trautv. Imag. Descr. Fl. Ross. 34, t. 23 (1844 844). (To be continued.) 210 Rhodora [NovEMBER TWO SUMMERS OF BOTANIZING IN NEWFOUNDLAND. M. L. Ferna.p. (Continued from p. 204.) TitaEa aquatica L. Sandy and peaty margin of pond back of barrier beach, Argentia, F apes Long & Dunbar, no. 26,737, new to Nenfocadiaan! AXIFRAGA RIVULARIS L. Net quartzite rocks and seepy banks ee a Deer Pond Brook, Highlands of St. John, Fernald & 0. 28,453; wet mossy cliffs, lower “seed Pond Brook, Gilbert & Hoth, no. 28,455; cool damp pockets in rocky crests, Cape t, Quirpon Island, Fernald & Long; no. "98,454. See pp. 116, 121. Previously reported vey from St. so — S. STELLARIS L., var. comosa Poir. Cres wet scab cliffs along upper Deer Pond Brisk: Highlands a = 7 d & Gye no. 28,464, new to Newfoundland. See p. VASS) ARNASSIA Korzesur C. Mossy teins and damp turfy slopes, Sacred Island, F Hers & — no. 28,475; mossy and turfy cits and » Anse ai uvages, Pistolet Bay, F dare 1926} Fernald,—Botanizing in Newfoundland 211 Hill, Highlands of St. John, rbeleees & Long, no. 28,477; new to Newfoundland. See pp. 122, 123, P. PARVIFLORA DC. Extending ae on the West Coast to St. Barbe Bay. See _ 79. PaRNASSIA multis ta (Ledeb.),n. comb. P. palustris, 8. multiseta Ledeb. Fl. Ross. i. 363 (1842). Ledebour’s only character for his P. palustris, 8. multiseta was the numerous (15-23) setae or filaments to each staminodial scale. Ledebour cited two specimens, one collected by Frisch near Baschu- row on the Argun River, the other collected by Rieder in Kamtchatka. Material of Rieder’s collection is in the Gray Herbarium and in its petals as well as in the staminodial scales it is quite like the American plant which passes as P. palustris but unlike true P. palustris of Europe and much of Asia. The contrasts between the two are indicated below. P. patusrris L. Calyx-lobes igtenslgr tins lanceolate, ae or narrowly obovate, obtuse, 4-7 m ong, 7—-13-nerved: petals rounded or a at tip, iicdonialy 7-13-nerved, with brown- ish lines: anthers 2.3-3 mm. long: staminodial scale longer than broad, abruptly contracted to a long narrow claw, with 9-15 filaments: capsule with a short slender beak: seeds linear-oblong, dark-brown. P. muutisera (Ledeb.) Fern. Calyx-lobes subherbaceous, linear to poe ens subacute, (4-) 6-11 mm. long, 3-7-nerved: petals narrowed to rounded, only rarely emarginate at tip, faintly 5-9- nerved, with tre lines: anthers 1.5-2.3 mm. long: staminodial scale as broad as or broader than long, tapering to a short broad base, with cat filaments: capsule beéakless: seeds short-oblong to oval, borough, no. 14,280. NEWFOUNDLAND: sion des on slaty hills back of Little Quirpon, Fernald & Gilbert, 28,481; mars x of strand near Isthmus Cove, Pistolet Bay, W gan Gilbert & Hotch- kiss, no. 28,480; sandy, gravelly and turfy u border of beach, Cook Boni Pistolet Bay, Fernald, Gilbert & ‘Hotchkiss, no. 28,482; H ahi 4 no. 28,483. Onrarto: Maitland, Huron Co, 1800, W. 212 Rhodora [NovEMBER White; Pic River, Loring; clay soil, Onaman River, Thunder Bay Distr., 1912, Pulling, Micuican: Isle Royale, Cooper, no. 122. Maniropa: Churchill, J. M. Macoun, no. 79,233. MInnEsoTa: Turtle August, 1892, Sheldon; Clithwall, July, 1897, Campbell. Norta Dakota: boggy meadows, Devil’s Lake, July 30, 1838, Nicollet; low meadows, Turtle Mountains, July 23, 1902, Lumell. MACKENZIE: Slave River, July 30, 1892, Elizabeth Ta ylor. Sas- KATCHEWAN: without locality, 1858, Bourgeau; 20 miles west of Yorkton, Herriot, no. 70,747. A.Berta: Milk River, John Macoun, no. 10,569; Banff, Canby, no. 85. British Cotumpra: Rocky Mts. lat. 49° N., 1861, Lyall. YuxKon: Whitehorse, Eastwood, no, 624; Dawson, 1896, Kusche: Carmacks, Eastwood, no. 570. ALASKA: along river, Skagway, Eastwood, no. 771; Kodiak, Kellogg, no. 226; et Island, J. M. Macoun, no. 86; lower Kuskakwim River, August, 8, F. C. Hinckley; bank of Anvik River, Chapman, no. 22; near F oa ‘Selkirk, Yukon River, Schwatka, no. 99; Fort Yukon, July 13, 1902, . Brooks; Cape Blossom, 1884, McLe enegan; Cape Nome, 1900, Bled sdell; St. Michaels, ee £6, Bannister; Kotzebue Sound, Chamisso; Eschscholtz Bay, Chamis. KamrcuatKa: without local- ity, Rieder (co-type of P. pelea 8. multiseta); Bering Island, 1883, Steineger. IBBALDIA PROCUMBENS L. Wet quartzite rocks and gravel along brook, Southwest Gulch, Highlands of St. John, Fernald, Wiegand, reg Gilbert & Hotchiiss, no. 28,505, new to Newfoundland. See Porentiiia pecttnata Raf. Dry limestone cliffs, western face of Doctor Hill, Highlands of St. John, Fernald & Long, no. 28, Li oo known i in Newfoundland only at Chimney Cove. See p. P. nrvea L. Typical P. nivea with the leaflets griseous-sericeous above: dry rocky and gravelly limestone barrens, Cape Norman, on western slope, under summit, Bard Harbor Hill, Highlands of St. John, Wiegand, Gilbert & Hotchkiss, no. 28,526; dry white limestone bluff opposite western escarpment of Bard Harbor Hill, Fernald & Long, no. 28,533. Ar aa collected on Pointe Riche and on Cape St. George. See p. 117. P. NIVEA, var. MACROPHYLLA Ser. The commoner American variety, with dark-green upper leaf-surfaces. Dominant on calcare- ous cliffs, slopes and barrens from Sacred Bay to St. John Bay. Erevionsty collected on Table Mt., Port-d-Port Bay. p. 125. ustica, n. Sp., ‘caudex crassus multiceps caudicu- ae floriferis_ pecetben tibus subscaposis 1-9 em. longis 1-3- floris villoso-lanatis; foliis radicalibus icv petiolatis 1-2.5, em. longis uni- vel bijugis; s se era infimis flavido-brunneis aliquando 1926] Fernald,—Botanizing in Newfoundland 213 purpurascentibus scariosis glabris breviter adnatis auriculis anguste ovatis; foliolis superioribus duobus inferioribus multo majoribus cuneato-obovatis profunde incisis, segmentis utrinque 2-3 lineari- oblongis obtusis margine revoluto subtus albis densissime villoso- lanatis vel tomentosis supra densissime villoso-lanatis vel glabres- centibus; calycibus villoso-lanatis 0.6-1 em. latis 5-6 mm. longis, sepalis externis oblongis obtusis brevioribus quam internis anguste ovatis subacutis; petalis rotundato-obovatis apice truncatis vel emarginatis — Vix longioribu us ochroleucis; staminibus 20, fila- mentis vix 1 mm. longis, antheris subrotundatis 0.4 mm. longis; PO RIE ina —NeEw OUN NDLAND: in ee of dry limestone barrens, northern half of erat Cape, Pistolet Bay, July 17, 1925, Fernald, Wie gand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 28,518 (ryPE i in Gray Herb.); dry gravel of limestone barrens, southern half of Burnt Cape, August 5, 1995, Fernald & Long, no. 28, Very close to the arctic P. pulchella R. Br. and especially to var. Sommerfeltii (Lehm.) Th. Wolf (P. Sommerfeltii Lehm.). But those ants, which Wolf and the students of the floras of Spitzbergen and of Greenland merge as variants of one species, are characterisized by long straight lustrous sericeous pubescence of leaves, stems and calyx, the pubescence of P. usticapensis being opaque and densely villous-lanate. P. pulchella, furthermore, has the segments of the leaf- lets narrower and more acute than in P. usticapensis; its sepals (both inner and outer) less obtuse; its petals yellow rather than creamy and its style-bases more papillose. On Burnt Cape the new species was associated in the dry limestone gravel with a remarkable series of highly localized species, including Festuca supina Schur, Poa alpina var. brevifolia Gaudin, Carex glacialis Mackenzie, Arenaria cylindrocarpa Fern., Lesquerella arctica (Wormsk.) Wats., Braya americana (Hook.) Fern. and Crepis nana Richardson; while in more turfy or peaty spots were Carex concinna R. Br., enaria straminea Fern., Astragalus alpinus L. and A. eucosmus Robinson, Oxy- tropis foliolosa Hook., Arctostaphylos rubra (Rehder & Wilson) Fern., Antennaria eucosma Fernald & Wieg., A. albicans Fern., A. spathulata var. continentis Fern. & St. John, and local species of Arnica and Taraxacum. See p. 103. PoreEnTILLA NorvEGICcA L., labradorica (Lehm.), n. comb. P. labradorica Lehm., Ind. Sem. Hee. Bot. Hamb. 1849: 12 (1849). P. monspeliensis, var. labradorica (Lehm.) Fernald, Ruopora, x. 50 (1908). P. norvegica, var. risen forma labradorica (Lehm.) Wolf, 214 Rhodora [NovEMBER Mon. Gatt. Pot. 404 (1908).—A fairly marked geographic variety. In Newfoundland known only from turfy shore, Big Brook, Fernald & Long, no. 28,536. P. norvegica and P. monspeliensis were published by Linnaeus on the same page. They were maintained as varieties of P. norvegica by Lehmann in 1856 and it was apparently not until 1898 that they were united (by Rydberg) under the name P. monspeliensis. By the International Rules the name P. norvegica must stand. P. atpestris Hall. f. Dominant on peaty or turfy calcareous slopes and eg: from Sacred Bay to Ingornachoix Bay; also on Cape St. George (Mackenzie & Griscom). See pp. 91, 98 ibe in; vitae L., var. VESTITA Area Fernald & Wiegand. Wet quartzite rocks and gravel along brook, Southwest Gulch, HA paren of St. John, Fernald, W Sel Long, ‘Gilbert & Hotchkiss, 28,580, the second tele in Newfoundland. See pp. S eege GALUs stragulus, sp., depressus, caulibus filiformibus ramosissimis repentibus irdirulos m. diametro formantibus; ramulis filiformibus adscendentibus 0.3-1.7 ee longis glabris vel sparse strigosis; stipulis oblongo-lanceolatis vel ovatis 2-4 mm. longis; foliis petiolatis divergentibus, petiolis rhachibusque compresso filiformibus strigosis 2-6 cm. longis, foliolis 4~9-jugis ellipticis retusis 0.2-1 em. longis 1-5 mm. latis glabris vel subtus strigosis; pedunculis filiformibus axillaribus 0.2-9 cm. longis glabris vel strigosis; floribus lilacinis patentibus 1-10, racemo sub anthesi congesto fructifero plus minusve laxo 1~4 em. longo; bracteis subscariosis lanceolato-attenuatis pe strigosis dentibus anguste deltoideis vel lanceolatis subulato-attenuatis brevioribus 1-2 mm. longis, tubo oblique campanulato 2.5-3 mm. longo; corollis 0.8-1.2 em. longis; vexillo rotundato-obovato obcordato culata, auricula oblonga obtusa 1 mm. longa; carinis 9-10 mm. longis lamina unguem aequante subacutis pars 2 mm. latis; ovario hirtello, legumine sessile vesicario oblongo-ovoideo nigro-strigoso 6-8 mm. mgo 2.5-4 mm. diametro uniloculari apiculato, a ag erecto; seminibus 5-8 inequilateralibus olivaceis vel brunneis 1.3-1.8 mm longis—NEWFOUNDLAND: sandy and turfy upper “aa of limestone , Cook oat Pistolet ay July 18, 1925, Fernald & Gilbert, no. 28,583; August 13, 1925, Fernald, Gilbert & Hotchkiss, no. 28,584 (ryPE in Gray He erb.). A beautiful species of the subgenus Homalobus, forming dense carpets and very attractive when in flower, the peduncles varying from only a few millimeters to nearly a decimeter in length, and » when well developed, close racemes of lilac flowers. A. 1926] Fernald,—Botanizing in Newfoundland 215 stragulus is closely related to A. yukonis Jones, Revis. N. A. Astrag. 89 (1923), a species of similar habit from the Yukon. Jones’s des- cription is very contradictory, he stating in the key that A. yukonis has “Leaves sessile, very many. Leaflets 6-8 pairs,” but in the detailed description “Leaves. . . all rather long-petioled Leaflets 3-5 pairs.” Fortunately one of the numbers he dies, Eastwood, no. 626, is before me and this differs in many details from A. stragulus. Its leaflets are fewer, in Miss Eastwood’s material ranging from 3-6 pairs, narrower, cinereous and only exceptionally retuse (in A. stragulus 4-9 pairs, retuse, and glabrous above); its peduncles are all elongate, 0.5-1 dm. long (in A. stragulus very variable, 0.2-9 em. long); the racemes promptly becoming lax and remotely flowered (in A. stragulus only the lowest flowers becoming slightly remote in age); the flowers smaller and strongly ascending (the larger flowers of A. stragulus widely divergent); the standard only about 5 mm. long and with strongly revolute margins (the flat and broader standard of A. stragulus 0.8-1 cm. long); the keel much shorter but with a longer and obtuse erect tip; the pods said to be 2-3-seeded, the tip not strongly bent upward (in A. stragulus the 5-8-seeded pod with the point abruptly upturned and erect). See p. 106. AareiGhine atpinus L. Dry limestone barrens, northern half of Burnt Cape, loge Wiegand, Pease, Long, Griscom, Gilbert & Ke EucosMuSs Robinson. Dominant on tl slopes and barrens or on slate, trap or limestone talus and gravel from Quirpon Island to Boat Harbor. See p. 104. The corolla is eaute deep-violet but it is, exceptionally white in A. rvcosmus, forma albinus, n. f., petalis lacteis. NEWFOUNDLAND: ee pom at- base of Ha-Ha Mountain, Ha-Ha Bay, July 17, 1925, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,588 (ryPE in Gray Herb.). In typical A. eucosmus the calyx is densely strigose with usually dark hairs, but on the banks of the Exploits River in east-central Newfoundland there occurs an extreme with glabrous calyx. This may be distinguished as A. gucosmus, var. facinorum, n. var., a forma typica recedit tee remotis, calyci cibus glabris. NEWFOUNDLAND: ledges and talus, bank of Exploits River below the falls, Grand Falls, July 3, 216 Rhodora [NovEMBER 1911, Fernald, Wiegand, Bartram & Darlington, no. 5795 (TYPE in Herb.). A. Buaxer Eggleston. Limestone cliffs, ledges and talus, western escarpments of Bard Harbor Hill, Highlands of St. John, Wiega nd, Gilbert & Hotchkiss, no. 28,597; F ernald & Long, no. 28 593; a species of northern New England new to Newfoundland. See p. 117 OxyTroPis FOLIOLOSA Hook. Dry gravelly limestone barrens, northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 28,599; a limestone gravel, Schooner (Brandy) Island, Pease & Long; wi 28,600; sandy and turfy upper rder of limestone beach, Cook Point, Fernald & Gilbert, no. "28,601; the first from east of a Rocky Mts. See pp. 103, 105, 106. O. CAMPESTRIS (L. . JOHANNENSIS Fernald. Not previ- a recognized ae Nsicundiand, though already collected. ewfoundland specimens are as follows. Dry crests of ee cliffs, Anse aux Sauvages, Pistolet Bay, Fernald, Wiegand & Long, no. 28,604; shelves, crests and talus of diorite cliffs, Ha-Ha Mountain oy Ha-Ha Point, Pease & Griscom, no. 28,602, Fernald & Long, no. 603; dry limestone cliffs and talus, western face of Doctor Hill, Highleade of St. John, Fernald & Long, no. 28,605; conglomerate limestone and pr ta sandstone cliffs and ledges, Cow Head, Fernald & Wiegand, no. 3633; high sea-cliffs, Chimney Cove, Wag- horne. See pp. 104, 125. HEDYsaRUM AaLPpInuM L. Dominant on turfy or peaty slopes or on calcareous cliffs and talus the length of the Straits from Quirpon nei S Savage Cove. : 104. Previously known from and from the limestones of Port-a-Port and Cape St. ae H. ALPINUM, var. AMERICANUM Michx. Calcareous cliffs and talus, western escarpments of Doctor Hill and Bard Harbor Hill, Highlands of St. John, Fernald & Long, nos. 28,626, 28,627, the first from north of Bay of Islands. See p. 125. LATHYRUS MARITIMUS (L.) Bigel., var. aLEUTICUS Greene. Gravelly and turfy strand near Isthmus Cove, Pistolet Bay, Wiegand, Gilbert & H , no. 28,631, the first from south of Labrador. 120 CALLITRICHE HETEROPHYLLA Pursh. Pools near the eastern cod of the Straits, from Quirpon — to Sacred Island, an extension north from the Exploits. 120. C. HERMAPHRODITICA L. (C. ole L.). In dead water near tide-limit, East Brook, St. Barbe Bay, Wiegand & cope no. 28,643, the first first station north of the Humber. _ NEMopP. S$ MuCRoNATA (L.) Trel. Local and scarce on slopes of peed Highlands of St. John, our first stations north of Bay of Lak SPICATUM Lam Slopes of the Highlands of St. John, our Sie adaik macs OF Tone Bay. IOLA NEPHROPHYLLA Greene. ei g in damp peaty soil, northward to the Straits. See p. 90. 1926} - Fernald,—Botanizing in Newfoundland 217 V. SELKIRKII Pursh. Peaty slopes and woodland knolls, Quirpon Island and Highlands of St. John, extensions northward from Bay of Islands. V. patustris L. Swale at mouth of Eddies Cove Brook, Fernald, Wiegand & Long, no. 28,662, the first Newfoundland station. See p. 108. V. rentroutia Gray. Thickets bordering limestone barrens, southern half of Burnt Cape, Fernald & Long, nos, 28,666, 28,667, the first from Newfoundland, where var. BRAINERDII (Greene) Fern. is generally distributed along the West Coast. EPILoBIUM ANGUSTIFOLIUM L., var. INTERMEDIUM (Wormsk.) Fern. Turfy clearings and pastures, Quirpon Island, Fernald & Long, no. 28,676; previously known in Newfoundland only from Grand Lake E. nesopuitum Fern. Mossy larch swamp on flat north of Doctor Hill, St. John Bay, Fernald & spore = . 28,684, northern exten- sion from the Exploits Valley. See p LAIEANUM Fern. RHopora, se se (1925). Widely dis- persed in southern Newfoundland. For citation of specimens see Fern., 1. c. p. 84. . “PALUST E L., var. MANDJURICUM Hausskn. Springy swamp, limestone strand, Bard Harbor, St. John Bay, Fernald & Long, no. 28,696, Wiegand ‘& Gilbert, no. 28,699; the first records from America. See eae E. USTRE,. var. LABRADORICUM Hausskn. Mossy barren meh side hak of Cape Ray, Fernald, Long & Dunbar, no. 26,856; w slaty cliffs, John Kanes’ s Ladder, western face of Doctor Hill, itigh- lands of St. John, Fernald & Long, no. 28,686a; previously known i in aden raare only from tundra at Little River, Mackenzie & Griscom, no. 11, E. P ronae var. LAPPoNIcuM Wahlenb. Dripping mossy quart- zite cliffs and lates upper Des? Pond Brook, Highlands of St. John, Fernald & Long, nos. 28,697, 28,698; springy ditch bordering spruce thicket, Bett Cove, Fi ernald & Long, no. 28,701; new to Newfound- p. 116. E. wyom1ncensE A. Nels. Swales, bogs and glades on limestones of the West Coast: Flower Cove, Fernald, Long & Dunbar, nos. 26,857. 26,859; Brig Bay, Fernald, Long ‘é Dunbar, no. 26,858 Ingornachoix Bay, Fernald & Wiegand, no. 3728; Cow Head, Feranld & Wiegand, no. 3727; Cape St. George, Mackenzie & Griscom, no. 10,355; the first records from east of the Rocky Mts. See p. 60. E. pavuricum Fisch. Peaty and turfy brooksides, borders ‘of pools or damp depressions in limestone barrens, Quirpon Island to St. John Island: Quirpon Island, Fernald & Long, n 0. 28,708; Big Brook, Pease & Griscom, no. 28,703; Sandy (Poverty) Cove: "Fernald, Long & 218 Rhodora [NovEMBER Gilbert, no. 28,706; Savage Cove, Fernald, Pease & Long, no. 28,704; St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,707; the first from eastern America. See pp. 100, E. LEPTOCARPUM Hausskn., var. (?) Macounit — " Gravelly shore of pond in limestone rock-berrens near Rock Marsh, Flower Cove, Fernald, Pease & Long, no. 28,705; wet peat alee trail in coniferous woods, lower southwestern slope of Bard Harbor Hill, Highlands of St. John, Fernald, Wiegand, Long, ed & Hotchkiss, no. 28,687; the first in eastern America. See E. BREVISTYLUM Barbey. Springy slopes and ee at head of Mauve Bay, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. “ eas springy ditch ee — thicket, Savage vine a ernald 28,713; the firs eastern America. See p. 1 E. BoREALE Hausskn. Springs ditch bordering spruce prac are Savage Cove, Fernald & Long, no. 28,712; glades and rich thickets, ao ay! of Bard Harbor Hill, Highlands of St. John, Fernald & Long, . 28,719, 28,722, 28,723; bushy swale on flat north of Doctor of Labrador and Tabletop Mts., Gaspé. See pp. 99, 108. EpiLopium scalare, n. sp., perenne, rhizom oe elongato obliquo crassiusculo, ad collum gemmas bulbiformes sessiles hypogaeas obovoideas 5-10 mm. longas edente; caule e basi arcuata erecto 0.5- 3.5 dm. alee crasso simplici vel plerum mque in parte media superiore ramoso, ramis revibus arcte adscendentibus fastigiatis, glabro subdense foliato insttaBors: lineis 2 elevatis pilosiusculis e marginibus foliorum decurrentibus notato; foliis inferioribus oppositis reliquis alternatis subapproximatis pallide viridibus subearnosis utrinque angustatis subacutis mediis 3-6 cm. longis 1.2-2.5 cm. latis margine denticulis minimis callosis vix repandis mts _— anguste ovoideis subacutis basi sparse pilosis; floribus erectis 5-6 mm. longis; calycis laciniis ovato-lanceolatis acutiusculis ike Taooonans petalis purpureis calyce 3 longioribus; stigmate clavato; capsulis junioribus pirsiaipe a8 sabes tis deinde 4-6 cm. longis pallide viridibus, pedi- cellis brevissimis 2-3 mm. longis; seminibus anguste ovoideo-oblongis 1.4 mm. longis, 0.4 mm. latis, testa tenuiter denseque papillosa.— NEWFOUNDLAND: wet slaty cliffs, John Kanes’s Ladder, western face 1926] Fernald,—Botanizing in Newfoundland 219 of Doctor Hill, Highlands of St. John, August 24, 1925, Fernald & Long, no. 28, 728 (ryPE in Gray Herb.). Nearest related to E. saximontanum Hausskn. of the cordilleran re- gion of northwestern America. That, however, as described - by Haussknecht, has a slender rhizome (in E. scalare rather thick and stout); simple stem “a span” high (in E. scalare fastigiate or bushy- branched and taller); remote leaves only 2-3 cm. long and at most 1 em. wide (in E. scalare 3-6 cm. long, 1.2-2.5 em. wide); the buds and calices glabrous (in E. scalare crisp-pilose), the petals pale (in E. scalare purple); and the young capsules glabrous (in EF. scalare crisp- pilose). The seeds of E. scalare are apparently very similar to those of E. saximontanum. The latter species is not well represented in American herbaria, and I have been forced to interpret it merely from Haussknecht’s description. Epilobium scalare may eventually prove to be identical with E. saximontanum or some species of northwestern America; but I have been unable to place it satisfactorily with any of them. The flora of western Newfoundland is rich in Epilobium—seventeen species and ten varieties thus far known—and these include besides the cir- cumpolar species such western American types as E. wyomingense A. Nelson, E. boreale Hausskn., E. brevistylum Barbey, E. Drum- mondii Hausskn., E. leptocarpum var. Macounii Trel., and E. Behringi- anum Hausskn. See p. 126. E. tactrFLoruM Hausskn. Brooksides on slaty hills back of Little Quirpon, Fernald & Gilbert, no. 28,730; mossy brooksides, Sacred Island, Fernald & Long, no. ‘98, 731; cold wet chimneys in trap — Anse aux Sauvages, Pistolet Bay, Fernald, Wiegand 9; 28,732; wet quartzite rocks and seepy banks along upper Deer Pond Brook ‘and in Southwest Gulch, Highlands of St. John, Fe & Long, no. 28,739, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. g , no. and adjacent Alaska, here found for the first time in eastern America See ares 109. E. inuM L. (E. anagallidifolium Lam.). Glades and damp ‘packets northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, wt & Hotchkiss, no. POS, 736, new to Newfoundland. See p. 103 MyRrIoPHYLLUM ALTERNIFLORUM DC. Frequent from Brig Bay to Savage Cove, an extension northward from Bay of Islends. See p. 127. 220 Rhodora [NovEMBER M. EXALBESCENS Fern. Common from Brig Bay to Pistolet Bay, an extension north from Pointe Riche. See p. 127 SANICULA MARILANDICA L., var. borealis, n. var. a var. typica recedit floribus stamineis majoribus sepalis 1.3-2 mm. longis pedicellis crassis subereo-angulatis vel -clavatis; fructu 6-8 mm. longo.— Newfoundland to northern New York. NEWFOUNDLAND: spruce thickets, Boat Harbor, Straits of Belle Isle, July 19, 1925, F ernald, Wiegan Long, no. 28,765; meadow below limestone escarpment, western face of Bard Harbor Hill, Highlands of St. John, August 27, 1925, Fernald & Long, no. 28,766; damp talus of limestone sea-cliffs, Pointe Riche, August 4, 1910, Fernald & Wiegand, no. 3765; wet runs and boggy spots, Ingornachoix Bay, August 2, 1910, F ernald & Wieg- and, no. 3764; marsh at base of Gros Morne, Bonne Bay, August 13 and 14, 1919, R. H. Kimball, no. 74; park-like openings in woods, Birchy Cove (Curling), July 21, 1910, Fernald & Wiegand, no. 3763; woods, Wild Cove, July 16, 1896, Waghorne; banks of Exploits River near Badger Brook, August 13, 1894, B. L. Robinson; ledges and talus, Grand Falls, July 22, 1911, Fernald, Wiegand, Bartram & Darlington, no. 5950; shaded bank of Salmon River, Glenwood, July 12, 1911, Fernald, Wiegand & ington, no. 5951; damp bushy ravine 1n limestone tableland, Table Mountain, Port-a-Port Bay, August 16, 910, Fernald iegand, no. 3 EBEC: alluvial woods, Riviére-aux-Rénards, Juil. 14, 1923, Victorin et al, no. 17,318; coni- erous forest, “Low’s Trail” from River Ste. Anne des Monts to 1903, Pease, no. 2260; Little Island, Shediac Bay, August 5, 1914, Hubbard. Mane: Fort Fairfield, July 16, 1902, Williams, Collins Fernald. New York: Canton, July 17, 1914, Phelps, no. 726. In typical Sanicula marilandica of the eastern United States and the southern edges of Canada the staminate flowers are smaller, with sepals 1-1.5 mm. long, the pedicels filiform and usually more elongate than in the extreme, and the fruits average slightly shorter, varying from 5-7 mm. in length. Var. borealis usually has a tendency 1926] Fernald,—Botanizing in Newfoundland 221 to firmer or somewhat fleshy or leathery leaves; but in none of the characters is it absolutely separable. In northern New England some of the plants are transitional, but all the material from farther north, in Gaspé and Newfoundland, seems fairly constant in its thickened pedicels and larger calyx and, when mature, in its usually larger fruit. OsMORHIZA DIVARICATA Nutt. Alluvial woods north of Doctor Hill, St. John Bay, Fernald & Long, no. 28,771, new to Newfound- land. See p. 125. CICUTA BULBIFERA L. Locally in swales along the Straits ary Flower Cove to Boat Harbor, an extension north from Bay of Isla Heracleum Sphondylium L. A dominant weed of fields and fenee: rows, Trepassey, Fernald, Long & Dunbar, no. 26,886. P wg . 84. ContosELinuM GMELINI (Bray) Steud. Nom. ed. 2, i. 403 (1840). Selinum Gmelini Bray, Denkschr. Bot. Ges. Regensb. i, pt. 2: 36 (1818) Certain collections from Newfoundland and the Gaspé Peninsula, with the upper stipules gradually narrowed from base to summit, rather than broadly dilated as in C. chinense (L.) BSP., and with the leaves very finely dissected, seem to be quite inseparable from the plant of arctic Eurasia which there regularly passes as C. Gmelini (see p. 117). C. Gmelini (Bray) Steud. is quite distinct from the plant of western North America which was christened C. Gmelini (C. & S.) Coult. & Rose, Cont. U. S. Nat. Herb. vii. 150 (1900), based upon Ligusticum Gmelini Cham. & Schl. Linnaea, i. 391 (1826). In view of Steudel’s combination, properly made in a well-known work and clearly listed in Index Kewensis, there was no call for the later combination. The plant of western America, which has larger fruits and much longer and coarser involucels than C. Gmelini (Bray) Steud., should be called ContosELinum Benthami (Wats.),n. comb. Selinum Benthami Wats. Bibl. Ind. 432 (1878). Ligusticwum Gmelint Cham. = be Linnaea, i. 391 (1826). C. Gmelini Coult. & Rose, Contrib. U. S. Nat. Herb. vii. 150 (1900), not Steud. (1840). _ The specific characters in Conioselinum are not well worked out but everyone who has watched the genus in the region of Gaspé and Newfoundland realizes that with us it is a complex group. In this region there seem to be at least four species: C. chinense (L.) BSP., C. Gmelini (Bray) Steud., C. pumilum Rose, and a fourth species from a single collection from Tabletop Mts., Gaspé) with 222 - Rhodora [NOVEMBER coarse and elongate involucels, near to if not identical with C. Benth- ami (Wats.) Fern. of Alaska; but until more time can be given to study of their technical characters it is wisest to defer the presenta- tion of a formal key to them. ANGELICA laurentiana, n. sp., A. atropurpureae similis; caulibus sphericis primariis 1.3-2.5 dm. latis radiis 28-45 subrigidis crassis dense griseo-puberulis floribus exclusis 5-10 ¢ m. longis; involucellis nullis vel bracteis lineari-setaceis vel Gasariblaricebinte pedicellos basi apiceque rotundato 5.5-10.5 mm. longo glabro; jugis tribus dorsalibu8 alatis, alis dorsalibus laterales subaequantibus. —North- western Newfoundland and eastern Saguenay County, Quebec. NEWFOUNDLAND: trap cliffs and talus, Anse aux Savio. Pistolet nie August 11, 1925, Fernald, Wiegand & Long, nos. 28,784 (TYPE ray Herb 5. 28,785; rocky meadows and brook-bottoms, upper Thee Pond ha Highlands of St. John, July 28, 1925, August 20, 1925, Fernald & Long, nos. 28,782, 28,783. QuEBEC: meadow, Tabatiére, eomgeabe August 14, 1915, St. John, no. 90,626 (dis- tributed as Coelopleurum a Angelica laurentiana is at once distinguished in mature fruit from A. atropurpurea L. by the thin wings of the dorsal ridges, the dorsal ridges in A. atropurpurea being low, blunt-edged and wingless. In A. atropurpurea the ripe fruit is 5.5-7.5 mm. long, in A. lawrentiana it may be as short but is commonly 8-10.5 mm. long. In A. atropurpurea the stem is usually suffused with purple and the umbels are inclin to be spherical, with the outer rays strongly reflexed, and the upper segments of the lowest leaves are commonly distinct or only slightly confluent. In A. laurentiana the stem may be purplish but it is more commonly green; the rays of the umbel ascend or are at most horizontally divergent and the upper segments of the lowest leaves are strongly confluent. In A. atropurpurea the involucels, when ian: are commonly much shorter than the pedicels; in A. laurentt- ana they are often nearly as long as or even longer than the pedicels. In its fruit A. laurentiana is nearer to A. ampla A. Nelson of the Rocky Mountain region of Wyoming and Colorado; but the latter has the fruits smaller and with lower scarcely winged dorsal ridges. In their key Coulter & Rose! define A. atropurpurea as having “ In- ‘Contrib. U. 8. Nat. Herb. vii. 154 (1900). 1926] Fernald,—Botanizing in Newfoundland 223 florescence glabrous” and in their detailed description states that the umbel is “15 to 25-rayed.”” Such an account is difficult to reconcile with the plants, for every specimen of A. atropurpurea in the Gray Herbarium and the herbarium of the New England Botanical Club has the rays and pedicels copiously puberulent and the primary or leading umbels all have 22-46 rays.!. See p. 117 PyRoLa SECUNDA L., forma eucycla, n. i a a suet Pag Eire rotudatis vel rotundo-ellipticis. —NEWFOUNDLAND woods north of Doctor Hill, St. John Bay, iicoost 24, 1995, ? Fernald < Long, no. 28,801. Forma eucycla is an extreme form of P. secunda, var. vulgaris Herder, and should not be confused with var. obtusata Turez. Typical P. secunda has the bracts at the base of the stem lanceolate, strongly involute and firm; the firm leaf-blades dark-green and rather lustrous, crenate-serrate, 1.5-6 cm. long; the scapes 1-2 dm. high; racemes 6-20-flowered, in anthesis 2-5, in fruit up to 8.5 cm. long; the petals greenish-yellow; the mature style 5-9 mm. long. The usually very distinct var. obtusata differs in having the basal bracts oblong to ovate, but slightly involute, membranaceous; the leaf-blades membra- naceous, pale-green, hardly lustrous, crenate, 0.8-3 cm. long; the scapes 0.5-1.5 dm. high; the racemes 2-10-flowered, in anthesis 0.5-3, in fruit up to 4 em. long; petals creamy-white; mature style 4-6 mm. long. In all its technical points forma eucycla belongs with typical P. secunda, which, however, has the leaves narrowed at tip. RHODODENDRON CANADENSE (L.) BSP. Shrubby barren near the pimp Marsh back of Bard Harbor, St. John Bay, Gilbert & Hotchkiss, . 28,824, an extension north from Bay of Is fe nds. pada ODOCE CAERULEA (L.) Bab. Abundant on peaty and turfy quartzite, upper slopes of Bard Harbor - Highlands of St. John, the third Nefoundiac! station. See p. CassioPe uypnoipeEs (L.) D. Don. Pury and mossy quartzite rocks along Mans Humbug Brook, Highlands of St. John, very searce, Fernald & Long, no. 28,839, the second Newfoundland station. See p. 124, where I spoke of it as new to — overlooking Eames & Godfrey’ s discovery of it on Blomi ae ae RUBRA (Rehder & ena ‘Fernald. Turfy or peaty limestone barrens along the Straits: Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 28,853, 1 The first station cited by Coulter & Rose is likewise difficult to reconcile with ‘Vermont: Alpine region of the White Mountains, Oakes. 224 Rhodora [NovEMBER Fernald & Long, no. 28,855; Cape Norman, Wiegand, Griscom & — no. 28,854, Wiegand & Long, no. 28,856; Big Brook, Fernald & Long, no. 28,852; Yankee Point, Zs yee, Wiegand & Long, no. 28,857; new to Newfo undland. See VAccINIUM NUBIGENUM Fernald. Dry s nae at margin of quart- zite barren near summit of Bard Harbor Hill and heathy banks near upper Deer Pond Brook, Highlands of St. John, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,872, Fernald & Long, nos. 28,873, 28,874; previously known only from the Shickshock Mts. See pp 116, 124. V. ovaLiFoLiuM Sm. Open spruce thickets, Boat Harbor, Fernald, Wiegand & Long, no. 28, 875; abundant in the thickets and gulches of - the “Highlands of St. John. See pp. 107, 116, 124. Previously known on the North Peninsula only from Croque (Banks, LaPylaie). ATICE LABRADORICA (Wallr.) Hubb. & Blake, var. SUBMUTICA Blake. Dominant on limestone barrens from Pistolet Bay to St. John Bay; previously known on the West Coast south to Lewis Hills. FARINOSA L. Fairly typical P. farinosa is represented by specimens from crests of trap cliffs, eb Island, Wiegand, Gilbert & Hotchkiss, no. 28,908; turfy and peaty pockets in limestone ledges, Sandy (Poverty) Cove, F ernald, Resi 4 Dunbar, no. 26,955; lower Cove, M. E. Priest, no. H1. See pp. P. FARI Mou: var. AMERICANA Torr. Wet limestone ledges and as iscom, Gilbert & Hotchkiss. no. 28,905; gravelly limestone shore, Schooner (Brandy) Island, Pease & Long g, no. 28,907; aap barrens, Yankee Point, Wieg and, & Hotchkiss no. 28,904; limestone barrens, Dog Peninsula, F. ernald, Wiegand, Long, Gilbert t Hotchkiss, no. 28 913. . FARINOSA, var. MACROPODA Fern. Abundant on the West ? SIBIRICA Jacq., var. ARCTICA Pax. Gravelly and peaty lime- stone barrens back of Big 2 Fernald & Long, no. 28,915; sandy and gravelly margin of Big Brook, Fernald, Wiegand, Long & & Gilbert, no. 28,916; swale near mouth of pessael Watts Bight, Pease, Griscom, Gilbert & Hotchkiss, no. 28,901; wet banks, turfy limestone barrens, Cape Norman, cit Griscom & Hotchkiss, no. 28,918; wet hollows pond-shores in Ren ep gravel-barrens, Cook Point, Fe & Gilbert, no. 28,917, Fernald, Gilbert & Hotchkiss, no. 28,919; the first records fro m eastern America. See pp. 98, 99, 105. P. EGaLIksensis Wormsk. Abundant in swales or on peat in the us from Sacred Island to St. John’s Island. See pp- a 60, 75, 118. GENTIANA PROPINQUA Richardson. Dominant on turfy and 1926] Fernald,—Botanizing in Newfoundland 225 gravelly limestone shores and cliffs from Quirpon Island to Ingorna- choix Bay. See pp. 54, 60, 79, 80. yee mium purpureum < Completely occupying a garden-site at head of Mauve Bay, Fernald, ee Long, Gilbert & Hotchkiss, no. 28,958, new to Newfoundla LIMOSELLA AQUATICA L. Shallow fresh pool back of beach, Harbour wat Fernald, Long & Dunbar, no. 27,013. Otherwise known in eastern America only from the southeastern corner of the Labrador Peninsula. See GRATIOLA AUREA Muhl. Gravelly margin of Junction Pond, Whitbourne, Fernald, Long & Dunbar, no. 27,014; discovered at the same station in 1911, Fernald & Wiegand, no. 6143; the only eae own in Newfoundland. 86. e material collect August 8, 1911, and August 25, 1924, is very dwarf and shows no sign of flowerin g. (To be continued) 234 Rhodora [DECEMBER TWO SUMMERS OF BOTANIZING IN NEWFOUNDLAND M. L. FERNALD. (Continued from p. 225.) VERONICA ALPINA L., var. UNALASCENSIS C. & 5. Wet quartzite rocks and gravel along brook, Southwest Gulch, Highlands of St. John, Fernald, Wiegand, eee? Gilbert & Hotchkiss, no. 28,965, new to Newfoundland. See V. numirusa Dickson. Wet limestone barrens, slopes and cliffs, general on the West Coas i Eup The collections of the past two seasons render a complete revision of the American species necessary. This cannot now be presen Bartsta aLpina L. Dam p turf or peaty limestone barrens or talus from Pistolet Bay seine to Half-way Brook; new to New- foundland. See pp. 105, 106 CULARIS FLAMMEA barrens one mile back of Savage Cove, Fernald & Long, no. 0,034; pt and a limestone barrens back of Big roe Fernald had 1926] Fernald,—Botanizing in Newfoundland 235 com, Gilbert & Hotchkiss, no. 29,036; new to Newfoundland. See pp. 96 i PALUSTRIS L. Long known from the region of St. John’s; seen from the train in ia and sloughs southward to Petty Harbor. - p . syLvaTica L. Collected for the first time in America in 1911 by F ernald & Wiegand. Characteristic of peaty s slopes and gee woodland glades Da Carbonear to St. John’s and vicinity. See UTRICULARIA GEMINISCAPA Benj. Shallow peaty pool back of strand, Bard Harbor, St. John Bay, Fernald & Long, no.. 29,042, northward extension from Bay of OROBANCHE (APHYLLON) terrae-novae, n. sp., 0. unifloram simu- lans; calycibus campanulatis 0.7-1.3 cm. longis pilosis, lobis ee deltoideis acuminatis vel caudatis 4-8 mm. longis; corollis 1.7-2.5 em. longis arcuatis lacteis vel pallide Sinciuiay laciniis oblongis valde ie ciliis 0.5 mm. longis; floribus inodoris—NEwFrounpD- LAND: gravelly shore of dead-water in Gander River, Glenwood, July 12 rand 13, 1911, Fernald & Wiegand, no. 6212; springy places in ledges and gravel, north bank of Exploits River, Grand Falls, July 4, 1911, Fernald, Wiegand, Bartram & Darlington, no. 6211; under Picea canadensis, cool springy glade, interior of saithard half of Burnt Cape, Pistolet Bay, August 5, 1925, Fernald & Long, no. 29,051, ‘swampy spruce (Picea canadensis) woods and thickets, Bard arbor, St. John Bay, July 29, 1925, Wiegand, Gilbert & “Hotchkiss, no. 29,049 (TyPE in Gray Herb.); mossy glades in spruce woods, western slope of Bard Harbor Hill, August 26, 1925, Fernald & Long, no. 29,053; mossy glades in spruce woods, north of Doctor Hill, St. John Bay, August 24, 1925, Fernald & Long, no. 29,052; turfy slopes and steep lane St. John’s Island, St. John Bay, July 31,1925, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 29,050; boggy thickets back of Birchy Cove (Curling), July 7, 1910, Fernald & Wiegand, no. 4018; talus slope of the marble region ‘between Mt. Musgave and Humber Mouth, July 18, 1910, Fernald & Wiegand, no. 4019; woods, Coal fe: July 21; 1896, Waghorne; Log Cabin, Harry’s River, July Hi, . L. Dam ecu terrae-novae is perhaps merely a geographic variety, but all Newfoundland material, in good flower, has the oblong segments of the corolla with much longer ciliation than in the continental 0. uniflora L., in which the segments are broader and usually obovate. O. uniflora has deliciously fragrant flowers, O. terrae-novae odorless, and the calyx of the latter is more copiously pilose than in the former. In 0. terrae-novae the lower lip has 2 linear yellow lines at the throat; whether they occur in OQ. uniflora cannot at present be stated. In his very detailed Monographie der Gattung Orobanche, Beck von 236 Rhodora [DecEMBER Mannagetta recognizes only two members of the section Aphyllon. The herbarium-material indicates that the section is much more complex, and certainly the species of Pacific North America with linear-subulate or setaceous calyx-segments and purple-violet corol- las is quite distinct from the eastern O. uniflora. The western plant is OroBancHE (APHYLLON) Sedi (Suksd.), n. comb. Aphyllon Sedi Suksd. Deutsch. Bot. Monatss. xviii. 155 (1900). A. uniflorum, var. occidentale Greene, Man. Bay Reg. 285 (1894). Thalesia purpurea Heller, Bull. Torr. Bot. Cl. xxiv. 313, t. 310 (1897), not O. purpurea Jacq. (4762). A. minutum Suksd. 1. ¢. (1900). not O. minuta Beck von Man. (1890). Thalesia minuta (Guksd.) Rydb. Bull. Torr. Bot. 1909). UM SAXATILE L. Stony margins of spring-rills by roadside aty barrens and about cold spring-head at border of spruce thicket in bog-barrens, Trepassey, Fernald, Long & Dunbar, nos. 27,085, 27,086, its first stations in America. p. 83. Gatium BranDeGEI Gray. Springy slopes and biocksaee at head of Mauve Bay, Fernald, Wiegand, ae olgeadt & Hotchkiss, no. 29,070, new to Newfoundland. See p. Mrrenenia REPENS L. Border of wet oe thicket among gneiss hills, very rare, Port aux Basques, Fernald, Long & Dunbar, = 27,097, the first Lathenetdated record from Newfoundland. See p. 5! EupPAToRIUM MACULATUM L., var. FoLIosum (Fern.) Wiee Meadow below calcareous sandstone escarpments, vasa face of Bard Harbor Hill and meadow by brook, Bard Harbo , St. John Bay, Fernald & Long, no. 29,093, Wiegand & Gilet, no. 29,094, an extension north from ay Si e River See p Sotmpaco LePmipa DC. below Paks escarpment, western face of Bard Hazbor I Tall "Highlands of St. John, Fernald & Long, nos. 29,117, - 118, the second station for the typical form of the species S. GRAMINIFOLIA Po.) Salisb., var. Nurratyiu (Greene) Fern. Dry gravelly slopes near Goose Po nd, Whitbourne, Fernald, Long & Dunbar, no. 27,129. Previously bnows only from Clarenville. RIGERON ACRIS L., var. arcuans, n. nom. EL. alpinus B. Hook. FI. be ae ii. 18 (1834). E. acris, var. debilis Gray, Syn. Fl. i. pt. 2: 220 (1884) as to Labrador and Ha ken n Bay ats only. E. _ pel Pitonia, iii. 164 (1897), not E. alpinus elata Hook. igocephalus Fern: Wiegand, RHoporA, xii. "226 FAvtOy. in Shee part, but not as to type. E. acris, var. oligocephalus was described two quite different plants were confused under it. The type from Blanc Sablon has much. larger heads than the others and later collections from the . its of Belle Isle show it to be specifically — from the smaller-headed plant which Hooker called E. 1926] Fernald,—Botanizing in Newfoundland 237 alpinus 8. and which Greene well described as E. elatus. The latter name is quite inappropriate and, as was shown in the original dis- cussion of E. acris, var. oligocephalus, Greene was in error in identify- ing his plant with E. alpinus, ¥ elata Hook. The type of E. acris, var. oligocephalus belongs, according to Dr. Ostenfeld and Mr. Erling Porsild who have independently been over the Newfoundland collections of 1925 with me, to E. porEauis (Vierh.) Simmons, Acta Univ. Lund., n. s. ix. no. 19: 127 (1913). Outside of Greenland E. borealis is known merica only from LaBrapor: limestone and calcareous aan terraces, Blane Sablon, Fernald & Wisin no. 4138 (type of EF. acris, var. spn hi enka NEWFOUNDLAND: Quirpon Hassoe September 8, 3, A. G. Huntsman; slaty cliffs and talus, Cape Raven, August 12, 5: iho, Wiegand, Long, Gilbert & Hotchkiss, no. 29,139. See pp. ANT fai EUCOSMA Fern. & Wieg. Dominant on turfy lime- stone wert nd slopes from Pistolet Bay westward to Four-Mile Cove; previously known only from Table Mt., Port-a-Port Bay and from Cape St. George. See pp. 98, 105. . CANA (Fern. & Wieg.) Fern. Dominant on gravelly mere ad barrens, Pistolet Bay to Ingornachoix Bay, and more locally to Bay St. George. See pp. 62, 96, 103, 105. A. VEXILLIFERA Fern. urfy limestone barrens, Cook Point, Fernald & Gilbert, no. 29,171; dry mixed gravel on crest southeast of Boat Harbor, Fernald, Wiegand & Long, no. gp N Previously known ei from the Shickshock Mts. See pp. 105, 107 ANTENNARIA Longii n. sp., humifusa, stolonibus foliosis confertis cateeuibhe (ad 1 em. longis); foliis basilaribus late spathulatis vel anguste cuneato-obovatis apice rotundatis vix mucronatis 5-11 mm. longis 2.5-4 mm. latis supra epee: tomento denso minuto; caule florifero 1-6 rarissime deinde —9 cm. alto; foliis caulinis 4-8 subdistantibus, imis oblanceolatis siitaue weds linearibus 0.7-1.5 apice scarioso oblongo-lanceolato 1.5-2 mm. longo munitis; capitulis femineis 1-6 corymbosis turbinato-campanulatis; involucro 7-10 mm. alto basi lanato; bracteis 2-3-seriatis subaequalibus tenuissimis, mascula ignota.—Straits ue ‘Belle Isle, NEwFro : dry lime- stone gravel, Schooner (or Brady) Island, Pistolet Ba Bay, July 18, 1925, Pease & , no. 29,177 (TYPE in Gray H erb.); vol limestone rock-barrens, i gand & Long no. 29,178; gravelly limestone barren, Four-Mile Cove, July 20, 1925, Fernald, Wiegand & Long, no. 29,179; dry gravel of limestone 238 Rhodora [DecEMBER barrens, Cape Norman, August 13, 1925, Wiegand & Long, nos. 29,180, 29,181. A. Longit is nearest related to A. vezillifera Fernald, but is at once distinguished by its larger heads; broader involucral bracts, the outer series with strongly squarrose tips; and the few appendaged cauline leaves. A. vevxillifera, originally described from the Shickshock Mts. of Gaspé but now known from northwestern Newfoundland as well, has the fulvous hemispheric-campanulate involucre only 6-7 mm. long; its uniformly ascending bracts narrower than in A. Longit; and all but the gore cauline leaves tipped by a long scarious ap- pendage. See p. STRAMINEA oS rm. Local on the West Coast, previously known northward to Pointe Riche. Our recent collections are: dry lime- oR barrens, northern half of Burnt Cape, Fernald, Wiegand, Pease, riscom, Gilbert & Hotchkiss, no. 29,168; turfy crests of lime- sone barrens, Shoat Cove, Pease & Griscom, no. 29,176; dry gravelly Gilbert & Hotchkiss, no. 29,175; open spots in limestone barrens near Ice Point, St. Barbe Bay, Wareund Gilbert & Hotchkiss, no. 29,167; dry limestone barrens, Brig Bay, Fernald, Long & Dunbar, no. 27,187; dry gravelly limestone barrens, St. John’s Island, Fernald, Wiegand, Nel Gilbert, & Hotchkiss, no. 29,170. See pp. 79, 97, 105. A. ICANS Fern. Very local on the West Coast, prev iously Go only fron Table Mt., Port-A-Port Bay. The new stations are: dry limestone barrens, northern. half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 29,164; aie pooner e & no. ;, ry. gravel of limestone barren, Cook Point, Fernald & pang no. re dry limestone gravel near Ice Fone St. Barbe Bay, Wiegand, Gilbert & otchkiss, no. 29,161. See p Agee ANTENNARIA Wiegandii, n. Place stolonibus foliosis (ad mediis ae are 7-10 mm. a 1- 1 5 mm. latis apice unguiculato- subulatis, superioribus 4-6 apice searioso lanceolato 2-3 mm. longo orvum basi rotundatis; involucro 6 mm. sh basi lanato; bracteis 2-3- seriatis, exterioribus anguste oblongis obtusis adscendentibus basi castaneis apice fulvis, interioribus lanceolatis fulvis; corolla 4.5 mm. longis apice purpurascentibus; acheaniis 1.2 mm. longis; planta mascula ignota.—NEWFOUNDLAND: turfy limestone barrens, St. John’s Island, St. John n Bay, July 31, 1925, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 29,153. 1926] Fernald,—Botanizing in Newfoundland 239 A. Wiegandit is the nearest beige yet found south of northern Labrador to the arctic A. alpina (L.) Gaertn. In the distinetly mucronate basal leaves it is very similar to A. alpina, all the other small-leaved species of Newfoundland and Gaspé having blunt basal leaves. A. alpina, however, has much narrower and more elongate rosette-leaves which are narrowed to tip; the upper cauline leaves have broader appendages; the corymb, except in var. ungavensis, is denser; the heads are larger (involucre 7-9 mm. high); and the achenes are longer (1.3-1.5 mm. long). In the green upper surfaces of the basal leaves A. Wiegandii re- sembles A. spathulata Fernald, but it is at once distinguished by the smaller leaves with less rounded, but more mucronate tips; more crowded and assurgent stolons; looser inflorescence, with smaller heads (involucres of A. spathulata 7-10 mm. high) with much narrower and darker involucral bracts, and shorter corollas and achenes (in A. spathulata the corollas 5.5-6 mm. long, the achenes 1.4-1.7 mm. long). A, spaTHuLaTA Fern. Turfy limestone barrens, northern pelt of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom & Hotchkiss, no. 29,182; turfy slopes and steep banks bordering id stone barrens, St. John’s Island, Fernald, Wiegand, Long, Gilbert & H — no. 29,183; northern extension from Pointe Riche. See p- re oo var. CONTINENTIS Fern. & St. John. Turfy limestone barrens, northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 29,184; previously known only from Natashquan, se 0., ’ p. 105. ANAPHALIS MARGARITACEA (L.) B. H., var. SUBALPINA Gray. Gravelly bank near sea, Cape Onion, F. anal no. 29,186, a ollie extension from Bonne Bay; the solitary colony locally famous as a curiosity. GNAPHALIUM NORVEGICUM Gunn. Wet quartzite rocks and seepy banks in gulches of Highlands of St. John: Deer Pond Brook, Fernald & Long, nos. 29,187, 29,189; Southwest Gulch, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 29,190. Turfy barrens and slopes, Sacred Island, Wiegand, Gilbert oe Hotchlkiss, no. 29,188. New to Newfoundland. See pp. 116, 1 TANACETUM HURONENSE } Nutt., var. TERRAE-NOVAE Fern ant on turfy or gravelly limestone barrens and slopes from Pistolet Bay to Ingornachoix Bay. See pp. 54, 80, 1 ARTEMISIA BOREALIS Pall. Dry gravelly limestone barrens, St. John’s Island, Fernald, W “tegand, Long, Gilbert & Hotchkiss, no. ; previo ously known in Newfoundland only ee Bi Mt., Port-d-Port Bay and from Cape St. George. See p. 240 Rhodora [DEcEMBER ARNICA TERRAE-NOVAE Fern. Turfy limestone barrens, Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, Gilbert & Hotchkiss, no. 29,209, Fernald & Long, no. 29,210 14; turfy limestone barrens, Cook Point, Fernald & Gilbert, no. 29,214; peaty and bushy areas on limestone barrens one mile back of Savage Cove, Fernald & Long, no. 29,207, Fernald, Pease & Long, no. 29,208; previously known only from Cape St. George. See pp. 96, 103, 105. A. PULCHELLA Fern. Turfy limestone barrens, Cape Norman, Wiegand, Griscom & Hotchkiss, no. 29,213; mossy talus of diorite cliffs, Ha-Ha Mountain, goon ld & Long, no. 29 215: previously known only from Table Mt., Port-a-Port Bay. a pp. 96, 106, 120. A. CHIONOPAPPA Fern. harty limestone barrens, Burnt Cape, Fernald, Wiegand, Pease, Long, Griscom, oe & Hotchkiss, no. 29,209 14, Fernald & Long, no. 99, 210; turfy limestone barrens, St. John’ s Island, Fernald, Wiegand, cs Gilbert & Hotchkiss, no. 29,211; dry white limestone luff oppos nee escarpment of Bard Harbor Hill, Highlands of St. coe Fernald & Long, no. 29,2 previously known in Newfoundland only from Table are Pata ort Bay and from Cape St. George. See pp. 96, 103, 11s. A. Griscomi Fern. Crests of turfy slopes and steep banks border- ing limestone barrens, St. John’s Island, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 29,216; previously known only from the Shickshock Mts., Gaspé. See p. SENECIO PAUCIFLORUS Pursh. Tu oe slopes of slaty hills, Little Quirpon, Wiegand, Gilbert & Hotchkiss, no. 29,218; peaty and turfy brookside east of the central valley of Quirpon Island, Fernald & Long, no. 29,219; slaty cliffs and talus, Cape Raven, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 29,221; mossy brookside, Sacred ome Fernald & Long, no. 29,220; calcareous cliffs and talus, western s of Doctor Hill and Bard Harbor Hill, Highlands of St. John, Fen ern inald & Long, nos. 29,222, ase the first authenticated records from Newfoundland. See pp. Taraxacum. The icine ie of Taraxacum are very numerous in the ancient flora of unglaciated northwestern Newfound- land. Some, like T.Larpontcum Kihlm. (p. 60) and T. cpRATOPHORUM (Ledeb.) DC. (p. 76), are well known circumpolar species; others, like T. Lacerum Greene (p. 60) and T. pumerorum Greene are species previously known only from the Rocky Mts.; T. tarmosum DC. p. 80) is endemic; and several others are apparently local endemics. A series is being studied by Dr. Handel-Mazzetti and a detailed report upon them cannot yet be given. ca (MvLGEDIuM) terrae- , biennis, caule erecto glabro robusto simplici 1 m. alto; foliis sabes ‘runcinato-pinnatifidis elliptico-ovatis, mediis 1.5-2 dm. longis 1-1.2 dm. latis acuminati lobis oblique ovatis horizontaliter divergentibus grosse dentatis, 1926] Fernald,—Botanizing in Newfoundland 241 superioribus apice caudatis; capitulis ee involucris urceolato-campanulatis valde ealyculatis 1.2-1.5 altis, bracteis exterioribus ovatis acutis Seteracbak lanceolatis ie apice arcua- tis; corollis coerulescentibus deinde purpurascentibus; achaeniis oblique lanceolato-oblongis rufescentibus 5-6 mm. longis 1-1. si mm latis in rostrum firmum 1 mm. longum attenuatis; pappo n NEWFOUNDLAND: meadow below limestone escarpment, aes tae of Bard Harbor Hill, Highlands of St. John, August 21, 1925, Fernald & Long, no. 29,290. Lactuca terrae-novae is known from only a single plant. When Mr. Long and I found it we took three specimens, but upon maturing in press two of them prove to be typical L. spicata (Lam.) Hitch., which is frequent in western Newfoundland; but the third specimen departs at once from the others in its white pappus, orange-brown or reddish unmottled achenes and definite slender but firm beak; L. spicata having sordid pappus, grayish-brown to blackish con- spicuously mottled achenes and stouter neck. It may prove, when the new plant is better known, that it is an extreme of L. spicata. If so, the characters of pappus and beak, which have so long proved constant in the genus will have most singularly lost their constancy. No species of Lactuca other than these two are known in Newfound- land; L. terrae-novae cannot, therefore, be considered a hybrid. See p. 124, CrEPIS NANA Richardson. Dry limestone barrens, very scarce, northern half of Burnt Cape, Fernald, Wiegand, Pease, Long, rece Gilbert & Hotchkiss, no. 29,278; prev iously known in eastern only from the Torngat region of northern Labrador. See pp. 103, 104, HIERACIUM GROENLANDICUM Almq. Thicket by Big Brook, Fernald & Long, no. 29,267; thickets on slaty hills bak of Little Quirpon, Fernald & Gilbert, no. 29,268; also noted —* on Burnt Cape; the mature plant from Little eo is a clos match for authentic Greenland material. the Exploits Valley. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY LXXVII im | RECORDS PRELIMINARY TO A GENERAL TREATMENT OF THE EUPATORIEAE,—VI #44864 By B. L. Roprnson PUBLISHED B THE GRAY HERBARIUM OF HARVARD UNIVERSITY CAMBRIDGE, oe a Ak de iy ae ri mice be + Ce inna ; ve) sono ; YY 864 RECORDS PRELIMINARY TO A GENERAL TREATMENT OF THE EUPATORIEAE,—VI. THE present paper presents such new Eupatorieae as have been found in collections examined by the writer during the past year. To the diagnoses needful to put these on record, there have been added a few notes regarding species and varieties appearing to require further elaboration or new interpretation. The chief material studied during the preparation of this paper has been as follows: 1. More than 1400 sheets of American Compositae- Eupatorieae from the herbarium of the University of Copenhagen, an important loan (mentioned in the preceding issue of these notes) bringing to America for the first time authentic material of many species founded upon the collections of Liebmann, Oersted, and Warming. 2. A loan of about 130 Eupatorieae for study and identifi- cation from the Museum of Natural History at Paris, a collection rich in Cordilleran species of rarity and interest. 3. About 150 Eupatoriums and related plants from the Dr. Ekman’s intensive exploration of Cuba, a collection forwarded to the writer for study and determination through the kindness of Prof. I. Urban of the Botanical Museum at Berlin and found to contain one new genus and many new species of unusual interest. 4. Portions (not previously elabor- ated) of the admirable collections made by Messrs. Pennell, Killip & Hazen in western Colombia in 1922, by Messrs. Macbride and Featherstone in Peru in 1922 and continued by Mr. Macbride in 1923. 5. Some Eupatorieae from Ecuador, sent by the Rev. Father L. Mille, containing plants of much interest obtained in part by himself and in part from the duplicates of the herbarium of the late Father Sodiro. 6. Remarkable collections of Eupatorieae made by Mr. P. C. Standley and assistants in Costa Rica in 1925 and 1926, including probably the best representation of the epiphytic Eupator- iums ever secured, and including also a singular trifoliolate species. 7 Dr. F. W. Pennell’s recent Peruvian and Chilian collection of 1924-25 bringing several novelties and excellent material of some plants heretofore known only from fragmentary specimens obtained by the early collectors more than a century ago. 8. The admirable collection of Dr. I. M. Johnston from northern Chile, bringing an interesting novelty in the genus Ophryosporus. 9. A loan from the Field Museum including novelties collected by Dr. Weberbauer in Peru and by Mr. Steinbach in Bolivia. 10. Portions of several large loans mentioned in previous issues of these notes. 11. Several 4 ROBINSON Eupatoricae, collected recently in the mountains of the State of Chia- pas, Mexico, by Dr. C. A. Purpus and forwarded to the writer by Dr. J. N. Rose of the U. S. National Museum. The wealth of this new material, requiring both critical identifica- tion and the description of many novelties, has been so great as seriously to delay several revisions partially in hand, and in conse- quence their completion must be reserved for another season. PiquEeRtIaA GALIOIDES DC. Prod. v. 105 (1836); Robinson, Proc. Am. Acad. xlii. 11 (1906). This has been known only from a fragmen- tary specimen in the Prodromus Herbarium of DeCandolle. It was collected by Haenke in 1834 somewhere in the mountains of Peru, but no more precise data have been available as to the source of the plant in question. It is accordingly well worthy of record that two sheets of this species have been found in the Peruvian collection of Dr. Francis W. Pennell, which localize the plant for the first time, since its characterization as a species. Their labels yield the follow- ing information: herb, with greenish-brown flowers, on open rocky slope, at Quive, Dept. of Lima, alt. 800-1500 m., June 9, 1925, Pennell, no. 14,308 (Gr.); plant suffrutescent below, with greenish- yellow flowers, on open rocky slopes, along the Rio Chillén, near Viscos, Dept. of Lima, alt. 1800-2000 m., June 10-15, 1925, Pennell, no. 14,475 (Gr.). The flower-color, as stated by the collector, seems to refer chiefly to the phyllaries and is not wholly constant, being probably, at least to some extent, correlated with the degree of maturity, perhaps also to exposure. In no. 14,308 the leaves are in part quite as alternate as in the fragmentary type, but elsewhere on the same individual they are sub-opposite or even precisely opposite. In no. 14,475 all but the uppermost are opposite. It is obvious that differences in leaf-arrangement have here little diagnostic significance. Thoug the heads are sessile the inflorescence is in this species looser than in the related Piquerias. hryosporus (§ Buophryosporus) Johnstonii, spec. nov., cae fee intricate ramosus Sere subglobosus 1 m. altus; ioe | o aeatars vel Gteers basi cuneatis glabris vel a se carnosis 5-7 mm. longis 2-3 mm. latis breviter crassiuscule petiolatis; RECORDS OF THE EUPATORIEAE,—VI 5 inflorescentiis terminalibus et lateralibus spiciformibus plerisque 2-5 cm. longis et 9-14 mm. crassis folioso-bracteatis; capitulis parvis 3-4- floris ca. 7 mm. longis sessilibus; involucri squamis oblongis viridibus m. longis; corollis flavescentibus tubulatis ca. 4 mm, longis sine faucibus distinctis extus puberulo-scabratis ; dentibus limbi ca. 0.6 m. longis; achaeniis atrobrunneis 2.7 mm. longis apicem versus shanniiten: pappi setis ca. 30 albidis barbatis corollam subaquanti- us.—CHILE: Prov. Antofagasta, Dept. Taltal: compact bush 1 m. tall, inrocky wash below springs, Aguada Panul, Dec. 4, 1925, Johnston, no. 5424 (type, in Gray Herb.); side of rocky narrow quebrada, Miguel Diaz, Dec. 3, 1925, Johnston, no. 5314 (Gr.); on rocky slopes in quebrada, Aguada Cardon, Nov. 30, 1925, Johnston, no. 5259 (Gr.). This species is clearly a near relative of O. triangularis Meyen, but differs in its fewer-flowered heads, smoothish leaves and phyllaries, - more intricate branching and in the form of the inflorescence. Dr. I. M. Johnston, the discoverer of the species, states that at each of the three stations, where it was observed and collected, it grew in company with 0. triangularis and was readily distinguishable as a different plant by its smaller heads, its somewhat different habit and especially by the form of the inflorescence. In 0. triangularis the leafy spikes, or rather spike-like panicles, are symmetrical, erect, parallel, and distinctly terminal, while in the species here described the inflorescences are more obviously branched, the lateral branchlets spreading; the heads in 0. ease are slender-pedicelled, while in O. Johnstonti, they are, with rare exceptions, sessile. Ophryosporus (§ malities Steinbachii, Ov., videtur fruticosus patenter ramosus; caulibus peaches) cove juventate molliter puberulis aetate Bhbtatie atropurpureis; inter- nodiis 3-8 em. vel ultra longitudine; foliis oppositis vel rare subalterni anguste ovatis graciliter petiolatis; attenuato-acuminatis ieee basi obtusis vel rotundatis supra glaberrimis subtus paullo pallidiori- bus parce pubentibus paullo supra basin 3-nervatis 3-6 cm. longis 1.4-1.8 em. latis; nervis venisque supra depressis subtus prominenti- bus; petiolo 6-10 mm. longo puberulo; paniculis terminalibus et lateralibus pyramidalibus plerisque 5-10 em. longis 4-6 cm. crassis basin versus foliaceo-bracteatis breviter pubescentibus vel tomentellis; capitulis 44.5 mm. longis 5-floris; involucri squamis ca. 6 plerisque oblongis apice rotundatis eroso-ciliatis dorso glabris ca. 3 mm, longis et 1 mm. latis, intimis paullo angustioribus; corollis extus glanduloso- puberulis ca. 3 mm. longis viridescenti-albidis sursum gradatim 6 ROBINSON ampliatis sine differentiatione exacta in tubum et fauces; antheris exappendiculatis pubentibus; styli ramis valde recurvantibus apice nigrescentibus clavellatis; achaeniis 1.8-2 mm. longis apicem versus paullo papilloso-scabratis sed in angulis non hispidulis; pappi setis ca. 24 breviter barbellatis nec plumosis.—BoLrv1a: common in low thicket, Incachaca, Prov. Sacaba, Dept. Cochabamba, alt. 2500 m., Aug. 30, 1921, José Steinbach, no. 5726 (tyPE, Field Mus., phot. and small fragm. Gr.). This species is clearly related to O. Cumingii (Sch.-Bip). Benth., which, however, has longer relatively narrower leaves more cuneate at base, a more leafy inflorescence, and more definite, subeylindric corolla-throat. 0. Steinbachii differs. from O. Kuntzei Hieron. in having shorter, blunter phyllaries, a somewhat looser inflorescence and smaller leaves wholly glabrous above. It is perhaps most ‘nearly allied to O. clavulatus Griseb. but that geographically some- what remote species has a more abruptly expanded corolla-throat, phyllaries slightly hirtellous dorsally, larger leaves, and achenes distinctly hispid on the angles. Stevia Mandonii Sch.-Bip. Bull. Soc. Bot. Fr. xii. 81 (1865) and Linnaea, xxxiv. 535 (1865-66) without char. This dwarf Andean species, originally collected by Gilbert Mandon in 1860 near Mt. Sorata, at the considerable altitude of 3800 m., and now rediscovered both by Dr. Weberbauer and by Dr. Pennell i in the Andes of the Department of Puno, Peru, at a similar or even greater height, seems never to have been described, and to have been left, like most of Schultz’s proposed novelties, a mere supposition to» be verified, characterized and keyed by others. The species appears to be a - one, indeed rather exceptionally well marked in this very difficult genus, by its unusually dwarf habit, small, elliptic, smoothish leaves and sessile glomerules. It may be put on legimate footing by the following character, drawn from an isotype (Mandon’ s no. 246 in the Gray Herbarium): Humilis prostrata perennis; radice e fibris fusco-brunneis numerosis subsimplicibus elongatis ca. 1 mm. diametro consistenti; caulibus gracilibus prostratis nodosis glandulari-puberulis usque ad 1 dm. longis ramosis vel simplicibus; ramis brevibus flexuosis adscendentibus apicem versus foliosis 2-4 cm. longis; foliis oppositis sessilibus vel brevissime petiolatis ellipticis apice obtusis vel saepius rotundatis basi rotundatis vel breviter cuneatis margine integerrimis vel supra mediam partem utroque 1-3-serrato-dentatis textura subcoriaceis utrinque viridibus concoloribus punctatis supra glabrius- culis subtus parce hirtellis 10-15 mm. longis 5-7 mm. latis; petiolo RECORDS OF THE EUPATORIEAE,—VI 7 vix 1 mm. longo; glomerulis terminalibus inter folia sessilibus 1.4-2 em. diametro; capitulis 5-floris ca. 12 mm. longis 6 mm. diametro, involucri cylindrati squamis 5 oblongis obtusis, exterioribus ca. 7.5 mm. longis et 1.8 mm. latis hispidulis, interioribus ca. 8.5 mm. longis et 2 mm. latis costa excepta glabriusculis tenuioribus; corollis ut videtur in tubo rubro-purpureis et in limbo roseo-albis 6.5 mm. longis; lobis limbi 5 oblongis obtusis patentibus ca. 1.7 mm. longis et 1 mm. latis; achaeniis 3.5 mm. longis sursum praecipue in angulis hispidulis; pappi aristis in achaeniis quatuor 19 ca. 7 mm. longis barbatis basi gradatim paullo ampliatis et lacerato-ciliolatis apicem versus purpureis; aristis in achaenio quinto 9 ca. 5 mm. longis et squamis 2-3 hyalinis laceratis ca. 0.7 mm. longis.—Botivia: Depart- ment La Paz: Prov. Larecaja, on hills of the alpine region near Anco- huma in the neighborhood of Sorata, alt. 3800 m., Mar.—Apr. 1860, Mandon, no. 246 (Gr., K.); Department of Tarija (?) at Escayache, alt, 3500 m., Fiebrig, no. 3416 (Gr.). Peru: Department of Puno: common in open mats at Pucard, alt. 3700 m., Weberbauer, no. 431 (Berl., phot. and fragm. Gr.); prostrate perennial herb with pinkish white flowers, on rocky clay slope over limestone at Chuquibambilla, alt. 3900-4000 m., Pennell, no. 13,372 (Gr.). This species is closest to S. Bangii Rusby, which, however, is a more erect plant, often as much as 2-3 or even 4 dm. high, with smaller lanceolate or lance-oval leaves, more exserted glomerules, heads about 1 em. long, and incurved crisped non-glandular puberulence (wrongly stated by Rusby to be stellate). The two species are undoubtedly near relatives and it is possible S. Mandonii may ultimately be shown to represent merely an alpine and strongly modified form of S. Bangii, but this seems unlikely and until further material is secured the two can be kept sharply separated. Mandon’s no. 246 bis, con- fused with his species by Schultz, really came from the neighborhood of La Paz and is precisely the erect, smaller leaved S. Bangii. S. humilis Hieron., an alpine Bolivian Stevia of somewhat similar foliage, has a totally different and much shorter pappus. SrEVIA SORATENSIS Hieron. in Engl. Bot. Jahrb. xxviii. 560 (1901). This striking species with very copious gland-tipped pubescence, deeply toothed leaves, and abundant, dense corymbs of purple flowers was founded upon Mandon’s no. 245, collected in 1860 in dry places in thickets, alt. 2650 m., near Mt. Sorata, Bolivia. A some- what less densely glandular-pubescent plant with similar character and seemingly conspecific, was collected on Sorata, May 1920, by Ernesto Giinther, no. 8 (U.S.). It is worthy of note that what appears 8 ROBINSON to be a form of this beautiful species has now been collected also in Perv, as follows: open rocky bank, at Checapupe, Dept. of Cuzco, alt. 3400-3450 m., Apr. 22, 1925, Pennell, no. 13,536 (Gr.). Dr. Pennell states that the plant is a perennial herb with “true purple” flowers. The corollas (as well as the upper part of the phyllaries) retain, even in the dried state, a deep, exceedingly rich purple color. In pubescence, habit, foliage, branching and density of inflorescence, brevity of pedicels, basal thickening of the midribs of the phyllaries, in fact in a multitude of details the correspondence between the Peruvian and Bolivian plants is very exact. The only differences noted are the deeper coloration of the flowers (perhaps wholly the result of fading in the much older material of Mandon and of Giinther), the slightly larger flowers and more puberulent achenes. In the presence of such close correspondence in other respects it is believed that these differences are merely formal. : Eupatorium (§ Subimbricata) amethystinum, spec. nov erectum dense pilosum; caule tereti ad 8 mm. vel ultra diametro distincte lignoso cum pilis aliis brevibus glanduloso-capitulatis patentibus atropurpureis aliis usque 1-2 mm. longis moniliformibus tenuissimis induto; internodiis 6-8 cm. longis; foliis oppositis ovati- lanceolatis fere a basi cordata ad apicem acutum gradatim attenuatis conspicue et regulariter crenatis (dentibus plerisque 1—1.5 mm. altis et ca. 2 mm. latis) supra bullatissimis tomentellis viridibus subtus sericeo-lanatis subcinereis 6-11 cm. longis 2.5-5 cm. latis a basi palmatim breviter 5-nervatis aliter penniveniis; textura membrana- ceis; petiolo 2.5-3 em. longo dense piloso; corymbo terminali composito aniusculo ca. 2 dm. diametro glanduloso-piloso; bracteolis fili-. formi-linearibus; capitulis ca. 12-floris ca. 12 mm. altis et 6 mm. iametro sessilibus vel breviter pedicellatis; involucri squamis bene 3—4-seriatim imbricatis peracutis, exterioribus viridibus laxe villosulis, interioribus sublaevibus lucidulis pulcherrime violaceo-amethystinis; corollis glabris graciliter tubulatis a basi ad limbum gradatim paullulo ampliatis ca. 6 mm. longis sine faucibus distinctis; dentibus limbi deltoideis ca. 0.6 mm. longis; achaeniis (immaturis) ca. 3.5 mm. longis in costis sursum hispidulis; pappi setis ca. 44 albis capillaribus bar- bellatis ca. 6 mm. longis—Ecvuapor: vicinity of Los Juntas, Sept. 28, 29, 1918, Rose, Pachano & Rose, no. 23,200 (TyPx, in Gray Herb.). This species most nearly approaches E. glutinosum Lam. but differs in many respects, as for instance, in its broader, smaller, more highly colored, more attenuate, and decidedly more imbricate phyllaries, the pubescent upper surface of its leaves, etc. RECORDS OF THE EUPATORIEAE,—VI y E. AMYGDALINUM Lam., var. OXYCHLAENUM (DC.) Bak. in Mart. Fl. Bras. vi. pt. 2, 314 (1876). To the synonymy of this variety may be reduced the uncharacterized and doubtfully mentioned E. goyazense Glaziou, Bull. Soc. Bot. Fr. lvi. Mém: 3, 386 (1909), if one may judge correctly from an isotype (no. 21,590) of Glaziou’s pro- posed species now in the herbarium of the Botanical Museum at Copenhagen. E. (§ Subimbricata) andinum, spec. nov., ut videtur herbaceum (basi ignota) breviter patenterque purpureo-pubescens (pilis nodulosis nec capituliferis) erectum supra paniculatim ramosum 6 dm. vel ultra altitudine; caule ad 3 mm. diametro tereti brunneo; medulla alba; foliis caulinis oppositis basin versus in partem petioliformem alatam cuneato-angustatis et ad insertionem sessilem cordato- amplexicaulibus; parte laminae ampliata ovata acuta remote dentata tenui membranacea utrinque breviter pubescenti pennivenia sed cum venis proximalibus paullo robustioribus et approximatis ca. 7 em. longa ca. 3 cm. lata; parte petioliformi integra ca. 1.5 em. longa 2-5 mm. lata; capitulis graciliter pedicellatis ca. 7-8 mm. altis et 9 mm. diametro ca. 70-floris; involucri squamis 3—4-seriatim imbricatis valde inaequalibus, extimis ovatis obtusis ca. 1.2 mm. longis et 0.6 mm. latis herbaceis dorso pubescentibus; intermediis ovato-oblongis estes subherbaceis apicem versus pubentibus 3-4 mm. longis 1.2 latis; intimis linearibus attenuatis glabris stramineis 5 mm. ee 0.7 mm. latis; receptaculo convexo glabro 1.5 mm. — corollis flavido-albis gracilibus glabris ca. 4 mm. longis; tubo propri 2.7 mm. longo; faucibus turbinatis ca. 0.8 mm. altis; dentibus limbi oblongo-deltoideis 0.5 mm. longis; achaeniis 1.5 mm. longis pallidis glabris basi calearatis.—PEru: Dept. Cuzco, Oct. 1839-Feb., 1840, Claude Gay, no. 959 (TyPE in hb. Mus. Nat. Hist. Paris, phot. and small fragm. Gr.) In many characters near E. pilluanense Hieron. but with florets much more numerous, leaves toothed instead of entire, inner involucral scales acute, and achenes pale. In habit and in its low-convex receptacle this species somewhat recalls E. nemorosum Klatt, but that has still more numerous florets, more slender tubular corollas with eats less faucal enlargement, blackish achenes and other dif- feren E. ( 6 ‘Subimbricata) atrescens, spec. nov., fruticosum 2-3 m. tum in exsiccatione brunneo-nigrescens; caule tereti breviter dense fulvescenteque tomentoso medulloso, internodiis 3-5 cm. longis; foliis oppositis lanceolato-oblongis utroque acutis vel acuminatis subremote * 10 ROBINSON cuspidato-denticulatis penniveniis 5-13 em. longis 1.5-3.4 cm. latis supra minute granulato-puberulis post exsiccationem nigrescentibus subtus post exsiccationem olivaceis fulvo-puberulis; petiolo 1.2-1.6 em. longo; panicula composita terminali; partialibus pyramidatis densis divaricatis; pedicellis 0-2 mm. longis; capitulis ca. 10-floris ca. 9 mm. longis et 5 mm. diametro; involucri squamulis ca. 16 bene imbricatis ca. quadriseriatis acutis obscure ciliolatis plerisque 4—7- nervatis purpurascenti-brunneis, extimis deltoideo-ovatis 2 mm. longis 1 mm. latis, intermediis oblongo-ovatis 4-5 mm. longis 1.8-2 mm. latis, intimis anguste oblongis vel oblanceolato-linearibus 6 mm. longis 0.8-1.5 mm. latis; corollis 6 mm. longis glabris, tubo proprio gracili 3 mm. longo sursum in fauces turbinatas 2 mm. altas ampliato, dentibus limbi purpureis deltoideis 0.6 mm. longis; achaeniis immaturis glabris ca. 2.2 mm. longis; pappi setis ca. 75 sordide albis sublaevibus ca. 5.5 mm. longis.—Ecuapor: Proy. Loja: between San Lucas and Offa, alt. 2200-3100 m., Sept. 7, 1923, Prof. A. S. Hitchcock, no. 21,514 (type, U.S., phot. and fragm. Gr.). Apparently related to E. Hitchcockii Robinson and similarly turning dark in drying but readily distinguished by its much narrower and thinner leaves, dark-fulvous tomentum on stems etc., its smoother, distinctly nerved phyllaries and much shorter icels. E. (§ Subimbricata-Eximbricata) chalceorithales, spec. nov., fruticosum patenter oppositirameum griseo-tomentellum; caule et ramis gracilibus teretibus brunnescenti-griseis; internodiis plerisque 1-2 em. longis; foliis (rameis) oppositis brevissime petiolatis 1 in axillis non proliferis ovato-lanceolatis utroque angustatis apice obtusis vel acutiusculis basi acutis margine utroque latere 2-3-serrato-crenatis ima a basi 3-nervatis pallide viridibus subtus vix pallidioribus utrinque dense sed brevissime puberulo-tomentellis 1.4—2 cm. longis 6-10 mm. latis; petiolo 1-2 mm. longo; corymbis sessilibus terminalibus basi foliaceo-bracteatis numerosis 2.5-5 cm. diametro conjunctim pani- culam subpyramidatam formantibus; capitulis ca. 9-floris 6-7 mm. altis; involucri turbinati squamis ca. 11 valde inaequalibus 3-4- seriatim gradatis sed paullo inter se imbricatis fulvescenti-griseis; intimis oblongo-linearibus obtusis ca. 3 mm. longis et 0.6 mm. latis 2-1 Sneten tis; intermediis brevioribus oblongo-lanceolatis vel lanceo- i ibus etiam 2-3-nervatis; extimis angustissimis lineari- lanceolatis acutissimis laxis in bracteolas gradatim transientibus; in fauces paullo ampliatas gradatim transienti; achaeniis ca. 2.8 mm. ees in angulis minute ciliatis; pappi setis he 35 albidis corollam RECORDS OF THE EUPATORIEAE,—VI 11 subaequantibus.—Cusa: Province Oriente: in eruptive soil on dry hills in the Sierra del Cobre at Cobre, 6 Oct. 1916, Ekman, no. 7811 (ryPE, in Gray Herb.). This species possesses much the habit and many of the character- istics of E. sinuatum Lam. and is doubtless closely related to it. However, E. sinuatum has relatively wider, much more obtuse leaves, which are prevailingly proliferous in the axils. Furthermore, their lobing or dentation, where present, tends to be more spreading. Moreover the outer phyllaries in E. sinuatum are ovate and obtuse or rounded, while they are lance-linear and acute in the plant here described. E. (§ Eximbricata) confluentis, spec. nov., herbaceum vers- similiter perenne erectum griseo-tomentellum vel -puberulum 4 dm. vel ultra altitudine; caule tereti costulato fistuloso 6 mm. diametro apicem versus adscendenter ramoso; foliis oppositis graciliter petio- latis ovato-oblongis vel oblongo-lanceolatis acutis argute serratis (dentibus 0.5-2 mm. longis 2-4 mm. basi latis) supra crispe puberulis subtus pallidioribus griseo-tomentellis 4-8 cm. longis 1.5-4 cm. latis basi plerumque abrupte acutatis fere a basi obscure 3(—7)-nervatis deinde penniveniis, petiolo 1.5-2 cm. longo griseo-tomentello; corym- bis in ramis terminalibus compactis valde convexis 2-2.5 em. (juvent- . ate) diametro; pedicellis 2-3 mm. longis griseo-tomentellis non glandulosis; capitulis ca. 85-floris ca. 4 mm. altis et 6 mm. diametro; involucri subeampanulati squamis ca. 16 lanceolatis acuminatis subaequalibus vix imbricatis dorso griseo-pubescentibus ca. 4 mm. longis; corollis ut videtur albis sparse granulatis; tubo proprio 1.6 mm. longo; faucibus gradatim paullo ampliatis ca. 1.4 mm. longis; dentibus limbi patentibus deltoideis 0.4 mm. longis; achaeniis nigris sparse granulatis 1.3 mm. longis basi pallide callosis; pappi setis albis tenuissime capillaribus ca. 2.8 mm. longis.—E. steviaefolium Britton, Bull. Torr. Bot. Club, xviii. 334 (1891),not DC. E. trichobasis Robinson, Contrib. Gray Herb. n. s. Ixi. 74 (1920) in part (as to plant of Rusby), not Bak.—Bo via: junction of the Rivers Beni and Madre de Dios, Aug., 1886, Dr. H. H. Rusby, no. 1656 (TYPE, in Gray Herb., isotypes N. Y., K., U. S., Mo.). The writer, when some years ago revising the Eupatoriums of Bolivia, doubtfully associated this plant with some material collected by Dr. Otto Kuntze, in the province of Santa Cruz and identified by Prof. Hieronymus as E. trichobasis Bak. Through the kindness of Prof. Diels, director of the Botanical Garden and Museum at Berlin-Dahlem, he has recently been able to obtain two photegraphs 12 ROBINSON and some fragments of the type material of FE. trichobasis Bak., a species which proves, as suspected, to be quite a distinct plant, with larger heads, more velvety, purplish phyllaries, shorter leaves, and a glandular pubescence. Careful investigation of the Bolivian plant here described shows its receptacle nearly flat and essentially glabrous. The species is therefore referred to § Eaximbricata, in vane among the Bolivian representatives of that section it appears o have no very close relative. alla it may here be recorded that the original material of Sello representing Baker’s E. trichobasis bears in the Botanical Museum of Berlin the number 4628 (duly labeled in hand of Baker), not 136 as, through some clerical error, entered in the Flora Brasilien- sis, vi. pt. 2, 364 (1876). Re-examination of the specimen, collected by Dr. Kuntze in the Department of Santa Cruz, and once more kindly loaned to the writer from the herbarium of the New York Botanical Garden, shows it identical with Rusby’s no. 1656, the plant here described as E. confluentis. E. trichobasis Bak. should, there- * fore, be oo from the flora of Bolivia, being as yet known only from Braz E. § Suitibnabeionte) dejectum, spec. nov., ut videtur fruti- culosum vel subherbaceum erectum 2-3 dm. sititudian: caule gracili tereti subvirgato folioso vel tardius foliis delapsis subnudo brunneo revissime oo griseo-puberulo apicem versus patenter corym- ; foliis alternis et suboppositis plerisque reflexis lanceo- latis sciaiesinita attenuatis sed ad apicem verum obtusiusculis basi cuneato-attenuatis sessilibus vel breviter petiolatis; limbo 9-21 mm. longo 2-5 mm. lato margine revoluto a basi 3-nervato supra puberulo viridi subtus breviter pubescenti vix pallidiori, petiolo gracili vix 2 mm. longo; corymbi compositi 5-15 cm. diametro ramis alternis patenter adscendentibus foliosis; capitulis pedicellatis ca. 15-floris 12 mm. altis 7 mm. diametro; involucri anguste campanulati squamis 3-4-seriatim imbricatis acutiusculis dorso tenuissime tomentellis, extimis ovatis ca. 2 mm. longis, intermediis ovato-oblongis, intimis lanceolato-linearibus ca. 8 mm. longis; corollis ca. 7 mm. longis obscure puberulis, tubo vebpeio 3 mm. longo, faucibus gradatim dilatatis etiam ca. 3 mm. longis, Paeoe limbi lanceolatis acutis 1.3 mm. longis; achaeniis gracilibus 5 mm, longis in faciebus granulatis, in angulis hispidulis; pappi setis ca. 30 subplumoso-barbellatis 6.5 mm. longis Savido-albis. —Bourvia or Arcentina: “ Miraflor,” A. _ seem originally to have been only the words “Herb. Mus. Paris” RECORDS OF THE EUPATORIEAE,—VI 13 (in print) and “ Ab D’Orbigny”’ (in manuscript). To these there were added apparently at a later date and in different hand the determina- tion “ Bulbostylis” and the locality “ Bolivie.”” However, on a separ- ate slip (presumably a sort of field label) glued to the same sheet appears the further data “ Miraflor No. 1331.” As no Miraflor or hispanice Miraflores has been found in Bolivia it seems probable that the Miraflor on D’Orbigny’s field label referred to the place of that name in Northern Argentina. E. dejectum (so named from its reflexed as it were pendant leaves) is most closely related to E. patagonicum Klatt and E. patens D. Don. It differs from both by its more numerous florets (about 15 in the head instead of 5-10), its entire leaves, very short petioles, narrow leaf-blade, subplumose rather than obscurely barbellate pappus-bristles, and the hispidity on the angles of the achenes. Unhappily the base of the plant is not shown in the material at hand, so it is not quite certain whether it is essentially herbaceous as it appears in the specimens or whether these are branches of a larger shrub. Their own radial rather than dorsiventral branching rather suggests an upright scarcely woody herb. In its involucre, pubescence and in many other respects E. dejectum approaches the Brazilian E. Bakerianum Robinson, Contrib. Gray Herb. Ixxv. 5, (1925), but that has very much larger leaves, more paniculate inflorescence, fewer-flowered heads and a very different corolla, which is somewhat constricted just below the limb and which has teeth less than half as long as those of E. dejectum. E. pecumpens (Gardn.) Bak. in Mart. Fl. Bras. vi. pt. 2, 344 (1876). Examination of the material now available of this species shows that it falls into two usually distinguishable though seemingly to some extent intergrading varieties, which may be defined as follows: Var. typicum, caulibus solum basin versus vix per tertiam partem longitudinis foliatis supra in pedunculos saepius nudiusculos perlongos productis; foliis ovatis vel obovatis plerisque obtusis crenato-serratis. Var. stenum, var. nov., caulibus basin versus magis ramosis et fere ad mediam partem foliatis et supra plerisque 3-5-capituliferis; foliis lanceolatis a media parte ad apicem utroque argutius 1—2-serrato- dentatis.—E. polyanthum Glaziou, Bull. Soc. Bot. Fr. lv. Mém. 3, 384 (1909), not Sch.-Bip.—Brazi: Campos da Bocaina, Sao Paulo, suffrutescent, flowers violet, Feb. 10, 1876, Glaziou, no. 8124 (Copenh., fragm. Gr.); open hills, Campo do Jordao, Sao Paulo, Apr. 23, 1922, Prof. & Mrs. E. W. D. Holway, no. 1762 (Gr.), in part; Campos do 14 ROBINSON Jordao, Sao Paulo, alt. about 1600 m., Feb. 12, 1924, L. H. & E. Z. Bailey, no. 830 (TyPE, in Gray Herb.). This plant, though charac- teristic, at times grows with the typical variety and has by two col- lectors been gathered with it. Especially Holway’s no. 1762 suggests probable intergradation. The author would place this species in § Cylindrocephala rather than § Prazelis where placed by E. prptopictyon Robinson, Proc. Am. Acad. liv. 242 (1918). This species was founded on a single specimen (in the Kew Herbarium) labelled as having been collected in Colombia by Lobb. Unfortun- ately the label gave no further data. Mr. E. P. Killip of the United States National Museum has recently informed me that from his studies of the Passifloras collected by Lobb and from an examination of specimens in the Kew Herbarium he had become convinced that some, if not all, of Lobb’s specimens labelled as from Colombia came in reality from Peru. In connection with this matter and in confir- mation of Mr. Killip’s interesting theory, it may here be recorded that the characteristic E. diplodictyon, never rediscovered in Colombia, has lately been collected by Mr. J. F. Macbride in Central Peru, his label bearing the following data: “No. 4296. Liana with wistaria- color fls., about 8000 feet. Herbage pungent, June 5-7, 1923, Mufia. Trail to Tambo de Vaca, Peru.” Thus Peruvian material corres- ponds remarkably with the original plant of Lobb, now represented at the Gray Herbarium by two photostat copies (positive and nega- tive). The occurrence of this plant in Peru and the failure of recent expeditions to discover any trace of it in Colombia adds new evidence in substantiation of Mr. Killip’s ingenious theory. E. (§ Eximbricata) dispar, spec. nov., annuum erectum patenter sordide articulato-villosum 5-6 dm. altum adscendenter ramosum; radice fibrosa; caule tereti basi paullo lignescenti fistuloso 4 mm. diametro; internodiis 5-60 mm. longis; pilis indumenti plerisque brevibus cum aliquibus multo longioribus attenuatis articulatis; foliis disparato-oppositis petiolatis deltoideo-ovatis acutis grosse crenato-serratis basi patenter cordatis supra viridibus puberulis et oes villosis annem —— villoso-hirsutulis ieee RECORDS OF THE EUPATORIEAE,—VI 15 vel purpureo-tinctis dorso tomentellis ca. 4mm. longis et 1.2mm. latis; corollis albis 3 mm. longis extus praecipue in tubo proprio (1 mm. longo) et dentibus limbi (ca. 0.8 mm. longis) papilloso-scabratis ; achaeniis nigris gracilibus basin versus attenuatis in angulis sursum cliffs above river at “ Pillahuata,’”’ Cerro de Cusilluyoe, alt. 2200-2400 m., May 3-6, 1925, Pennell, no. 13,967 (type, in Gray Herb.). This well marked species may be placed near E. vallincola DC., but differs in its annual nature, densely puberulent, as well as villous leaves, which are of smaller size and more deeply cordate, its broader, less numerous phyllaries, more copious pappus, ete. The name dispar is used in the sense of unpaired, and alludes to the members of the leaf-pairs, which are often separated by an interval of nearly 1 cm. E. Domseyanum DC. Prod. v. 167 (1836) ; Robinson, Proc. Am. Acad. lv. 11-12, 82 (1919). This species, described from material _ collected by Dombey, was in the original diagnosis recorded merely as from South America. Some years ago, when the writer (I. ¢.) was revising the Eupatoriums of Peru, he ventured to place provisionally E. Dombeyanum in his key and include it in the enumeration of species for that country. It is now possible to confirm its occurrence in Peru, for specimens from the Peruvian collection of Pavon, now in the Herbarium of the University of Geneva, prove to be identical with the De Candollean type, agreeing closely in all respects. Though most of Pavon’s Peruvian material lacks precise data of collection this bears a label of some detail giving the habitat. as “in Provinciae Chancay collibus altis nec dis ad Teguam et Huauram,” in other words the coastal region north of Callao, which is in fact one of the parts of Peru visited by Dombey. E. (§ Eximbricata) Ekmanii, spec. nov., ut videtur herbaceum perenne (basi ignota) 7 dm. vel ultra altum; caule tereti post exsic- cationem multicostulato ca. 2.3 mm. diametro albo-medulloso atro- brunneo paullo purpurascenti breviter patenterque purpureo-pubes- centi, capillis attenuatis moniliformibus; internodiis usque ad 1 dm. longis; foliis oppositis breviter petiolatis ovato-oblongis acutis vel acuminatis subintegris vel irregulariter undulato-dentatis (dentibus utroque 0-3 er 0. die? man. altis inter se distantibus) basin versus integerrimis b papil- loso-scabridulis qgiwate prominulenter reticulatis subtus atroviolas- centibus prominulenter | reticulatis in nervis minute pubescentibus in superficie cum glandulis minutis sparsis lucentibus ornatis paullo 16 ROBINSON supra basin 3-nervatis 4.5-6.5 cm. longis 2-2.5 cm. latis firmius membranaceis; petiolo ca. 4 mm. longo atroviolaceo dense fusco- pubescenti; foliis superioribus multo minoribus bracteiformibus; panicula terminali laxissima ca. 2 dm. diametro; ramis gracilibus patenter adscendentibus; pedicellis usque ad 2.5 cm. longis cum bracteolis lineari-subulatis 2-3 mm. longis ornatis; capitulis ca. 11- floris 7 mm. longis; involucri squamis interioribus ca. 7 linearibus acutis subaequalibus 4.5-6 mm. longis atropurpureis minute fusco- hispidulis et cum glandulis globosis lucentibus ornatis; squamis exterioribus 3-5 multo brevioribus angustissimis; corollis tubulatis sursum paullo gradatim ampliatis (sine faucibus distinctis) ca. 3.5 mm. longis dense atomiferis verisimiliter atropurpureis; dentibus limbi brevissimis recurvatis; achaeniis 2.5 mm. longis deorsum attenuatis granulatis; pappi setis ca. 25 sublaevibus sordide albis ca. 3 mm. longis.— Cusa: Province Oriente: in pine wood on northern slope at El Yungue near Baracoa, 20 Nov. 1914, Ekman, no. 3552 (tyPE, in Bot. Mus. Berlin, phot. and fragm. Gr.). This species stands probably nearest E. incisum Griseb. but differs conspicuously, in its lax inflorescence and sparingly undulate-dentate instead of sharply incised leaves which have on both surfaces an elevated reticulation. E. Havanense HBK. Noy. Gen. et Spec. iv. 128 (1820). This pecies, common in many parts of Cuba as well as in Mexico and southwestern United States, has a fair constancy in habit and floral characters. It shows, however, some fairly marked variations both in leaf-contour and in pubescence. These may be indicated as follows: - Var. typicum, inflorescentia saepius plus minusve puberula vel tomentella excepta glaberrimum; foliis basi rotundatis vel subtrun- catis.—E. havanense HBK. |. c. Var. cardiomorphum, var. nov., foliis kasi distincte cordatis, aliter var. typico simile.—CuBA: Pi cesliies Pinar del Rio; jurassic limestone rocks, Sierra de Vifiales in Loma de la Bandera, 9 Mar., 1924, Ekman, no. 18,666 (ryPE, in Bot. Mus. Berl., phot. and fragm. Gr.). The available material of this rather aiikiae variety is un- . fortunately overmature, the inflorescences are withered, and the involucres dried, recurved and empty. However, it seems sufficiently clear that it belongs.to the species with which it is here associated, though exhibiting a leaf-form not previously recorded. Var. pubinerve, var. nov., foliis (leviter crenatis vel obtuse ser- ’ ratis) supra parce pulverulento-puberulis subtus secus nervis patenter RECORDS OF THE EUPATORIEAE,—VI 17 pubescentibus et in superficie parce pubentibus; axibus juvenilibus et petiolis puberulis vel etiam tomentellis—Cusa: without indicated locality, Wright, no. 3598 (Gr.); open places in shrubby thickets on north slope of Pico Turquino, about 1700 m. alt., Sierra Maestre, Ekman, no. 5416 (rype, in herb. Bot. Mus. Berlin, fragm. Gr.); and in the same range on top of Punta de Palmamocha, south of Yara, alt. 1400 m., 15 July, 1922, Ekman, no. 14,302 (Brl., fragm. Gr.). In Ekman’s specimens of this plant the leaves are all rather small rameal ones, but Wright’s no. 3598 shows fully developed cauline leaves with precisely similar pubescence. Var. inerve as to pubes- cence resembles the plant described long ago from Texas as E. agerati- folium B. texense Torr. & Gray, Fl»ii. 90 (1841), but later completely reduced by Gray, Syn. Fl. i. pt. 2, 100 (1884). However, the Texan form has larger, more acute, and much more deeply toothed leaves than the Cuban plant here described, in which the leaves are obtusish to rounded at tip, rounded at the sides, and very shallowly crenate- dentate. The species runs into several other more or less distinguish- able local forms or varieties but their differences are of trifling nature and by no means constant. E. (§ Eximbricata) helianthemoides, spec. nov., fruticosum ramosum crispe puberulum vel brevissime griseo-tomentellum; ramis patenter adscendentibus gracilibus foliosis; internodiis 3-20 mm. longis; foliis oppositis brevissime petiolatis oblongo-linearibus vel anguste lanceolatis obtusis integris basi rotundatis utrinque griseo- pulverulentis vel subtus brevissime puberulis basi breviter 3-nervatis aliter penniveniis firmiter membranaceis 2-3 cm. longis 4-6 mm. latis; petiolo 1-2 mm. longo; corymbis ad apices ramorum sessilibus plerumque 4—12-capitulatis planiusculis 2-4 cm. diametro; pedicellis 4-13 mm. longis; capitulis ca. 15-floris 6-7 mm. altis; involucri squamis ca. 14 lanceolato-linearibus attenuatis granulatis, interioribus subaequalibus ca. 4.5 mm. longis, exterioribus 3-4 minoribus angustis; corollis ut videtur pallidis 3.5 mm. longis cum glandulis sphaericis ad- sparsis; tubo proprio ca. 2 mm. longo; faucibus turbinatis ca. 1.5 mm. longis; dentibus limbi deltoideis ca. 0.6 mm. longis; achaeniis nigris granulatis 2.3 mm. longis basi breviter attenuatis et pallide callosis; pappi setis 25-30 albidis tenuibus laeviusculis ca. 2.7 mm. longis.—CuBa: Province Oriente; hillside, Santiago Harbor, 14-15 Mar., 1909, NV. L. Britton, no. 1890 (Gr.); steep bank, Santiago Bay, 10-25 Mar., 1912, NV. L. & E. G. Britton & J. F. Cowell, no. 12,697 (Gr.); on bank of river Aguadores; Santiago, June 14, 1918, Ekman, no. 9228 (Gr.); vicinity of Santiago Bay, July, 1920, Bro. Clement, no. 288 (Type, in Gray Herb.). 18 ROBINSON This plant has much the habit of E. Iindenianum A. Rich. and the specimens earlier examined were doubtfully referred to that species. However, the amount of material now at hand shows consistent and ample differences for its separation. In E. Lindenianum the leaves are larger, more ovate-lanceolate, attenuate, decidedly discolorous, and grayish-tomentellous beneath, while the phyllaries are but 2-2.5 mm. in length, oval and obtuse. Both species appear to be common about Santiago. E. (§ Subimbricata) holguinense, spec. nov., fruticosum ad 1.5 m. altum oppositirameum brevissime denseque pannoso-lanula- tum; caulibus teretibus tardius glabratis post exsiccationem levissime costulatis 3-4 mm. diametro albo-medullosis; internodiis 2-8 em. longis; ramis patenter adscendentibus; foliis oppositis lanceolatis vel ovato-lanceolatis vel rarius suboblongis gradatim acuminatis vel rarius obtusis integerrimis vel repando-denticulatis basi rotundatis fere a basi 3(—5)-nervatis supra primo griseo-puberulis mox glabres- centibus bullatissimis griseo-viridibus et lucidis subtus primo breviter albo-tomentellis tardius griseo-albidis prominulenti-reticulatis 2.5-5 (-8.3) em. longis 6-20(-38) mm. latis subcoriaceis; petiolo brevissime albido-tomentello 3-5 mm. longo; corymbis compositis planiusculis; capitulis subglomeratis breviter pedicellatis ca. 16-20-floris ca. 7 mm altis; involucri squamis 3—4-seriatim imbricatis lanceolatis vel oblongis appressis dorso griseo-pubescentibus 2-3-nervatis et cum glandulis lobosis aureis sparsis ornatis; exterioribus acutis, interioribus sub- obtusis; corollis (verisimiliter albis vel pallidis) tubulatis vix sursum ampliatis 3.2 mm. longis usque ad mediam partem cum glandulis sessilibus globosis adspersis supra glaberrimis; achaeniis juvenilibus 2.2 mm. longis deorsum paullo decrescentibus in angulis paullo scabratis; pappi setis ca. 25 vix barbellato-scabratis ca. 3 mm. longis sordide albis,—Cusa: Province Oriente, on slopes of the Cerro de Fraile, about 275 m. alt., near Holguin, 29 Oct., 1914, Ekman, no. 3271 (Type, in Gray Herb. ‘ not rare in pine-savannas on the Sierra de Nipe near the Rio Piloto, 1 Sept., 1914, Ekman, no. 2697 (Gr.); in stony pine woods near the Rio Piedra on the Sierra de Nipe, 3-4 July, 1914, Ekman, no. 1810 (Brl., fragm. Gr.). Earlier and not wholly satisfactory specimens of this plant were collected in the same region by the late J. A. Shafer, namely his numbers 1219 (from the top of Loma Pilon), 1238 (by water course on barren savannas, southeast of Holguin), 1753 (shrubby, 1.5 m. high, flowers white, lower hills near Pinales), and 3307 (in thickets, on serpentine formation, along trail from Puedra Gorda to Woodfred, RECORDS OF THE EUPATORIEAE,—VI 19 in the Sierra Nipe). These specimens were at the time doubtfully referred by the writer to E. lantanifoliwm Griseb., which, however, has cordate-ovate much more obtuse leaves, tawny beneath, as well as much more acute and less distinctly nerved phyllaries. E. holguinense, as now judged from much more complete material, can be seen to approach more nearly E. hypoleueum Griseb., but that has more broadly ovate and more decidedly toothed leaves, which even at full maturity are dull, not lucid above. It also has more num- erous florets, 30-40 in a head. E. numiLE (Benth.) Hieron. ex Therese, Princess of Bavaria, Beihefte Bot. Centralbl. xiii. 68 (1902); Robinson, Proc. Am. Acad. liv. 322. (1918). Conoclinium humile Benth. Pl. Hartw. 199 (1845). The original transfer of this species from Conoclinium to Eupatorium, at the point indicated, though made in 1902, was unfortunately recorded so obscurely as to type or other indication of a new combina- tion of names as to escape completely the notice of the expert indexers employed both on the supplements of the Index Kewensis and of those engaged in carrying forward the Card Index of New Genera, Species, and Varieties of American Plants. In consequence its existence was unknown to the writer, when some years ago preparing a revision of the Eupatoriums of Colombia. Accordingly, attention may be here called to the fact that the correct authority for the name in question is (Benth.) Hieron. E. (§ Subimbricata) hygrohylaeum, spec. nov., herbaceum perenne erectum 3 dm. vel ultra altum; caule tereti obscure tomentello 4 mm. vel ultra diametro albo-medulloso; internodiis usque ad 8 cm. longis; foliis oppositis petiolatis ovatis acuminatis serratis basi rotundatis pinnatim ca. 1l-nervatis utrinque laete viridibus supra glabris subtus vix pallidioribus in’ nervis venisque puberulis: inter nervis prominulenter reticulatis ca. 8 cm. longis et 5 cm. latis; petiolis ca. 1.4 em. longis puberulis; corymbo terminali planiusculo ca. 3 dm. diametro; bracteis bracteolisque linearibus parvis vel minutis; pedicellis filiformibus 2-9 mm. longis; capitulis ca. 30-floris 7 mm. altis 6 mm. diametro; involucri squamis ca. 36 pulcherrime imbricatis ca. 4-seriatis. apice rotundatis dorso viridibus albo-striatis puberulis margine lacerato-ciliolatis, exterioribus ovatis, interioribus ovato- oblongis vel linearibus; receptaculo leviter convexo; corollis albis gla- bris 3 mm. longis tubulatis sursum paullo ampliatis sine faucibus distinctis; dentibus limbi ca. 0.5 mm. longis; achaeniis maturitate nigris 1.5 mm. longis in angulis pallidis parcissime hirtellis vel glabris in faciebus praesertim apicem versus paullo granulatis; pappi setis 20 ROBINSON ca. 35 albis sublaevibus ca. 2.7 mm. longis—Costa Rica: Province San José: wet forest, erect herb; flowers dirty white, La Hondura, alt. 1200-1500 m., Mar. 9, 1926, Standley & Valerio, no. 51,933 (TYPE, in U. S. Nat. Herk,, phot. and fragm. Gr. This species may be readily fetoknited among the Costa Rican members of the genus by its very numerous involucral scales which are rounded at the tip, dorsally striped, and lacerate-ciliolate on the edge. The involucre, though of smaller size recalls those of some of the Colombian and Ecuadorian members of § Hebeclinium, but the receptacle is here glabrous. E. (§ Eximbricata) juninense, spec. nov., fruticosum scandens (?) fulvido-tomentellum; caule tereti usque ad 8 mm. diametro tardius glabrato griseo multicostulato fistuloso; internodiis 3-14 cm. longis; foliis oppositis petiolatis oblongis utroque acuminatis minute cuspidato-denticulatis (dentibus ca. 0.3 mm. altis et 5 mm. inter se distantibus) primo aspectu integris supra puberulis post exsiccationem brunneo-olivaceis subtus pallidioribus fulvescenti-velutinis penniveniis 7-12 em. longis 4-6 em. latis membranaceis; petiolis 8-15 mm. longis; corymbis terminalibus patenter trichotomis convexis ca. 8 cm diametro; bracteolis subulatis; pedicellis plerisque 2-5 mm. longis; capitulis ca. 33-floris 6.5 mm. altis 6 mm. diametro; involucri squamis ca. 17 lanceolatis attenuatis dorso brunnescenti-pubescentibus 3- nervatis, exterioribus paucis gradatim minoribus, interioribus sub- aequalibis vix imbricatis, ca. 5 mm. longis et 1.2 mm. latis, intimis 2-3 oblanceolato-linearibus laeviusculis; receptaculo plano paullo alveo- lato glabro; corollis albidis viridescentibus tubulatis 4 mm. longis glabris sine faucibus distinctis; dentibus limbi anguste deltoideis re- flexis; achaeniis Sipressentibus 2 mm. longis glabris; pappi setis ca. 25 sordide albis 3.8 mm. longis siblaevibus sursum non incrassatis.— Peru: Dept. Junin: on tree; flowers whitish, turning greenish, La Merced, about 610 m. alt., Aug. 10-24, 1923, J. Francis Macbride, no. 5583 (TYPE, in Field lie: ., isotype Gr.). This species has foliage similar to that of the Mexican E. mirador- ense Hieron., which, however, has quite different, very pubescent corollas and other technical distinctions. E. juninense is, as to its involucre, somewhat intermediate between § § Eximbricata and Sub- imbricata, but seems better referred to the former, since its fairly uniform inner phyllaries are little imbricated, though a few of the outer ones are graduated Among the previously described Eupatoriums of Peru this is related pee most nearly to E. endytum Robinson, a species RECORDS OF THE EUPATORIEAE,—VI 21 similarly transitional between § § Subimbricata and Eximbricata, but showing slightly greater approach to § Subimbricata. E. endytum has leaves usually rounded at base, longer petioles, a leafy-bracted inflorescence, and elliptic-lanceolate phyllaries. Perhaps the best difference is to be found in the indument, which in E. endytum is on the lower surface of the leaf-costa and nerves compactly woolly- crisped and of a rusty color, while in E. juninense the hairs are straight, setuliform and of a khaki color with no ferruginous tinge. E. § Subimbricata) Killipii spec. nov., fruticosum ramosum; caule tereti griseo obscure costulato 4 mm. vl ultra diametro; foliis oa—* petiolatis ovatis crenato-serratis acuminatis sed ad apicem erum obtusis basin rotundatam vel patenter cordatam versus fu teeiaotalia firmiter membranaceis utrinque viridibus a basi 5- nervatis in nervis puberulis aliter glabris leviter reticulatis minute crebreque pellucido-punctatis 4.5-8 em. longis 3-4.5 em. latis; petiolo 5-10 mm. longo sordide tomentello, paniculis terminalibus divari- catim ramosis ca. 5 cm. altis et 10 cm. diametro rotundatis sacheveanenl puberulis; ramis ramulisque paniculae gracilibus; bracteis parvis bracteolis minutis subulatis; capitulis graciliter pedicellatis vel plerisque ad apices ramulorum binis vel trinis et arcte sessilibus ca. 21-floris ca. 4 mm. longis et 2.5 mm. diametro; involucri campanulati bene imbricati et gradati squamis pluriseriatis obtusis viridibus pul- cherrime 3-5-nervatis, exterioribus ovatis interioribus lanceolato- oblongis ciliolatis; corollis albis 2.5 mm. longis (immaturis) graciliter tubulatis sine faucibus distinctis; achaeniis (valde immaturis) 0.7 mm. longis; pappi setis ca. 35 corollam aequantibus albis,—CoLoMBIA: Department of El Valle: forest in Dagua Valley, Cisneros, alt. 300 to 500 m., 21 Sep., 1922, Ellsworth P. Killip, no. 11,488 (TyP£, in Gray Herb.). This neat shrub in many respects resembles the still obscure E. ’ smilacinum HBK. but that was described by Kunth as a twining herb with sulcate, angled stems and hispid leaves. In E. smilacinum (as shown by a photograph of the type), the heads are more densely clustered in globular glomerules; furthermore the more coarsely crenate-serrate leaf-blade is obtuse at base rather than rounded or subcordate as in E. Killipii; finally the leaf-blade in the former is about 4-5 times as long as the petiole, while in the latter it is 8-9 times as long. Both species in their involucral characters form a transition between § § Dole diomphals and Subimbricata. E. § phala) laevigatulum, spec. nov., herbace vel a patenter oppositi-ramosum ubique | eeu 22 ROBINSON primo aspectu glaberrimum sed obscure praesertim in partibus juvenilibus puberulum 5-10 dm. altum ad inflorescentiam foliatum; caule subtereti post exsiccationem pluricostulato erecto glabro vel -_ apicem versus minute nigro-puberulo; foliis oppositis graciliter petio- latis ovato-lanceolatis plerisque integerrimis rariter uno vel altero latere 1(-3)-dentatis ad apicem longe attenuatis basi cuneatis 4-7.5 cm. longis 7-20 mm. latis paullo supra basin 3-nervatis membranaceis utrinque glabris vel obscure puberulis; petiolo 5-15 mm. longo; corymbis terminalibus planis compositis laxis; corymbulis subdensis plerisque 5-10-capitulatis; capitulis 7-13-floris 11 mm. longis 5 mm. diametro saepe juventate in pedicello curvato nutantibus; involucri squamis ca. 15-19 substramineis plerisque 5-nervatis obtusis obscure ciliatis deciduis dorso glabris; extimis ovatis 1.6 mm. longis 1 mm. latis, intermediis 3 mm. longis 1.4 mm. latis, intimis paullo acutioribus ca. 5.5 mm. longis 0.7 mm. latis; corolla glabra ca. 5 mm. longa lilacino- caerulea, tubo proprio sursum gradatim paullo ampliato, faucibus dis- tinctis nullis, dentibus limbi brevissimis ca. 0.3 mm. longis; achaeniis gracilibus atrobrunneis deorsum decrescentibus 4 mm. longis in costis breviter sursum hispidulis basi breviter pallido-calcaratis in summa parte cum annulo pallido pappifero coronatis; pappi setis 32-39 flavido-albis scabratis corollam subaequantibus.—E. graciliflorum Chod. Bull. Herb. Boiss. ser. 2, ii. (146). 306 (1902), iii. (133) 705 (1903), not DC. E. Orbignyanum Hassl. in Fedde, Rep. Spec. Nov. xiv. 279 (1916), not Klatt—ArcentIn® Repus.ic: Corpus in Mis- siones, 1826, Bonpland, (Par., phot. and fragm. Gr.). ParaGuay: without locality Dr. Edward Palmer, while on the Page Expedition for the exploration of the River La Plata, during the years 1853- (U. S.); on the banks of Lake Ypacari, 1898-99, Dr. E. Hassler, no. 3936 (Gr., Univ. Geneva); in swamp near Paraguari, Dec., 1900, Pr. E. Hassler, no. 6537 (Univ. Geneva); Santa Elisa in Gran Chaco, — lat. S. 23° 10’, Mar. 1903, Rojas, no. 2817 (Gr., Univ. Geneva); in the region of Lake Ypacari, Mar., 1913, Dr. E. Hassler, no. 11,584 (ryPE, Gr., isotype, Copenh.); in woods near Salado, Chodat, no. 393 (Univ. Geneva); Puerto San Pedro, Chodat, no. 397 (Univ: Geneva). . For some years this plant has been known to the writer from fairly good specimens but all efforts to place it satisfactorily in any described. species have failed. It is distinctly more slender, more smooth and more entire-leaved than any form of E. odoratum L. As Dr. Hassler rightly observed it fails to agree with De Candolle’s E. graciliflorum to which it was first referred by Prof. Chodat. However, it is equally RECORDS OF THE EUPATORIEAE,—VI 23 unsatisfactorily placed as by Hassler in E. Orbignyanum Klatt for a fragment of the type of that species contained in the betiexiiany of the late Dr. Klatt—a collection purchased by the Gray Herbarium —shows that E. Orbignyanum had a more rounded inflorescence, distinctly more acute inner phyllaries, and smaller leaves which are conspicuously glandular punctate. As several specimens have now accumulated, collected at different points in Paraguay and north- eastern Argentina at divers times extending over almost a century and as they cannot be convincingly placed in any species hitherto described it seems best to put them on record as constituting a new species even though belonging to a very perplexing group. E. taevicatum Lam. Encycl. ii. 408 (1788). This characteristic species widely distributed in the warmer parts of America is for the most part entirely glabrous and is apt to be so described without _qualification.. The first intimation that it might at times have pubescent forms seems to be found in Baker’s var. tomentosum, pub- lished (as tomentosa) in Mart. FI. Bras. vi. pt. 2, 287 (1876) a gray- velvety plant collected in Minas Geraes. In looking over an excellent representation of EF. laevigatum recently lent me from the Botanical Museum of Copenhagen, I find several sheets of material (all collected in Minas Geraes though at a places and dates) which, being sparingly but definitely hairy, can by no means be placed either with the typical, wholly glabrous se or Se in Baker’s much more hairy, indeed gray-tomentose varie provide a definite place for these, the following may be put on nuke E. LaEvicatum, forma subcapillatum, forma nov., halitu - formae typicae simile et similiter variabile, differt inflorescentia et caule (sursum) et petiolis et nervis venisque in pagina inferiori folio- rum omnibus parce vel subdense hirtis——Braziu: Caldas, Minas Geraes, A. F. Regnell, no. II]. 685 (ryPE, in Bot. Mus. Copenh., phot. and fragm. Gr.); Facenda Mocambo, April, Warming, no. 494 (Copenh., fragm. Gr.); Lagoa Santa, Warming, no. 491 (Copenh.) ; a plant of the plains, common, Sitio, Schenck, no. 3176 (Copenh.). . LAxuM Gardn. in Hook. Lond. Jour. Bot. v. 476 (1846). When treating Eupatorium in the Flora Brasiliensis, vi. pt. 2, 307 (1876) Baker reduced Gardner’s E. larum to a variety of the then very broadly interpreted E. guadalupense Spreng. It has since become clear that this disposition will hold neither on classificatory nor nomenclatorial grounds. E. guadalupense Spreng. has been definitely shown by Urban, Symb. Ant. iv. 625 (1911) to have been nothing more nor less than Pluchea purpurascens (Sw.) DC. Nor can this lax- 24 ROBINSON panicled plant of eastern Brazil be satisfactorily placed varietally as by Baker, |. c., in the species then interpreted as “ FE. guadelupense”’ and now more properly called E. microstemon Cass. which has slightly smaller heads, narrower and thinner phyllaries, a smaller and more tubular corolla without very definite throat and with a glabrous limb. From the nearly related E. pycnocephalum Less., E. laxum Gardn. also differs clearly. Its inflorescence is much looser, its branches more slender, involucre less closely imbricated, and outer phyllaries more acute. It is true, no one of these features yields an entirely satisfactory differential trait, but they and various other minor characters are sufficiently definite to justify keeping the Brazilian E. laxum out of the chiefly cordilleran E. pycnocephalum which ranges from Mexico to Venezuela and Peru. To the writer it does not seem worth while to separate the plants placed by Baker in his E. guadalupense, vars. laxa and Gardneri Even in Regnell’s no. I. 232, cited by Baker as of his var. laxum, some of the leaves are quite markedly cordate, while the indument is exactly similar in kind to the pubescence of his var. Gardneri and very nearly as dense. The two varieties seem to be at best mere forms or chance combinations of characteristics which are subject to independent variation. Very closely related is the earlier E. remotifolium DC. Prodr. v. 165 (1836) placed by Baker quite unaccountably in the synonymy of E. vitalbae DC. to which it has no likeness whatever. Vauthier’s no. 259, the type of E. remotifolium has exceptionally long internodes, much surpassing the leaves, but this seems likely to have been merely _ a causal state for otherwise identical material collected for instance in Minas Geraes by Widgren (Copenh.) and in mountain woods at Petropolis by Ball (Gr.) have internodes considerably shorter, while preserving all the other traits of DeCandolle’s species. E. (§ Eximbricata) lobuliferum, spec. nov., ut videtur herbaceum perenne multicaule 6-9 dm. altum puberulum; catilibug erectis tereti- bus atropurpureis post exsiccationem paullo costulatis ca. 4 mm. diametro albo-medullosis; internodiis 8-11 cm. longis; foliis oppositis petiolatis late ovatis acuminatis multilobulatis basi patenter cordatis , eis 6-8 cm. longis 5-6 cm. latis; lobulis utroque latere 6-9 saepius in margine posteriori 1-dentatis paullo prorsum curvatis 4-9 mm. longis et latis obtusis vel acutiusculis; corymbis in caule ramis- que terminalibus planiusculis subdensis multicapitulatis 4-8 cm. RECORDS OF THE EUPATORIEAE,—VI 25 diametro conjunctim inflorescentiam compositam foliaceo-bracteatam usque ad 2 dm. diametro formantibus; pedicellis filiformibus fulvo- puberulis 3-7 mm. longis cum bracteolis 1-3 lineari-subulatis ornatis; capitulis maturitate 1 cm. altis et 1 em. diametro ea. 50-floris; involucri campanulati squamis lanceolato-linearibus ca. 17 stramineo-viridibus attenuatis dorso hirtellis; receptaculo plano nudo; corollis gracilibus limbum versus paullo hirtellis aliter glabris; tubo proprio 2 mm. longo; faucibus cylindratis 3 mm. longis; dentibus limbi lanceolatis acutis valde patentibus vel reflexis ca. 1 mm. longis et 0.5 mm. latis; achaeniis maturitate nigrescentibus sursum in faciebus et praecipue in angulis hispidulis ca. 2.6 mm. longis; pappi setis ca. 15 delicatule capillaribus albis 4-4.8 mm. longis.—Prrvu: Department of Lima: in clumps 6-9 dm. high on rocky eastern slope; flowers white, Matucana, alt. 2440 m., Apr. 12-May 3, 1922, Macbride & Featherstone, no. 481 (TYPE, Field Mus., isotype Gr.). This species is in its more significant characters very close to E. Sternbergianum DC., but differs in having more numerously flowered heads of somewhat larger size and larger leaves which are distinctly longis plerumque curvatis basin versus attenuatis atrogriseis in costis sursum breviter ciliatis et in superficiebus subpuberulis; pappi setis ca. 35 albidis minute scabratis corollam subaequantibus.—CvuBa: 26 ROBINSON Province Oriente: near Bayate in the valley of the Rio Bayate, Sierra de Nipe, 13 July, 1914, Ekman, no. 1976 (Brl., fragm. Gr.). This plant is clearly related to the somewhat variable E. plucheoides Griseb. but has much smaller, fewer-flowered heads, more obtuse phyllaries and quite a different pubescence. In E. plucheoides the upper surface of the leaves is at first appressed-villous. In the present plant it has scattered subsessile globose glands but is other- wise free from pubescence. The stem furthermore has a very short pubescence of curled hairs, while in E. plucheoides it is very densely appressed villous-tomentose especially about the nodes. In some points the present plant agrees with the description of the still obscure E. libanoticum Sch.-Bip., but as Schultz states that that has 55-flowered heads. and obtuse leaves about 2 inches by 1 inch of ovate form, it seems quite impossible that the two should prove conspecific. E. (§ Cylindrocephala) odoratum L., var. mallotophyllum, var. nov., differt a var. typico praecique foliis brevioribus plerisque 2.5- 3.5 em. longis utrique dense griseo-tomentosis—CoLomB1A: Caqueza, Dept. Cundinamarca, 1700 m., Jan. 1856, Triana, no. 1202 (TYPE, Par., fragm. Gr.); Gachetd, Dept. Quesada, Triana, no. 1203 (Par., fragm. Gr.). ¥ So great has been the confusion both in the classification and the nomenclature of this species that a new variety is here recorded with some hesitation. However, even after many related species have been segregated (largely through the efforts of the late Prof. Hierony- mus) there is still left in E. odoratum material of considerable diversity not as yet distinguished by varietal names. Thus the velvety plant here characterized certainly merits some separate designation from the often nearly smooth forms of this common and widely distributed species. Triana’s two specimens, 1202 and 1203, are not precisely identical, though their differences may well be merely ecological resulting from response to different degrees of exposure or due in part perhaps to different stages of maturity. _In 1202 the phyllaries are shorter, more stramineo-scarious, paler, and tend to be more obtuse. In his 1203 they are darker, firmer in texture, longer, narrower, and at least the inner more acute. In foliage these two forms appear entirely alike and their involucral differences are no greater than those found else- where in the species. E. Pennettn Robinson, Proc. Am. Acad. liv. 254 (1918). The type material of this species, collected by Rusby & Pennell at the RECORDS OF THE EUPATORIEAE,—VI 27 foot of the Cordillera Oriental, near Neiva, Dept. Huila, Colombia, had as stated in the original description leaves which were subglabrous beneath. A very similar plant, collected in a hillside field at Salento, Dept. Caldas, in the Cordillera Central, alt. 1800-2100 m., by Pennell, Killip & Hazen, no. 8751 (Gr.), has the leaves very definitely hairy on the veins and nerves beneath. In all other respects the plants appear so closely similar that it seems best to regard them as conspecific variations of minor importance. The extension of the range to the Cordillera Central is of interest in a species previously observed only in the Cordillera Oriental and is, therefore, here put on record. E. (§ Eximbricata) petrophilum, spec. nov., herbaceum annuum erectum ca. 5 dm. altum fere a basi ramosum molliter villosum; radice principi subsimplici deorsum decrescenti, secundariis paucis fibrosis; caule tereti brunnescenti multicostulato ca. 4 mm. diametro; internodiis plerisque 1.5-3 cm. longis; ramis gracilibus oppositis adscendentibus plerumque simplicibus; foliis ovatis vel ovato- rhombeis acutis incise crenato-serratis basin versus integris et cuneatis 3-nerviis concoloribus supra adpresse villosulis subtus, longe et laxe -villoso-pubescentibus textura teneris 2.5-4 cm. longis 1-2.5 ¢m. latis; petiolis 5-9 mm. longis sursum alatis; corymbis caulem et ramos terminantibus pluribus primo densis subglobosis ad maturita- tem laxioribus et planiusculis; capitulis pedicellatis ca. 23-floris; ca. 6-8 mm. longis; involucri companulati squamis ca. 15 obovatis vel oblanceolatis obtusissimis vel apice rotundatis longitudine sub- aequalibus extus dense villosis margine lacerato-ciliatis, maximis (immaturis) 5 mm. longis et 2.2 mm. latis; receptaculo parvo planius- culo; corollis glabris ca. 4.mm. longis; tubo proprio gracili 1 mm. longo, faucibus cylindratis paullo ampliatis ca. 1.5 mm. longis; dentibus limbi deltoideis 0.5 mm. longis; achaeniis (valde immaturis) ca. 1.5 mm. longis cum glandulis globosis rubris ornatis; pappi setis ca. 30 albis capillaribus laevibus corollam subaequantibus; antheris apice distincte appendiculatis—E. steviaefolium;Glaziou, Bull. Soc. Bot. Fr. lv. Mém. 3; 386 (1909), in part, nat DC.—Braziu: Itatiaia, among rocks on the campo, Rio Janeiroy Dr. A. Glaziou, no. 6572 (ryeE, Bot. Mus. Copenhagen, fragm. , This plant differs from E. steviaefolium DC. in having rhombic- ovate leaves (becoming 2.5 cm. wide) instead of lanceolate leaves (at most 1 cm. wide)... © Its petioles (somewhat winged toward the summit) are 5-9 mm. long, while in E. steviaefolium they are exceedingly short (1-3 mm. in length). The leaf-blade is much more deeply and incisely crenate- 28 ROBINSON serrate than in E. steviaefolium and the achenes much more densely beset with red glands. In other respects it appears closely similar, except that it is clearly an annual, while EF. steviaefolium is certainly at least sometimes a perennial herb if not a shrub. Unfortunately Glaziou’s no. 6572 is very immature. It is possible that better developed specimens will disclose other differential traits. At all events this plant cannot be satisfactorily left in E. steviaefoliwm DC., the only question being whether it is best treated as a broad- leaved variety or as a distinct species. In the absence of any inter- mediates the latter course seems the more appropriate. E. (§ Subimbricata) pithecobium, spec. nov., frattioomean epiphyticum ab arboribus dependens obscure crispeque puberulum caulibus 12-15 dm. longis crassis curvatis solum apicem versus foliiferis pallide grisea obtectis alterni- rameis vel dichotomis; foliis oppositis vel alternis petiolatis oblongis vel oblongo-lanceolatis gradatim acuminatis argute serrato-dentatis (dentibus utroque latere 5-8 ca. 0.5-2 mm. altis et 1-1.5 cm. inter se distantibus) basi cuneatis penniveniis utrinque viridibus puncticulatis subtus praecipue in costa hirtellis 5-12 em. longis 2-5 cm. latis ut videtur textura subcarnosis; petiolis 1-1.5 cm. longis gracilibus; corymbis terminalibus ca. 12-capitulatis folia subtendentia subae- quantibus; pedicellis filiformibus saepe rectis 1-1.5 cm. longis; capi- tulis ca. 22-floris ca. 11 mm. longis et 8 mm. crassis; involucri campanu- lati squamis ca. 18 hirtellis 3—4-seriatim gp extimis lanceolatis attenuatis ca. 3.5-5 mm. longis et 1-1.5 mm. latis, intermediis oblongis apice acutis vel obtusis vel rotundatis et re roseis ca. 6 mm. longis et 1.7 mm. latis, intimis 1-3 spatulatis angustis; corollis roseis 6 mm. longis glabris a basi ad limbum gradatim ampliatis sine faucibus distinctis; dentibus limbi ovatis obtusis ca. 0.6 mm. longis; achaeniis ca. 3 mm. longis in angulis minute scabratis; pappi setis ca. 40 albis capillaribus ad apicem non incrassatis—Costa Rica: Province of San José: a shrub pendent from tree in dense wet forest, flowers pale pink, eas de la Chonta, northeast of Santa Maria de Dota, alt. 2000-2100 , Dec. 18, 1925, Standley, no. 42,181 (rypE, in Gray Herb.); near Fine La Cima, above Los Lotes, north of El Copey, alt. 2100-2400 -» Dec. 21-22, 1925, Standley, no. 42,727 (Gr.). species remarkable from its epiphytic habit, a trait which it shares with E. araliaefolium Less., E. eximium Robinson, and prob- ably E. —— Kuntze, E. artirodes Robinson and E. sciaphilum E. § Eximbricata) piurae piurae, spec. nov., fruticosum adscendenter ramosum > iaaiiesteieentellinin: caulibus cabkebetibadils ramis saepe RECORDS OF THE EUPATORIEAE,—VI 29 subfastigiatis primo foliosis tardius basin versus delapsu foliorum denudatis nodulosis; internodiis 7-10 mm. longis; nodis subarticulatis; foliis oppositis elliptico-ovatis petiolatis obtusis crenato-serratis basi rotundatis supra basin subtrinerviis vel distincte penniveniis crassius- culis supra in nervis paullo puberulis aliter glabris rugosis post exsiccationem nigrescentibus subtus multo pallidioribus sordide lanu- latis 1.5-2.2 em. longis 9-14 mm. latis; petiolo 2-4 mm. longo; corym- bis (immaturis) compactis terminalibus convexis 3-4.5 cm. diametro multicapitulatis; capitulis ca. 15-floris ca. 10 mm. longis; pedicellis ca. 2 mm. longis; involucri squamis ca. 14 subaequalibus oblongis apice rotundatis margine ciliolatis dorso brunneo-tomentellis, plerisque 2 mm. latis, extimis 2-3 brevioribus ca. 3-4 mm. longis et 1 mm. latis, intimis 1-2 angustis; receptaculo parvo planiusculo; corollae glabrae tubo proprio gracili 2 mm. longo, faucibus cylindratis ca. 4 mm. longis, limbi dentibus anguste deltoideis ca. 0.6 mm. longis; antheris apice cum appendice oblongo hyalino in- structis; achaeniis ca. 2.8 mm. longis in angulis dense glanduliferis; pappi setis 45-50 ca. 5.5 mm. longis albidis inaequalibus vix scabratis in apices attenuatis——Prru: above Huancabamba, Prov. Huanca- bamba, Dept. Piura, alt. 3200-3300 m., Apr., 1912, Dr. A. Weber- bauer, no. 6045 (type, Field Mus. Nat. Hist., phot. and fragm. Gr.). This characteristic plant may perhaps best be placed near E. cutervense Hieron. which it resembles somewhat in habit and small short-petioled leaves. However, it differs much in its dense in- florescence, in leaf-form, in its lanulate indument on the lower surface of the leaves, and fewer-flowered, smaller heads. E. poLycEPHALUM Sch.-Bip. ex Bak. in Mart. Fl. Bras. vi. pt. 2, Lond. Jour. Bot. v. 472 (1846). When preparing his treatment of Eupatorium for the Flora Brasiliensis Baker maintained Gardner’s species and mentioned E. polycephalum Sch.-Bip. in its synonymy. He also reduced the earlier E. bracteatum Hook. & Arn. Comp. Bot. Mag. i. 239 (Mar., 1836) to the position of a variety of E. tozziaefolium DC. Prod. v. 146 (Sept., 1836). It is evident that this course is not in accord with the International Rules in Nomenclature. E. bracteatum Hook. & Arn. is now known to have had some months of priority over E. tozziaefolium DC. and cannot be reduced to it, even in varietal status. The matter is some- what complicated by the circumstance that E. congestum Hook. & Arn. Comp. Bot. Mag. i. 239 (Mar., 1836) was also closely of this 30 ROBINSON affinity and was also considered by Baker conspecific with E. tozziae- folium DC., being indeed regarded as representing the typical variety of that species, while E. bracteatum Hook. & Arn. was treated by Baker as a variety. It may be added that E. bracteatum has a slight page-priority over E. congestum. To the writer, who has studied photographs and bits of the types concerned, it seems pretty clear that their differences are sufficient to prove them separate species. There- fore, it becomes necessary to make the’ following nomenclatorial readjustment. E. concestum Hook. & Arn. Comp. Bot. Mag. i. 239 (Mar., 1836). E. tozziaefolium DC. Prod. v. 146 (Sept., 1836). BRACTEATUM Hook. & Arn. Comp. Bot. Mag. i. 239 (Mar., 1836). E. potycerHatuM Sch.-Bip. ex Bak. in Mart. Fl. Bras. vi. pt. 2, 338, t. 88 (1876). Name published in synonymy but here raised to full specific rank being associated with Baker’s adequate diagnosis of E. bracteatum Gardn. Under the specific name E. polycephalum Sch.- Bip. may be placed the synonym E. bracteatum Gardn. in Hook. Lond. Jour. Bot. v. 472 (1846). Of the species E. polycephalum three varie- ties may be distinguished: Var. typicum, foliis dentato-crenatis vel saepe subintegris. Var. pinnatifidum (Bak.), comb. nov., foliis leviter pinnatifido- lobulatis.—E. bracteatum, var. pinnatifida Bak. |. c. 339 (1876). Var. racemosum (Bak.), comb. nov., foliis angustius oblanceo- latis majus venulosis, indumento laxiori cum glandulis intersperso.— E. bracteatum, var. racemosa Bak. |. c. 339 (1876). E. (§ Eximbricata) prinodes, spec. nov., fruticosum; caulibus teretibus a cortice laevi pallide griseo tectis apicem versus saepius copiose ramosis; ramis adscendenter patentibus juventate pubentibus deinde glabratis; foliis oppositis breviter petiolatis ovato-oblongis acutis pungenter dentatis (dentibus plerisque 1-2 mm. altis et 5-9 mm. inter se distantibus) basi rotundatis vel brevissime cordatis paullo supra basin 3-5-nervatis 3.5-6.5 em. longis 1.5-2.8 em. latis textura coriaceis utrinque viridibus subnitidis reticulato-venulosis supra glaberrimis subtus vix pallidioribus aliquando in costa prope basin obsolete puberulis; petiolo crassiusculo saepius brunnescenti 3-5 mm. longo plerumque patenter villosulo; corymbis ad apicem ramorum subsessilibus 3-8 em. diametro convexis pauci- vel saepius multicapitulatis glabris vel paullo pubentibus; capitulis ca. 20-floris 7-8 mm. longis; pedicellis nigrescentibus 3-8(-12) mm. longis; involucri squamis ca. 18 anguste linearib attenuatis glat is p re we RECORDS OF THE EUPATORIEAE,—VI 31 tibus; receptaculo parvo planiusculo nudo; corollis fere tubulatis 2.8 mm. longis limbum versus resinoso-granulatis aliter glabris sursum gradatim ampliatis sine faucibus distinctis; dentibus limbi deltoideis recurvatis; antheris apice bene appendiculatis; achaeniis olivaceis 2.4~2.7 mm. longis deorsum decrescentibus in costis sursum ciliolatis; pappi setis ca. 35 albis sublaevibus ca. 2.2 mm. longis attenuatis.— Cusa: Province Oriente: in shady places below Yunipero in the Sierra de Nipe at the foot of Loma Mensura, alt. 725 m., 11 July, 1919, Ekman, no. 9709 (ryPE, in Gray Herb.); in thickets, Loma de Estella, Sierra de Nipe, 18 June, 1915, Ekman, no. 5797 (Gr.); Loma Mensura in Sierra de Nipe at the foot of the mountains, 19 Oct., 1914, Ekman, no. 3167 (Gr. In its instesly toothed short-petioled leaves of firm texture as well as in its narrow, firm and acute phyllaries this species bears some slight resemblance to E. incisum Griseb., but that is essentially herbaceous, has ovate rather than ovate-oblong leaves (pubescent on both surfaces) and pedunculate cor S. E. pycNocEPHALUM Less. Linnaea vi. 404 (1831). Although long known to exhibit considerable variation of leaf-contour, density of inflorescense, etc., this species has been treated thus far as a poly- morphous aggregate, and little effort has been made to name its forms, most of which are probably mere inconstant combinations of characters subject to independent variation. Somewhat better marked a to be the following: ucanum, var. nov., foliis quam apud formam typicam amit saepius 2-3 cm. ‘iniasies corymbulis paniculatim dispositis plerumque 2-5(-7)-capitulatis quam apud formam typicam laxioribus; pedicellis plerisque 2-7 mm. longis; corollis albis.—CoLoMBIA: Department of El Cauca: woodland, between Rio Ortega and “El Ramal,” Cordillera Occidental, alt. 1900-2200 m. ” July 2, 1922, Pennell & Killip, no. 8072 (ryPE, in Gray Herb.). This variety recalls some plants of eastern Brazil in which the heads are more loosely disposed than elsewhere characteristic of the species, but it cannot be closely matched with. any of them and it appears to merit record. Indeed it is probable that it would be regarded as fully worthy of specific rank were it not for the already mentioned variability of E. pycnocephalum with which it agrees surprisingly | in many details of its involucre, ete. E. Eximbricata) rhodopappum, spec. nov., herbaceum perenne ut videtur adscendens 4 dm. vel ultra altitudine; caule tereti brunnescenti-griseo minute villosulo (villis tenuibus subappressis 32 ROBINSON purpureo-articulatis) ; medulla post exsiccationem viridescenti-flavida; internodiis plerisque 3-6 cm. longis; foliis oppositis graciliter petiolatis ovato-oblongis acuminatis argute serratis (dentibus ca. 1 mm. altis 3-5 mm. inter se distantibus cuspidatis) basi obtusis vel rotundatis textura membranaceis supra basin 5(~7)-nervatis untrique supra in nervis obsolete puberulis aliter glabris depresse reticulatis subtus sordide in nervis et venulis hirtellis 4~6.7 em. longis 1.5-3 cm. latis; petiolo 1.5-1.8 em. longo crispe et subappresse puberulo non glandu- loso; ecorymbis terminalibus convexis laxiusculis plerisque 2.5-5 cm diametro; pedicellis filiformibus 8-24 mm. longis dense _hirtellis capitulis ca. 30-floris 7 mm. altis 8 mm. diametro; involucri campanu- lati squamis subaequalibus ca. 16, exterioribus lanceolato-linearibus acutis fuscis dense hirtellis 2.5-3 mm. longis vix costulatis, interioribus lineari-oblongis obtusiusculis tenuioribus ca. 3 mm. longis et 0.7 mm. latis plerumque 2-costulatis substramineis; corollis ut videtur albis limbum versus hirtellis; tubo proprio gracili ca. 1.6 mm. longo; faucibus campanulati-subcylindratis ca. 1.7 mm. longis; dentibus limbi deltoideis ca. 0.5 mm. longis; achaeniis (immaturis) ca. 1.5 mm. longis; pappi setis ca. 17 pulcherrime (saltem post exsiccationem) roseis ca. 3 mm. longis basin versus albidis.—CoLomB1a: Department of Caldas: open rocky stream bank, Rio San Rafael, below Cerro Tatama, Cordillera Occidental, alt. 2500-2800.m., Dr. Francis W. Pennell, no. 10,424 (rype, in Gray Herb.). This species, obviously close to E. pichinchense HBK., differs in its ovate-oblong rather than broadly ovate leaves, its considerably shorter and less spreading inflorescence, somewhat less numerously flowered heads, and bright roseate pappus. Its greenish-yellow, chambered pith is also a peculiar character not observed elsewhere in Eupatorium. ke effort has failed to place this plant in any hitherto known E. (§ Bublnitelests) goat ala spec. nov., multicaule decumbens basin versus lignescens gracile debile wabestenn’ caulibus teretibus viridibus plus minusve violascenti-tinctis costulatis usque ad 2 m. longis et 4 mm. diametro albo-medullosis; internodiis 2-16 em. longis; foliis oppositis petiolatis deltoideo-ovatis vel ovato- lanceolatis acuminatis grosse crenato-dentatis basi cordatis vel sub- truncatis vel rotundatis utrinque viridibus et pubescentibus subtus vix pallidioribus textura membranaceis a basi 3(-5)-nervatis 4-6.5 cm. longis 1.7-4 cm. latis translucente reticulatis; petiolis gracilibus 1.5-2.5 cm. longis villosulis (villis p' purpureo-art etienlatin): corymbis veestaraeensse ar lateralibus laxiusculis plus minusve irregularibus; RECORDS OF THE EUPATORIEAE,—VI 33 bracteolis lanceolatis 3-7 mm. longis; pedicellis filiformibus 8-10 mm. longis; capitulis ca. 20-floris 6 mm. altis 3.5 mm. diametro; involucri squamis ca. 20 ca. triseriatim imbricatis roseo-griseis tenuibus dorso dense hirtellis, exterioribus lanceolatis attenuatis, intermediis acumi- natis, interioribus anguste oblongis acutis vel acutiusculis ca, 6 mm. longis et 1 mm. latis; receptaculo plano glabro; corollis roseis vel lilacinis limbum versus paullo hispidulis aliter glabris; tubo proprio ca. 1.3 mm. longo; faucibus paullo ampliatis cylindratis ca. 3 mm. longis; dentibus limbi deltoideis ca. 0.6 mm. longis; achaeniis glaberrimis 2 mm. longis maturitate atrobrunneis vel nigris cum costis pallidis; pappi setis ca. 27 basi in annulo brevissime connatis.—PErRv: Depart- ment of Huanuco: in canyon, stems many, woody below, weak, be- coming 1.2 m. long if among other shrubs; flowers old rose, fragrant, Yanahuanca, alt. 3050 m., June 16-22, 1922, Macbride & Feather- stone, no. 1228 (Type, in Field Mus., isotype Gr.); dense thicket floor, stems 0.6-1.2(-1.8) m. long, flowers pink; vernacular name “macha macha,” Mito, alt. 2750 m., July 8-22, 1922, Macbride & Featherstone, no. 1508 (Field Mus., Gr.) ; open thicket on stream bank, 0.6-1.5 m. high, flowers lavender, 24 km. southeast of Huanuco, alt. 3200 m., May 31-June 3, 1922, Macbride & Featherstone, no. 7 (Field Mus., Gr.). This species, admirably shown in copious material of three collec- tions, approaches most nearly E. gracilentum Robinson, but that is a considerably smaller plant with leaves (only 1.8-3 cm. long) more copiously pubescent, in fact tomentellous beneath; corolla somewhat shorter, and achenes conspicuously ciliated on the angles. E. (§ Eximbricata) rhypodes Robinson, Proc. Am. Acad. lv. 28 (1919). This species is found to be capable of the following formal subdivision. Forma «. typicum, manifeste viscidum; caule et ramis et pedi- cellis_ glanduloso-tomentosis; pilis capituliferis; inflorescentia sub- axa—Ecvapor: Vicinity of Las Juntas, Rose, Pachano & Rose, no. 23,179 (Gr., U. S., N. Y.). Forma ¢. agloeum, forma nov., multo minus visccsum vel omnino non viscosum; caule etc. setuloso-tomentoso; pilis attenuatis nec capituliferis; inflorescentia densiuscula.—PERv: Palea[Palea?], Dombey (Par., phot. and small frag. Gr.). The label of Dombey bears the words “guaxmi guaxmi” presumably a memorandum of a local vernacular name. E. (§ Eximbricata) rhytidodes, spec. nov., fruticcsum 2 m. altum; caulibus gracilibus flexuosis teretibus 3 mm. diametro albo- 34 ROBINSON medullosis juventate parce setuliferis tardius glabratis griseo-brun- nescentibus; internodiis cm. longis; foliis oppositis petiolatis exsiccationem) atrobrunneis glabris lucidulis valde rugosis subtus pal- lide viridibus prominulenter reticulatis in nervis venisque patenter setuliferis textura firmiter membranaceis 5-6 cm. longis 3-4.5 em. — latis; petiolis gracilibus 1.5-2.5 cm. longis; sales terminalibus compositis rotundatis foliaceo-bracteatis; partialibus 3-6 cm. diametro valde convexis; pedicellis plerisque 4-9 mm. longis; capitulis ca. 24- floris ca. 7 mm. altis et 6 mm. diametro; ohne squamis lanceolato- linearibus acutis plerumque 3-costulatis brunneis apicem versus paullo hirtellis, i ca. 4.5m m. longis, exteriori- bus 2-3 parvis brevioribus; corollis albis; tubo proprio gracili 2 mm. longis; faucibus cylindratis 2 mm. longis; dentibus limbi anguste deltoideis ca. 0.7 mm. longis dorso cum setulis 1-3 minutis instructis; achaeniis maturitate nigrescentibus, 2.2 mm. longis parce in angulis hirtellis; pappi setis ca. 26 delicatule capillaribus flavescenti-albis minute scabratis ca. 3 mm. longis sursum non incrassatis.—PERU: Dept. Huanuco: very slender-stemmed shrub in sunny thicket; flowers white, fragrant, Mufia, alt. 2135 m., May 23-June 4, 1923, J. Francis Macbride, no. 3985 (Type, in F id Mus., isotype Gr.). This species seems well marked among the icdean Eupatoriums. Its round-ovate firmish leaves are peculiarly rugulose on their glabrous upper surface and are setulose on the prominent nerves and veins beneath E. (§ Subimbricata) rivulorum, spec. nov., herbaceum perenne gracile ramosum ubique minute puberulum; caule subtereti obscure striato 1.5-2 mm. diametro viridi vel purpurascenti debili albo- medulloso; internodiis 2-8 cm. longis; foliis oppositis petiolatis lanceolatis gradatim acutatis apice attenuatis vel acutis vel obtusius- culis margine a media parte serratis vel crenato-serratis (dentibus paucis utroque 2-7) basi cuneatis integris ima a asi 3-nervatis utrinque viridibus obscure puberulis concoloribus 2-3.5 em. longis 8-15 mm. latis membranaceis; petiolo gracili 5-10 mm. longo; pani- cula terminali laxiuscula irregulariter saepe alternatim ramosa foli- aceo-bracteata 1.5-3 dm. alta et crassa; pedicellis 3-12 mm. longis; capitulis ca. 40-floris ca. 6 mm. altis et crassis; involucri campan- ulati squamis ca. 30 subtriseriatim imbricatis viridibus acutis plerumque 2-costulatis obscure puberulis, extimis anguste lanceolati- linearibus attenuatis 2 mm. longis, intermediis naan peracutis, intimis lineari-oblongis breviter acutatis ca. 4 mm. lon RECORDS OF THE EUPATORIEAE,—VI 35 et 0.8 mm. latis; corollis roseo-lilacinis tubulatis sursum paullo gradatim ampliatis puberulis 3.3 mm. longis; dentibus limbi brevibus late deltoideis vix 0.2 mm. longis; achaeniis nigris 1.3 mm. longis sursum praecipue in costis brevissime hispidulis basi pallide callosis; pappi setis ca. 12 delicatule capillaribus albis ca. 3 mm. longis.— Mexico: State of Chiapas: on rocks in the creeks east of Monserrate, March, 1925, Dr. C. A. Purpus, no. 34 (Type, in U. S. Nat. Herb. and Gray Herb.). his species in habit, habitat, foliage, and in some details of pubes- cence recalls Fleischmannia arguta (HBK.) Robinson, but has smaller heads in a more developed panicle, a more imbricated involucre, with more ovate intermediate phyllaries and also indefinite (mostly 11-12) pappus-bristles. The species, if referred to Fleischmannia, would by its indefinite (though not very numerous) pappus-bristles break down the slight distinction between that genus and Eupa- torium. If, on the other hand, it is referred to Eupatorium (from which on technical grounds it cannot be readily separated) its close similarity to Fleischmannia must render the further separation of that genus rather artificial. Neither disposition is entirely satis- factory E. 6 Subimbricata) Ruckeri, spec. nov., grande ut videtur herbaceum perenne inflorescentia paullo puberula excepta aspectu glaberrimum; caule inde subtereti sursum paullo angulato fistuloso; foliis oppositis graciliter petiolatis ovatis obtusis vel apice angustatis et deinde rotundatis in lateribus grosse crenato-lobulatis vel grosse dentatis basi obtusis vel rotundatis utrinque obscurissime in nervis puberulis aliter glaberrimis subcarnosis ca. 8-10 mm. supra basin 3-nervatis 7-13 em. longis 4-8 em. latis impunctatis; dentibus pleris- que 2-9 mm. altis et basi latis apice rotundatis; petiolis 2.3-4.7 cm. longis exalatis, eis ejusdem jugi basi paullo connatis; corymbo ter- minali denso rotundato 3 cm. vel ultra diametro breviter sordideque puberulo; bracteis linearibus ca. 6 mm. longis; capitulis ca. 60-floris 9 mm. altis 10 mm. diametro ok late involucri campanulati squa- mis viridibus subcarnosis oblongis obtusis vel apice rotundatis, intimis 5.5 mm. longis 1.5 mm. latis eroso-ciliolatis, exterioribus paucis gradatim brevioribus lanceolato-linearibus; corollis pallidis glabris limbum versus purpureo-roseis ca. 5 mm. longis tubulatis paullo limbum versus gradatim ampliatis sine faucibus distinctis; antheris apice cum appendice membranacea obtusissima ornatis; achaeniis immaturis ca. 2.3 mm. longis glabris; pappi setis ca. 4 mm. longis.— Cusa: Province Oriente: Sierra Maestra, on rocks at Casa Piedra, 36 ROBINSON Rio Yara, Aug. 10, 1922, E. L. Ekman, no. 14,834 (Brl., phot. and fragm. Gr.); a “tall weed” in windfall, Loma Cardero, Pico Tarquino, Oct. 12-14, 1924, G. C. Rucker, no. 45 (TyPx, in herb. of N. Y. Bot. Gard., phot. and fragm. Gr.). This species bears considerable ee to E. Shafert Robinson and E. iodostylum Robinson, but has heads with many more florets and strikingly lobulate, ovate, blunt-based leaves on long slender wingless petioles. Dr. Ekman’s specimen showed sterile shoots only and proved indeterminate until through the kindness of Dr. Britton the author was able to see the rather scanty but flowering material collected by Mr. Rucker and now in the herbarium of the New York Botanical Garden. E. SANCTOPAULENSE Robinson, Contrib. Gray Herb. n. ser. Ixviil. 32 (1923). Since the description of this Brazilian species (of § Praxe- lis) founded upon Saint-Hilaire’s no. 1401 three additional collections of the plant have come to the attention of the author, being as follows: Sao Paulo, 1840-46, Gustave Perdonnet, no. 274 (hb. Univ. Geneva); meadows, Santa Ana, Sao Paulo, Feb. 22, 1922, purple-flowered, Prof. & Mrs. E. W. D. Holway, no. 1587 (Gr.); and between Ipanema and Sorocaba, Campo of Ipiranga, Sado Paulo, Glaziou, no. 16,190 (Copenh.). This newly examined material agrees closely with the original specimen of Saint-Hilaire and yields some data regarding the base of the plant which was not shown in the type material. Glaziou’s excellent specimen shows that the plant is an essentially erect, slender perennial herb 5-6 dm. high with 4-7 stems from a hard woody globu- lar corm 6-8 mm. thick and provided with several subsimple tough slender fibrous roots. This number is reported in Glaziou’s list, Bull. Soe. Bot. Fr. lvi. Mém. 3, 389 (1909) as E. kleinioides HBK., but of course is readily distinguished from that species by its involucre with blunter phyllaries, as well as its perennial nature. E. (§ Subimbricata) scabrifolium, spec. nov., ligncsum valde elongatum scandens dumeta alnorum penetrans usque ad 15 m longum; caule juventate patenter pallide pubescenti et fuliginoso- sordido tardius glabrato et cum cortice pallide grisea lenticellata ob- tecto; ligno albo denso; medulla alba parva; internodiis 8-15 cm. longis; foliis oppositis petiolatis lanceolatis vel lanceolato-ovatis longe attenuatis margine revolutis vel integris vel paullo undulatis vel plus minusve cuspidato-denticulatis basi breviter cuneatis vel subrotunda- tis 7-11 cm. longis 1.5-4 cm. latis in parte tertia proximali pinnatim 7-ll-nervatis supra scaberrimis bullato-rugulosis subtus prominulenter reticulatis tenuiter pubescentibus textura chartaceo- RECORDS OF THE EUPATORIEAE,—VI 37 coriaceis subconcoloribus; petiolo gracili 5-12 mm. longo; corymbis terminalibus trifidis 10-14 em. diametro subdensis fuliginoso-sordidis; capitulis sessilibus vel breviter (0.5-1.5 mm. longe) pedicellatis ca. 9-floris ca. 7 mm. longis; involucri squamis ca. 3-seriatim imbricatis apice rotundatis plerisque 3-costatis fusco-ciliolatis dorso paullo _ puberulis apicem versus saepe purpureo-tinctis, exterioribus late ovatis, interioribus ellipticis vel oblanceolatis; corollis limbum versus lilaceis gracilibus; tubo ca. 4.2 mm. longo sine faucibus distinctis; dentibus limbi deltoideis 0.6 mm. longis et 0.5 mm. latis extus minute granulatis; receptaculo parvo plano glabro; achaeniis maturitate nigris 2 mm. longis basi pallide callosis in angulis minute pubentibus in facie- bus parce granulatis; pappi setis 35-40 stramineo-albis attenuatis vix scabratis ca. 3.8 mm. longis.— Peru: Dept. Huanuco: liana, much branched above to 8 m., flowers lilac; in alders of dense stream jungle, Mito, alt. 2745 m., July 23-Aug. 14, 1922, Macbride & Featherstone, no. 1873 (TyPE, in Field Mus., isotype Gr.); liana-shrub in montafia, flowers purple, Yanano, alt. 1830 m., June 29, 19234 Machride, no. 4937 (Field Mus., Gr.). This species differs from E. trachyphyllum Hieron. in its much broader, firm, and purple-stained phyllaries, and shorter, more deeply colored corollas. From E. inulaefolium HBK. it differs in the leaf- form, which is not so rhombic-ovate, in the leaf-indument which is both shorter and harsher, and in the broader more deeply-colored phyllaries. It appears furthermore to be of quite different habit, being a liana, rather than an erect shru E. (§ Eximbricata) silvaticum, spec. nov., fruticosum ramis gracilibus vix 2 mm. diametro curvato-adscendentibus striato-costu- latis griseo-brunneis vel atropurpureis brevissime puberulis et juvent- ate cum glandulis lucidis sessilibus ornatis; internodiis 1.4-3.7 em. longis; foliis ovato-lanceolatis gradatim acutatis breviter cuspidato- serratis basin versus integris basi subacutis vel rotundatis textura coriaceis supra viridibus punctatis cum atomis lucidis ornatis alite glabris levissime sed dense reticulatis subtus subolivaceis densissime punctatis glabris laxe reticulatis supra basin 3-nervatis 2.5-3.2 em. longis 8-11 mm. latis; petiolo 2~3 mm. longo supra canaliculato; corymbis terminalibus laxiusculis 3-6 cm. altis 5-9 em. diametro; bracteis parvis lanceolato-linearibus 1-9 mm. longis; _pedicellis dorso saepe purpurascentibus puberulis, squamis exterioribus 2-4 38 ROBINSON linearibus acutis multo brevioribys; corollis ca. 3 mm. longis pallidis glanduliferis tubulatis sine faucibus distinctis; dentibus limbi brevibus vix 0.3 mm. longitudine; achaeniis ca. 1.5 mm. longis minute puberulis; pappi setis ca. 25 albidis sublaevibus corollam subaequantibus.)— Cusa: Province Oriente: in the edge of woods, Sierra de Nipe, 4 Oct., 1919, Ekman, no. 9811 (rTyPx, in Gray Herb.). This species in most characters closely resembles E. polystictum Urb., but differs so decidedly in its ovate-lanceolate acute and less reticulated leaves and its longer pedicels with consequent looser in- florescence, as to have a rather different habit. E. polystictum is now represented in herbaria by a fairly good series of specimens from several localities. While these show some minor variations of dimen- sions, texture, etc., no material hitherto obtained approaches in leaf- form or in inflorescence the plant here described. It therefore seems best, in the absence of evident intergradation, to treat it as a distinct © species. E. (§ Eximbricata) simillimum, spec. nov., fruticosum vel arboreum 4-6 m. altum in inflorescentia et in nervis foliorum immatur- orum puberulum vel brevissime sordide tomentellum aliter glaber- rimum; caule 1-3 dm. crasso; ramis pallide griseis subteretibus cum lenticellis ornatis; ramulis brunnescentibus costato-angulatis; foliis oppositis petiolatis ovato-oblongis acuminatis obtuse serratis basi plerumque rotundatis mox glabratis membranaceis conspicuiter supra basin pinnatim ca. 5-nervatis tenuibus post exsiccationem subnigres- centibus 8-12 cm. longis 3-5.5 em. latis; petiolo 1-1.8 cm. longo; inflorescentia ampla terminali thysoidea saepe composita basi vel ad mediam partem foliaceo-bracteata; pedicellis gracilibus puberulis saepe curvatis 3-10 mm. longis; capitulis ca. 15-floris 1 cm. altis; involucri squamis ca. 11, extimis 2-3 lanceolatis brevioribus ca. 3.8 mm. longis attenuatis subherbaceis, intermediis oblongis obtusis mucronatis obscure 3-nerviis 4.7-5.5 mm. longis 1.2-1.4 mm. latis, intimis 1-2 linearibus obtusis vix 0.7 mm. latis; receptaculo plano nudo; corollis ca. 5 mm. longis sparse glanduliferis albidis; tubo proprio ca. 2.7 mm. longo; faucibus vix amplatis ca. 2 mm. longis; dentibus limbi brevibus; achaeniis ca. 3 mm. longis gracilibus in costis paullo glandulari-scabratis; pappi setis ca. 40 inaequalibus corollam sub- aequantibus albidis.—E. laeve, var. macrophyllum Chod. Bull, Herb. Boiss. ser. 2, ii. 310 (1902). E. laeve, var. pubens Robinson, Contrib. Gray Herb. Ixv. 49 (1922).—Paraguay: A low tree, in bushy woods, Feb. 7, 1879, Lorentz, no. 11 (Berl.); in woods, Cordillera de Altos, Dec., 1898, Hassler, no. 3585 (hb. Univ. Geneva); same locality, RECORDS OF THE EUPATORIEAE,—VI 39 Oct. 20, 1902, Fiebrig, no. 277 (typE, Gray Herb.); in the region of Lake Ypacaray, March, 1913, Hassler, no. 12,147 (Gr., Copenh.); in woods, Paraguari, Aug., 1914, Chodat, no. 403 (hb. Univ. Geneva). This tree of Paraguay bears a close habital resemblance to E. laeve DC., a recognized dye-plant of Brazil. However, the two were long ago seen to be somewhat different and attempts have been made to give at least varietal distinction to the Paraguay plant upon leaf- form or pubescence. More thorough examination, now possible through the availability of more copious material, proves them to have a number of seemingly constant differences, indicating that they are in reality separate though nearly related species. In the Brazilian E. laeve the leaves are for the most part smaller, more oblong than ovate and more sharply toothed (as correctly shown in plate 93 illustrating the species in the Flora Brasiliensis); the inflorescence is strictly glabrous; the heads are slightly smaller (about 9 mm. long) and the florets tend to be more numerous (about 20); the corollas are shorter (about 4 mm. long), the pappus more distinctly yellowish; and the phyllaries narrower and all acute, as well as of firmer texture. E. (§ Subimbricata-Eximbricata) sinaloense, spec. nov., erec- tum gracile herbaceum verisimiliter perenne, ca. 7 dm. altum; caule tereti viridi-striato crispe puberulo ca. 2 mm. diametro albo-medulloso; foliis oppositis graciliter petiolatis ovatis acuminatis argute grosseque serrato-dentatis basi abrupte angustatis imam ad basin breviter acutatis vel subacuminatis untringue viridibus supra sparse pubenti- bus subtus in nervis venisque crispe puberulis ima a basi 3-nervatis usque ad 6 cm. longis et 3 em. latis tenuibus membranaceis; petiolo plerumque 1.5-2 em. longo; corymbis multis parvis 1-3 cm. diametro densiusculis convexis ramos superiores terminantibus conjunctim in paniculam laxam elongatam folioso-bracteatam dispositis; pedicellis filiformibus 2-3 mm. longis; capitulis ca. 25-floris ca. 4.2 mm. longis et 3.3 mm. diametro; involucri turbinato-campanulati squamis ca. 15 valde inaequalibus acutis viridibus, extimis lanceolatis gradatis, intimis lineari-oblongis saepius 2-costulatis; receptaculo paullo con- vexo nudo; corollis glabris albis 2 mm. longis; tubo proprio brevissimo vix 0.5 mm. longo; faucibus subcylindratis ca. 1.2 mm. longis sursum paullo ampliatis; dentibus limbi 0.4 mm. longis ovato-deltoideis delicatule papillosis ; achaeniis ca. 1.2 mm. longis in costis sursum lara — setis ca. 22 capillaribus albis ca. 1.5 mm. longis.— Mex rro Colorado near Capradea, Sinaloa, Nov. 1, 1904, T. S. Scoadbed pertiny in Gray Herb.); also Sinaloa without more precise locality, 1922, J. G. Ortega, no. 4426 (Gr., U. S.). 40 ROBINSON _ A species somewhat resembling the common E. pycnocephalum Less. but easily distinguished by its more delicate habit, smaller heads and acute phyllaries. Mr. Brandegee distributed his specimens as a form of E. pawperculum Gray, a species which, however, ‘has lanceolate less coarsely toothed leaves usually 3-nerved from above the base, distinctly larger heads, and more pubescent campanulate and less calyculate involucre. FE. sinaloense has an involucre with graded scales but with little imbrication and falls somewhat between the typical development of § § Subimbricata and Eximbricata. E. (§ Subimbricata) Standleyi, spec. nov., herbaceum 1.2-3 m. altum primo erectum deinde in arbores conscendens inflorescentia excepta glabrum vel subglabrum; caule subtereti crasso fistuloso post exsiccationem costulato; foliis oppositis petiolatis ovato-oblongis subacutis vel obtusis margine grosse dentatis basi breviter cuneatis penniveniis subcarnosis utrinque glabris supra viridibus subtus paullo pallidioribus 12-16 cm. longis 7-9 cm. latis; petiolo 3-4.5 cm. longo supra sulcato; panicula terminali 2-3 dm. lata composita sub- laxa puberula; capitulis ca. 6-floris 8 mm. altis 4 mm. diametro pedicellatis a bracteola lineari ca. 2 mm. longa suffultis; involucri squamis ca. 8 valde inaequalibus acutis vel subobtusis dorso glabris " margine ciliolatis, extimis ovatis ca. 2.6 mm. longis 1.7 mm. latis, intermediis 2.5 mm. longis 1.4 mm. latis, intimis 5.5-7.5 mm. ‘longis 0.7-1.2 mm. latis; corollis roseis 5 mm. longis glabris; tubo proprio gracili 2.2 mm. longo; faucibus ampliatis subcylindratis 2.2 mm. longis; dentibus limbi deltoideis 0.6 mm. longis; achaeniis gracilibus glabris 2.5 mm. longis; pappi setis ca. 40 corollam longitudine aequan- tibus flavescentibus sublaevibus.—Cosra Rica: brushy slope, a shru 3 m. high, climbing over a tree; flowers pink, Alto de La Estrella, Province of Cartago, Mar. 26, 27, 1924. Standley, no. 39,288. (ryPE, in U. S. Nat. Herb., isotype Gr.); also as herb 1.2 m. high in wet forest, La Hondura, Province of San José, alt. 1300-1700 m., Mar. 16, 1924, Standley, no. 37,816 (U. S., small fragm. Gr.). This species, very clearly distinct from any known to the writer from north of the Isthmus of Panama, bears a curiously close resem- blance in habit, texture, toothing, venation of its leaves, in the general nature of its involucre and even in its yellowish pappus to the recently discovered E. Hitchcockii Robinson of central Ecuador which, how- ever, has somewhat broader leaves, heads which are about 20-flowered, and a more tubular corolla. There seems to be no doubt that the resemblance is merely habital and that the two are distinct species. It is a pleasure to commemorate in naming this new een of Costa RECORDS OF THE EUPATORIEAE,—VI 41 Rica the very notable services of its collector, Mr. Paul C. Standley, both in the exploration and record of the tropical American floras. E. (§ Conoclinium?) tatamense, spec. nov., herbaceum scandens dense tomentosum; caule tereti ad 5 mm. Giemetin: indumento patenti uno latere atropurpureo; pilis (saltem juventate) attenuatis purpureo-nodulosis; internodiis plerisque 6-7 cm. longis; foliis op- positis vel aliquando alternis petiolatis ovatis acuminatis crenato- dentatis basi sinuato-cordatis uno latere vel utroque angulatis vel breviter unilobatis a basi 3-nervatis supra sordide viridibus dense papilloso-puberulis maturitate rugulosis subtus griseo-viridibus tomentosis reticulato-venosis 9-12 cm. longis 6-9 cm. latis crassius- cule membranaceis; petiolo ca. 4 cm. longo; panicula terminali patenter oppositiramea folioso-bracteata; pedicellis ca. 5 mm. longis; capitulis _ ca. 55-floris ca. 1 em. altis et crassis; involucri squamis 3—4-seriatim imbricatis sed non distincte gradatis lanceolatis caudato-attenuatis tenuibus laxe villoso-hirtis plerisque 3-6 mm. longis; corollis sordide purpureis glabris; tubo proprio gracili ca. 2.46 mm. longo; faucibus graciliter turbinato-subcylindratis 2.5 mm. longis; dentibus limbi ca. 0.4 mm. longis; achaeniis (immaturis) glabris pallide brunneis 2 mm. longis; pappi setis ca. 30 sublaevibus flavescenti-albis attenuatis basin versus firmiusculis.—CoLomstia: Dept. of Caldas: a climbing herb on a rocky stream-bank in forest, Rio San Rafael, below Cerro Tatama, alt. 2500-2800 m., Sept. 7-11, 1922, Dr. F. W. Pennell, no. 10,426 (ryPE, in Gray Herb.). This plant was first taken for a form of E. diplodictyon Robinson, but further study has shown that it differs from that species in its blunter-toothed, much more copiously and softly tomentose leaves which are less clearly and sharply double-netted, in its much more open paniculate inflorescence, smaller heads, relatively shorter, much more numerous phyllaries and more numerous florets. Furthermore, the deltoid-ovate leaves often have sae oxae lobes externally angled instead of rounded as in E. di In the rather immature heads as yet Geatitas for dissection the receptacle is only moderately convex, but it would presumably become more so at a later stage of development, and, as the affinities of the plant are seemingly with E. diplodictyon and E. lamiifolium HBK., it is provisionally, though with some doubt, referred to § Conoclinium. FE. (§ Exmericata) Tonpuzi Klatt, Beibl. z. Leopoldina 1895, 4 (1895). This smoothish shrub of Costa Rica, collected by Tonduz on the Cuesta de Tarrazu more than thirty years ago, has until now re- mained rather problematic. It was described by Klatt, apparently 42 ROBINSON from defective material, and its representation in the Klatt herbarium, now in possession of the Gray Herbarium, consisted merely of two detached dark brown leaves and the remains of an inflorescence long past anthesis from which all the florets had dropped. The beauti- ful collection of Mr. Standley and his assistants, just at hand, includes five sheets of a slender shrub or lignescent perennial herb, with simil- larly shaped and crenate-toothed leaves of like texture and also pellucid-reticulate, but with heads no less than 28-flowered, while Klatt states that the heads of E. Tonduzii were about 10-flowered. owever, a careful examination of the receptacle from one of the heads of the type material in the herbarium of Klatt shows as many as 28 scars from which florets have fallen, and it is clear that Klatt’s must have been made on the basis of insufficient examination or possibly from study of an abortive head. Klatt also states that the heads of E. Tonduzii are but 1 mm. long. This is pretty clearly a clerical or typographical error for 1 cm., as he later says that the achenes are 1 mm. long. So far as can be seen from the imperfect type material in the Klatt herbarium, E. Tonduzit is conspecific with the much more perfect material now obtained by Mr. Standley. As Klatt’s description proves in several respects in- correct even in relation to the material from which it was drawn, it seems worth while to place on record a new character for this species on the basis of the far better material now available: Fruticosum vel subherbaceum 9-18 dm. altum puberulum oppositirameum; caulibus gracilibus teretibus primo puberulis maturitate glabratis griseis; internodiis 3-7 cm. vel ultra —— as oppositis petiolatis utroque acuminatis crenatis vel crenato-s tis penniveniis utrinque viridibus in costa (saepe rubescenti) ac puberulis aliter glabris textura membranaceis translucente reticulatis 10-14 em. longis 3-5.2 cm. latis; petiolo 6-18 mm. longo; corymbis terminalibus compositis _ leviter convexis folia subtendentia vix aequantibus multicapitulatis; bracteolis filiformibus; pedicellis filiformibus plerisque 5-7 mm. longis; capitulis ca. 28-floris 7 mm. altis 6 mm. diametro; involucri squamis ca. 16 lineari-lanceolatis attenuatis viridibus apicem versus saepe purpureo-tinctis plerisque 2-costulatis ca. 4 mm. longis et 0.7 mm. latis; receptaculo plano glabro; corollis roseo-albis ca. 4 mm. longis limbum versus hirtellis aliter glabris; tubo proprio gracili 1.7 mm. longo; faucibus cylindratis ampliatis 1.7 mm. longis; dentibus limbi ca. 0.6 mm. longis deltoideis; achaeniis nigrescentibus 2 mm. longis in angulis sursum paullo scabratis; pappi setis ca. 20 albis delicatule capillaribus vix scabratis. —Costa Rica: Province of San José: RECORDS OF THE EUPATORIEAE,—VI 43 shrub 1.2-1.8 m. high, common in moist thicket; flowers pinkish white, Laguna de la Chonta, northeast of Santa Maria de Dota, alt. 2000-2100 m., Dec. 18, 1925, Paul C. Standley, no. 42,233 (Gray Herb.); shrub, 9-12 dm. high, flowers white, in oak forest near Quebradillas about 7 km. north of Santa Maria de Dota, alt. about 1800 m., Dec. 24, 1925, Standley, nos. 42,871 (Gr.) and 42,888 (Gr.); herb 9 dm. high with white flowers, in oak forest, vicinity of Santa Maria de Dota, alt. 1500-1800 m., Dec. 14-26, 1925, Standley, no. 41,627 (Gr.); shrub 9-12 dm. high, with pale pink flowers, in wet forest near Finca La Cima, above Los Lotes, north of El Copey, alt. 2100-2400 m., Dec. 21-22, 1925, Standley, no. 42,556 (Gr.). An attractive species recalling E. ligustrinum DC., but with thinner impunctate leaves, translucent veins, and much more numerous phyllaries and florets. E. (§ Eximbricata) triptychum, spec. nov., fruticosum magnum novellis et inflorescentia exceptis glabrum; caulibus pallide brunneis lucidis 5 mm. vel ultra crassis; cortice tenui post exsiccationem longitudinaliter suleato cum lenticellis sparsis paullo scabrato; inter- nodiis 10 cm. vel ultra longitudine; foliis oppositis trifoliolatis; petiolo 5-6.5 cm. longo basin versus plus minusve obcompresso et dilatato; petiolulis subaequalibus 10-14 mm. longis supra obscure puberulis; foliolis oblongis acuminatis basi rotundatis vel vix acutatis margine obsolete crenato-undulatis tenuissimis penniveniis utrinque glabris 6-10 cm. longis ca. 3:5 em. latis; corymbis terminalibus planius- culis compositis subdensis multicapitulatis puberulis ca. 18 cm. diametro; pedicellis filiformibus 6-8 mm. longis; capitulis ca. 12-floris ca. 9 mm. altis et 7 mm. diametro; involucri squamis ca. 12 lineari- oblongis ca. 4-4.5 mm. longis apice rotundatis brevissime ciliolatis plerisque 2-3(-5)-nervatis dorso glabriusculis, extimis 2-3 multo brevioribus; corollis glabris 6 mm. longis; tubo proprio 0.7 mm. longo; faucibus infundibuliformibus ca. 1.6 mm. longis; dentibus limbi del- toideis ca. 0.8 mm. longis; achaenis nigris 2.5 mm. longis deorsum paullo attenuatis in costis pallidis obscure puberulis; pappi setis paullo inaequalibus sordide albis 4-5 mm. longis sublaevibus.—Costa Rica: large shrub in wet forest, El Mufieco, south of Navarro, Province of Cartago; alt. about 1400 m., Feb. 8, 9, 1924, Paul C. Standley, no. 33,458 (U. S., phot. and fragm. Gr.). This species is remarkable for its trifoliolate leaves with clearly petiolulate leaflets. It exhibits no close relationship to any other species. The corolla-throat is unusual, consisting first of a slender tubular portion little differentiated from the proper tube and about 3 mm. long. This at the summit expands 44 ROBINSON into a turbinate portion about 1.6 mm. high. The anthers bear at the apex the ovate membranaceous appendage usual in the genus. E. (§ Eximbricata) Weddellii, spec. nov., subherbaceum vel basin versus paullo fruticulosum 1—4 dm. altum; caule subtereti brun- neo vel purpurascenti crispe puberulo (capillis plerisque atropur- pureis) a basi usque ad inflorescentiam oppositirameo et folioso; foliis oppositis graciliter petiolatis ovatis ad apicem obtusum vel rotundatum angustatis utroque 3—4-crenato-dentatis basi rotundatis membranaceis a basi 3-nervatis supra viridibus incurvo-puberulis subtus paullo pallidioribus crebre punctatis et in nervis incurvo- puberulis 6-15 mm. longis 4-6 mm. latis; petiolo filiformi flexuoso 3-7 mm. longo; inflorescentia composita in ramis subcorymbosa terminali; pedicellis filiformibus plerumque 1-3.5 em. longis 1-2-capitulatis; capitulis ca. 50-floris 1 cm. altis 12 mm. diametro; involucri squamis ca. 20 subaequilongis eroso-ciliolatis, extimis ovato-ellipticis apice rotundatis dorso puberulis vix nervatis subherbaceis ca. 6.5 mm longis ca. 3 mm. latis, intermediis anguste oblongis obtusis sub- scariosis 2-nervatis 7.5 mm. longis 2 mm. latis, intimis oblanceolato- linearibus subacutis 1-nervatis 8 mm. longis 1 mm. latis; receptaculo paullo convexo nudo; corollis glabris ca. 5 mm. longis roseis; tubo proprio 1 mm. longo; faucibus cylindratis 3.5 mm. longis; dentibus limbi deltoideis 0.6 mm. longis; antheris distincte ad apicem appendi- culatis; achaeniis pallidis deorsum decrescentibus 2.5 mm. longis in angulis hispidis; pappi setis ca. 30 inaequalibus corolla fere dimidio brevioribus paullo barbellatis—Prru: on grassy hills near Quiaca, Prov. Carabaya, June to July., 1848, H. Algernon Weddell, no. 4654 (TYPE, Par., phot. and small fragm. Gr.). This species has much the habit of E. scopulorum Wedd. which however has smaller heads, narrower acutish more distinctly nerved phyllaries, impunctate leaves, a longer proper tube of the corolla an a pappus nearly equalling the corolla. The related E. F ‘unckir Robinson of Colombia has 80-flowered heads, acute phyllaries and hexagonal stems. E. g Pxtabriesta) xanthochlorum, spec. nov., ut videtur erectum perenne 5 dm. vel ultra altitudine basi ignota; caule tereti purpurascenti obscure adpresseque puberulo albo-medul- loso 4 mm. diametro; internodiis usque ad 1 dm. longis; foliis oppositis graciliter petiolatis rhombeo-ovatis acuminatis acute et grosse ser- a cuneatis | basin versus a ima a _basi 3-5-nervatis : oes minute praecipue secundum 1 nervos suleatos adpresse sed plerum enhtnue a RECORDS OF THE EUPATORIEAE,—VI 45 . 6-10 em. longis 3-5 em. latis textura firmius membranaceis; petiolo 24 cm. longo; corymbis terminalibus compositis densis multicapitu- latis planiusculis vel modice convexis 14-22 cm. diametro breviter subadpresse pubescentibus; pedicellis gracilibus plerisque 3-8 mm. longis; capitulis ca. 55-floris ca. 5 mm. altis et latis; involucri squamis ca. 20 lanceolato-oblongis acutis vel acutiusculis plerisque subaequali- bus pallide flavido-viridibus dorso dense pubescentibus plerisque 2-costulatis ca. 3.5 mm. longis, extimis 2-3 lanceolatis multo minoribus acuminatis; corollis albis glabris; tubo proprio gracili 1.7 mm. longo: faucibus cylindratis distincte ampliatis 1.2 mm. longis; dentibus limbi deltoideis 0.3 mm. longis; achaeniis maturitate nigris basi pallide callosis 1.3 mm. longis glabris; pappi setis ca. 12 delicatule ceapillaribus deciduis.—Mexico: State of Chiapas: on moist rocks in ravines of mountains west of Monserrate, March, 1925, Dr. C. A. Purpus, no. 74 €rype, in U. S. Nat. Herb. and Gray Herb.); on rocks in mountain streams, Monserrate, April, 1925, Purpus, no. 10,074 . S., Gr.); on rocks in creeks near Monserrate, April, 1925, Purpus, no. 176 (U. S., Gr.). This species shows most resemblance to E. pazcuarense HBK. and E. photinum Robinson, but may be readily distinguished from them by its yellowish green color, slightly firmer leaves with coarse and sharper serration, depressed veinlets and clear translucent puncta- tion. The fact that it was obtained several times by Dr. Purpus would suggest that it must be fairly abundant locally, though not apparently collected by previous explorers of Chiapas. SPANIOPAPPUS, gen. nov. Compositarum-Eupatoricarum. Capit- ula homogama tubiflora. Involucrum campanulatum; squamis sub- aequalibus 2-3-seriatim imbricatis subherbaceis. Receptaculum modice convexum nudum. Corollae regulares aequales, tubo grada- tim paullo ampliato sine faucibus distinctis, limbo 5-dentato. An- cula ovata hyalina ornatae. Styli rami longissimi filiformes vix apicem versus paullo clavellati. Achaenia anguste prismatica 5- angulata 5-costata. Pappus plerumque e seto unico elongato capil- lari erecto cum setulis paucis brevissimis patentibus compositus, rarius e setis 2-3 inaequalibus plus minusve elongatis et setulis brevissimis obscuris.—Herba glabriuscula ageratoidea. Folia op- posita (superne reducta et alterna) ovata, breviter petiolata. Capitula corymbosa. S$. Ekmanii, spec. nov., perennis herbaceus decumbens 2-3 dm. altus; caule simplici geld ken: usque ad 4 mm. diametro 46 ROBINSON juve entate obscure hirtello (capillis moniliformibus atropurpureis ineurvis gracillimis) tardius glabrato in media parte dense folioso supra nudiusculo; internodiis plerisque 5-8 (rariter ad 35) mm. longis; foliis oppositis breviter petiolatis oblongo-ovatis obtusis crenatis basi cuneatis subcarnosis pinnatim 5-nervatis utrinque viridi- bus et glabris pellucide punctatis et venulosis 3.5-6.8 em. longis 2-3 em. latis; petiolo 2-4 mm. longo obcompresso latiusculo; foliis sup- erioribus paucis (2-4) multo minoribus bracteiformibus alternis; corymbo terminali planiusculo 3-10-capitulato 3-6 em. diametro; capitulis 50-70-floris 10 mm. altis 14 mm. diametro; involucri cam- panulati squamis ca. 25-30 subaequalibus lanceolato-oblongis obtusis apicem versus eroso-ciliolatis crassiusculis ca. 5 mm. longis et. 2-2.3 mm. latis obscure 3-nervatis; corollis glabris tubulatis 5 mm. longis sursum paullo ampliatis limbi dentibus deltoideis; achaeniis penta- gonis 5-costatis olivaceis glabris 3.5 mm. longis in costis multo pallidi- oribus; pappi setis 1-3(+4) valde inaequalibus 1-2 mm. longis ceapillaribus erectis albis vix scabratis; setulis pluribus intermediis patentioribus vix 0.1-0.2 mm. longis etiam capillaribus nec squami- formibus.—CuBa: Provirice Oriente: rare, in very dense low woods at top of the mountain spur El Yungue near Baracoa, Ekman, no. 3975 (TYPE, in Gray Herb.). This plant differs from Eupatorium in its pappus, which is reduced to 1-3 or rarely 4 unequal short erect bristles (opposite, i. e. above, the angles of the achenes) and several intermediate setules very much shorter and more spreading. The presence of this rather striking, even if sparse pappus, seems to show that this is not a casual calvous phase of some normally pappus-bearing plant. From Ageratum the present plant may be distinguished by its terete pappus-bristles, which fail to take on the scale-like nature or expanded base characteristic of the pappus in Ageratum. Nor has the present plant the strongly ribbed phyllaries, which are prevalent in Ageratum. Its capillary non-plumose pappus, consisting of very unequal bristles, would suggest possible affinity between Spaniopappus and the Brazilian Dyssothriz or the Mexican and Central American Fleisch- mannia, but “its habit ss ie approach either of these genera. Furthermore its sparse pappus does not agree with the relatively dense and decidedly writ pappus of Dyssothriz nor the longer and definitely 5-setose pappus of Fleischmannia. The inclusion of the present plant in any one of these older genera would much increase the difficulty of mans them. It therefore seems best to give the RECORDS OF THE EUPATORIEAE,—VI 47 plant independent generic standing. It has a fairly characteristic habit and seems to be a shy denizen of an exceptionally secluded habitat, where discovered by Dr. E. L. Ekman whose name it is a pleasure to associate with this rare and interesting Cuban endemic— one of many brought to scientific knowledge by his patient and effective explorations. The generic name is, of course, from oxéytog scanty, and caxtoc, coma. Mrixanta Ascuersontt Hieron. in Engl. Bot. Jahrb. xxviii. 577 (1901); Robinson, Contrib. Gray Herb. n. ser. lxiv. 45 (1922). This species has been hitherto known only from the type collection in the Cordillera Occidental near Popayan, Colombia, Lehmann, no. 5979. Specimens, seemingly conspecific, have now been received from Peru as follows: wet sunny thicket; fls. dirty white, aromatic, Yanano, alt. 1830 m., Macbride, no. 3775 (Field Mus., Gr.); fls. yellowish green, not fragrant, Huacachi, estacion near Mufia, alt. 1980 m., Macbride, no. 4151 (Field Mus., Gr.); wet rocky uplands, Tambo de Vaca, alt. 3965 m., Macbride, no. 4906 (Field Mus., Gr.). In foliage, inflores- cence, dark punctation, nature of pubescence, exact proportions of corolla, and many other traits the correspondence. between the Colombian type and this newly observed Peruvian material is exceed- ingly close. Ina single point a difference has been observed. In the Colombian plant the pappus is of a reddish or salmon tinge, in the Peruvian, a dirty yellowish white. This may, at least to some extent, be due to the age of the specimens. It is a well known fact that pappus in herbarium specimens may with age turn to much deeper tones than when first dried. This extension in the range of M. Aschersonii is noteworthy and of an unusual kind. The species has not thus far been recorded from Ecuador, but is now to be expected in that country. In Mr. Macbride’s second Peruvian collection there is also a closely related plant, agreeing with M. Aschersonii in all important features, but possessing denticulate leaves and having longer and very much more copious pubescence. is may be put on record as follows: Var. uta, var. nov., caule dense patenterqué subsetuloso- tomentello; indumento (fide Macbridei) rubro; foliis distincte cuspidato-denticulatis supra crispe puberulis subtus pallide (post exsiccationem) tomentellis; aliter var. typicae (cum caule obscure glanduloso-puberulo et foliis integerrimis) similis.—Prru: Dept. Junin: on sunny brush; flowers pale green; pubescence red; La Merced, Hacienda Schunke, about 1220 m. alt., Aug. 27-Sept. 1, 1923, J. Francis Maebride, no. 5777 (rps, in Field Mus., isotype Gr.). 48 ROBINSON M. (§ Thyrsigera) brunnescens, spec. nov., fruticosa volubilis; caulibus perflexuosis subteretibus multi-costulatis 2-3 mm. vel ultra diametro brunnescenti-purpurascentibus crispe puberulis vel patenter villosulis maturitate subglabratis; foliis oppositis breviter petiolatis lanceolato-oblongis integris vel obsolete undulato-dentatis acuminatis basi acutis supra in costa obscure puberulis aliter glabris lucidis post exsiccationem olivaceis vel brunnescentibus subtus paullo pallidioribus obscuris laxe praecipue in costa et nervis brunnescente pubescentibus 5-6 em. longis 15-24 em. latis subcoriaceis paullo supra basin pin- natim 5-nervatis, nervis venisque supra immersis et subtus exsertis; paniculis elongatis compositis, partialibus primo subcylindricis densis deinde laxioribus graciliter conicis; capitulis glomeratis breviter pedicellatis statu juvenili ca. 6 mm. longis; pedicellis (valde im- maturis) usque ad 2 mm. longis; bracteolis ovatis obtusis vel saepe acutis ca. 3-4 mm. longis et 2 mm. latis; involucri squamis 4.7—5 mm. longis ca. 1.8 mm. latis oblongis obtusis brunnescenti-stramineis, exterioribus paullo pubescentibus, interioribus subglabris sed apicem versus obscure puberulis et ciliolatis; corollae tubo brevi ca. 0.7 mm. longo, faucibus turbinatis, dentibus lanceolatis ca. 1.3 mm. longis; achaeniis nimium immaturis ut videtur glabriusculis; pappi setis ca. 50 pallide roseo-salmoneis.—M. laevis Glaziou, Bull. Soc. Bot. Fr. lvi. Mém. 3, 392 (1909), in part, not DC.—Braziu: Burchell, no. 4774 (Type, in Gray Herb.); also at Barbacena and Caraca, Minas Geraes, A. Glaziou, no. 15,129 (Copenh., fragm. Gr.). Traced by Baker’s key of the Brazilian Mikanias, this plant falls near M. Candolleana Gardn., for, like that species, it has ecordate, lance-oblong leaves, pubescent beneath, and phyllaries 4.7-5 mm. long (even when immature), therefore over 2 lines in length. From M. Candolleana the plant here described differs in its smaller, much thicker leaves, lucid on their upper surface, in its smaller heads, enser and more irregular inflorescence, and in its salmon-pinkish not white pappus-bristles. M. laevis DC., to which Glaziou, |. ¢., referred his no. 15,129, is a later name fof M. . trinervis Hook. & Arn., a species with much larger, ovate, glabrous leaves, and mes panicles, more slender heads and relatively longer pedicels, et e It is to be regretted that neither the specimen of Burchell nor that of Glaziou shows the mature florets, but the distinctness of the species seems assured by the other characters clearly shown and it may there- fore be put on record. It is also regrettable that Glaziou in his list has repeated, probably by clerical or typographical error, the no. 15,129, citing it also on page 396 as M. ligustrifolia DC., with which, of course, this plant has no connection. RECORDS OF THE EUPATIOREAE,—VI 49 M. (§ Thyrsigera) caldasana, spec. nov., fruticosa scandens; caule tereti post exsiccationem leviter eostulate usque ad 6 diametro albo-medulloso fulvo-tomentello; internodiis plerisque 5—7 em. longis; foliis ovatis obtusis mucronulatis integris vel obscure remoteque cuspidato-denticulatis basi rotundatis pinnatim 5(-7)- nervatis untrinque viridibus supra scabris et in nervis dense fulvo- tomentellis subtus paullo pallidioribus hispidulis impunctatis in nervis etiam fulvo-tomentellis 7-9.5 cm. longis 5-7 cm. latis mem- branaceis; venulis supra impressis subtus modice exsertis; petiolis 2-2.5 em. longis teretibus fulvo-tomentellis; eis ejusdem jugi inter se ab annulo transverso paullo incrassato sed exappendiculato conjunc- tis; panicula thyrsiformi 1.5 dm. longa ca. 8 cm. diametro ovoidea oblongo-linearibus 7-8 mm. longis 1-1.5 mm. latis obtusis brunneis dorso breviter hispidulis; corollis flavido-albis ca. 5 mm. longis obscure granulatis; tubo proprio 1.8 mm. longo gracili; faucibus turbinato- subcampanulatis ca. 2.7 mm. longis; dentibus limbi deltoideis 0.5 mm. longis; achaeniis gracilibus nigris minute glanduliferis 5 mm. longis vix 0.5 mm. diametro; pappi setis ca. 35 albidis laeviusculis 44.5 mm. lon gis. COLOMBIA: Department of Caldas: climbing shrub in edge of forest, alt. 3100-3400 m., “ Alaska,” above Salento, Cordillera Central, Aug. 7, 1922, Dr. Francis W. Pennell, no. 9364 (TYPE, in Gray Herb.). In some respects this species would seem to stand near the still im- a glomerate arrangement of its heads in the thyrse. However, M. multinervia is a smooth plant, with elliptical and serrate, rather than ovate and subentire leaves. Its phyllaries are ovate-oblong instead of almost linear as in the present species. M.caldasana is probably even more nearly related to M. Aristet Robinson, but differs in its somewhat larger, more ovate, nearly entire and somewhat more distinctly reticulated leaves, its more spherical glomerules of heads in the denser thyrse, fulvous, shorter, and denser indument, and narrower phyl- —_ Thyrsigera) caustolepis, spec. nov., fruticosa scandens ast pubescens et minute nigro-papillata; caule tereti multi- costulato fusco 4 mm. vel ultra diametro dense incurvo-puberulis et dense nigro-pustulato albo-medulloso; internodiis ad 1.5 dm. longis; foliis oppositis petiolatis ovatis acutis integris basi rotundatis pin- natim 7-9-nervatis textura membranaceis utrinque viridibus et 50 ROBINSON dense nigro-atomiferis supra hirtellis subtus in nervis pubescentibus 6-7 em. longis 44.5 em. latis; nervis et venulis subtus leviter prominu- lentibus; petiolo 1.5-2 em. longo; panicula magna pyramidata foliaceo- bracteata terminali 5 dm. alta et 3 dm. crassa; ramis divaricatis; capitulis (immaturis) ca. 5-6 mm. longis breviter pedicellatis ad apices ramulorum aggregatis; bracteolis oblongis ca. 2 mm. longis ad vel prope apicem pedicelli affixis; involucri squamis oblongis viridibus 7-9-striatulo-costulatis dorso glabriusculis apice rotundatis apicem versus prope marginem nigro-glandulosis quasi ustis; corollis albis glabris ca. 4 mm. longis (immaturis) ; tubo ca. 1.5 mm. longo; faucibus 1 mm. longis; dentibus limbi lanceolatis acutis 1.5 mm. longis et 0.6 mm. basi latis; achaeniis (immaturis) ca. 1.5 mm. longis; pappi setis ca. 36 sursum scabratis et paullo incrassatis albis corollam sub- aequantibus. —Peru: Dept. Cuzco: climbing white-flowered shrub in forest near Rio Yanamayo, below “ Pillahuata,”’ alt. 2000-2300 m., May 4-5, 1925, Pennell, no. 14,070 (rypr, in Gray Herb.). The available material of this species, though immature and not showing the leaves below the large inflorescence, is still so clearly different from any previously known species as to justify characteriza- tion and published record. The nervation of the leaves is about intermediate between the conditions usually described as pinnate- nerved and pinnate-veined. The species may be placed near M. trachodes Robinson, though clearly distinct from it, having dark- punctate leaves of much thinner texture and much less hairy above, very different phyllaries, etc. The name caustolepis alludes, of course, to the dark glandular margin of the phyllaries which gives them a scorched or singed appearance. M. (§ GtoposaE) DESMOCEPHALA Robinson, Contrib. Gray Herb. n. s. lxiv. 7 (1922). This well marked species was originally des- cribed from Bolivian material. However in Mr. Macbride’s second Peruvian collection there is a specimen which corresponds closely with the original Bolivian material in all details of inflorescence, involucre, corolla, and achenes, but differs noticeably in the form, texture and surface of the leaves, seeming therefore to warrant varietal separation as follows: Var. typica, caule minute granulato; foliis membranaceis opacis supra minute papillosis subtus parce pubentibus.—M. desmocephala Robinson, 1. e—Bottvia: Province Yungas, ~~ no. 1740, and in vies San Carlos, near Mapiri, linchliea: x ar. peruviana, var. nov., caule siheees ae ote juvenili fide Macbridei subroseato sed post exsiccationem RECORDS OF THE EUPATORIEAE,—VI 51 fuscescenti) ; foliis maturiate chartaceo-subcoriaceis supra lucidis rugosis depresse puncticulatis subtus ut apud var. typicam.—PERU: ept. Junin: on sunny brush, La Merced, Hacienda Schunke, alt. 1220 m., Aug. 27-Sept. 1, 1923, J. Francis Macbride, no. 5776 (TYPE, Field Mus., isotype Gr.). Until much more material is available, both from Bolivia and Peru, the classificatory value of the differences here indicated must remain doubtful. They may prove specific or, on the other hand, merely formal. However, the two plants, rather widely separated geo- graphically and showing such distinctions, should scarcely be merged without comment M. GLoBuLiFERA Rusby ex Robinson, Contrib. Gray Herb. n. s. Ixi. 15 (1920); Ixiv. 50 (1922). An inflorescence closely simulating that of this species was collected by Mr. E. P. Killip in a forest at “La Gallera,’’ Micay Valley, Department of El Cauca, in the Cordil- lera Occidental of Colombia, alt. 1800-1900 m., July 1, 1922, no. 7917 (Gr.). Unfortunately no attached foliage was secured and there was associated with the inflorescence a stem bearing alternate ovate leaves and doubtless belonging to some quite different plant, possibly a Tournefortia. Determined from inflorescence alone, the identification of the Mikania is subject to some doubt, especially as the only previously known station for this species was in the widely removed region of Santa Marta and at considerably lower altitude. M. (§ Thyrsigera) granulata, spec. nov., herbacea vel fruticosa scandens hirtella vel glabriuscula; caule tereti post exsiccationem costulato purpureo fistuloso usque ad 4 mm. vel ultra diametro op- positirameo; foliis oppositis petiolatis ovatis acuminatis obscure remoteque cuspidato-denticulatis basi saepius rotundatis aliquando subcordatis vel obtusis firmiuscule membranaceis utrinque viridibus et (sub lente) Jagger cor aa supra basin pinnatim 5(—7)-nervatis usque ad 7 cm. longis et 5 cm. latis pellucide reticulato-venosis; petiolis gracilibus ficsuale iL aie supra sulcatis 1-2 cm. longis; paniculis thysoideis compositis basi foliaceo-bracteatis ad 4 dm longis et 2 dm. crassis; capitulis ca. 1 cm. longis sessilibus plerumque trinis ad apices paniculae ramulorum et in jugo unico paullo infra in quoque ramulo dispositis; bracteolis lanceolato-linearibus acutis ca. 2 mm. longis; involucri squamis oblongo-linearibus brun- neis ca. 6 mm. longis apicem obtusum vel rotundatum versus tenui- oribus ciliolatis dorso paullo pubentibus vix nervatis; corollis albis glabris ca. 4 mm. longis; tubo proprio 1.5 mm. longo gracili; faucibus subturbinato-campanulatis ca. 2 mm. longis; dentibus limbi deltoideis 52 ROBINSON 0.5 mm. longis; achaeniis gracilibus atrobrunneis 5 mm. longis in ' summa parte et in costis parce pubentibus; pappi setis ca. 40 brun- nescenti-albis apicem versus paullo incrassatis—CoLomBIA: Depart- ment of El Cauca: in forest at “Canaan,’’ Mt. Purace, Cordillera Central, alt. 3100-3300 m., June 11-13, 1922, Pennell & Killip, no. 6626 (TYPE, in Gray Herb. ., lsotype Acad. Philad.); also in field at same locality, no. 6499 (Gr.). This species differs from M. Stuebelii Hieron. conspicuously in the form and texture of its leaves, which examined under a lens can be seen to be roughened by a slight granulation on both surfaces. The species also bears in its inflorescence much resemblance to M. Jame- sonit Robinson of Ecuador, but that has much more deeply and regularly dentate leaves of slightly different form. Two varieties of M. granulata, seemingly referable to a single species, may here be recognized, thus: Var. typica, foliis praecipue subtus in costa pubescentibus margine distincte sed remote cuspidato-denticulatis textura membranaceis vel submembranaceis; petiolis puberulis——Exsiccatae as above stated. Var. integra, var. nov., foliis etiam in costa glabris margine integris textura subchartaceis; petiolis glabris—CoLtomp1a: Depart- ment of Caldas: climbing shrub in forest, at “ Pinares,”’ above Salento, alt. 2600-2800 m., Aug. 2-10, 1922, Francis W. Pennell, no. 9308 (TYPE, in Gray Herb., isotype Acad. Philad.). In both varieties the pubescence seems remarkably inconstant on axes of different orders even in the same individual. Thus at a given node in the panicle the lateral branches may become decidedly pubescent, while the central axis remains quite glabrous, or all may ata given node become pubescent. In no. 6626 a portion of a main stem is for a distance of 2 or 3 cm. above a node almost shaggy, being covered by coarse spreading trichomes 1-2 mm. in length, while the lateral axes on the same individual are essentially smooth above the nodes. M. (§ Thyrsigera) hylibates, spec. nov., fruticosa grandis neandens vereneeecitin sordide breviter aye excepta) — nervatis coriaceis 7-11.5 cm. longis 3-5 cm. latis; petiolis ut videtur E ibus supra sulcatis ca. 4 cm. longis; paniculis lateralibus et RECORDS OF THE EUPATORIEAE,—VI ay: terminalibus conjunctim inflorescentiam elongatam foliaceo-bractea- tam formantibus; capitulis (paullo immaturis) 8 mm. longis; sessilibus plerumque oppositis vel terminalibus; bracteolis oblongis herbaceis 2.5 mm. longis; involucri squamis oblongis obtusis pallido-brunneis vix nervatis apicem versus dorso puberulis ca. 4.5 mm. longis 1-1.5 mm. latis; corollis (immaturis) glabris albis 5.5 mm. longis tubulatis apicem versus paullo ampliatis; achaeniis valde immaturis 2 mm. longis; pappi setis ca. 40 sordide albis corollam subaequantibus vix scabratis.—Costa Rica: large woody vine in wet forest; leaves thick; flowers white, El Mufieco, south of Navarro, Province of Cartago, alt. about 1400 m., Feb. 8, 9, 1924, Paul C. Standley, no. 33,848 (type, U. S., phot. and fragm. Gr.). A species very close to the Brazilian M. laevigata Sch.-Bip., which, however, to judge from Dusén’s no. 15,607 has leaves nerved from somewhat above the base, a denser panicle, somewhat smaller heads and relatively shorter bractlets. In the presence of these differences and its wide geo- graphic — it seems best to treat this Costa Rican plant as an independent s M. § OR Killipii, spec. nov., fruticosa scandens; caule tereti juventate obscure crispeque puberulo purpurascenti vel in uno latere violascenti-brunnescenti; foliis oppositis graciliter petiolatis ovatis acuminatis integris basi obtusis vel rotundatis utrinque laete viridibus glabris subtus minute nigro-puncticulatis supra basin 3(-5)-nervatis 6-8 cm. longis 3.5-4.5 em. latis textura tenuiter caigrdesartinae petiolo 2.5-3 em. longo; se omeoeniad inter- bracteata crispe puberula; spicis plerisque 3-6 cm. longis 1.5-2 cm. crassis; capitulis patenter subadscendentibus ca. 12 mm. longis sessilibus; bracteolis subulati-linearibus ca. 2 mm. longis; involucri — oblanceolato-oblongis viridibus tenuibus obtusis vel sub- cutis apicem versus eroso-serrulatis, et obscure puberulis aliter glabris obscure nervatis basi calloso-subgibbosis 8 mm. longis 2 mm. latis; corollis viridescenti-albis; tubo gracili curvato glabro ca. 5 mm. longo; faucibus vix ullis; lobulis limbi oblongis subacutis patentibus ca. 1.7 mm. longis; achaeniis (immaturis) subglabris 4 mm. longis; pappi setis ca. 110 pallide salmoneis 5-6 mm. longis capillaribus laeviusculis—CotomBi1A: Department of El Cauca: thicket along stream, at Coconuco in the Cordillera Central, alt. 2300-2500 m., June eet 1922, Ellsworth P. Killip, no. 6823 (tTyPx, in Gray Herb., isotype in herb. of Phil. Acad. Nat. Sci.). 54 ROBINSON This species is obviously more nearly related to M. bogotensis Benth. than to any of the other Mikanias of Colombia, but differs in having heads, florets, and achenes about twice as large, its leaves (so far as shown) are not cordate and they are somewhat more ob- viously dark-punctate beneath. Furthermore, the forking of the midnerve occurs in M. bogotensis at the very base of the leaf, while in the species here characterized, the midnerve (after giving off near the base one or two minor lateral branches) itself divides into three nerves at a point 4-5 mm. above the base. It is a pleasure to commemorate in the name of this species the effective hei exploration of its collector. M. piciformes) micayensis, spec. nov., fruticosa volubilis on Saye caule subtereti post exsiccationem costulato brunneo —3 mm. vel ultra diametro albo-medulloso; internodiis usque ad 2 dm. longis; foliis oppositis petiolatis late ovatis vel obovatis apice rotundatis vel breviter subacuminatis integris basi obtusis vel acutius- culis 5.5-10 em. longis 3-9 cm. latis supra (post exsiccationem) atrobrunneis glabris subtus multo pallidioribus brunneo-griseis puncticulatis pinnatim 5(—7)-nervatis textura firmiter coriaceis; nervis supra sulcatis subtus prominentibus prorsum Ancurvis et inter ae anastomosantibus; saepius plus minusve protractis; petiolo crasso supra leviter suleato; appendicibus stipuliformibus lunulatis ca. 1 mm. altis suberosis; panicula terminali obsolete puberula ad 4 dm. longa 1 dm. crassa basin versus foliaceo-bracteata sursum nudiuscula; spicis 1-4 em. longis multicapitulatis; capitulis ca. 5 mm. longis arcte sessilibus; bracteolis ovatis vel lanceolatis ca. 1.2 mm. longis; involucri squamis anguste oblongis apice rotundatis puberulis 3 mm. longis; corollis. glabris; tubo proprio gracili 1.5 mm. longo; faucibus turbinatis paullo atomiferis 1 mm. longis; dentibus limbi anguste deltoideis acutiusculis ca. 0.7 mm. longis; achaeniis 1.8 mm. longis atrobrunneis obscure eemdviar-stonnteris:; costis pallidioribus; pappi setis ca. 40 apicem versus paullulo incrassatis, —CoOLOMBIA: Department of El Cauca: a climbing shrub in clearing, “La Gallera,”” Micay Valley of the Cordillera Occidental, alt. 1500- oe m., July 1, 1922, Ellsworth P. Killip, no. 7894 (ryPx, in Gray e This species is most nearly related to M. leiostachya Benth., but has leaves which are obtuse or rounded instead of sharply acuminate at tip, of somewhat firmer texture, more discolorous, and reticulated in a somewhat different manner, the areolae being more elongated RECORDS OF THE EUPATORIEAE,—VI 55 and the transverse veins between the nerves being much less definite and regular. M. (§ Thyrsigerae) Millei, spec. nov., fruticcsa robusta volubilis in inflorescentia brunnescenti-tomentella aliter glabra vel glabrius- cula; caule subtereti post exsiccationem multicostulato fistuloso ca. 4-5 mm. diametro; foliis oppositis petiolatis late ellipticis integris vel paullo undulato-subdenticulatis margine revolutis apice et basi rotundatis utrinque glabris et leviter reticulatis supra sublucidis subtus distincte pallidioribus opacis supra basin pinnatim 7-nervatis textura rigidiuscule coriaceis 6-8 em. longis 4—5 em. latis; nervis sub- tus valde prominentibus supra sulcatis a venis multis transverse conjunctis; petiolis robustis supra sulcatis plerisque 1—-1.6 cm. longis, eis ejusdem jugi ab annulo (vel ala brevissima) primo tenui deinde paullo suberoso transverse conjunctis; internodiis 5-7 cm. longis; inflorescentiis terminalibus et caulis apicem versus lateralibus con- vexis densis conjunctim paniculam ovoideam vel etiam planiusculam formantibus; capitulis ca. 8.5 mm. longis sessilibus congestis; bracteo- lis lanceolatis sessilibus ca. 5.5 mm. longis attenuatis acutis 3-5- costatis; involucri squamis anguste oblongis apice rotundatis dorso convexis atrobrunneis ecostatis ca. 6.5 mm. longis et 1.5 mm. latis praecipue basin versus tomentellis vel lanceolatis; corollis glabris; tubo proprio gracili 2 mm. longo; faucibus cylindratis paullo ampliatis ca. 3 mm. longis; dentibus limbi anguste deltoideis acutis ca. 1 mm. longis et 0.5 mm. latis; achaeniis (juvenilibus) 2.4 mm. longis; pappi setis ca. 30 salmoneo-albidis ca. 5.5 mm. longis sursum paullulo incrassatis laevibus.—Ecvuapor: in silvis subandinis ad Tambillo alt. 3200 m., Aug., 1917, Father Luis Mille, no. 525 (ryper, in Gray Herb.). This species with its densely clustered heads, sometimes in rather flattish inflorescences, forms a transition between §§ Thyrsigerae and Corymbosae, but on the whole appears more naturally placed in the former. It is a pleasure to dedicate this well marked plant to its discoverer, the Rev. Father Luis Mille, S. J., whose patient observa- tion and discriminating collecting have done much to aid in the elucidation of the difficult flora of Ecuador. According to Father Mille it bears in the region of Tambillo the vernacular name of “ Po- roto Angu.”’ M. (§ Thyrsigera) mucronulifera, spec. nov., herl acea perennis scandens patenter brunneo-pubescens; caule tereti gracili purpureo- brunneo post exsiccationem pluricostulato fistuloso ca. 3 mm. diame- tro; internodiis 5-18 cm. longis; foliis oppositis graciliter petiolatis rhombeo-ovatis obtuse acuminatis et breviter mucronulatis utroque 56 ROBINSON supra mediam partem minute 2-3-cuspidato-denticulatis (dentibus ca. 0.5 mm. longis 6-8 mm. inter se distantibus) basi integris cuneatis membranaceis post exsiccationem atrobrunneis supra basin pinnatim _ 5-nervatis ciliolatis et utrinque sordide pubescentibus ca. 6 cm. longis et 3-4 em. latis; petiolis 8-12 mm. longis dense pubescentibus; eis ejusdem jugi ab‘annulo vix incrassato inappendiculato conjunctis; paniculis terminalibus sublaxis ca. 2 dm. altis et crassis, ramis curvato- adscendentibus vel divaricatis; capitulis 8-9 mm. longis plerisque ternis in apices ramulorum sessilibus; bracteolis linearibus acutis ca. 2 mm. longis; involucri squamis linearibus brunneis apicem versus paullo angustatis tenuioribus eroso-ciliatis in apicem verum rotunda- tis ca. 6 mm. longis 1.3 mm. latis apicem versus paullo dorso puberulis; corollis albis ca. 4 mm. longis; tubo proprio 1 mm. longo; faucibus turbinato-campanulatis ca. 2.5 mm. altis et 1.7 mm. latis; dentibus limbi deltoideis ca. 0.5 mm. longis; antheris apice modice appendi- culatis; achaeniis gracilibus maturitate nigrescentibus sublaevibus 3.5-3.8 mm. longis vix 0.7 mm. crassis; pappi setis ca. 40 brunnescenti- albis ca. 4 mm. longis apicem versus distincte incrassatis scabratis.— Cotompra: Department of El Cauca: open land, near Rio San Joaquin, at “La Gallera,’’ Micay Valley, alt. 1400-1500 m., in the Cordillera Occidental, June 29-30, 1922, Ellsworth P. Killip, no. 7858 (TYPE, in Gray Herb.). This species is probably nearest M. Stuebelii Hieron., but differs in its rather copious dark and spreading pubescence and essentially (1901). When revising the Mikanias of Colombia some years ago, the writer (Contrib. Gray Herb. n. s. Ixiv. 44), having seen no material of this species, was obliged to interpret it from the original diagnosis. : This was happily pretty full and well drawn, but in one respect mis- leading, since it suggested that the species belonged in the corymbose section of the genus. Some years later the writer was able to obtain, through the kindness of Prof. Diels, Director of the Botanical Garden and Museum at Berlin-Dahlem, a photograph of the type and there has been recently received at the Gray Herbarium from the Royal Gardens at Kew, in continuation of exchange, an isotype of the species, namely Lehmann’s no. 5964, collected in the Cordillera Occidental near Popayan at 1200-1800 m. These sources of informa- tion show clearly that the species does not belong in § Corymbosae (as interpreted by the writer), but in § Thyrsigerae, with thyrsoid- paniculate and only very obscurely if at all corymbose inflorescence. RECORDS OF THE EUPATORIEAE,—VI 57 The extensive collections of Dr. Pennell’s party which visited the Andes of Popayan in 1922, contain no specimen precisely matching Lehmann’s type of M. nigropunctulata, but include three numbers, also from the Cordillera Occidental (though obtained at a somewhat greater altitude) which differ somewhat in leaf-breadth, leaf-texture, and length of petiole, but agree with M. nigropunctulata in all es- sential floral and achenial characters, and show also so many close similarities in nervation, pubescence, punctation, etc., as to leave little doubt that they represent mere varieties of a somewhat variable species. The plants in question may be treated as follows: M. NIGROPUNCTULATA Hieron., var. typica, foliis ovatis usque ad 15 cm. longis et 7 cm. latis integris subchartaceis; petiolis modice gracilibus usque ad 2 cm. longis,—M. nigropunctulata Hieron. 1. ec. Var. cryptodonta, var. nov., foliis lanceolato-oblongis ca. 12 cm longis et 44.5 cm. latis obscure obsoleteque calloso-denticulatis; petiolis robustioribus 4—8(—10) mm. longis—Cotoms1a: Department El Cauca: climbing shrub in glen of forest, alt. 2400-2700 m., “San José,”” San Antonio, in the Cordillera Occidental, June 28, 1922, Pennell & Killip, no. 7342 (type, in Gray Herb.); also no. 7314 (Gr.) secured the same day on a bushy slope at the same locality. The presence of teeth on the leaf-margins can be detected only by examina- tion with a lens Var. phaenodonta, var. nov., foliis oblongo-lanceolatis caudato- attenuatis distincte denticulatis (Siiitiias ca. 0.5 mm. altis et inter se 2-10 mm. distantibus) subchartaceis 12-14 cm. longis 3-4 em. latis; petiolis vix 4 mm. longis.—Co.LomB1a: Department El Cauca: climb- ing shrub in bushy forest, alt. 2400-2700 m., “San José,” San Antonia, June 28, 1922, Pennell & Killip, no. 7430 (Type, in Gray Herb.). M. (§ Thyrsigera) nubigena, spec. nov., suffrutescens scandens inflorescentia puberula excepta omnino glabra; caulibus teretibus 3 mm. vel ultra diametro laevibus albo-medullosis; internodiis usque ad 8 cm. longis; foliis oppositis petiolatis ovato-oblongis caudato- i viridibus subconcoloribus reticulatis (venulis utrinque pane bus) coriaceis 8-10 cm. longis 3.5~4.7 em. latis; petiolis 2-2.5 cm. longi glabris supra sulcatis, eis ejusdem jugi ab annulo elevato pera conjunctis; panicula terminali composita foliaceo-bracteata usque ad 3 dm. longa et 2 dm. diametro laxiuscula sordide in ramulis puberula; capitulis sessilibus plerisque oppositis ca. 7 mm. longis (paullo ante —— involucri squamis anguste oblongis herbaceis apice rotunda- 58 ROBINSON tis ciliolatis fere enervatis dorso glabris basin versus valde incrassatis callosis pallidis et variolatis; corollis (statu juvenili) 4.5 mm. longis tubulatis vix apicem versus ampliatis; limbi dentibus angustis acutis; achaeniis immaturis 2 mm. longis; pappi setis ca. 30 corollam sub- aequantibus flavescenti-albis sublaevibus,—Cosra Rica: suffrutescent vine; flowers green, in wet forest, Las Nubes, Province of San José, alt. 1500-1900 m., Mar. 20-22, 1924, Paul C. Standley, no. 38,509 (ryPE, U. S., isotype Gr.). This species, having heads sessile on the branchlets of the panicle, forms a perceptible transition to § Spiciformes but its spicate tendency is so slight that the species seems best placed in § Thyrsigerae. It has much the habit and many of the characters of the Brazilian M. laevigata Sch.-Bip. but that has oblong rather than ovate-oblong, acute rather than caudate-acuminate leaves, shorter petioles, and denser inflorescence. Its involucral scales are similarly callose- thickened at the base but not variolate. § Thyrsigerae) pulverulenta Sodiro in herb., fruticosa scandens sordide puberula vel etiam in inflorescentiis tomentella; caule subtereti pluricostulato 5 mm. vel ultra diametro primo fulves- centi-tomentello deinde subglabrato vel glabrato fistuloso; internodiis 6-14 cm. longis; foliis oppositis petiolatis ellipticis vel lanceolato- oblongis acuminatis integris basi rotundatis margine revolutis textura chartaceis supra praecipue in nervis puberulis sublucidis planis vel paullulo rugulosis subtus pallidioribus prominenter reticulato- venosis supra basin 5(—7)-nervatis minute punctatis vel papillosis et in nervis patenter crispeque pubescentibus 7-10.5 em. longis 3.4-5 em. latis; petiolis ca. 1 em. longis sordide tomentellis, eis ejusdem jugi ab annulo tenui exappendiculato vix prominulenti transverse con- junctis; paniculis pyramidatis in inflorescentiam compositam ter- minalem usque ad 6 dm. longam foliaceo-bracteatam conjunctis; capitulis sessilibus plerisque ad apices paniculae ramulorum ternatim glomeratis 6-7 mm. longis; involucri squamis subobovati-oblongis apice rotundatis ca. 3 mm. longis et 1 mm. latis viridibus dorso. sordide puberulis et granulosis; bracteola lineari-lanceolata ca. 1.3 mm. longa; corollis ca. 5 mm. longis extus puberulis; tubo proprio 1.6 mm. longo; faucibus gradatim paullo ampliatis ca. 2.2 mm. longis; dentibus limbi anguste deltoideis acutis ca. 1 mm. longis; achaeniis nigris deorsum decrescentibus ca. 2 mm. longis glabris lucidis; pappi setis ca. 23 flavescenti-albis apicem versus incrassatis sublaevibus.— Ecuapor: in subandine woods near Tambillo, alt. 2800 m., Aug., 1918, Father Luis Mille, no. 524 (ryPr, in Gray Herb.). RECORDS OF THE EUPATORIEAE,—VI 59 This species is certainly close to the Colombian M. sylvatica Klatt (of which the Ecuadorian M. lanceolata Hieron. may prove a synonym), but M. sylvatica (as well as M. lanceolata) has narrower, more coriace- ous, and much smoother leaves, with less sharply prominulent veins and a somewhat different type of pubescence. In M. sylvatica the lower surface of the leaves is sometimes almost glabrous, at others sparingly villous with appressed or subappressed slender straight hairs. In M. pulverulenta the lower surface of the leaves, at least upon the nerves and chief veins, is shortly but distinctly shaggy with a fulvous crisped pubescence and also somewhat obscurely granuli- ferous (whence doubtless the specific name selected by Father Sodiro). In M. pulverulenta, two varieties may be distinguished as follows: Var. typica, foliis ellipticis breviter acuminatis basi rotundatis bis longioribus quam latis subtus cease in sewer in venulis vix puberulis.— Distrib. and exsicc. as stated abov Var. lasiophlebia, var. nov., foliis lanceoato-ob ong caudato- acuminatis ter longioribus quam latis basi plus minusve acutatis subtus flavescenti-viridibus in venulis at a a DOR: in subandine forests of the a Valley on Mt. Pichincha, alt. 2800 m., 1919, Father Luis Mille, no. 526 (ryPE, in Gray Herb.). Netwithstnadnie the marked foliar diteuancs stated, these plants show such coincidence in inflorescence and many minor characters as to discourage their treatment as independent species. Much more copious eae is needed to draw specific lines with confidence in this — (§ Thyrsigerae) Standleyi, spec. nov., suffruticosa scandens in aus nigrescens; caule tereti multi-costulato primo crispe fulvido-puberulo tardius glabrato; internodiis usque ad 14 cm. vel ultra longitudine; foliis oppositis longe petiolatis ovatis acuminatis obsolete cuspidato-denticulatis basi late cordatis sed ad apicem petioli breviter acuminatis supra glabris subtus crispe puberulis - a basi 3(-5)-nervatis subtus prominulenter reticulatis 8-10 c longis 6-7 cm. latis cen are petiolis gracilibus flexuosis veri- similiter prehensilibus 4-5 cm. longis; eis ejusdem jugi basi per cellis plerisque 4-5 mm. longis; bracteola lanceolata scuminata 2-3 mm. longa in apice pedicelli affixa; involucri squamis 6.5 mm. longis ‘atrobrunneis vix costulatis apicem versus dorso puberulis, exteriori- bus subacutatis ca. 2 mm. latis, interioribus apice rotundatis 2.5 mm. latis subscareoso-marginatis; corollis albidis ca. 6 mm. longis; tubo 60 ROBINSON proprio gracili 4 mm. longo glabro; faucibus brevissimis extus puberu- lis; dentibus limbi ovatis acutis marginatis ca. 2.5 mm. longis et 1.2 mm. latis; antheris longe exsertis; achaeniis glabris deorsum paullo decrescentibus olivaceo-brunneis 4-5 mm. longis; pappi setis ca. 70 ca. 5.5 mm. longis fulvescenti-albidis.—Costa Rica: in wet forest; flowers dirty white, La Estrella, Province of Cartago, Mar. 26-27, 1924, Paul C. Standley, no. 39,155 (U. S., phot. and fragm. Gr.). In some respects recalling the Peruvian M. incasina Robinson, but differing in its shorter, broader, relatively longer-petioled leaves, and in its phyllaries which are less modified toward the tip and show no tendency to spread abruptly as in the case of the South American cies . (§ Thyrsigerae) stereodes, spec. nov., fruticosa volubilis; caule tereti multicostulato ca. 4 mm. diametro fulvido-tomentello scabriusculo; medulla flavescenti-alba; internodiis 5-15 cm. longis; -Tamis patentibus; foliis oppositis breviter petiolatis ovatis acuminatis integris vel obsolete remoteque cuspidato-denticulatis basi obtusis vel subrotundatis supra breviter scabro-pubescentibus tardius (nervis exceptis) subglabratis minutissime reticulato-rugulosis subtus opacis 1 *1 Sh 8 Bal | ® 4 x rs. | ed ro “ pp I gro-f mi- nulenter reticulatis textura rigidiusculis a basi fere ad mediam partem pinnatim 7-9-nervatis 6-10 cm. longis 2.5-6 cm. latis; petiolis 1-1.4 em. longis flexuosis tomentellis, eis ejusdem jugi leviter annulo inappendiculato transverso conjunctis; paniculis ramos terminanti- bus ovoideis 6-25 cm. longis 5-10 em. diametro fulvo-tomentellis patenter ramosis; bracteolis lanceolato-linearibus; pedicellis 0.7-2.8 mm. longis tomentosis; involucri squamis oblongis apice obtusis vel rotundatis 3.5 mm. longis 1 mm. latis, exterioribus dorso dense appresseque pubescentibus ecostatis; interioribus glabriusculis; corol- lis parce villosulis ca. 4.5 mm. longis; tubo proprio ca. 1 mm. longo; faucibus distincte ampliatis cylindratis ca. 3 mm. longis; dentibus limbi anguste deltoideis ca. 0.7 mm. longis; achaeniis olivaceis laevius- culis ca. 4 mm.-longis et 0.8 mm. crassis; pappi setis 23-30 brun- nescenti-albis ca. 4 mm. longis.—Prru: Villacabamba, Hacienda on Rio Chinchao, alt. 1830 m., July 17-26, 1923, Macbride, no. 5163 (TYPE, Field Mus., isotype Gr.). In many respects near M. trachodes Robinson but having longer achenes, and a very different corolla with throat much longer in Yelation to the proper tube and limb. Furthermore the veinlets in M. trachodes are prominulent on the upper surface of the leaves, while in the species here described they are immersed or even some- _ what depressed. RECORDS OF THE EUPATORIEAE,—VI 61 Kanmia Poni Bak. in Mart. Fl. Bras. vi. pt. 2, 370 (1876). During the examination of material recently lent by the Botanical Museum of Copenhagen for study at the Gray Herbarium, there were found two sheets collected by Warming and provisionally referred to Mikania, which proved on comparison to be identical with the little known Kanimia Pohlii Bak. Examination of this new material yields a few added characters, which may be put on record. Baker states without qualification that the leaves are opposite, but even in the type (Pohl’s no. 536) they are only subopposite, some of the pairs being more or less separated, while in the obviously identical material of Warming, namely his no. 645 and an unnumbered speci- men, both from Lagoa Santa, the leaves are practically alternate, those of the same pair being separated often by an interval nearly as long as the distance to the nearest leaf of the next pair. Baker states that the pappus-bristles are 60 or more. A careful count showed them in one floret to be 145. Schultz-Bipontinus, annotating an envelope containing bits of a dissection accompanying Warming’s no. 645, stated that the heads are 3-flowered, but they were clearly 4-flowered in a dissection made by the writer. Baker states that there are five involucral scales. Of course one of these is the bractlet which is broadly rhomboid-ovate about 8 mm. long, and 5 mm. wide, while the four phyllaries are elliptic-oblong, about 9 mm. long and 3-4 mm. wide. The corolla-dimensions, not very clearly stated by Baker, are as follows: proper tube 1.5 mm. long, the scarcely differentiated and but slightly enlarged turbinate throat about 1 mm. long; the teeth of the limb lanceolate, acutish, 3 mm. long, 1 mm. wide. The whole corolla is granulated on its outer surface. Among Ekman’s Cuban Eupatoriums, sent to the writer for identi- fication, there was found a new Salmea of eupatorioid habit. It may be put on record as follows: Salmea umbratilis, spec. nov., fruticosa non scandens 5 dm. vel ultra altitudine; caulibus teretibus cortice griseo-brunnea tectis 2 mm. diametro adscendenter ramosis; internodiis 3-5.5 cm. longis juventate prorsum hispidulis; foliis oppositis breviter petiolatis subdeltoideo- ovatis obtusis vel acutiusculis u utroque plerumque 1(—2)-dentatis basi rotundatis supra in costis et venis principibus puberulis subtus hispidulis paullo supra basin 3-nervatis utrinque reticulatis 2.5-3.5 em. longis 1.8-2.3 em. latis subtus pallidioribus textura firmiuscule membranaceis; nervis bus; petiolo 4-6 mm. longo; ‘corymbis terminalibus (3-4 cm. longe) pedunculatis convexis ca. 6-capitulatis; pedicellis 3-6 mm. —— 62 ROBINSON capitulis ca. 7 mm. altis 1 cm. diametro; involucri campanulati— squamis ovalibus apice rotundatis persistentibus; paleis oblanceolatis apicem versus erosis; corollis in tubo cum glandulis globularibus ornatis; achaeniis sursum in marginibus villoso-ciliatis apice biarista- tis.—CuBa: Province Pinar del Rio: shady places at about 500 m. on steep limestone cliffs, region of the Mozotes in the Sierra del Sitio Santo Tomas, 10 June, 1923, Ekman, no. 16,631 (TypE, in Gray Herb.). This plant is manifestly close to the little known S. glaberrima Wright, collected on the Guajaibon Mountain by Wright in the early sixties and rediscovered and again collected by Ekman in 1921, but S. glaberrima is really glabrous, it has larger leaves of a somewhat more rounded contour and subcordate at the pinnately 5-nerved base. Furthermore their upper surface (at time of anthesis) is reticulated with clearly elevated veinlets while those in S. umbratilis are impressed even in specimens collected after flowering. While it is possible that these two plants, so nearly related, may yet be found to intergrade, their differences are too marked to permit them to be classed as conspecific in our present ignorance of any such intergrading tendencies. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY. No. LXXVIII. ISSUED MAR. 15, 19er STUDIES IN THE BoRAGINACEAE.—VI. _ A REVISION OF THE SOUTH AMERICAN BORAGINOIDEAE. By Ivan M. Jounston. PUBLISHED BY THE GRAY HERBARIUM OF HARVARD UNIVERSITY CAMBRIDGE, MASS., U. 8. A. 1927. ia ee ees Je ee oe, A REVISION OF THE SOUTH AMERICAN BORAGINOIDEAE., By Ivan M. Jounston. In the present paper an attempt has been made to give a critical descriptive account of the American species of the subfamily Boragi- noideae now known from south of Panama. The species of this group were last brought together in the general account of the whole sub- family published in 1846 in the 10th volume of DeCandolle’s Pro- dromus. Since the appearance of this scholarly treatment great changes have come about, the generic classification has been well nigh remade and the great abundance of the group in the Chilean flora has become known, the recognized species being much more than doubled. The subfamily having the South American distributional center in Chile, the recent work dealing most extensively with these plants is Reiche’s account of the family published serially in the Anales de la Universidad de Chile (vol. exxi) in 1907 and 1908 and as part of his Flora de Chile (vol. v) in 1910. This treatment is very disappointing, for through the author’s careless and often manifestly inaccurate in- terpretations of the Philippian species, the types of which he might have studied, he has only added to that confusion in the literature on the Chilean flora that first arose from R. A. Philippi’s over ene multiplication of species The classification here presented has developed from a fairly de- tailed account of the Chilean species which I prepared in 1926 while studying in the Philippi herbarium at the Museo Nacional in Santi- ago. After my return from Chile I was loaned the South American material of the group from the collections of the United States Na- tional Herbarium, the New York Botanical Garden and the Field Museum of Chicago. With this extensive material to supplement the rich collections of the Gray Herbarium the extra-Chilean species were also studied, the account of the Chilean species thoroughly re- vised and the present paper written. Though as thorough and critical as the materials and facilities available permitted, the following treatment makes no pretense at finality, but is believed to afford a consistent and logical summary of our present knowledge, thus providing a sound foundation for the work still to be oer in further elucidation of the complexities of this interesting gro Without opportunity to study the extensive and ndialentad col- lections of Philippi in the Museo Nacional at Santiago the subjoined account of the South American Boraginoideae could never have been written. I am, therefore, under particular obligation to the Com- mittee on Sheldon Fellowships at Harvard University for the privilege 4 JOHNSTON of visiting Chile as a Sheldon Traveling Fellow. I am also extremely indebted for the very courteous treatment which I received from the staff of the Museo Nacional de Chile at Santiago. Prof. Carlos Porter, Prof. Marcial R. Espinoza, Sr. Luis Moreira, Sr. Gualterio Looser and Sr. Gilberto Montero were very friendly and helpful and did much to make my work in the museum both pleasant and profit- able. I am particularly indebted, however, to Prof. Francisco Fuentes, Curator of the Phanerogamic Herbarium, through whose friendliness and courtesy I was afforded every facility for studying the famous collection which is in his charge. While at Santiago I also received very courteous treatment from Prof. Victor M. Baeza, of the Instituto Pedagéico de Chile, being through his kindness permitted to study the fine herbarium of recent collections brought together by his efforts and those of Prof. Federico Johow and their students. Various persons have assisted me in important details. Dr. H. M. Hall of the Carnegie Institution and Mr. E. P. Killip of the United States National Museum generously procured for me photographs of various essential types in the European herbaria. Prof. H. Lecomte of the Muséum d’Histoire Naturelle, Paris, most kindly sent for my examination, fragments of certain obscure species of Clos and Weddell. Dr. A. Brand, the well known monographer of the family, in generous cooperation has provided me with fragments of and notes concerning his recently published South American species. I have also been much aided by notes and specimens received from Dr. A. W. Hill, Director of the Royal Gardens at Kew, Prof. Carl Skottsberg of the Géteborgs Botaniska Tradgard, Prof. C. C. Hosseus, of the Univer- sidad Nacional de Cérdoba, Prof. Lucien Hauman, formerly of the Universidad de Buenos Aires, Dr. Alberto Castellanos of the Museo Nacional of Buenos Aires, Prof. L. R. Parodi of the Universidad de Buenos Aires, and the late Carlos Spegazzini of La Plate. Of great aid in the present work, have been the large loans received from the United States National Herbarium through Dr. W. R. Maxon, from the New York Botanical Garden through Dr. N. L. Britton and from the Field Museum of Natural History of Chicago through Mr. D. C. Davies and Mr. J. F. Macbride. In preparation of this paper, as well as in the previous ones of the series, I have profited much from the suggestions and the able editorial assistance of Prof. B. L. Robin- son as well as from the help in bibliographic matters received from the librarian of the Gray Herbarium, Miss Ruth D. Sanderson. pees appreciating the aid of all these persons, who have so cour- y responded to my needs, I would here express my hearty ele for their generous assistance. SOUTH AMERICAN BORAGINOIDEAE 5 In the systematic account, which follows, I have cited all the ma- tezial which I have examined. I have tried to arrange the collections listed in geographical sequence grouping them under the countries and the major subdivisions (Provinces or Departments) of the coun- tries. Following the citation of each collection, I have appended in parenthesis initials indicating the herbarium in which I have seen material of the particular collection. The initials used and the her- baria they indicate are as follows—“G”’ for Gray Herbarium, “NY” for New York Botanical Garden, “US” for United States National Herbarium, “FM” for Field Museum of Chicago, “MS” for the Museo Nacional of Santiago and “IP” for the Instituto Pedagéico of Santiago ARTIFICIAL Key To SoutrH AMERICAN GENERA, Nutlet wing-margined or appendaged dorsally. — _Nutlets with appendages or teeth of the en ti pubescence; slender herbs with ee "Warermsek leaves opposite Nutlets with glochidiate aprendegss: leaves _ alternate. Nutlets equalling the s ch t¢ xed for nearly hair whole length along the entra PCE he EES a ee PA ee Ce eee 13. Lappula. Nutlets much longer agp the es gynobase to which they are affix a broad areo Nutlets spreading : divert ont _keeled teas — Bg or lateral and rom ere Se I a cee: 6. Cynoglossum. Nutlets “ascending, keeled ventrally above the ‘oni Pedic els eg ae in fruit; inflorescence paniculate, ted or naked: betie... oo. ici. sks. 14. Hackelia. Pedicels ele in fruit; iahoreseade corymbose, bract- een; abr Re | ana geting Sy sees es 15. Selkirkia. Nutlets neither wing-ma: pendaged y- Attachment-surfa sta of nutle Eh cuered strophiolate, sur- rounded by a tumid Corolla par ht Tg pasta ae that of a Solanum; stamens conspicuously exserted, appendaged dorsally.......... 7. Borago. Corolla salverform; stamens included, unappendaged.... .. 6. Anchusa te hairs. Inflorescence bracted; sales stipitate...... 3 Cle Mere Inflorescence bractless; nutlets angen 2S ee . Morit Wee a ee 1. Lithospermum. 6 JOHNSTON Nutlets attached laterally; at least lower leaves opposite Corolla blue; stigmas geminate............... 2. Antiphytum. Corolla not blue; stigmas solitary and simple. rolla yellow or orange, throat unappendaged. 12. Amsinckia, Corolla white, throat appendaged. Nutlets with pericarpial wall fused above mid- e to form a definite medial ventral keel. 11. Plagiobothrys. Nutlets with a medial ventral groove formed by the non-fusion of the pericarpial wall. Calyx cut to base, distinct from the bracts. 9. Cryptantha. Calyx cylindrical, merely toothed, with floral bracts decurrent on its lower Pe ea ak la eke 10. Nesocaryum. I. Trise LirHosrerMEAE. Nutlets erect, straight or rarely bent, smooth or rough, unmargined; areola basal or occasionally suprabasal, near tip of cotyledon, un- margined, flat or nearly so, not at all strophiolate, usually sessile but occasionally stipitate; gynobase flat or low-pyramidal, not excavated; style cleft or entire; stigmas 2 or exceptionally 4, distinct or proximate or rarely fused, capitate or obscurely thickened; corolla yellow or orange or occasionally white or somewhat purplish_—Johnston, Contr. Gray Herb. Ixxiii. 43 (1924). Key To GENERA. Corolla with conspicuously oblique limb and unequal stamens...5. Echium. Corolla with horizontal limb and stamens. , gate lobes erect or sp: ng; nutlets usually 4, falling individually; calyx | brac inate hairs Nutlet-areola , Sessile; leaves all alternate... .. 1. Lithospermum. Nutlet-areola suprabasal, sessile or stipitate, at least the Se Ope. ks a 2. Antiphytum. Mature calyx cylindrical or ellipsoid, merely toothed with the teeth connivent over the fruit; nutlets solitary, falling away tightly invested by the calyx; calyx or bracts with uncinate pu ce. Inflorescence bracted; calyx lacking uncinate hairs; corolla- With stipitate arcola..-....2........... ......3. Thaumatocaryon. Inflorescence bractless; calyx with uncinate hairs; corolla- throat with circular villous spots or intruded append- i ages; nutlet with sessile Sava! aisok eel We eee 4, Moritzia. 1. Lithospermum L. Calyx usually divided. Corolla tubular or salverform; tube cylin- drical; lobes spreading, imbricate; throat with intruded appendages or SOUTH AMERICAN BORAGINOIDEAE 7 with pubescent or glandular areas; stamens affixed in tube, included; filaments short; anthers oblong, usually with apiculate connectives. Style filiform. Stigmas geminate. Ovules 4. Nutlets 4 or rarely fewer, erect, ovoid or angular, smooth or verrucose, affixed by a broad horizontal or slightly oblique basal areola. Gynobase flat or very broadly pyramidal.—Annual or perennial, herbaceous or fruticose plants with alternate leaves. Flowers white, yellow or violet, in bracted racemes.—Sp. Pl. 132 (1753) and Gen. Pl. 64 (1754). Batschia Gmel. Syst. ii. 315 (1791). Cyphorima Raf. Jour. Phys. Ixxxix. 98 (1819). Pentalophus A. DC. Prodr. x. 86 (1846). A large genus with about 30 species in North America and 20 in the Old World. The South American clearly have been derived from those of Mexico and Central America. Key To Speci&s. Nutlets om _ tg much roughened; mature calyx oblique at base; introduced annual............. 1. L. arvense. Nutlets smooth aba ae mature calyx Seewrnece tt not notice- oblique; indigenous perennia Goss: 15-36 mm. long, yellowish, tube surpassing the sr 1 eg or more, throat without protruding append- Corals 25-36 mage Rees, hispidulous; stems with erp! or ng pu MOG oe. yr L. cinerascens. Corolla Rai fea long, villous-strigose; stems ii Pee ee Ure, oy eC Cte ee 3. L. Macbridei. Contliag ‘3-10 mm long, — or ochroleucous, tube surpas- sing the calyx 5 pea r less, throat with pubescent pro- truding appe de 1 cm. Can: a 3-5 dm. tall; Colombia and Vene- SOS eit eevee ee 4. L. mediale. Ceehe 3-6 mm. long; plant less than 1.5 dm. tall; Peru. Corolla 5-6 mm _ long; leaves 10-20 mm. long....... 5. L. Shepardae. Corolla 3.5—-4 mm. lo ae é wee 4G em, Wn ns eS Ss 6. L. a Leaves 12-30 mm. pe iefeclies ese vcs wk uae 7. L. peruvianum. 1. Lithospermum arvense L. Annual; stems erect, 2-7 dm. -tall, one to several, simple or sparsely branched above, hoary-strigose; leaves 2-5 cm. long, 2-8 mm. broad, closely appressed-hispid, firm, veinless, with violist midrib, pale beneath; basal leaves rosulate, oblanceolate or spathulate, those of the mid-stem acute and lanceolate or linear; racemes bracted, becoming loosely flowered; calyx parted into subulate-linear lobes, hispid; mature calyx with the erect or ascending lobes becoming 8-13 mm. long, the very short tube oblique pale and chartaceous; pedicels short and stout, ca. 1 mm. long; corolla white or yellowish or even purplish, 5-7 mm. long, tubular-funnel- form, the tube glabrous within and scarcely if at all surpassing the 8 JOHNSTON calyx; corolla-throat gradually expanded, without protuberances, merely puberulent; corolla-lobes ascending, ovate, obtuse; nutlets brown, dull, roughened, tuberculate or rough-wrinkled or pitted, ca. 3 mm. long, triquetrous-ovate.—Sp. Pl. 132 (1753); Johnston, Contr. Gray Herb. Ixx. 21 (1924). ARGENTINA. Buenos Arrzs: Sierra de la Ventana, Dec. 1911 , Hau uman (G); vicinity of Buenos Aires, Ezlikman 10 (G); Villa Soldati, Nov. 1924, Parodi 5961 (G). A European plant, widely oe? in the United States, which can now be reported from the Argen 2. L. cinerascens (A. DC.) Sokactin. Perennial (?); stem ca. 2.5 dm. tall, simple, erect, cinerescent, somewhat shaggy, very loosely appressed or spreading short-hispid; leaves lanceolate, firm, 1.5-2.5 cm. long, 4-6 mm. long, acute, erect, crowded, densely appressed- hispid, cinerescent, somewhat bulbous-pustulate, base sessile and rounded, midrib prominent beneath, margins slightly involute; racemes terminal, bracteate, few-flowered; calyx divided into nar- rowly linear acute lobes, appressed short-hispid, in flower 5-8 mm. long; pedicels ca. 3 mm. long, slender; corolla yellowish, tubular with a slight contraction towards the base, 2.5-3(-3.6?) em. long, hispi- dulous outside, lobes spreading and ca. 3 mm. long, throat weakly differentiated; nutlets ovoid, white, shiny, sparsely punctate, ca. 3 mm. long; style subexserted, very long and slender.—Contr. Gray Herb. Ixxv. 40 (1925). Macromeria cinerascens A. DC. Prodr. x. 69 (1846) ; Johnston, 1. c. Ixx. 15 (1924). PERU. Amazonas: Chachapoyas, Matthews (G, photo. of TYPE). I know this species only from the original description and from an excellent photograph of the type specimen. From these sources the above description has been drawn. The species has obvious affinities with L. Macbridei of central Peru, but is also related to L. Palmeri Wats. and L. viride Greene of northern Mexico and adjacent United _ States and with L. guatemalense Donn. Sm. and L. oblongifolium: Greenm. of Guatemala and southern Mexico. These more northern species besides being widely separated from it geographically, have r more herbaceous veiny leaves and so are readily separable feiss L. cinerascens and its Peruvian relative, L. Macbridci. 3. L. Macbridei, sp. nov. odio beaoesinad fa fruticulosis a erectis 2-6 dm. altis e radic cineres- ibus SOUTH AMERICAN BORAGINOIDEAE 9 margine leviter reeurvis; racemis terminalibus paucifloris bracteatis; calycibus strigosis floriferis 4-5 mm. longis fructiferis 5-6 mm. longis; pedicellis maturitate ca. 1 cm. longis faleatis strictis; corollis viridi- flavis tubuliformibus 15-18 mm. longis, tubo attentuate cylindrato 3-4 mm. crasso calycem ca. 1 cm. superanti intus glaberrimo extus villoso-strigoso; lobis divaricatis vel recurvatis orbicularibus 2.5-3 mm. longis; faucibus exappendiculatis obscurissime puberulenti- pubescentibus; staminibus supra medium tubi affixis, filamentis quam antherae lineares 2 mm. longae brevioribus; nuculis ovoideis albis lucidis 2.83.2 mm. longis calyce evidente brevioribus; stylo longissimo gracillimo. PERU. Huvanuco: densely shrubby northeastern slo Pdcnarge ss, 3000 m. alt., June, 1922, Macbride & Featherstone 1185 (FM ; GU, Un, nioereeeys steep grassy slopes, Chinchapalca ahve Mito, 2850 m. m. > ake, "July 1922, Macbride & Pee ee 1592 (FM); grassy eastern 5 slope, Evidently allied to L. cinerascens of northern Peru but differing in having smaller corollas and in being closely appressed-strigose rather than loosely or spreadingly hispid. The corolla-structure is quite like that of the North American L. viride Greene and L. Palmeri Wats. 4. L. mediale Johnston. Perennial; stems several, erect, 3-5 dm. tall, simple or occasionally very sparsely branched, cinerescent, ap- pressed pallid-hispid or hispid-strigose and very densely so on the younger parts; leaves oblong-lanceolate or oblong-ovate, 3-6(-9) cm. long, 4-13(-17) mm. broad, firm, acute, appressed short-hispid, finely pustulate, midrib prominent beneath; racemes terminal, leafy; calyx parted into narrowly linear appressed villous-hispid lobes; pedicels strictly ascending, 1-3(—5) mm. long; corolla white or ochroleucous, ca. 1 em. long; tube cylindrical, almost twice the length of the calyx, villulose outside, glabrous within, ca. 2-2.8 mm. thick; lobes 1-2 mm. long, widely spreading or divergent, orbicular; throat with 5 prominent oblong pubescent appendages; filaments attached above middle of the tube, shorter than the anthers, anthers almost 1 mm. long; nutlets 2-3, white and shiny, 2.5-3 mm. long, clearly surpassed in length by calyx and style-—Contr. Gray Herb. Ixx. 28 (1924). VENEZUELA. Mé&rma: between Mucurulea and Mucuchies, 2700 m. alt., Jan. 1922, eS te — ners GOLOMB A. ba: ore, Zipacon, cased m. alt., railroad Aug. 1917, Rusby © Pt Posed Hi pa @; meadow on sabana te, 2600 m. ig Oct. i917, Pennell 2468 (G, NY); Fucatativa, oo 669 NY); vicinity ore Falls, ——! vce (NY); S: n Antonio, 1905, artaleeek A (G). InpDEFINITE: carne Scooney. Ariste- Joseph B-44 (US); El Dintel, Jan. 1920, Rite Tomek A-525 (G.) 10 JOHNSTON This species also occurs in Central America.. Its relations are among the white-flowered Mexican species. 5. L. Shepardae, sp. nov. Perenne fruticulosum nanum breviter cinereo-strigosum e radice subsimplici crassa oriens; caulibus pro- stratis vel ascendentibus 4-12 cm. longis ramosis foliosis; foliis alternis subcoriaceis acutis vel obtusis subtus cum costa prominenti spathu- lato-obovatis vel anguste oblanceolatis 1-2 cm. longis 2-4 mm. latis basem versus contractis, strigosis et inconspicue tuberculato-pustu- latis margine leviter recurvatis, superioribus paullo reductis; racemis terminalibus brevibus foliaceo-bracteatis; lobis calycis oblongis vel linearibus; corolla alba evidenti 5-6 mm. longa, tubo 3 mm. longo cylindrico calycem evidente superanti, lobis rotundis divaricatis 2-2.5 mm. longis, limba 5-7 mm. lata, faucibus cum appendiculis 5 pubescentibus oblongis prominentibus; nuculis 4 ovoideis lucidis laevibus 2-2.3 mm. ingle: t in maturitate calycis lobos atque stylum paulum superantibus. PERU. Puno: rocky clay soil on puna, Chuquibambilla, 3900-4000 m alt., April 1925, Ponnal: 13357 (G, FM); mountain-side near Puno, 3750 m. alt., Dee. 1920, Mrs. Reid S. Shepard 114 (G, TYPE). Apparently most related to L. Gayanum from which it differs in its larger leaves and flowers, its shorter sparser pubescence and more southern range. The corolla of L. Gayanum is salverform and has a narrow limb. In L. Shepardae the limb is very broad, in fact much broader than the tube is long, and the corolla hence appears to be rotate. 6. L. Gayanum (Wedd.), comb. nov. Perennial from a strong root; stems much branched from the base, fruticulose, procumbent or ascending, 3-5 cm. long, leafy, densely pale-strigose; leaves lanceolate to lance-elliptical or lance-spathulate, obtusish or acute, subcoriaceous with a prominent midrib beneath, densely appressed pale short-hispid, 4-6 mm. long, 1.5-2 mm. broad, the basal leaves slightly larger and shortly petiolate; racemes apical, short, few-flowered, bracteate throughout; calyx with linear acute lobes; pedicels obscurely devel- oped; corolla white, 3.5-4 mm. long, tube ca. 2.5 mm. long and only slightly surpassing the calyx, lobes rotund and spreading, throat with 5 protruding oblong pubescent appendages; nutlets 4 or by abortion fewer, ovoid, shiny, becoming as long as style and calyx-lobes.— Eritrichium Gayanum Wedd. Chlor. pistes ii. 88 (1859). L. eas Krause in Engler’s Jahrb. xxxvii. 636 (1906). peeRu.. Some: Andes of Cusco, Gay (G, fragment of the type of SOUTH AMERICAN BORAGINOIDEAE ll A species of central soe Beep ee eo by its dwarf habit, very small leaves and small flow 7. L. peruvianum A. DC. Perennial, hispid; stem branched, terete, woody at base, with spreading hairs; leaves linear, acute, 1.2-3 cm. long, 2-4 mm. broad, crowded, tuberculate-pustulate; racemes few-flowered, leafy-bracted; calyx 5-parted into acute linear lobes, appressed-pubescent; corolla white (?), 4 mm. long, appressed-pilulose outside, funnelform above the middle, tube scarcely surpassing the calyx, lobes ovate and obtuse, throat with pubescent protuberant appendages; anthers included, not reaching the appendages, not mu- cronate, filaments short, nutlets usually solitary by abortion, 2 mm. long, ovoid, white and shiny, punctate.—Prodr. x. 77 (1846) The description given above is adapted from the original given by DeCandolle. I know the plant only from description. The type was collected at Chachapoyas in northern Peru by Matthews. The species seems to be related to L. Gayanum. DovuBTFUL oR EXCLUDED SPECIEs. L. aggregatum R. & P. Fl. Peruv. ii. 4 (1799); Lehm. Asperif. ii. 313 (1818).—Apparently a species of Coldenia, C. PARONYCHIOIDES Ph., cf. Johnston, Contr. Gray Herb. Ixx. 60 (1924). (1818). L. incanum R. & P. FI. Peruv. ii. 4 (1799); not L. incanum Forst. f. (1797).—Tarma, Peru. Described as an annual with cunei- form leaves, solitary axillary flowers and shiny ovate acuminate nutlets L. dckenen R. & P. Fl. Peruv. ii. 5, t. 111¢ (1799).—This is a Coldenia, C. picHotoma (R. & P.) Lehm 2. Antiphytum DC. Calyx cut to base or nearly so. Corolla-tube cylindrical; lobes im- bricate, rounded, spreading or laxly ascending; throat with intruded ciliate or papillate appendages; stamens included, attached at middle of tube, filaments slender, short; anthers oblong. Style slender, included in tube; stigmas geminate. Ovules 4. Nutlets 4, erect, ovoid or angulate, rugose or verrucose; areola inframedial or supra- basal, at lower end of ventral keel or terminating a downward-pro- jecting stipitate prolongation of the keel springing from just above the base of the nutlet. Gynobase flat to depressed-pyramidal.—Perennial fruticose or fruticulose plants. Leaves opposite at least at the very base of the plant, linear to lanceolate. Racemes bracted. Corolla 12 JOHNSTON blue, yellowish or white—DC. in Meisner, Gen. i. 280; ii. 188 (1840); Johnston, Contr. Gray Herb. Ixviii. 48 (1923). Amblynotopsis Macbr. Contr. Gray Herb. xlviii. 41 (1916 The definition of this genus is still subject to revision. As here defined it is a group most related to Lithospermum and consists of several fruticose perennials of Uruguay and perhaps adjacent Brazil, as well as a group of slightly less fruticose Mexican plants to which the generic name Amblynotopsis is applicable. The material which I have studied of true Antiphytum, that is to say the South American plant, is too meagre and incomplete to yield precise knowledge as to their characteristics. Since no one has yet pointed out funda- mental characters which set off the Mexican species from the true Antiphyta of South America I have drawn my generic diagnosis to cover both. Key To SPEcIzEs. Leaves linear-lanceolate, —s em. long, 3-6 mm. broad, ob- securely nerved, margins flat; racemes paniculately’ rel iia eee var oul Ge Aik ke ae hee eas . A. cruciatum. ——- linear, 1. 5-2.5 em. long, 1-2 mm. broad, ares ess, s revolute; racemes geminate and terminal.2. A. stoechadifolium. ft: Aikignetin cruciatum (Cham.) DC. Perennial, fruticose, 3-5 dm. tall; stems erect, appressed-hispid, canescent, ascendingly branched, decorticating in age; leaves linear-lanceolate or lanceolate, 1.5+4 em. long, 3-6 mm. broad, appressed-hispid, sparsely and incon- spicuously veined, acutish, attenuate apically, opposite with narrowly connate subsheathing bases, becoming reduced up the stem, margin weakly if at all revolute, midrib prominent and hispid underneath; racemes paniculately disposed above, bracted throughout, the bracts lanceolate, lateral and not surpassing the calyx; calyx (at anthesis) ca. 3 mm. long, appressed-hispid, 5-parted, the lobes lanceolate, unequal with the abaxial pair the largest; pedicels very short; corolla-tube short, surpassed by calyx; corolla-lobes rotund, pubescent outside, about length of tube; throat with rotund fimbriate appendages; sta- mens included, anthers elliptical; filaments very short, affixed at pi sh of tube; genty unknown; style short, stout, with emarginate 21 (1846); Fresen. in Mart. Fl. Brasil. viii. pt. 1, 64 yi peda Contr. Gray Herb. Ixviii. 51 (1923). Anchusa eruciata Cham. Linnaea iv. 438 (1829 I have seen no material referable ie this species. The above de- scription is adapted from the original one given by Chamisso. The _ species seems to be congeneric with, and closely related to A. stoechadi- sh lace differs, according to Chamisso, in its solitary terminal SOUTH AMERICAN BORAGINOIDEAE 13 cymes, larger flowers, very pale rev olute-margined leaves and different habit. The type of A. cruciatum is given as having come from south- ern Brazil. I suspect it is from northern Uruguay or adjacent Brazil. 2. A. stoechadifolium (Cham.) DC. Perennial, fruticose, 1.5-3 dm. tall; stems erect with strictly ascending branches, appressed pale- hirsute, cinerescent, stipitate-glandular above, in age decorticating and becoming smooth and brown; leaves linear, acutish, 1.5-2.5 mm. long, 1-2 mm. broad, veinless, opposite with narrowly connate sub- sheathing bases, becoming reduced and loosely disposed up the stem, upper surface appressed-hirsute, occasionally somewhat glandular, the lower surface densely villous-strigose and somewhat silvery, frequently covered by the meeting of the prominent hirsute midrib and the revolute margins; racemes geminate, 2-4 cm. long, few-flowered, strictly ascending, bracted throughout; bracts opposite or beside the flowers, broadly lanceolate; calyx cut to middle or beyond into erect or loosely connivent lanceolate slightly unequal lobes, densely ap- pressed-hirsute, 5-6 mm. long at maturity; pedicels stout, 1-2 mm long; corolla blue, ca. 5 mm. long; tube ca. 2.5 mm. long, 2.5-3 mm. thick at summit, tapering to ca. 1.5 mm. thick at base, shortly sur- assed by calyx, glabrous outside, inside glabrous except for the villous tufts associated with the 5 wart-like appendages at the base; limb 6 mm. broad, concave; lobes rounded, ca. 2 mm. long, appressed short- villous on middle outside; throat with 5 erect squarish trapeziform emarginate appendages (ca. 1 mm. long) of intruded tissue, margin and outer face of appendages covered with stout white hairs; stamens included, attached at middle of tube; anthers oblong, ca. 1 mm. long; filaments 14-24 length of anthers; nutlets 4, ca. 2 mm. long, ovate, tuberculate-rugose, back rounded, edges obtuse, sia a Sab Vv entral _ keel prolonged down below to form a stout the nutlets, ca. 2.5 mm. long, bearing geminate stigmas. = Prody. *. 121 (1846); Fresen. in Mart. Fl. Brasil. viii. pt. 1, 64 (1857); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 119, fig. 47e-f (1893). Anchusa stoechadifolia Cham. Linnaea iv. 439 (1829). Myosotis Berroi Arech. Anal. Mus. Nac. Montevideo ser. 2, i. 69, fig. 5-6 (1911) and op. cit. ser. 1, vii. (FI. Urug. iv.) 181, fig. 13-14 (1911). BRAZIL: without locality, Sellow 1840 (G, isotype of A. stoechadifolia). The above description is based upon the specimen cited, supple- mented by some details from Chamisso’s lengthy original description. I know Myosotis Berrot only from description, but feel quite confident that it is synonymous. It was collected in the Department of Rivera, in northern Uruguay, which Sellow is known to have traversed and 14 JOHNSTON’ explored during the early part of 1823. It is, hence, not impossible that the types of Anchusa stoechadifolia and Myosotis Berroi were both obtained in northern Uruguay and perhaps in the same region. DovuBTFUL SPECIES. Myosotis uruGuAYENsIs Arech. Anal. Mus. Nac. Montevideo ser. 2, i. 68, fig. 4 (1911) and op. cit. ser. 1, vii. (Fl. Urug. iv.) 180, fig. 12 (1911).—From Dept. of Tacuarembo in northern Uruguay; a fruticose plant probably of this genus and apparently nearest to A. cruciatum but differing in its alternate leaves and tubular calyx. I have seen no material of this species. 3. Thaumatocaryon Baill. Calyx campanulate, toothed, in age accrescent, ellipsoid, closely investing and with the short teeth connivent over the mature fruit, closely subtended by uncinate-pubescent bracts. Corolla-tube broadly cylindrical, surpassing the calyx; lobes ovate or orbicular, spreading, imbricate, obtuse; throat densely pubescent along 5 obscure horizontal plaits; stamens included, attached at middle of tube, fila- ments slender slightly longer than the obtuse oblong anthers. Style filiform, short; stigma deeply bilobed. Ovules 4. Nutlets solitary by abortion, ovoid, erect, smooth or minutely tuberculate, ventral keel strong and above the base abruptly drawn out into a stout strongly deflexed stipe bearing the areola. Gynobase flat or low- convex.—Perennial herbs with at least the very basal leaves opposite, leafy. Inflorescence of pedunculate bracted forked spikes. Corolla white (?) or blue.—Bull. Mens. Soc. Linn. Paris 839 (1890). A well marked genus of western and southern Brazil which, although in the past confused with Antiphytum, is evidently most closely related * to Moritzia. Key To Species. Stem-leaves all opposite; ae hee ; ry Bare th and shiny 1. T. tetraquetrum. ves alternate above the base; corolla white (?); nutlets (at least in no. 3) minutely low-tuberculate and dull. 5 mm. broad; ce spreading; basal seers oe peieent tle 2. T. dasyanthum. Corolla-limb 2.5-3 mm. broad; pubescence strigose, closel ; ther tufted nor y tent. .3. ?-. Sellowianum. _1. Thaumatocaryon tetraquetrum (Cham.) Johnston. Peren- nial; stems erect 3-8 dm. tall, simple or with short branches above, puberulent and hispidulous above but glabrate below; leaves acute, SOUTH AMERICAN BORAGINOIDEAE es scabrous, sparsely hispidulous, the hairs on the upper face with dis- coid bases; lower leaves tufted, the blade 9-15 em. long, 2.5-4.5 em. broad, broadly lanceolate, broadest above the middle, gradually con- tracted below into a narrowly winged petiole 7-15 cm. long; cauline leaves all opposite, broadly lanceolate or lance-elliptical or narrowly ovate, 5-10 cm. long, 1.5-3 em. broad, sessile by a rounded base; peduncles densely puberulent and appressed-hispidulous; calyx at anthesis 2 mm. long, cut to middle into 5 lance-triangular lobes; mature calyx densely villulose, narrowly ovoid, 3-4 mm. long, 2 mm. thick, sessile, teeth short, 1 mm. long; corolla blue, 5-6 mm. long; limb spreading, 4.5-5.5 mm. broad; lobes orbicular, 1.5-2 mm. long, strigose and puberulent outside; tube 2.5-3 mm. long, 1.5-2 mm. thick, cylindrical, narrowed at the base, slabrous within, colored and puberulent above the middle outside; throat densely villous along horizontal plaits; filaments filiform, attached at middle of tube, ca. 0.8 mm. long; anthers oblong, ca. 0.6 mm. long, included; nutlet nar- rowly ovoid, smooth and shiny, 3 mm. long, with the strong ventral keel prolonged as a stout abruptly incurving stipe; style very short, not reaching to middle of nutlet.—Contr. Gray Herb. Ixx. 12 (1924). Anchusa tetraquetra Cham. Linnaea viii. 113 (1833). Antiphytum tetraquetrum DC. Prodr. x. 122 (1846); Fresen. in Mart. FI. Brasil. viii. pt. 1, 64 (1857); Johnston, I. ¢. Ixviii. 51 (1923). T. Hilarii Baill. Bull. Mens. Soc. Linn. Paris 839 (1890) ; Johnston, |. ¢. Ixx. 12 (1924). Antiphytum Bornmuelleri Pilger in Fedde, — ili. 24 (1906). Antiphytum Bornmuelleri, var. asperior Pilger, |. ec. BRAZIL. ParanA: marshy places, Curityba, 900 m — Dusén 15631 (G) INDEFINITE: no locality ee Sellow (US, ISOTYPE 0 . tetraquetra ee a Sierra de Mara Nov., Hassler 5231. ‘O): “in regione fluminis Alto Parana.” Fiebrig 5885 and 6131 (G). This plant frequents wet places in eastern Paraguay, Chodat, Bull. Soc. Bot. Genéve ser. 2, xii. 158 (1920) and Bull. Herb. Boiss. ser. 2, ii. 817 (1902), and southern Brazil. Baillon, Pilger and Chodat have reported the plant from Rio Grande do Sul. Glaziou, Bull. Soc. Bot. France lvii. Mem. 3, 480 (1910), reports it, perhaps upon a misde- termination, from Curalinho in Bahia. The plant from Rio Grande do Sul has been twice described. Un- fortunately I have seen no material from this state and have been forced to judge the proposed segregate upon the basis of descriptions alone. With the exception that the leaves are described as slightly broader and perhaps slightly cordate at the base, there appear to be no differences given that serve to separate it from true A. tetraquetrum. 2. T. dasyanthum (Cham.) Johnston. Perennial; stems several, 16 JOHNSTON erect or ascending, 6-8 dm. tall, simple, leafy, scabrous, sparsely hispid and pubescent with spreading or retrorse hairs; leaves scabrous, hispid, somewhat glabrescent in age, the hairs with persistent discoid bases; lower leaves persistent at anthesis, opposite, 10-15 cm. long, ‘ em. broad, lanceolate or ovate-lanceolate, gradually con- tracted into a long winged petiole; middle leaves oblanceolate or spathulate, petiolate, 6-8 cm. long, alternate; upper leaves oblong, sessile by a triangular base, alternate, 3-5 mm. long; peduncles shaggy, densely retrorse-spreading hispidulous ae sparsely hispid; calyx (at anthesis) very densely appressed short-villous, 1.8 mm. long, cut almost to middle into 5 lanceolate lobes, sessile; corolla white (?); limb spreading, ca. 5 mm. broad; lobes ovate, pubescent outside, ca. 1.8 mm. long; tube 2.5-3 mm. long, almost 2 mm. thick, cylindrical, appressed-pubescent outside, glabrous within; throat densely villous along horizontal plaits; filaments subulate, affixed at middle of tube, ca. 1 mm. long; anthers oblong, ca. 0.7 mm. long; ovules 4, apparently only one maturing; mature nutlets unknown; stigma deeply bilobed. —Contr. Gray Herb. Ixx. 12 (1924). Anchusa dasyantha Cham. Lin- naea iv. 437 (1829); DC. Prodr. x. 52 (1854). Moritzia dasyantha Fresen. in Mart. Fl. Brasil. viii. pt. 1, 63 (1857). BRAZIL: no locality given, Sellow 7964 (US, 1soryPE). 3. T. Sellowianum (Cham.) Johnston. Perennial; stems erect, several, simple, ca. 4.5 dm. tall, strigose with the hairs antrorsely hispid and very appressed; leaves closely strigose, the hairs all with discoid bases though these sometimes obscure, oblanceolate to oblong or elliptical, 2-7 cm. long, 1-2.3 mm. broad, obtuse or acute, lower ones opposite and apparently neither tufted nor persistent, contracted into a winged petiole, upper ones alternate and sessile by rounded bases; peduncles densely strigose or appressed-hispidulous; mature calyx ca. 3 mm. long, narrowly ovoid, sessile, densely puberulent, teeth lan- ceolate and connivent; corolla white (?); limb spreading, ca. 3.5 mm. broad; lobes ca. 1.5 mm. long, ovate, very pubescent outside; tube ca. 2 mm. long, cylindrical, ca. 1.3 mm. thick, appressed-pubescent outside, glabrous within; throat densely villous along horizontal plaits; filaments attached at about middle of tube, filiform, ca. 0.5 mm. long; anthers ca. 0.4 mm. long, oblong; mature nutlet 2.5-3 mm. long, narrowly ovoid, back rounded, densely covered with minute very low lucid tuberculations, dulled, with the strong anterior keel suprabasally drawn out into a stout terete abruptly deflexed stipe; style reaching to about half height of nutlet; stigma deeply bilobed.— Contr. Gray Herb. Ixx. 13 (1924). Anchusa Sellowiana Cham. Lin- SOUTH AMERICAN BORAGINOIDEAE 17 naea viii. 115 (1833); A. DC. Prodr: x. 51 (1846). Mortizia Sellowiana Fresen. in Mart. Fl. Brasil. viii. pt. 1, 63, t. 10, fig. 14 (1857). BRAZIL: no locality given, Sellow (US, tsoType). 4. Moritzia DC. Calyx cylindrical or ovoid, toothed, in age accrescent, sessile or pedicellate, ribs armed with stout uncinate hairs. Corolla with tube cylindrical, equalling or surpassing the calyx; lobes triangular or ovate to oblong, spreading, imbricate, obtuse; throat with hairy spots or hairy intruded appendages; stamens affixed above middle of tube, filaments 1-2 times length of the oblong partially to definitely exserted anthers. Style filiform, short to long; stigma bilobed. Ovules 4. Nutlets solitary by abortion, erect, ovoid, smooth to muricate, ab- ruptly contracted below with a small basal areola at the lower end of the strong ventral keel. Gynobase flat——Perennial herbs. Leaves alternate, the basal ones large and tufted, cauline ones relatively small. Inflorescence of bractless peduncled forked unilateral spikes or ra- cemes. Corolla blue-—DC. in Meisner, Gen. i. 280; ii. 188 (1840). Meratia A. DC. Prodr. x. 104 (1846). A very natural genus of Colombia and Venezuela and of southern Brazil. It is obviously most related to Thaumatocaryon but differs in habit, inflorescence, pubescence of the calyx, attachment of the nutlets and structures of the corolla-throat. ae TO ae Style — 1-2 mm. long, n the calyx; lowermost cal ensian liseli, te, the he “pelicula 2-7 mm. long; Sg with anasto mosing usually forked veins; tone sr ube ca. 2 = long. — the calyx; stamens eer Colom ia and | Vanes ee ene pa gaia ge ees M. Lindenii. _— lon; alien "Ton, gnert surpassing the calyx; sick y sessile; leaves with simple distinct veins; coro tube +s = ‘log, 6 surpassing the calyx; stamens short- Corolia-throns with 5 bearded spots; leaves conspicuously appressed-hispid o = poe surfaces ens inserted in he throat and werent tly exserted; JP tesereneres ca. 2 mm. long and about twice length of mHbere 5 sce 2. M. Dusenit. Corolla-throat with 5 intruded ligulate bearded appendages; leaves minutely strigose or occasionally sparsely hispid below, at first glance appa : i a about # height of — waa gery 82 ang ns passed = 8 mm. Ene 5 about equate hes anthers in engl Sy > os era . M. ciliata. 1. Moritzia Lindenii (A. DC.) Benth. Perennial; stems several, erect, simple, 3-8 dm. tall, short-strigose, the hairs usually sparse and 18 JOHNSTON retrorse below the middle and abundant and antrorse above, usually also somewhat hispid; leaves acute, appressed short-hirsute with the rib and veins and margins frequently hispid, bicolored, midrib and branching veins prominent beneath and usually impressed above; lower leaves tufted, 15-20 cm. long, 2-2.8 cm. broad, oblanceolate, gradually contracted into a winged petiole, hairs on upper face all with pustulate bases; cauline leaves lanceolate to lance-oblong, 3-5 cm. long, 8-15 mm. broad, sessile by a rounded base, scarcely reduced up the stem; racemes 2-5 cm. long, disposed in a terminal corymbose panicle; mature calyx ovoid-cylindrical, ca. 4 mm. long, ca. 2 mm. thick, densely armed with coarse uncinate hairs arising from coarse tuberculate bases, more or less pentangular, with 5 erect triangular lobes ca. 1 mm. long; pedicels densely appressed short-hispid, upper ones 1-2 mm. long, the lowermost somewhat longer becoming as much as 7 mm. long; corolla blue, with a very fine tuberculate-mealy surface; limb spreading, 6-9 mm. broad; lobes semicircular or ovate, rounded, sparsely strigose outside; tube ca. 2 mm. long, slightly am- pliated upwards, about equalling the calyx; throat with 5 circular spots very densely bearded with short obese hairs; stamens attached at about 34 height of tube; anthers oblong, ca. 0.9 mm. long and about twice length of filaments, included in throat or only partially exserted; _ hutlet ovoid, obscurely angulate, white, densely tuberculate-muricate, 2-2.5 mm. long, ernpey contracted below to the basal attachment; style reaching to about ? height of nutlet, 1-2 mm. long—Ben ace. Giirke in E. & P. Nat. Pflanzenf. i iv. Abt. 3a, 121 (1893); iolincion, Contr. Gray Herb. Ixx. 11 (1924). Meratia Lindenii A. DC. Prodr. x. 104 (1846). Mériwa: Paramo de la Sal, 3400 m. alt., Jahn 515 and VEN COLOMBIA. CUNDINAMARCA: wet mo de Cru Pies near gins stg Boer tary ~ait., Deed. 2052 (G, oe NY. US). Ce oia: t_ grassy o, Cerro Tatama, alt. 3400-3500 alt, , Pennell 10576 (G); oie ae hae ius del Quindio, 3500-3800 m. ak nell & Haz A species of the higher mountains of western Venezuela and northern Colombia. The type is given as from Caracas, Venezuela. 2. M. Dusenii, sp. nov. Perennis; caulibus pluribus erectis sim- plicibus 2.5-3.5 dm. altis minute pubescentibus et hispidis; foliis acutis obscure bicoloribus ubique puberulentis et conspicue adpresse hispidis subtus cum costa et nervis simplicibus prominentibus, * foliis basalibus late lanceolatis 1-2 dm. longis 3-5 cm. latis basin versus in petiolum alatum brevem attenuatis, medio-caulinis oblongis vel late lanceolatis 4-5 em. longis 8-14 mm. latis per basem triangularem SOUTH AMERICAN BORAGINOIDEAE 19 sessilibus; spicis 1-3 cm. longis densifloris paniculatim dispositis; calycibus maturitate cylindricis late sessilibus dense breviterque pu- bescentibus .et cum setis paucis uncinatis munitis; corolla caerulea extus supra medium pubescenti, limba patenti 8-10 mm. lata, lobis ca. 4 mm. longis oblongis paullo attenuatis ad apicem abrupte obtusis vel obtuse rotundatis extus dense pubescentibus, tubo paullo gradatim ampliato ca. 8 mm. longo 2.5-3 mm. lato calycem valde superanti, faucibus cum locis 5 circularibus dense villosis sed exappendiculatis ; staminibus in apicem tubo affixis exsertis, filamentis 2 mm. longis, antheris oblongis ca. 1 mm. longis; nuculis maturis ignotis; stylo longo gracili 8-10 mm. longo. BRAZIL. Parand: campo, Curityba, 900 m. alt., Oct. 30, 1915, Dusén 17273 (G, TYPE). A very distinct species which is somewhat intermediate between M. ciliata and M. Lindenii but clearly most related to the former. It agrees with M. Lindenii in having merely bearded spots in the throat rather than ligulate bearded intruded appendages. With M. ciliata it agrees in having broadly sessile calyces armed with few uncinate hairs, in its long corolla-tube and style, and in its simply nerved leaves the cauline ones of which are strongly reduced up the stem. From its close relative it differs in having much less distinctly bi- colored leaves, both surfaces of which are conspicuously and uniformly appressed-hispid. Furthermore the hairs on the upper surface of the foliage are provided with evident discoid bases. The lower leaves, also, are broader, shortly stalked and widest at about the middle. The corolla differs in its larger size and in having an unappendaged throat, from which develop longer more protruding filaments. I have seen no authentic material of M. ciliata, var. hirsuta Fresen., Mart. Fl. Brasil. viii. pt. 1, 63 (1857), which is described as differing from typical M. ciliata in being densely strigose on both leaf-surfaces and coarsely hirsute on stems and peduncles. This variety was based upon a collection by Sellow cited as from ‘*Brasil.”’ Possibly it is based upon material referable to M. Dusenii. 3. M. ciliata (Cham.) DC. Perennial; stems several, erect, simple, 3-4 dm. tall, with a short pubescence obscured by a shaggy white coat of much longer hispid-villous hairs; leaves acute, minutely strigose and below sparsely hispid on midrib and veins, conspicuously ciliate on margins, noticeably bicolored, none of the pubescence with discoid bases, veins unbranched; lower leaves 1.5-3 dm. long, 2.5-3.5 cm. broad, oblanceolate, gradually contracted into a long winged stalk; cauline leaves lanceolate to linear with middle ones 4-6 cm. 20 JOHNSTON long and 8-12 mm. broad, sessile by a rounded base, strongly reduced up the stem; spikes 1-3 cm. long, very dense, paniculately disposed; mature calyx cylindrical, ca. 5 mm. long, 2 mm. thick, densely short pubescent, along the angles occasionally with a few stout uncinate hairs, broadly sessile, lobes blunt, triangular and ca. 1 mm. long; corolla blue; limb spreading, 4-5 mm. broad; lobes triangular, blunt, densely pubescent outside; tube 4-5 mm. long, weakly dilated upwards evidently surpassing the calyx; throat with 5 very villous ligulate intruded appendages 1—1.5 mm. long; stamens at about # height of tube; anthers narrowly oblong, ca. 0.8 mm. long and about equalling the filaments in length, exserted, reaching almost to the tip of the appendages; nutlet obliquely subovoid, 2.5-3 mm. long, pale, shiny, contracted to a stout stipitate base; style 5-7 mm. long, greatly sur- passing the corolla as well as fruit—DC. in Meisner, Gen. ii. 188 (1840) ; DC. Prodr. x. 56 (1846); Fresen. in Mart. FI. Brasil. viii. pt. 1, 63, t. 10, fig. 13 (1857); Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a. 121, fig. 47 l-m (1893); Johnston, Contr. Gray Herb. lxx. 11 (1924). Anchusa ciliata Cham. Linnaea iv. 440 (1829). BRAZIL. Rio Granpe po Su: open woods, Puerto Alegre, Reineck (G). INDEFINITE: without looality: Sellow 672 (US, IsoTYPE). 5. Echium [Tournef.] L. Calyx 5-parted with linear or lanceolate segments. Corolla tubular- funnelform, oblique and irregular; lobes imbricate, ascending, unequal; throat naked, broad and gaping; stamens inserted low in the tube, usually unequal and exserted; anthers ovate or oblong, on slender filaments. Style filiform, usually exserted, commonly bilobed; stigmas 2, small. Ovules 4. Nutlets 4, erect, rugose or verrucose, ovoid or oblong, with a broad basal areola. Gynobase flat or low- pyramidal.—Herbs or shrubs with alternate leaves. Racemes uni- lateral, bracteate. ayer violet to pink or white.—Sp. Pl. 139 (1753) and Gen. Pl. 68 (17. A large and complex Mike ‘of the Old World with two introduced species in South America. Key To Species. Basal leaves ovate or oe oblo — nerved; upper leaves narrowed f 4 goes, ain date pies hit base; corolla of PS Soule texture, ahidicees or but —S Wee ee ee oly oh ec es 2 aaaaecaaee po harlot eee sencindaonerg SOUTH AMERICAN BORAGINOIDEAE 21 1. Echium plantagineum L. Biennial, stems erect or ascending, one to several, 3-8 dm. tall, branched from base, appressed hispid- villous and sometimes sparsely hispid; leaves appressed-hispid, the lower ones 8-16 cm. long, oblong-ovate, petiolate, with strong midrib and evident veins, usually persistent at anthesis, the cauline ones strongly reduced up the stem, oblong, narrowed from a broad cordate semiamplexicaul base; racemes elongate, loosely disposed; corolla violet-blue, 2-3 cm. long, rather delicate in texture, glabrous or sparsely setose, tube shorter than the calyx; stamens with a pair con- spicuously exserted, a pair almost as long as the lower corolla-lobe and an odd one definitely included; nutlets ca. 2.5 mm. long.—Mant. Alt. 202 (1771); St. Hil. Pl. Usuel. Brasil. t. 25 (1825); Fresen. in Mart. FI. Brasil. viii. pt. 1, 61 (1857); Arech. Anal. Mus. Nac. Mon- tevideo ser. 1, vii (FI. Urug. iv.) 182 (1911); Johnston, Contr. Gray Herb. Ixx. 10 (1924). E. violaceum of authors; Lacaita, Jour. Linn. Soc. xliv. 421 (1919). E. bonariense Poir. Encye. viii. 674 (1808); Lehm. Asperif. ii. 462 (1818). E. violaceum, var. medium Kuntze, Rev. Gen. iii. pt. 2, 204 (1898). ARGENTINA. Buenos Aires: Barracas al Sur, Oct. 1902, Venturi 73 (G); Sierra de Ventana, Hauthal 57 (NY, TYPE of E. violaceum, var. medium); Buenos Aires, Parker (G); Azul, Nov. 1887, Osten (G). Entrerios: Con- cepeion del Uruguay, Nov. 1875, Lorentz (G). URUGUAY. Monreviveo: Pocitos Malvin, Feb. 1924, Herter 80a (G); Montevideo, 1838, [sabelle (NY). Widely introduced over temperate South America west of the Andes. It has been reported from southern Brazil, Uruguay and central and northern Argentina. The species was introduced into the Argentine apparently at a very early date, since Poiret described his E. bonariense in 1808 basing it upon material collected at Buenos Aires by Commer- son in 1767. 2. E. vulgare L. Biennial; stems erect, solitary or occasionally several, 3-9 dm. tall, finely hispid-villous as well as shaggy coarse- hirsute; leaves appressed-hispid and along the margin and midrib somewhat hirsute, with a strong midrib but very obscure or absent veins; lower leaves 8-16 cm. long, oblanceolate, broadly stalked, forming a rosette which withers away at anthesis; cauline leaves re- duced up the stem, the middle ones linear-lanceolate, 3-9 cm. long, contracted to a rounded sessile base; racemes short, lateral, disposed in a long narrow thyrse or open panicle; corolla bright blue, rarely rose or white, pubescent and sparsely setose, rather firm in texture, 10-15 mm. long, tube about equalling the calyx; stamens very un- equal, the two pairs slightly unequal but both surpassing the lower 22 JOHNSTON corolla-lobe, the odd stamen included; nutlets ca. 2 mm. long.—Sp. Pl. 139 (1753); Reiche, Anal. Univ. Chile exxi. 835 (1908) and Fi. Chile v. 240 (1910); Johnston, Contr. Gray Herb. Ixx. 10 (1924). CHILE. Vatoprvia: Antilhue, 1916, Baeza (IP). Ma uxco: Collipulli, 1915, Baeza (IP). Concercion: Cabrero, 1913, Baeza (IP). UBLE: Alico, 1878, Puga (MS). Santraco: Mapocho, 1909, Rudolph (IP). This European weed is apparently widely dispersed over south- central Chile. It is to be expected from the Argentine though I have seen no specimens from there. Various authors have reported the species in that country but from their discussions it seems probable that the records are in large part, at least, based upon misdeter- minations of EF. plantagineum. Il. Trrpe ANCHUSEAE. Nutlets erect or oblique or horizontal, straight or curved, smooth or rugose; areola basal or lateral, along the side or near tip of coty- ledons, surrounded by a tumid annular rim, strophiolate, more or less plug-shaped or rarely merely high-convex; gynobase usually convex, commonly with broad cavities left by the strophiolate plugs; style lobed or simple; stigmas 1 or 2, distinct or almostly completely fused; corolla blue or rarely white or yellow— Johnston, Contr. Gray. Herb. Ixxili. 52 (1924); Gusuleac, Pub. Soc. Nat. RomAnia no. 6, 1-16 (1923). This tribe is native of the Old World and is represented in America only by introduced species. Key To GENERA. Stamens unappendaged, erect, distinct; corolla-tube well de- WON ee a: We ge eo ery ee ee 6. Anchusa. Stamens organ dorsally; anthers connivent; corolla-tube 2 Ver BROFb SS. oe ee ry se ea, ea eu ae 7. Borago. 6. Anchusa L. Calyx 5-toothed or -parted. Corolla regular or slightly zygo- morphic; tube straight or bent, cylindrical; lobes spreading, imbricate; throat with oblong or deltoid usually hairy intruded appendages; stamens affixed at middle of tube, included. Style filiform. Stigma solitary and emarginate or geminate. Ovules 4. Nutlets 4, erect, oblique or incurved, usually rugose; areola basal or suprabasal, sur- ded by a thick annular rim, prolonged down as a strophiolate plug. Gynobase flat or low-convex.—Annual or perennial herbs, hirsute to villous. Leaves alternate. Corolla bluish purple to white or rarely pink. Inflorescence bracteate—Sp. Pl. 133 (1753) and Gen. SOUTH AMERICAN BORAGINOIDEAE 23 Pl. 64 (1754). Lycopsis L. Sp. Pl. 138 (1753) and Gen. Pl. 68 (1754). A large genus the distribution of which centers in the Mediterranean region. Key To SpEcIEs. Corolla regular, with straight tube and it ee limb...1. A. officinalis. Corolla irregular, with bent tube and oblique limb......... 2. A. arvensis. 1. Anchusa officinalis L. Biennial; stems erect, 3-6 dm. tall, simple or ascendingly long-branched, hispid; leaves scabrous, short- hispid, with straight or obscurely undulate margins; lower leaves lanceolate, acute, attenuate into a long winged petiole, 1-2 dm. long; middle leaves oblong to oblong-lanceolate, tapering from a broad rounded or cordate base; racemes solitary or geminate, 5-20 cm. long, dense; mature calyx globose, appressed-hispid, with short narrowly triangular lobes; pedicels 1-2 mm. long; corolla regular, purple, sal- verform, the limb concave, not oblique, 5-8 mm. broad, tube 3-6 mm. long, straight, surpassing the calyx; nutlets very strongly incurved, depressed, rugose, 3-4 mm. long.—Sp. Pl. 133 (1753); Johnston, Contr. Gray Herb. Ixx. 9 (1924). ARGENTINA. Buenos Arres: Nijfiez, Feb. 1919, Molfino (G). This European species appears to have been introduced in South America only in Argentina. Berg, Anal. Soc. Cient. Argent. iii. 198 (1877), miata its presence in the vicinity of Buenos Aires fifty years a 2. c arvensis (L.) Bieb. Annual; stems coarse, shaggy coarse- hirsute, ascendingly long-branched from the base, 1-6 dm. tall; leaves hirsute, pustulate, margins undulate or late; lower stem- leaves oblanceolate, 5-15 cm. long, gontencted into a winged petiole; middle and upper cauline leaves broader, oblong or oblong-lanceolate, with a broad rounded clasping base; racemes solitary or geminate, 5-15 em. long; mature calyx cut nearly to base into hirsute lanceolate lobes; pedicels 2-4 mm. long; corolla zygomorphic, salverform, blue, tube longer than the calyx, 4-5 mm. long, bent above the middle, limb concave, oblique, 4-6 mm. broad; nutlets strongly incurved, rugose, 3-4 mm. long.—Fl. Taur.-Cauc. i. 123 (1808) ; Gusuleac, Pub. Soc. Nat. Romania no. 6, 13 (1923). Lycopsis a . Sp. Pl. 139 (1753); Clos in Gay, Fl. Chile iv. 477 (1849); Johnston, Contr. Gray Herb. Ixx. 8 (1924). I have seen no material of this Old World species from South America. Although to be expected there I have found only one def- 24 JOHNSTON inite record attributing the species to South America, this made long ago by Hooker & Arnott, Bot. Beechey Voy. 37 (1830), who reported it as introduced at Valparaiso, Chile. 7. Borago [Tournef.] L. Calyx 5-parted. Corolla rotate or very broadly campanulate; tube . very short; lobes imbricate, acute; throat appendaged; stamens in- serted on the tube, exserted; filaments broad, prolonged as an ap- pendage beyond the insertion of the linear connivent anthers. Style filiform. Stigma emarginate. Ovules 4. Nutlets 4, ovoid or oblong, erect; areola with conspicuous annular rim surrounding the strophi- olate plug. Gynobase flat or weakly convex.—Annual or perennial hirsute herbs with alternate leaves—Sp. Pl. 137 (1753) and Gen. Pl. 67 (1754). A small very natural genus of the Mediterranean region. 1. Borago officinalis L. Annual; stems usually coarse, erect or ascending with spreading branches, 3-6 dm. tall, coarsely shaggy- hirsute; leaves coarse, veiny, sparsely appressed-hispid and setose- hirsute, scabrous; lower leaves obovate or oblong, contracted into a long coarse winged petiole, blade 1-1.5 dm. long; middle leaves nar- rower, with the shorter winged petioles eared at base; upper leaves sessile; flowers in loose racemes usually corymbosely disposed; mature calyx hirsute, 1-2 cm. long, cut into erect or ascending lanceolate lobes; pedicels very long, hirsute, 2-4 cm. long; corolla clear blue, rotate, 1.8-2.4 cm. broad, lobes ovate and acute or acuminate; stamens dark and very conspicuous; nutlets 5-6 mm. long, on sides finely ribbed longitudinally, tuberculate at summit.—Sp. Pl. 137 (1753). ECUADOR. Phe Rose 22102 (G). BOLIVIA. i eens ie — 129 (G). CHILE. Coe Gigoux A medicinal plant, ae ‘evan in herb-gardens in Latin America, and occasionally appearing as an escape. iil. TRIBE ERITRICHIEAE. Nutlets erect or suberect or rarely horizontal, straight or slightly bent, smooth or roughened or appendaged, margined or immarginate; areola lateral or basal, at side of cotyledons or near their tips; gynobase pyramidal or columnar or rarely flat; style entire; stigma 1, entire or emarginate, capitate or disciform; corolla blue or white or less com- prin yellow or orange.—Johnston, Contr. Gray Herb. Ixxiii. 57 4 SOUTH AMERICAN BORAGINOIDEAE 25 This large tribe centers in Asia and western America. Brand, Fedde Repert. xxi. 249 (1925), has recently. substituted the name Cryptantheae for the well established tribal name, Fritrichieae, arguing that the latter is inappropriate since the genus Eritrichium is now considered a small genus and since it is not characteristic of the tribe. I am maintaining the old tribal name as it has had wide use since it appeared in Bentham & Hooker’s Genera, ii. 836, in 1876, and because Brand’s arguments if applied consistently to Borago and the family and subfamily names would destroy several of the major nomenclatorial landmarks in the family. Key To GENERA, Corolla-lobes coe ae oa = bud; nutlets smooth, basally affix gtiome discon oss ee ae ore rs 8. Myosotis. Corolladobes imbricate in ee bud. Nutlets — . definite medial ventral groove formed by the of the. pericarpial walls. - Corcila-tabs at base within naked or with small Ae Seti mature calyx ‘*t to beyond the pit and commonly to the vent Se bracts present or absent, peas ‘distinet TAURO MAIS oc es ee oe ree Lt Cryptantha. Corolla-tube don middle —— ae with 5 pairs oF elongate vertical knife- : re ni calyx — decidedly —, merely gweried Bre ts decur- rent on “ek Sealine with the lower re al calyx 10. Nesocaryum. Nutlets with ae } este ial walls fused at yo above the middle and iinaky 40 rming a medial ventral keel. << al madela oo ti Be by a flange or oon of append- Corolla w white, appendaged in the throat; eg cnireay ah ot eins Vt ee eee ee Plagiobothrys. Colin os orange or yellow, unappendaged in the gitat: eg ey ee es eo ore ey eee 12. Amsinckia. Back of i hore enuircled by a toothed or lacerate rim or nge or by a row of subulate glochidiate appendages. Nutlets equalling the subulate obase, attached for nearly their whole length along the crest of the on keel, lacking a definite areola; style usually surpassing the nutlets; pocemes bracted...... 13. Lappula. Nutlets twice surpassing the sto eae us a re oe Se age attached obliquely sannmeaiiite ovate areo! oe an ght towers recurved in fruit; leaves herba- 15. Selkirkia. Sa ee ee asses Goes any amare 1h, ak Se i a A i a ae a ew 8. Myosotis [Dill.] L. Calyx cut to beyond the middle into lanceolate or triangular lobes. Corolla with short tube; lobes contorted, rounded, spreading; throat 26 JOHNSTON with intruded appendages; stamens affixed on tube, included or ex- serted, the filaments filiform; anthers oblong or ovate, obtuse. Ovules 4. Nutlets 4, erect, ellipsoid, smooth and shiny, with a basal areola; gynobase flat or high convex. Style filiform. Stigma disciform.— Annual or perennial herbs. Leaves alternate. Racemes bractless or bracted. Corolla blue, white or rarely rose.—Sp. Pl. 131 (1753) and Gen. PI. 63 (1754). A large genus centering in the Old World. It is separated from practically all other genera of the subfamily by the contorted rather than imbricate aestivation of the corolla. Key to Sprciss. ey i leafy-bracted throughout; sv indigenous extreme southern Ameri Plant. pe and = sparsely strigose; onlin inks 3-4 s of mature neg much longer than ~ tube; aestiees Fe ie Soe eek ens ss 1. M. albiflora. Plant appressed short-bispid; poecila tinh ca. 2.5 mm. broad; lobes of ma Dice calyx much shorter than tube; nutlets RE re oe a . M. antarctica. Toliccuasencs Ea braces or occasionally bracted at the Corolla white, 1-1.5 mm. broad; calyx clightly rola i a BOR er eae asks . M. virginica. as, les of of wel ined so Oe er 4. M. azorica. face; racemes usually aaucely bractless. ...5. M. scorpioides. Corolla cage 3-6 mm. broad; style commonly sur- passed by th usually 1. Myosotis albiflora Banks & Solander. Perennial from a short rhizome; stems several, slender, prostrate or loosely ascending, short- strigose, leafy; leaves light green, thin, sparsely short-strigose; lower leaves 1-5 em. long, 3-8(-10) mm. broad, broadly oblanceolate, peti- olate; upper leaves obovate, reduced and continuing as bracts through the inflorescence; flowers in leafy false racemes, usually extra-axillary and opposite or beside or between the bracts; calyx sparsely strigose, lacking uncinate pubescence, at maturity cleft into lanceolate ascend- ing lobes; pedicels slender, usually not longer than the calyx; corolla white with yellow appendages, tube cylindrical and about equalling the calyx, limb widely ca pa 3-4 mm. broad; nutlets grayish, “Lie ae. long, back low-convex, face strongly angled; style a SOUTH AMERICAN BORAGINOIDEAE 27 little shorter than the nutlets.—Hook. f. Fl. Antarct. ii. 329 (1846) ; Clos in Gay, FI. Chile iv. 461 (1849); Reiche, Anal. Univ. Chile exxi. 834 (1908) and FI. Chile v. 239 (1910); Skottsb. Svenska Vet. Akad. Hand. lvi. pt. 5, 290, t. 23, fig. 8a, ¢ (1916); Johnston, Contr. Gray Herb. Ixx. 41 (1924). FEritrichium albiflorum Griseb. Abh. Gess. Wiss. Goett. vi. 131 (1854); Ph. Anal. Univ. Chile xe. 515 (1895). Allocarya albiflora Greene, Erythea iii. 57 (1895). CHILE. Mecatuanes: prostrate on beach at Puerto Americano, Chonos Island, lat. 45° S., Jan. 1857, Fonck 118 (MS); ete Bay, Isla Davin. a 52° 15/ s., Jan. 9, 1879, collector not given (MS); Porvenir, Fuegia, lat. 30’ s. Dusén 267 (G, Ms); Rio Azopardo, Fuegia, lat. 54° 27’ S., March 1606, Dusén 593 (G, Ms); forest region, Navarin Island, lat.. 55° S., Jan. 1920, Gusinde 258 (G). The type material of this species came from Bahia Good Luck in the western part of the Straits, lat. 53° 28’ S., and from southern Fuegia. Skottsberg, |. ¢., reports the species from Puerto Ramirez in Canal Smythe, lat. 52° 20’ S., Caleta Cutter in Canal Jeronimo, lat. 53° 20’ S., and Puerto Fortuno on Isla Londonderry, lat. 54° 53’ S., allin Megallanes. He also reports it from Estancia Mayer, lat. 51° 17’ S., which is on the Rio Coile in Santa Cruz in extreme southern Ar- gentina. The species was very poorly understood and much confused until Skottsberg published his notes concerning it. Spegazzini ap- parently confused it with M. virginica, Dusén did not distinguish it from M. antarctica and Grisebach and authors following him mistook it for Plagiobothrys calandrinioides. 2. M. antarctica Hook. f. Annual or perennial; stems several to many, prostrate or decumbent, stiffish, appressed short-hispid, leafy, internodes short; leaves dark green, appressed short-hispid, obtuse or rounded at apex; basal leaves broadly spathulate, 1-2 cm. long, 3-6 mm. broad; upper leaves reduced, oblong-obovate, broadly affixed, continuing as bracts through the inflorescence; flowers in leafy false racemes, usually extra-axillary and beside or between or opposite the bracts; calyx loosely appressed short-hispid, at maturity with erect lanceolate lobes much shorter than the ribbed campanulate tube; pedicels short; corolla white or bluish with yellowish appendages, tube about equalling the calyx or surpassing it, limb spreading, ca. 2.5 mm. broad; nutlets small, 1-1.3 mm. long, dark, back convex, face obtuse; style surpassing the nutlets.—FIl. Antarct. i. 57, t. 38 (1844) and Handb. N. Zealand FI. 193 (1864); Skottsb. Svenska Vet. Akad. Hand. lvi. pt. 5, 291, t. 23, fig. b, d (1916); Johnston, Contr. Gray Herb. Ixx. 41 (1924). M. antarctica, f. gracilior Skottsb, |. ¢ I have seen no American material of this species, which has been 28 JOHNSTON well known from New Zealand, Steward Island and Campbell Island in the eastern hemisphere. Skottsberg, |. c., however, who first gave a clear account of its American occurrence, collected it at Puerto Altamirano in Seno Skyring, lat. 52° 33’ S., and mentions material collected at Punta Arenas by Dusén (no. 173) and Lechler (no. 1132, in pt.) and in “ Magellans Land” in 1852 by Andersson. Dusén’s collection, the type of the forma gracilior, has been figured by Dusén, Svenska Exped. Magell. iii. no. 5, 134 (1900) as M. albiflora. The species is known from America only from the Chilean mainland just west and north of the Straits of Magellan. 3. M. virginica (L.) B.S. P. Annual or biennial, branched from the base, erect, 0.5—4 dm. tall, hispid-villous with the hairs spreading or appressed; leaves appressed hispid-villous, obtuse, the lower ones oblanceolate 1-4(—5) em. long and 3-9(-12) mm. broad, the upper ones linear or oblong and little reduced; calyx strigose and short- hispid with spreading frequently uncinate hairs, slightly but definitely irregular; calyx-lobes lanceolate, erect or connivent, longer than the tube, the lower one the longest; pedicels slender, at least equalling the length of the calyx, usually erect and appressed to the stem; corolla inconspicuous, white but commonly drying brown, subtubular or salverform with a funnelform throat and ascending lobes, limb 1-2 (-2.5) mm. broad; nutlets grayish, back convex, face obtusely angled above middle and convex below; style very short, much shorter than the nutlets—Prel. Cat. N. York 37 (1888); Johnston, Contr. Gray Herb. Ixx. 43 (1924) and 1. c. Ixxv. 42 (1925). Lycopsis virginica L. Sp. Pl. 139 (1753). M. verna Nutt. Gen. ii. add. (1818). M. versi- , var. americana A. DC. Prodr. x. 109 (1846). Eritrichium pam- peanum Speg. Contr. Fl. Sierra Vent. 44 (1896). ARGENTINA. Buenos ArrEs: common plone Streamlets in Sierra Ventana, Nov. 14, 1895, Spegazzini (G, “coryrus” of E. pampeanum). Cuasvr: hills near Nafofo-c bie “Marsh 12, 1935, Spepaisiins (G). This species, which ranges over the eastern half and the northwest- ern parts of the United States, is apparently indigenous in northern Patagonia and the Province of Buenos Aires. It appears to be, at least in part, the plant which Spegazzini, Pl. Patag. Aust. 551 (1897), as reported as Eritrichium albiflorum. In fact the specimen from Chabut cited above was sent me by Spegazzini as representing an un- published variety of that species. 4. M. azorica H. C. Wats. Plant annual or possibly perennial; stems several, rather coarse, erect or decumbent, simple or with numerous long ascending branches, conspicuously and abundantly SOUTH AMERICAN BORAGINOIDEAE 29 hispid-villous with the hairs sordid or white and decidedly retrorse at least towards base of plant; leaves sparsely appressed villous- strigose; lower leaves broadly oblanceolate or obovate-oblanceolate, 3-6(-10) em. long, 10-20(-25) mm. broad, petiolate; upper leaves gradually reduced, oblong-elliptic or ovate-lanceolate; racemes solitary or geminate, bractless or leafy-bracted towards base; calyx densely strigose and with spreading uncinate hairs, slightly but definitely irregular; calyx-lobes lanceolate, erect, longer than the tube, the lower one the longest; pedicels slender, longer than the calyx; corolla blue, 5-7 mm. broad, limb flat; nutlets shiny black, back convex, face strongly angled and keeled; style reaching to the tip of the nutlets or just beyond.—Bot. Mag. Ixx. t. 4122 (1844) COLOMBIA: cultivated in Bogotd, 1876, Bayon (US). BOLIVIA: hillsides, escaped, canyon of La Paz River, 1920, Shepard 174 Comte ies cultivated, Copiapé6, 1886, Gigoux AL cultivated, Santiago, 1886, Gigoux (G); Valdivia, 1896, Buchtien (US). This is the cultivated “no-me-olvides” of South American gardens. For a discussion of the status of the name, MW. azorica, see note by Johnston, Contr. Gray Herb. Ixx. 42 (1924). Although the name is unquestionably applicable to the plant treated here, there are several older but still obscure names that are probably synonymous and perhaps some one of them may be sufficiently verified to displace it. M. scorpioides L. Perennial; stems ascending from a trailing or creeping rhizomatous base, rooting at the nodes, 2-8 dm. long, sub- simple or loosely branched above, usually obscurely angulate, rather sparsely short-strigose; leaves sparsely strigose, the lower ones o lanceolate or spathulate, 4-8 cm. long, 8-12(-20) mm. broad, the upper ones gradually reduced and becoming linear-lanceolate to linear oblong; racemes geminate or solitary, almost always bractless; calyx sparsely strigose, lacking uncinate hairs, usually with deltoid lobes shorter than the tube; pedicels slender, longer than the mature calyx; corolla blue with yellow appendages, the limb flat, 5-9 mm. broad; nutlets convex dorsally, somewhat obtuse ventrally; style commonly much surpassing the nutlets, usually reaching to level of the calyx- sinuses.—Sp. Pl. 131 (1753); Rendle & Britt. Jour. Bot. xlv. 440 (1907); Johnston, Contr. Gray Herb. lxx. 41 (1924). M. scorpioides, var. palustris L. |. c. M. palustris Lam. FI. Fr. ii. 283 (1778). CHILE. Cxttor: wet grassy ditch behind beach, Ancud, Feb. 1925, Pennell 12523 (G). A native of Europe, widely introduced in wet places in temperate North America. Now first reported from South America. 30 JOHNSTON 6. M. laxa Lehm. Annual or soboliferous perennial; roots fas- ciculate and fibrous; stems slender, erect or decumbent, solitary, loosely branched commonly from the base, terete, sparsely strigose; leaves sparsely strigose, the lower ones oblanceolate or spathulate, 3-6(-8) cm. long, 6-10(-12) mm. wide, the upper ones gradually reduced, narrowly linear-lanceolate or linear-oblong; racemes geminate or solitary, with the principal one leafy-bracted just above the base; calyx sparsely strigose, lacking uncinate hairs, usually with narrowly ovate-triangular or almost cuneate lobes about equalling the tube in length; pedicels slender, longer than the mature calyx; corolla light- blue with yellow appendages, limb slightly concave, 3-6 mm. broad; style usually shorter than the nutlets and reaching much below the level of the calyx-sinuses.—Asperif. i. 83 (1818); Johnston, Contr. Gray Herb. Ixx. 41 (1924) where further synonymy. ARGENTINA. Buenos Arrss: vicinity of Buenos Aires, 1913, Rodri- quez 167 (G). CHILE. Vatoprvia: Valdivia, 1899, Buchtien (US). A native of Europe and North America, but probably introduced in South America. Dovustrut or ExciupEep SPECIES. M. Fottosa Lehm. Asperif. i. 99 (1818).—‘ America meridionalis” —Probably a species of Lithospermum and perhaps the same as the next. M. grandiflora HBK. Noy. Gen. et Sp. iii. 90, t. 199 (1818).—Al- though given as from Ecuador this appears to be LiraosPERMUM DISTICHUM Ort. and to have come, in fact, from Mexico, ef. Johnston, Contr. Gray Herb. Ixxv. 40 (19235). 9. Cryptantha Lehm. Calyx usually cut to the base, the lobes erect or connivent, linear or oblong. Corolla with usually a short cylindrical tube, with or without scales at the base inside; lobes imbricate, rounded, spreading; throat with intruded appendages. Style slender, short or long, included; stigma capitate. Ovules2—4. Nutlets 1-4, erect, ovate to triangular, roughened or smooth, margined or marginless, affixed laterally through a medial ventral and commonly basally forked groove to a usually columnar subulate or pyramidal gynobase.—Annual or perennial herbaceous or fruticulous plants, usually with coarse stiff pubescence. Leaves opposite at base, firm, veinless. Flowers white or very rarely yellow, in bractless or bracted spikes or racemes. Mature calyx in- SOUTH AMERICAN BORAGINOIDEAE 31 vesting the nutlets and falling away as a bur or persistent with the nutlets falling away individually—Del. Sem. Hort. Hamburg iv. Sem. Hort. Petrop. vii. 52 (1841). Piptocalyx Torr. in Wats. Bot. King Exped. 240 (1871). Oreocarya Greene, Pittonia i. 57 (1887). Eremocarya Greene, Pittonfa i. 58 (1887). Greeneocharis Giirke & Harms in E. & P. Nat. Pflanzenf., Gesamtreg. 462 (1899). Wheelerella Grant, Bull. So. Calif. Acad. v. 28 (1906). Johnstonella Brand in Fedde, Repert. xxi. 249 (1925). One of the largest and most perplexing genera of the Boraginaceae. It is exclusively American. About two-thirds of the approximately 150 species occur in western United States. In South America the genus is most abundant in Chile with outlying species in Peru, Bolivia and Argentina. As here defined, the genus includes the North American group of coarse perennials current as Oreocarya. My study of the South American species of Cryptantha has convinced me that the recognition of Oreocarya has been most arbitrary. That group appears to consist merely of the more or less coarse persistent members of Cryptantha § Krynitzkia. I am convinced that the characters of the sections of Cryptantha defined below are much more fundamental than are such “generic” characters as coarseness of habit and persistence of root which must be stressed in justifying the continued maintenance of Oreocarya. I do not believe that Oreocarya should be recognized unless the primitive, coarse, persistent species and the more evolved, slender, annual ones in the three sections of Cryptantha each be given generic recognition. Personally I do not approve of, and certainly will not be responsible for the segregation of 5 weak genera from the currently accepted Cryptantha simply to give logical justification for the maintenance of any mere habit-genus as indefinite as Oreocarya. Key To SEcTIONsS. Plants not producing cleistogamic flowers; North and South Nos oa og és een ee i 3 Krynitzkia (spp. 1-22). Plants producing cleistogamic flowers; South Ameri Clei a inwtes simple, borne in the middle od upper or frequently i in all the leaf-axils, occasionally replacing the hasm: ¢ flowers in the inflorescence siege: re eigen (spp. 23-30). Cleistogamic ers highly specialized, appearing ent ovoid ~ F eotenie structures (cleistogenes ao kes borne at the base of the plant below or just above the surfac ee aos a es II. enn (spp. 31-40). I. Section Krynitzx1a.—This is the largest and most difficult - section of Cryptantha and is distinguished from the other two sections 32 JOHNSTON of the genus by the complete lack of cleistogamic flowers. It reaches its maximum development in North America, where it is represented by about 100 species. There can be little doubt that in the section are found the most primitive members of the genus. One species, C. albida, occurs in both North and South America. The only other approach to this condition is the case of °C. parviflora. This species, while evidently separable, is very closely related to the North American C. Grayi and C. angustifolia. The species assembled in the series Barbigerae are as a group clearly related to those North American species I have collected under that heading, cf. Contr. Gray Herb. Ixxiv. 60-69 (1925). Iti is interesting to note that whereas the short style tes among the North American species of the section, the long style is somewhat more common among the southern members. In North America there are numerous species with smooth polished nutlets; in South America there are only four, all closely related and quite different from the northern forms. Cryp- tantha gnaphalioides is a peculiar shrubby perennial, which would be refered to the genus Oreocarya if it grew in North America. In other than its habit, however, it is much like C. dichita, C. hispida and C. phaceloides, species which in some regards are reminiscent of the Californian C. mohavensis. Key To Sprcres. baer pectocsty smooth and shiny, large, 2.8~-3.3 mm. long. e Plant perennial, shrubby a a. Gnaphalioides....1. C. gnaphalioides. Plant an annual herb. Pha celoid Corolla 7-10 mm. broad; leaves 713 Wii. Dros es Les 2. C. dichita. Corolla it Sg . broad; "leaves 2-3 mm. broad. Nutlets2..... SU ee ates eee a eee ee 3. C. hispida. Nutlets . Pie as oy ester ae geil 8 Bal 4. C. phaceloides. Nutlets more or less roughened and dull. Nutlets thick with deep plane aaying not ieee only im- sae on one another close e gynobase and hence y open prep nent seni anterior face occu- ved by a large triangular (apparently) excavated areola; gynobase narrowly pyramidal; northern Argentina. went Pere. Vee pl owe were eee ee Ss 5. C. albida. strongly compressed with ne i nace crowded and closely wd ay secoapt not separated by open sinuses, an ace with a very small shallow areola; gynobase us Edges of nutlets , with a definite thin margin; abaxial — more or emphasized, usually largest and a Angunifoline Nutlet oc wrarggraeendbas 5 mm. long; spikes bracted Peas ao, CVG OL Suk een ed eae Soc es 6. C. diplotricha, uw morp. 0.6-1 mm. long; es bract- less 0} seep gg eg ie SOUTH AMERICAN BORAGINOIDEAE 33 aioe of nutlets merely i. or rounded, not sharp, ostly marginless but more or less thickly margined in a few F pasion iti al nutlet more or less a eet ally the lar sat or most persistent. Barbigerae Corolla evident or conspicuous, 2-7 mm. bro Spikes phterore rsistent, a perennial or age dia annual; ud 5 been oming prostrate or trailin Plant loosely branched, forming a = en ose broad at, a ence canescent; ‘Cru 8. MH DRONE. eo a eae se a meee few — branched, forming a dense ¢ ence silvery; co rolla 2-4 mm. hesed: 9. C. argentea. dogg a a short-lived annual; free short, erect o Nutlets Ceaencecbons obscurely roughened 10. C. chaetocalyz. — sera a tuberculate Binto sot isi. i i at 11. C. filiformis. olla 2-4 mm eg nutlets oe or hetero morphous, 1 .3-1.8 mm. lon Style much surpassing the 4 portions Peru....12. C. granulosa. Style shortly surpassing the 2-4 nutlets; Chile. 13. C. taltalensis. Corolla 5-6 mm. broad; nutlets homomorphous, ca. fe ee eee Te 14. C. Romani. Corolla i pe Bene ce te s or minute, 0.5-1.5(-2) mm. broad. Spikes bracted; style noticeably surpassing the nutlets. Prdtue ¢ long, with narrow erect or slightly patil a Lieine: Chile and gece re . C, filaginea. Fruiting calyx 5-6 mm. long, with broad decay in 8-0) fo. 2 aU peemyacen eaperanal eles met . C. limensis. ae tater fd bes equalling the nutlet or just es Nutlets obscurely ¢ a eige or almost smooth. Fruiting calyx 8- m. long; nutlets lanceolate, 2.8—3.2 mm. ray rien coarse shaggy-hirsute; Andes of northern Chile................ 17. C. calycina. <=". calyx 2-3 mm. long; natiets lance-ovate, mm. long; 2; ian a strigose or appres- is BISEOWA 20S 18. C. patagonica. Nutlets 5 fre i acute Gs wrinkled. Back of nutlets oon nla, coarsely if at all transversely Fruiting salve 25 531 mm. long, lobes lance-linear or narrowly lanceolate; Peru......... . C. peruviana. Fruiting calyx 3-4 mm. long, lobes linear ce-linear; Chile and y ener lla "20, “C. globulifera. ith transv lin grooves ny long; plant stiff; Chile and Laie ela m2). C. diffusa. Spikes green; fruiting calyx 2-2.5 mm. lo ong; plant weak; Bolivia Peerage ts lew ces 22. C. debilis. 1. Cryptantha gnaphalioides (A. DC.) Reiche. Perennial, 2-4 dm. tall, becoming much branched and shrubby, commonly decidedly 34 JOHNSTON woody towards the base; stems erect, closely white-strigose and fre- quently sparsely hirsute also, cottony-strigose towards base, older parts with exfoliating bark; leaves linear to lance-linear, broadly at- tached, 1—4(—6) em. long, 1.5—2.5(-3) mm. broad, obtusish or acutish, more or less strigose and frequently somewhat hirsute, usually canes- cent, spreading, abruptly reduced above, with leaf-buds or fasciculate leaves in the axils, bases more or less persistent; racemes unbracted, geminate, 1-4 cm. long, not loose, pedwnculate; fruiting calyx tardily deciduous, oblong-ovate, 4.5-6 mm. long; mature calyx-lobes oblong- linear, obtusish, densely strigose or abundantly short-hirsute; pedicels 1-2 mm. long; corolla white, 5-9 mm. broad; ovules 4; nutlets 1-4, apparently with the axillary one always developing, ovate-oblong, smooth and shiny, light colored, frequently somewhat mottled with brown, 2.8—3.3 mm. long, 1.2-1.5 mm. broad, apex acute and slightly outcurved, base rounded, edges acute particularly above the middle; back of nutlet flattish or low convex, occasionally above the middle somewhat obtuse with a rather definite medial ridge; inner face of nutlet right-angled with the groove closed throughout; gynobase about } height of nutlets; style usually very conspicuously surpassing the nutlets, more than twice the length of the gynobase.—Anal. Univ. Chile cxxi. 822 (1908) and Fl. Chile v. 227 (1910). Eritrichium gnaphalioides A. DC. Prodr. x. 131 (1846); Clos in Gay, FI. Chile iv. 466 (1849); Ph. Fl. Atac. 39 (1860) and Viage Des. Atac. 213 (1860); Ph. Cat. Pl. Itin. Tarapaca 56 (1891). E. fruticosum Ph. Linnaea xxix. 15 (1857). 1 , Geisse rt of > June 1841, Gay 1074 (MS); foe del Tabaco, Sierra San Miguel, 3000 m. alt., Nov ov. 1925, ease: on 48. 4892 ( (G); brada San Miguel above Los Marayes, , Johnston 4919 (G); Quebrada de Paipote, Jan. 1885, F. Philipps ( Ms); Potrerillos, 1920, Harding 22848 (US, NY); vicinity of Potrerillos m. alt., Oct. 1925, Johnston 3696 (G); Quebrada oe oe ue. Chiea, st 1886, Gigoux (G). InpEFINiTE: no locality given, A common and very characteristic undershrub of gravelly slopes and plains between 2000 and 3000 m. alt. in the arid mountains of the Province of Atacama. Frequently the dominant perennial over large areas. The greatest abundance of the plant noted was on the barren llano near the highest point on the railroad between Copiap6 SOUTH AMERICAN BORAGINOIDEAE 35 and Pueblo Hundido, where at about 2000 m. alt. the plants of this species were extremely common and apparently the only kind of vegetation over many square kilometers. The plant is well known to the natives frequenting the mountains as “'Té burro,” “ Té del campo” or “Té blanco.” An infusion of the leaves is used as a beverage. Generally, however, this Cryptantha is used in tea-making only when the more generally favored “Té colorado” or Vivania rosea is unob- tainable. 2. C. dichita (Ph.), comb. nov. Very coarse annual, 1-3 dm. tall; stems erect or ascending, one to several, subsimple or with a few short ascending branches, short-hispid and (especially above) coarsely pungent tawny-hirsute, 2-4 mm. thick; leaves broad and coarse, obtuse, hirsute and short-hispid, at least the under surface conspicu- ously pustulate, the lower leaves 7-11 mm. broad and 5-7 cm. long; middle and upper cauline leaves oblong or lanceolate, 3-5 cm. long, 8-13 mm. broad; spikes coarse, dense, 1-2 cm. long, bractless or with a leafy bract at base; fruiting calyx ovate-oblong, coarse, ca. 5 mm. long, short pedicellate; fruiting calyx-lobes linear or spathulate-linear, erect, hirsute and hispid, midrib not prominent; corolla large, 7-10 mm. broad, white with a yellow throat; fruit 4-ovulate; nutlets 2, oblong-lanceolate, 3-3.3 mm. long, smooth and shiny, apex acute, back flat or low convex, sides angled, groove very narrow; gynobase 2-% height of nutlets; style very much surpassing nutlets, about twice length of gynobase.—Eritrichium dichita Ph. in Villanueva, Anal. Univ. Chile lili. 444 (1878), nomen; Ph. Anal. Univ. Chile xc. 516 (1895); Reiche, Anal, Univ. Chile exxi. 831 (1908) and FI. Chile v. 236 (1910). CHILE. Aracama: “Desierto de Atacama,” 1877, Villanueva (MS, TrPE; G, photo.); Potrerillos, 1920, Harding 22849 (US); near Los Alamos, Quebrada de Potrerillos, 2400 m. alt., Johnston 3695 (G). The two collections with precise locality-data both come from about Potrerillos in the Department of Chafiaral. Since Villanueva, who collected the type, mentions visiting the Quebrada Salado and such other localities in the Potrerillos region as Pedernales Lake, Dojia Ines Chica, etc., it is not improbable that the type of the species actually came from that general region. According to him the plant was called “dichita” and was eaten by animals. _ 3. C@. hispida (Ph.) Reiche. Annual; 1-2 dm. tall, coarsely hirsute, stiffly erect with a few short strictly ascending floriferous ranches above; leaves linear, strictly ascending, 1-4 em. long, 2-3 mm. broad, coarsely hirsute; spikes dense, bractless, ca. 1 cm. long; 36 JOHNSTON fruiting calyx ovate-oblong, 4-5 mm. long, short-pedicellate; mature calyx-lobes linear, erect, commonly hirsute, midrib not very promi- nent; corolla 5-7 mm. broad, white; fruit 4-ovulate; nutlets 1-2, smooth and shiny, pale, oblong-lanceolate, 3-3.3 mm. long, ca. 1.4 mm. broad, apex acute, base rounded, edges angled, back flat or broadly obtuse with a weak but definite medial ridge, ventrally right- angled, groove closed to base and unforked; gynobase 4-32 height of nutlets; style very long, much surpassing the nutlets, about twice the length of gynobase.—Anal. Univ. Chile exxi. 831 (1908) and FI. Chile v. 236 (1910). Eritrichum hispidum F\. Atac. 38 (1860) and Viage Des. Atac. 44 and 212 (1860). CHILE. Aracama: Quebrada de Dojia Ines Chica, Dept. Chafiaral, Jan. 1886, Gigour (G). Antoragasta: Alto de Varas, Dept. Taltal, 3700 m. alt., 1854, Philippi (MS, Type; G, photo.). After C. phaceloides, from which species it is scarcely separable, the closest relative of this species is C. dichita. It differs from that species in its very much more slender habit and smaller flowers and leaves. Although all the material I have seen falls readily and defi- nitely into one or the other of these two species, I should not be sur- prised if future collecting gave reason for treating C. dichita and C. hispida as subordinate to C. phaceloides. 4. C. phaceloides (Clos) Reiche. Annual, 1.5-3 dm. tall, hirsute, stiffly erect, loosely branched; leaves linear, rather few, 1-2 cm. long, 2-3 mm. broad, ascending, hirsute and pustulate; spikes geminate or ternate, spreading, 6-12 mm. long, bractless, densely flowered; fruiting calyx ovate, ca. 4 mm. long, with a short coarse pedicel; mature calyx- lobes lanceolate to linear, erect, appressed hispid and hirsute along the weakly prominent midrib; corolla 4-5 mm. broad, white; fruit 4-ovulate; nutlets 4, slightly heteromorphous with the axial one a trifle the largest, brown and somewhat mottled, oblong-lanceolate, 3.5-3.8 mm. long, ca. 1.5 mm. broad, apex acute, base obtuse, edges acute, back low-convex, groove closed with a small areola at the acute basal forking; gynobase 3—3 height of nutlets; style shortly surpassing the odd nutlet—Anal. Univ. Chile exxi. 813 (1908) and Fl. Chile v. 218 (1910). Eritrichium phaceloides Clos in Gay, Fl. Chile iv. 468 (1849), not tab. 52 bis! Krynitzkia phacelioides F. & M. ace. F. Ph. Cat. Pl. Chile 211 (1881); Gray, Proc. Am. Acad. xx. 280 (1885). ae CHILE: no locality given, Gay (G, frag. of the TYPE). | I know this species only from the original description and from some ents (a leaf and a spike with mature fruit and flowers), SOUTH AMERICAN BORAGINOIDEAE 37 apparently from the type, sent me through the courtesy of Professor H. Lecomte. The material is without precise locality-data. When the species was published, however, it was said to grow “eh Copiapo.”’ The region immediately about Copiap6é and especially that just to the south and west has been quite well explored botanically, yet the species does not appear to have been rediscovered. It is not im- probable that Gay got the specimens on his journey through the cor- dilleras southeast of Copiap6. The species is extremely close to C. his- pida which grows in the cordilleras in the northern part of the Province of Atacama. The plant described above is not that illustrated by A comparison of the original description of C. phaceloides with the plant subsequently illustrated as that species is sufficient to demonstrate that two species are concerned. The plant illustrated as E. phaceloides is obviously a rember of the section Geocarya and is ap- parently that which I have described as C. Gayi. 5. C. albida (HBK.) Johnston. Annual, 1-3 dm. tall; stems erect or loosely ascending, solitary or several, subsimple or paniculately branched above, very leafy, finely ashy-hispid; pubescence usually appressed below and spreading above; leaves narrowly oblanceolate or spathulate to linear, obtuse, appressed-hispid and somewhat pustulate, reduced up the stem but extending through the inflorescence as linear bracts; lower and middle cauline leaves 2-7 cm. long, 1.5-4 mm. broad; spikes solitary, numerous, loosely bracted, 1-7 em. long, be- coming remotely flowered; fruiting calyx ovate, 2-3 mm. long, sub- sessile, tardily deciduous; mature calyx-lobes ovate-oblong to nar- rowly lanceolate, acute, loosely connivent, margins appressed-hispid, midrib somewhat prominent and short-hirsute; corolla white, incon- spicuous, ca. 2.5 mm. broad; fruit 4-ovulate; nutlets usually 4, homo- morphous with the abaxial one subpersistent, triangular-ovate, 1-1.5 mm. long, broadly tuberculate and usually somewhat granulate, rather thick, not crowded, touching one another only close to the gynobase and hence individually separated by an obvious sinus, apex acute, base truncate or obtuse, back convex, sides straight and deep, ventrally with a large open triangular areola which appears to be ex- cavated at maturity through the breaking away of the attachment- surface from the nutlet-walls; gynobase narrowly pyramidal, a little shorter than the nutlets; style much surpassing the nutlets.—Contr. Gray Herb. Ixviii. 53 (1923); Johnston, Contr. Gray Herb. Ixxiv. 42 (1925). Myosotis albida HBK. Nov. Gen. et Sp. iii. 91 (1818). Tithospermum ramosum Lehm. Asperif. ii. 328 (1818). Eritrichium ramosum A. DC. Prodr. x. 132 (1846). Krynitzkia ramosa Gray, Proce. Am. Acad, xx. 274 (1885). C. ramosa Greene, Pittonia i. 11 38 JOHNSTON (1887). E. hispidum Buckley, Proc. Acad. Philad. 1861, pg. 462 (1861). EE. falcatum Hieron. [Sertum Sanjuan.] Bol. Acad. Nac. Cient. Cordoba iv. 64 (1882). C. falcata Johnston, Contr. Gray Herb. Ixviii. 54 (1923). K. mexicana Brandg. Zoe v. 182 (1904). C. argentinica Brand in Fedde, Repert. xx. 318 (1924). I have been quite unable to separate the Argentine plant from the better known one of Mexico. The matter of leaf-breadth which Brand gives as separating them is hopelessly inadequate. The habit, flowers and the fruiting structures agree surprisingly and I doubt if there are characters that will separate the Mexican from the Argentine material. It seems that C. argentinica Brand and E. falcatum Hieron. are clearly synonymous. The first collection cited by Brand and appar- ently that which he considers the type of his C. argentinica is Hiero- nymus & Niederlein 16 from the Sierra Velasco in La Rioja. The type is in the Berlin herbarium: The specimen of this same collection in the herbarium of the University of Cordoba, Argentina, must be considered the type of FE. faleatum, for as Dr. C. C. Hosseus, in lit., has emphasized to me Hieronymus definitely stated in publishing his species that he had used the material from the Sierra Velasco in drawing up his description since the other collection (that made by Echegaray at Leoncito in San Juan), which he referred to the species, was imperfect. Brand used the name C. falcata, apparently for a different species, but for which one I do not know. 6. C. diplotricha (Ph.) Reiche. Annual, 1-1.5 dm. tall; stems several, erect, frequently somewhat decumbent at base, appressed- hispid and sparsely hirsute with short pallid hairs; leaves linear or spathulate-linear, 1-2 em. long, 1.5-3 mm. broad, appressed short- hispid, somewhat pustulate, reduced up the stem but extending ough the i ce as bracts; spikes solitary.or rarely geminate, bracted, 1-4 em. long, very numerous, congested ; fruiting calyx ovate, ca. 2.5 mm. long, very hispid, subsessile; mature calyx-lobes lanceolate, erect or loosely connivent; corolla white, inconspicuous, tubular, ca. 0.8 mm. broad; fruit 4-ovulate; nutlets usually 4, homomorphous with apparently a slight emphasis on the abaxial one, lucid, triangular, dark with light-colored tuberculations or wrinkles, 1-1.5 mm. long, readily deciduous, crowded and closely juxtaposed with no open sinus between the individual nutlets, apex acute and slightly recurving, base trun- SOUTH AMERICAN BORAGINOIDEAE 39 cate, sides sharply margined, back usually convex; groove open or closed above but below expanded broadly into a shallow open tri- angular areola; gynobase subulate, about equalling length of nutlets; style clearly surpassing the nutlets——Anal. Univ. Chile cxxi. 821 (1908) and FI. Chile v. 226 (1910). Eritrichium diplotrichum Pl. Cat. Pl. Itin. Tarapaca 57 (1891). E. diplotrichum, var. humilis Ph. 1. ec. C. diplotricha, var. humilis Reiche, |. c. E. axillare Ph.\.c. C. axillaris Reiche, |. c. C. modesta Brand in Fedde, Repert. xx. 48 es , . dip ‘ photo.); between Breas and Salar, Jan. 1885, F. Philippi (MS, Tyre of E. axillare; G, photo.). on Rioga: Tambillos, 1600 m. alt., Hosseus 1711 (G, part of TyPE of C. modes Most related to C. albida but clearly different, particularly as to fruit. Both Brand and Reiche have placed the species in the section Eucryptantha, accrediting it with cleistogamic flowers. I have searched for closed-flowers but have been quite unable to find any. The corol- las are small and tubular, however, and perhaps might be mistaken for cleistogamic ones if not examined under high magnification. C. parviflora (Ph.) Reiche. Slender annual, 3-20 cm. tall; stems commonly one, usually loosely branched from near base, very leafy, more or less abundantly appressed chalky-hispid and frequently also somewhat spreading short-hirsute; leaves linear, 5-40(-50) mm. long, 1-2 mm. broad, appressed-hispid or spreading short-hirsute, somewhat pustulate beneath, weakly reduced up the stem and in plants beginning to fruit somewhat obscuring the spikes; spikes solitary or geminate, 1—-5(-7) em. long, numerous, bractless except for a leafy bract at base; fruiting calyx ovate or oblong-ovate, 1.5-2 mm. long, sessile, readily deciduous; mature calyx-lobes lanceolate or lance- oblong, connivent, densely short-hirsute; corolla minute, subtubular, 0.3-0.5 mm. broad, white; fruit 4-ovulate; nutlets usually 4, hetero- morphous, ovate or triangular-ovate, dark with pale tuberculations or broad murications, apex acute, base rounded, back convex, edges sharply margined, groove usually closed above but below dilated into a shallow triangular areola; odd-nutlet abaxial, largest and most persistent, ca. 1 mm. long; consimilar nutlets 0.6—-0.8 mm. long, readily deciduous; gynobase ca. } height of odd-nutlet; style much surpassing odd nutlet, a little shorter than the gynobase.—Anal. Univ. Chile exxi. 821 (1908) and Fl. Chile v. 226 (1910). Eritrichum parviflorum Ph. Fi. Atac. 39 (1860) and Viage Des. Atac. 10 and 213 (1860). E. microphyllum Ph. Cat. Pl. Itin. Tarapaca 57 (1891). C. micro- 40 JOHNSTON phylla Reiche, |. c. 828 and 1. c. 233. C. piscoensis Brand in Fedde, Repert. xx. 49 (1924). C. umbelliformis Biand: l. «. 317. °C. Selers Brand, |. c. 317. CH ILE. yr alaesany AMA: Chajiarcillo, Sica 1876, a collector given (MS; photo.); Bandurrias, Geisse (MS); Desert of Ataca a [? Bandurrias], then 58B (NY); Atacama Desert, 1884, Bax Roman (MS; G, photo.); gravel, She north of Copia, 800 m. alt., Johnston 5026 rat Calders: Dee. 1853, Philippi bot TYP . parvi, orum; G, photo.); coastal dunes, Ca ldera, ERU. Moaqvecva: a southeast of Moquegua, 1500-1600 m. alt., Weberbauer 73984 (G, FM). Arequipa: open sandy and rocky slopes, Tingo, 2100-2300 m. alt., Pouiiel 13112 (G, FM); ope gravelly slope above Arequipa, 2500 m. alt., "Pennell 13177 (G, FM); Yura, 1884, Borm (G). e slopes above Pisco, 1200-1300 m. alt., Wsberbeis 5375 (G, ‘part "of TYPE piscoensis). Lima: open rocky slope, saat e, 800-1000 m. alt., Penne 14500 “G ); mountains near Chosica, 1500 m alt. , Weberbauer 5337 (FM ISOTYPE of C. umbelliformis). This species, while evidently distinct, is very closely related to the North American C. angustifolia (Torr.) Greene and C. Grayi (V. & R.) Macbr., cf. Contr. Gray Herb. Ixxiv. 31 and 38-40 (1925). All the material cited above has heteromorphous nutlets with the abaxial one evidently the largest and most persistent. In this fact as well as in general form and appearance of the nutlets, the present species agrees closely with C. i tt though in size of parts and general sees it is much nearer C. Grayi. I have been quite unable fe separate specifically the Chilean and Peruvian plants. In a very general way it can be said that the Peruvian specimens are somewhat more strictly branched and more leafy particularly above. These, however, are but vague differences and, perhaps, hardly more than an impression. The type and only known collection of E. microphyllum unquestionably belongs here. The specimen is a poor one, with numerous simple flagellate branches and linear leaves ca. 6 mm. long. In habit it very much suggests C. pusilla (T. & G.) Greene, of North America. The nutlets are clearly heteromorphous. I have seen no authentic material of C. Seleri but assume it to be the hispid plant collected by Pennell and Borm in the vicinity of Arequipa. Cryptantha wmbelliformis I be- lieve to be merely a young state of the same form. Cryptantha pis- coensis has an appressed pubescence and is a somewhat coarser form. It has been collected at Quiva by — ting a distinct ies in my own collection, no. Possibly r 6281, from a dryish bench just Boks the fertile belt in the hills back SOUTH AMERICAN BORAGINOIDEAE 41 of Mollendo, Peru. This plant definitely has homomorphous nutlets. Though the nutlets are all consimilar, the abaxial one is more firmly attached to the gynobase than are the others. In other than the homomorphy of its nutlets the plant is quite like the Peruvian forms of C. parviflora. It has the appressed pubescence and the slightly coarse habit of the type of C. piscoensis . ©. subamplexicaulis (Ph.) Ratehie: Perennial or persistent annual with a firm branching root, erect and subsimple at first but later with several or many trailing branches from a loose fruticulose caudex and forming a loose prostrate mat 3-12 dm. broad and 1-2 dm. tall, hispid or hispid-villous and usually appressedly so and canescent; leaves linear to linear-lanceolate or oblong-lanceolate, 1-4 (-5) em. long, 2-4(-5) mm. broad, acutish, sessile by a broad usually cordulate or rounded base, more or less pustulate; spikes solitary or geminate, bractless, 3-12 cm. long, becoming remotely flowered; fruiting calyx ovate or ovate-oblong, 3-5 mm. long, contracted below into a short stout and angled pedicellate base; mature calyx-lobes lanceolate or linear, erect with tips more or less spreading, densely or at times sparsely appressed hispid-villous, the midrib prominent below middle and more or less hirsute; corolla evident, 5-6 mm. broad, white; fruit 4-ovulate; nutlets 4, homomorphous or somewhat heteromorphous with the axial nutlet (always the most persistent) slightly the largest and dullest, 1.5-1.9 mm. long, tuberculate or papillate, obscurely granulate, sometimes sparsely stipitate-glandular, apex acute, base truncate or obtuse, sides angled, back convex or ob- scurely obtuse; groove narrow or cloeed, usually with a small areola at the broad basal fork; gynobase 3—{ height of nutlets; style very much surpassing the nutlets and about twice length of gynobase.— Anal. Univ. Chile exxi. 826 (1908) and Fl. Chile v. 231 (1910). Eri- trichum subamplexicaule Ph. Fl. Atac. 39 (1860) and Viage Des. Atac. 25 and 213 (1860). CHILE. AntToracasta: dry shrubby: tine Cerro Perales near Taltal, Johnston 5633 (G); Paposo, Dec. 1853, Philippi (MS, TrPE; G, photo.); i ope bes Pa SO . pone in fertile belt, El Rincon just north of Paposo, Johnston 5543 (G); ope in fertile od leigh Aguada Panulcito, Johnston 5475 (G); sheltered places, i eveigien 1, Johnston 5448 (G); moist bench in quebrada Aguada de Mi Diaz, Faleiies nr G); dune slope just southwest of Aguada de Miguel Diaz, Johnston 5412 (G); ss ‘slope just back of Punta Reyes below Aguada de Miguel Dice Johnston 5411 (G). A well marked species in its typical form, characterized by its very remarkable habit. The plant grows on the fog-bathed middle slopes of the hills fronting the sea. The long branches trail through the 42 JOHNSTON grass or even clamber somewhat in the low bushes forming a loose mat commonly about 8 dm. broad. Although appearing to flower the first year, it seems to persist ordinarily at least for several seasons, forming a loosely branched depressed fruticulose caudex. The three collections from the vicinity of Miguel Diaz are not typical, having longer more hirsute calyces and non-canescent herbage. The two collections from the dunes are further atypical, being more or less ascendingly branched and not certainly persistent. It is not im- probable that future collecting will show the plants from Miguel Diaz to be worthy of special nomenclatorial recognition. As now known C. subamplexicaulis reaches its northern limit at Miguel Diaz. The species is reported from Iquique and Tocopilla by Reiche, but through misdeterminations of C. filiformis. C. argentea, sp. nov. Perennis prostrata dense argenteo- strigosa; ramis numerosissimis ramosis e caudice prostrato dense ramoso fruticoso orientibus in pulvinum 2-5 cm. altum 1-4 dm. latum densum congestis; foliis firmis 5-15 mm. longis 1-2 mm. latis dense argenteo-strigosis numerosis acutis basi rotundis vel cordulatis, superioribus paullo reductis, inferioribus conspicue marcescentibus; spicis geminatis vel solitariis 3-12 em. longis ebracteatis; calycibus fructiferis ovatis 2-3 mm. longis deciduis basem versus crasse angu- latimque attenuatis; lobis calycis oblongi-linearibus obtusis strigosis vel dense velutinis, supra medium plus minusve herbaceis et saepe recurvatis vel ascendentibus, infra medium erectis cum setis flaves- centibus paucis perbrevibus horridis; corolla conspicua alba 2-4 mm. lata; fructu 4-ovulato; nuculis 4 homomorphis cum nucula axillari subpersistenti 1.5-2 mm. longis tuberculatis vel papillatis minute granulatis apice acutis basi rotundatis dorso convexis ventre 2-2 longitudinis ad gynobasem adfixis, sulcis clausis vel anguste apertis in areolam deltoideam dilatatis; stylo nuculam longe superante quam gynobasis duplo longori. CHILE. Anroraaasta: crevices about head of high fog-bathed sea-cliffs near Aguada Cachina, Johnston 5734 (G); about head of high fog-bathed sea-clifis near Aguada Grande, Johnston 5814 (G, TyPE). This species is known only from the crevices about the head of the high line of huge fog-bathed sea-cliffs which stretch for about 25 km. northward from the Atacama-Antofagasta ‘provincial boundary be- tween Caleta de Pan de Azucar and Caleta Esmeralda. The plant is notable for its silvery, very dense strigose pubescence and particu- larly because of its very densely pulvinate habit. The root is strong, branched and indubitably perennial. The species is evidently most related to C. subamplezxicaulis of the fertile fog-bathed hillsides north SOUTH AMERICAN BORAGINOIDEAE 43 of Taltal, but is clearly different in its pulvinate habit, silvery pu- bescence, smaller corollas and inconspicuously hirsute calyces. 10. C. chaetocalyx (Ph.), comb. nov. Annual; stems several, 4-10 em. long, spreading, loosely branched, with a very short mostly ap- pressed slender hispid pubescence; leaves linear or oblong to lanceolate, obtusish, 1-1.5 em. long, 1-3(-4) mm. broad, scabrous, short-hirsute, evidently and abundantly pustulate; spikes 1-3 em. long, bractless, solitary or geminate, becoming loose in age; fruiting calyx ovate, 2-4 mm. long, contracted into a short stout angulate pedicellate base; mature calyx-lobes linear, above middle sparsely hirsute and green and spreading, midrib very prominent and pungent short-hirsute below middle; corolla white, 2.5-5 mm. broad; ovules 4; nutlets usu- ally 4, homomorphous with the axial one most persistent, 1.5-2 mm. long, oblong-ovate, very finely granulate, more or less obscurely rugulose or tuberculate, apex acute, sides sharp, base rounded, back convex without a medial ridge; gynobase ca. ? height of nutlets; style definitely surpassing the nutlets.—Eritrichum chaetocalyx Ph. Fl. Atac. 39 (1860) and Viage Des. Atac. 10 and 213 (1860). E. di- varicatum Ph. Anal. Univ. Chile xe. 534 (1895). C. divaricata Reiche, Anal. Univ. Chile cxxi. 827 (1908) and FI. Chile v. 232 (1910). E. pustulosum Ph. Anal. Univ. Chile xe. 537 (1895). CHILE. Aracama: Bandurrias, Geisse, (MS); Desert of Atacama [? Ban- durrias], — 154 (NY); Atacama Desert (Caldera sg ere sag S, FM); “Caldera ? Cones 4 no collecto: Dee. 1853, Philippi (MS, tyre of E. chaetocalyx; G, ‘ oes); Callers Bape. 1879, no collector given (MS, Type of E. pustulosum; G, photo to.); Caldera, pt. 1885, F. Paton (MS, tyPE of E. divaricatum; G, hota. ); Caldera, oux Although i in the original account of E. chaetocalyx the type is given as coming from near Pan de Azucar, lat. 26° 10’, I feel certain pag in fact it came from Caldera, lat. 27° 5’. There are no specimens in the Philippi herbarium from north of Caldera. I searched for the plant about Caleta de Pan de Azucar and did not find it. In the Philippi herbarium there is but a single collection labeled E. chaetocalyx which was collected by Philippi during his travels in the Atacama Desert in 1853-4. This is from Caldera and checks perfectly with the original description. It is notable that in the text of the narrative of his explorations, Philippi does not mention the species in connection with Pan de Azucar although he does so with Caldera. Hence I feel con- fident that the locality given with the original description is an error and probably a slip of the pen. The type almost certainly came from Caldera. The species is a very distinct one with characteristic coarse spreading stems and more or less obscurely roughened nutlets. Its 44 JOHNSTON relations are with C. granulosa and C. filiformis though it is quite distinct from both. 11. C. filiformis (Ph.) Reiche. Annual, 5-30 ecm. tall, usually loosely and ascendingly branched from the base, finely hispid and frequently appressedly so; leaves not numerous, lanceolate to lance- lmear or oblanceolate, 1-3 cm. long, 1-5(-9) mm. broad, obtuse or acutish, sessile, pustulate and hirsute; spikes solitary or geminate, 1-10 cm. long, bractless, becoming loosely flowered; fruiting calyces ovate or oblong-ovate, 3-5 mm. long, subsessile; mature calyx-lobes linear or lance-linear, obtusish, hirsute below the middle, above the middle usually herbaceous sparsely appressed short-hispid and spreading; corolla conspicuous, 4-7 mm. broad, white; fruit 4-ovulate; nutlets usually 4, heteromorphous, ovate or ovate-oblong, back obscurely obtuse, base truncate, sides angled, groove opened or closed and broadly forked near base; odd nutlet axial, evidently the largest, 1.5-2 mm. long, most firmly affixed, palest, tuberculate or muricate or papillate, more or less stipitate-glandular; homomorphous nutlets 1.3-1.5(-1.7) mm. long, tuberculate, brownish; gynobase about 3 length of nutlets; style much surpassing nutlets, about twice length of gynobase.—Anal. Univ. Chile exxi. 829 (1908) and FI. Chile v. 234 (1910). Eritrichum filiforme Ph. Fl. Atac. 39 (1860) and Viage Des. Atac. 213 (1860). E. mite Ph. Anal. Univ. Chile xe. 539 (1895). C. mitis Reiche, |. ¢. a Silom ee Hueso Parado, Taltal, 1853, ial (MS, (MS f 3 mite; G, photo ‘. Tocopilla, no collector piven, no. 13 (M 8); gravelly fog-bathed ce “Tocopila et. 1925, Johnston 8 (G). Tar ebrada H 2 ei 1904, ner (M3), Quebrada Huantajaya, Iquique, 7700 , Werdermann The type of C. filiformis is a very small slender plant given as from Hueso Parado at Taltal. Although with a different gross aspect, it has the fruit-characters, the calyx and pubescence of the Tocopilla and Iquique plants. I made very many collections of Cryptanthas in the region about Taltal but did not discover any plants there that could be mistaken for C. filiformis. If the plant actually came from the Taltal region it must be rare or very local. The known range of the species is very disrupted. The material at hand shows variation in size and shape of the calyx-lobes and to some extent in the outline , of the nutlets also. However, these variations are unimportant since are all present among the large suite of specimens I collected on a single hillside at Tocopilla. 12. ©. granulosa (R. & P.) Johnston. Annual, 5-30 emi. tall; stems erect, becoming loosely and ascendingly branched, finely hispid SOUTH AMERICAN BORAGINOIDEAE 45 with the hairs appressed or somewhat spreading; leaves linear, 2-3.5 em. long, 2-4 mm. broad (rarely, in C. latifolia, somewhat oblong and 5-10 mm. broad), obtuse, rounded to a sessile base, sparsely pustulose- setose; spikes solitary or geminate, bracted, 3-12 cm. long, becoming loosely flowered; bracts few to many, linear or oblong, 3-5(—10) mm. long; fruiting calyx ovate, 2-4 mm. long, subsessile; mature calyx- lobes linear or lance-linear, obtuse, sparsely appressed-hispid, below middle sparsely short-hirsute on the weakly prominent midrib, above middle more or less herbaceous and spreading; corolla evident, white, 2-4 mm. broad; ovules 4; nutlets 4, homomorphous or heteromorphous with the axial one (always subpersistent) slightly the largest, ovate- oblong, 1.3-1.8 mm. long, tuberculate or papillate and somewhat glandular, apex acute, base rounded or obtuse, edges angled, back CORY ER, groove closed or gradually dilated towards the base; gynobase 2-3 height of nutlets; style surpassing the nutlets, about equalling the renibake —Contr. Gray Herb. Ixviii. 54 (1923). Myosotis granu- losa R. & P. FI. Peruv. ii. 5 (1799). C. latifolia Johnston, Contr. Gray Herb. Ixxiii. 73 (1924). PERU. Lima: mountains near Chosica, Jege am alt., Weberbauer 5340 (FM); Mount San p ey near Lima, 250 m. alt., Weberbauer 5695 (FM); loose stony upper slopes of seaside hills, Choris near ~ m. alt., Macbride 5861 (FM, Trp of C. latifolia; G, isoTyPE). ARE sandy hills, pee 2 Hitchcock gill - US, G); ee stig of fertile ‘belt | in hills back of Mollendo, Johnston 2 (G . I have Gy ee C. latifolia is merely the luxuriant phase of C. granulosa. Mounted on the sheet with the type is a small plant, which is apparently typical of Ruiz & Pavon’s species. The plant from Mollendo seems to belong to the present species, but appears to be very much more abundantly bracted than is the typical form about ma. 13. C. taltalensis, sp. nov. Annua 1-3 dm. alta laxe | asaliter ramosa; ramis saepe gracilibus minute strigosis et plus minusve breviter hispidis; foliis linearibus vel rariter anguste lanceolatis paucis 1-2 em. longis 1-2(-3) mm. latis pustulatis breviter hispidis, superiori- bus paullo reductis; spicis solitariis vel geminatis 3-5 cm. longis bracteatis; bracteis linearibus caducis 3-5 mm. longis; calycibus fruc- tiferis congestis ascendentibus oblongis 3-4.5 mm. longis asymmetricis sessilibus faciliter deciduis basi conicis; lobis calycis linearibus erectis ascendenter hispidis cum costa prominenti; corolla evidenti 3-4 mm. lata alba; fructu 4-ovulato; nuculis 1-3 saepissime plures abortis heteromorphis ovatis apice acutis dorsi convexis margine angulatis, sulcis clausis vel anguste apertis basi in areolam deltoideam dilatatis; nucula axillari ca. 1.8 mm. longa subpersistenti semper maturanti 46 JOHNSTON papillata plus minusve glandulari opaca ventre 2 longitudinis ad gynobasem affixa; nuculis consimilibus tuberculatis deciduis ca. 1. mm. longis; stylo nuculam axillarem brevissime vel breviter superanti. CHILE. Anroracasta: gravelly slopes near crest of hills southeast of Taltal, Johnston 5121 (G, Type); rocky bottom of dry quebrada in hills south- east of Taltal, Johnston 5120 (G); rocky talus-slope in Quebrada amon, near Taltal, Johnston 5156 (G); dryish gravelly open places on sum- as ridge of Cerro Perales near Taltal, 1000 m. alt., Johnston 5634 and 5635 This species is most closely related to C. filaginea, from which it differs in its larger corollas and few, strongly heteromorphous nutlets. It was found at various points in the region about the port of Taltal, always associated with C. filaginea, though less common and less generally distributed than that widely ranging species. Although it appeared to be a very definite and recognizable entity in the field, sub- sequent study of the collected material shows it so closely related to C. filaginea that I have been tempted to propose it merely as a variety of that species. 14. C. Romanii, sp.nov. Annua laxe ramosa 8-15 cm. alta sparse strigosa hirsuta; foliis linearibus 1-3 em. longis 1-2.5 mm. latis pus- tulatis breviter hirsutis acutis, superioribus paullo reductis; spicis solitariis vel geminatis 2-3 cm. longis bracteatis; bracteis linearibus 4-6 mm. longis; calycibus fructiferis congestis ascendentibus oblongis 3-4.5 mm. longis asymmetricis subsessilibus deciduis basi conicis; lobis calycis linearibus erectis costatis ascendenter hirsutis; corolla conspicua 5-6 mm. lata alba; fructu 4-ovulato; nuculis 4 homomorphis cum nucula axillari subpersistenti 2 mm. longis oblongo-ovatis tuber- culatis obscure granulatis apice acutis dorse convexis margine angu- latis et incrassatis ventre 4 longitudinis ad gynobasem affixis; sulcis clausis vel anguste apertis basem versus breviter furcatis; stylo nuculas valde superanti. CHILE. Awnroracasta: Sierra Esmeralda, Dept. Taltal, 1884, San Roman (MS, trrxE; G, photo. and frag.). In gross habit this species much suggests C. hispida and in fact was so determined by Philippi. Itsrelations, however, are with C. filaginea and C. taltalensis, particularly the latter. It differs from both in its very large corollas. From C. taltalensis it also differs in its 4 homo- ohous nutlets. It differs from C. filaginea, which I collected in the Sierra Esmeralda, in its slightly coarser habit, very large corollas and slightly larger nutlets. _ 15. C. filaginea (Ph.) Reiche. Annual, 1-2 dm. tall, erect and shortly branched above or loosely branched from the base; stems SOUTH AMERICAN BORAGINOIDEAE 47 slender, finely strigose and hispid; leaves linear or lanceolate, 1-2.5 (-3) em. long, 1-2 mm. broad, spreading, obtuse, pustulate, finely short-hispid, weakly reduced up the stem; spikes solitary or geminate 2-5(-8) em. long, bracteate; bracts linear, 3-6 mm. long, caducous; fruiting calyx oblong, 3-4 mm. long, usually congested, readily de- ciduous, sessile, ascending; mature calyx-lobes linear, below the middle appressed short hispid and along the prominent midrib hirsute, above somewhat herbaceous and sparsely hairy and slightly spreading; corolla inconspicuous, white, subtubular or with a very narrow limb, 0.5-1.5(-2) mm. broad; fruit 4-ovulate; nutlets 4 or rarely fewer, subhomomorphous with the axial one slightly emphasized and most persistent, tuberculate, usually obscurely granulate, 1.2-1.5(-1.8) mm. long, ovate or oblong-ovate, apex acute, edges angled and fre- quently margined, base obtuse, back convex; groove narrow or closed, forked at base; gynobase 3-$ as high as the nutlets; style much surpassing the nutlets.—Anal. Univ. Chile exxi. 829 (1908) and FI. Chile v. 234 (1910). Eritrichium filagineum Ph. Anal. Univ. Chile xe. 536 (1895). rada San Taltal, Johnston 5157 and 5158 (G); Paso Malo north of Taltal, Johnston 5179 (G); gravelly bench, Estancia Vieja, Johnston 5204 (G); summit of Cerro : U. Arequipa: lower of fertile belt in hills back of Mollendo, Johnston 6280 (G); hills, Mollendo, Hitchcock 22384 (G, US). In the Museo Nacional at Santiago the original collection of this species is represented by six plants, two in the small herbarium of Federico Philippi and four in the main herbarium of the museum. It seems certain that the older Philippi described his species from the latter four plants. These are associated with a label giving the 48 JOHNSTON locality as “ Monte amargo,” a station on the railroad about half-way between Caldera and Copiapé. This collection of four plants is a mixture, three representing C. filaginea as here defined and one being a plant of C. parviflora just coming into fruit. I suspect that the plant of C. parviflora is the one mentioned by Philippi as having been collected by Borchers near Caldera. It seems very probable, also, that the specific name, filaginea, was suggested by this same odd plant. The species seems to reach its southern limit in the Caldera-Copiapé region. It is very generally distributed in the coastal region further north, in fact is the predominating and common species of the genus in the coastal portion of the Department of Taltal. It is related on one hand to C. taltalensis and on the other to C. limensis. The latter is a coarser plant and less branched and has broader firmer leaves and larger fruiting calyces. 16. C. limensis (A. DC.) Johnston. Annual, 5-15 em. tall; stems usually loosely branched from the base, finely appressed-hispid ; leaves oblanceolate to linear, 1-3 em. long, 2-4°mm. broad, obtusish, finely appressed-hispid, more or less pustulate; spikes solitary or geminate, evidently bracted, 3-5 cm. long; fruiting calyx 5-6 mm. long, ovate-oblong, subsessile; mature calyx-lobes linear to linear- oblanceolate, below the middle with a prominent very hirsute midrib, above broadened herbaceous spreading and merely appressed-hispid ; corolla inconspicuous, subtubular, white, ca. 1 mm. broad; fruit 4- ovulate; nutlets 4, homomorphous or sometimes slightly hetero- morphous, ovate-oblong, 1.5-1.7 mm. long, tuberculate, apex acute, base truncate, back convex, edges angled, groove closed or narrow but below gradually dilated into a suprabasal triangular areola, axial nutlet always most firmly affixed and occasionally a trifle the largest; gynobase about 3 height of nutlet; style surpassing the nutlets.— Contr. Gray Herb. Ixx. 46 (1924). Eritrichium limense A. DC. Prodr. x. 133 (1846). C. Macbridei Johnston, Contr. Gray Herb. Ixxii. 74 (1924). C. Woitschachii Brand in Fedde, Repert. xx. 317 (1924). PERU. Lima: loose ston i i i Lima, ea. 150 m. alt., Macbride 5860 (EME reewcr C. MoN Ton ere Lima, 1833, Abadia (G, photo. of TypE of E.limense). AREQUIPA: open places in and just beloy in hills back of Mollendo, Johnston 3531 (G). FINITE: “Peru,” Woitschach (G, fragment of type of C. Woitschachii). Very closely related to C. filaginea but a coarser plant with larger fruiting ealyces. Cryptantha Macbridei and C. Woitschachii are clearly synonyms of this species. de @. calycina (Ph.) Reiche. Annual, 1-3.5 dm. tall; stems commonly solitary with short ascending laterals above or with nu- SOUTH AMERICAN BORAGINOIDEAE 49 merous long erect branches from towards the base, shaggy coarse- hirsute; leaves linear or lance-linear, 1-4 em. long, 1.5-3 mm. broad, obtusish, hispid, somewhat pustulate; spikes geminate or ternate, bractless, glomerate or elongating and becoming very loosely flowered, 1.5-6(-10) cm. long; fruiting calyx strongly accrescent, 8-11 mm long, 2-2.5 mm. broad, short-pedicellate; mature calyx-lobes lance- linear or linear, proximate and erect above the fruit which they greatly surpass, midrib thick and very prominent and armed with coarse slightly tawny hirsute bristles, margins villous-hispid; corolla incon- spicuous, subtubular, white, 1-2 mm. broad, lobes erect; fruit 4- ovulate; nutlets usually 4, homomorphous with the axial one slightly the most persistent, lanceolate, 2.8-3.2 mm. long, 0.9-1.2 mm. broad, smooth or obscurely tuberculate particularly towards the apex, apex acuminate, back flat or broadly obtuse with, at most, an obscure low medial ridge, groove closed or very ened but usually open at the broad forking; gynobase subulate, ca. 3 height of nutlets; style equalling the nutlets or shortly surpassing them.—Anal. Univ. Chile. exxi. 825 (1908) and Fl. Chile v. 230 (1910). Eritrichum calycinum Ph. Fl. Atac. 38 (1860) and a Des. Atac. 80 and 212 (1860) Atacama: talus o Sancarron below Rucas, 3200 m. alt., CHILE. Johnston 6225 (G); dry rocky . of quebrada, Rio de Sat above Jun Encierro, 3250 alt., Johnston 6025 (G); Quebrada Alfalfa, Migue 1, 3400 m. alt., Johnston 4950 (G); gers a Jan. 1901, Reiche (MS). ANTOFAGASTA: Sando on, 2700 m. alt., Feb. 1854, Philippi (MS, TypE; G, photo.). An extremely well marked and very distinct species, which appears to be without any close, obvious relatives. The unusually large fruiting calyces and its very elongate nutlets make it readily recog- nizable. I found it usually associated with C. diffusa, commonly in the upper parts of the range of that species, though always much less on. 18. C. patagonica (Speg.) Johnston. Annual, 1-1.5(-3) dm. tall, drying dark; stems slender, erect or decumbent, usually several, commonly with short spreading branches, finely strigose or appressed short-hispid; leaves linear, obtuse, 1-3.5 cm. long, 1-2(-3) mm. broad, spreading, finely short appressed-hispid, finely pustulate beneath; spikes solitary or geminate, bractless, 1-3 cm. long, glomerate or elongating but not particularly loose-flowered; fruiting calyx ovate, 2-3 mm. long, 1.5-3 mm. broad, short-pedicellate; mature calyx-lobes narrowly lanceolate, connivent above, shortly surpassing the nutlets, 50 JOHNSTON midrib prominent and hirsute, margin appressed villous-hispid; corolla inconspicuous, white, ca. 1.5 mm. broad; fruit 4-ovulate; nutlets 4, homomorphous with the axial one most persistent, lance-ovate, ca. 2 mm. long, ca. 0.8 mm. broad, smooth or more often obscurely tuberculate particularly above the middle, apex acute, base truncate, sides angled, back usually convex below the middle and flat above; groove closed or very narrow, usually with a small areola at the fork; gynobase subulate, 2-4 height of nutlets; style equalling or shortly surpassing the nutlets——Contr. Gray Herb. Ixviii. 54 (1923). Am- sinckia patagonica Speg. Anal. Soc. Cient. Argentina liii. 137 (1902). ARGENTINA. Santa Cruz: very arid desert gato Rio Santa Cruz a a Gallego; Feb. 26, 1882, Spegazzin (G); arid comer — Rio Santa Feb. 7, 1882, Spegazzini (G, part o f TYPE); desert g Rio Chico, Reb, 2, 1882; Spegazzini (G); Patagonia, lat. 50° 53°, 1882, aN, pie & Tonint This species, which sets the southern limit of the genus at about lat. 52° S., is probaly most related to the very distinct C. calycina of the desert Andes of northern Chile. It has been much misunderstood and has passed under many names. The collection by Moreno & Tonini, above cited, is that reported as Cynoglossospermum humile by Kuntze, Rev. Gen. iii. pt. 2, 204 (1898). Spegazzini, Pl. Patag. Aust. 551-552 (1897), and Macloskie, Fl. Patag. 678-679 (1905), treated it as E. diffusum and E. parvulum 19. C. peruviana Johnston. Annual, slender, 1-2 dm. tall; stem simple or with several long ascending branches, finely short-hispid and frequently appressedly so; leaves narrowly linear, 1-3(-6 em. long, 1-2 mm. broad, numerous, finely hispid, inconspicuously pustulate, little reduced up the stem; spikes solitary or geminate, 1-5 em. long, bractless, glomerate then elongating; fruiting calyx ovate, 2.5-3 mm. long, subsessile; mature calyx-lobes lance-linear or narrowly lanceolate, erect, appressed hispid-villouis, somewhat hirsute along the weakly prominent midrib; corolla white, inconspicuous, subtubular with a very narrow limb 0.8-1.4 mm. broad; fruit 4- ovulate; nutlets 4 or rarely fewer, homomorphous with the axial one subpersistent and always developing, 1.5-1.8 mm. long, lance-ovate, very coarsely muricate or muricate-rugose, finely pag pale, apex acute, edges acute, back convex, groove closed or dilated towards the base; gynobase about 3 length of aatiets: style just surpassing the nutlets.—Contr. Gray Herb. Ixxiii. 74 (1924). C. Weberbaueri Brand in Fedde, Repert. xx. 318 (1924). C. cajabam- bensis Brand, |. c. 319. PERU. Ancasu: H. Weberbauer 3041 (G, Haciewla Cajabambe between Samanco Soae ie SOUTH AMERICAN BORAGINOIDEAE dl, 3200-3400 m. pts Weberbauer 2658 (G, frag. of Type of C. Weberbaueri). Lrua: loose rocks on dry slope above river, Rio Blanco, 3600 m. alt., Macbride & Ricthcsieke vir (FM, hi E of C. peruviana; G, IsOTYP FE). AREQUIPA: 200 oer. soil along stream-courses, Arequipa near ‘pase of El Misti, 3000-3: alt., Pennell 13235 = EM; ravines and hillsides on southern slopes of El Chachani north of Are uipa, 3350 m. alt., Hinkley 77 (G, US). Moqus- Fab. open mixed forsake, Torata, 2200-2500 m. alt., Weberbauer 7398 (G, FM A Peruvian plant which is very closely related to C. globulifera and perhaps is only a phase of it differing in its northern range, slender habit and slightly smaller flower-parts. It sets the northern limit for the distribution of Cryptantha in South America, in the Department of Somer occurring north almost to lat. 9° S. . ©. globulifera (Clos) Reiche. Annual, 1-3(-4) dm. tall; aa few or solitary, usually loosely branched, hispidulous; leaves linear, 14 em. long, 1-2 mm. broad, obtusish, short appressed-hispid, little reduced up the stem; spikes solitary or geminate, bractless, 1-6(-10) em. long, becoming loosely flowered in age; fruiting calyx ovate-oblang, 3-4 mm. long, ascending, base rounded and very shortly pedicellate; mature calyx-lobes linear or lance-linear, connivent above with the tips somewhat spreading, margins villous, hirsute along the prominent midrib, the hairs all clean and white or only slightly tawny; corolla inconspicuous, white, subtubular, ca. 1 mm. broad; fruit 4-ovulate; nutlets 14, homomorphous, ovate-oblong, 1.5-2 mm. long, finely granulate, pale, weakly but densely tuberculate with the tuber- culations in more or less broken transverse rows, apex acute, back convex, edges acute below middle and rounded above, groove nar- rowed or closed but usually with an open fork; gynobase $-% height of nutlets; style just surpassing or a little surpassed by the nutlets. —Anal. Univ. Chile exxi. 827 (1908) and FI. Chile v. 232 (1910). Eritrichium globuliferum Clos in Gay, Fl. Chile iv. 464 (1849). E. glareosum Ph. Linnaea xxxiii. 189 (1864). C. glareosa Greene, Pit- tonia i. 111 (1887); Reiche, 1. c. 820 and 1. c. 225. E. carrizalense Ph. Anal. Univ. Chile xc. 526 (1895). C. carrizalensis Reiche, |. c. 819 and |. c. 224. E. floribundum Ph. Anal. Univ. Chile xc. 532 (1895). C. floribunda Reiche, |. ¢. 825 and |. c. 230. E. paroulum Ph. Anal. Univ. Chile xe. 535 (1895). E. oliganthuwm Ph. Anal. Univ. Pees xe. 535 (1895). OC. oligantha Reiche, |. c. 819 and 1. ¢. 224. E. sphaerophorum Ph. Anal. Univ. Chile xc. 539 (1895). E. longisetum Ph. Linnaea xxxiii. 189 (1864) in part; only as to plant from San Felipe which became type of E. floribundum, cf. Anal. Univ. Chile xe. 538 (1895). ARGENTINA. Cuasurt: gravelly places along Rio Carren-leofi, March 5, 1900, Spegazzini (G). R1o Necro: vicinity of General Roca, 250-360 m sy JOHNSTON alt., Fischer 131 (G, US, FM). Menpoza: vicinity of Mendoza, Nov. 1913, Hauman 260 (G). CHILE. Sanriago: Las Arafias, Sept. 18 (MS, TrpE of E. sphaerophorum; G, photo.); between Caldera and Copiap6?, no collector given (MS, TyrpE of E. oliganthum; G. photo.); Piedra colgada, Sept. 1885, no collector given (MS; G, photo.); Vizeachitas, Oct. 14, 1914, Rose 19336 (US, NY). A somewhat variable but a natural and readily recognized species. It is not improbable that with the accumulation of study-material the species as here defined may be broken up into several minor species of restricted distribution. The material from Coquimbo is more tawny, coarser and more spreading than other forms. As here interpreted the species is notable since it is the only member of its genus generally ranging at low altitudes which occurs both in Argentina and Chile. Possibly the axial nutlet is slightly emphasized. However, in a number of cases, I found it aborted and the abaxial one developing. 21. ©. diffusa (Ph.), comb. nov. Annual, 8-20 cm. tall; stems commonly several to many, ascending, strictly branched, hispid; leaves linear or lance-linear, 1-3.5 em. long, 1-2.5 mm. broad, obtuse, numerous, weakly reduced up the stem, finely appressed-hispid, very obscurely if at all pustulate; spikes solitary or geminate, bractless, 1-5(-8) em. long, glomerate but in age becoming rather loosely flowered, tawny; fruiting calyx ovate or oblong-ovate, 3-4 mm. long, ascending, base rounded and very shortly pedicellate; mature calyx- lobes lanceolate to linear-lanceolate, connivent above, at times with tips spreading, margins villous, the prominent midrib hirsute, the hairs usually very conspicuously tawny; corolla inconspicuous, sub- tubular, white, ca. 1 mm. broad; fruit 4-ovulate; nutlets 1-4, homo- morphous, ovate-oblong, 1.5-2.2 mm. long, pale, finely granulate, obscurely tuberculate, marked by 6-12 more or less sinuous deep lineate transverse grooves and hence broadly wrinkled, apex acute, back flattish or obscurely obtuse or convex, edges angled, groove narrow or closed; gynobase 3-4 height of nutlets; style just sur- passed by or just surpassing the nutlets.—Eritrichum diffusum Ph. SOUTH AMERICAN BORAGINOIDEAE 53 Linnaea xxxiii. 191 (1864) and Anal. Univ. Chile xe. 523 (1895). E. difusum Ph. in Villanueva, Anal. Univ. Chile liii. 444 (1878), nomen. E. micranthum Ph. Fl. Atac. 38 (1860) and Viage Des. Atac. 80 and 212 (1860); not E. micranthum Torr. (1854) nor C. mi- crantha Johnston (1925). E. Borchersii Ph. Anal. Univ. Chile xe. 531 (1895). C. Borchersii Hauman, Anal. Soc. Cient. Argentina Ixxxvi. 302 (1918). E. globuliferum, var. Ph. Anal. Univ. Chile xe. 524 (1895). (?) C. famatinae Brand in Fedde, Repert. xx. 318 (1924). ARGENTINA. Menpvoza: Bajos del Inca, Jan. 15, 1886, Borchers 4951 (G); upper part of Qu a San Miguel, 2: alt., Johnston 4932 and 4933 (G s M s, Sierra San Miguel, 1100- 1500 m. alt., Johnston 4918 and 6282 (G); rocky draw, Potrerillos, 2800 m. alte Ss ohnston 1798 (G); abandoned field, Los Alamos, Quebrada de Potrerillos, 2400 m. alt., Johnston 3697 (G); Agua de Acerillo, Oct. 1877, Villanueva (MS). Anroracasta: Sandon, 2700 m. alt., Feb. type of E. micranthum; G, photo). Clearly related to C. globulifera but differing in its lineately trans- verse-grooved nutlets, usually conspicuously tawny spikes and high montane distribution. Though there is some intergradation between the two species it is rare. Commonly they are distinguished at a mere glance. In C. diffusa no particular nutlet seems to be empha- sized nor always developing. 22. G. debilis (Ph.) Reiche. A weak slender annual, 1-1.5 dm. tall, with a few very loose well developed branches, sparsely his- — pidulous; leaves linear or oblanceolate, spreading, distant, 2-3 cm. long, 2-4 mm. broad, quite herbaceous, sparsely hispidulous, obtuse, upper ones scarcely if at all reduced; spikes geminate or solitary, spreading, 2-3 cm. long, bractless, becoming very loosely flowered; fruiting calyx broadly ovate, 2-2.5 mm. long, 1.5-2 mm. broad, base broad and very shortly pedicellate; mature calyx-lobes lanceolate, herbaceous, sparsely appressed-villous, weakly hirsute along the non- prominent midrib; corolla inconspicuous, white, subtubular, ca. 1 mm. broad, fruit 4-ovulate; nutlets 4, homomorphous, ovate-oblong, 1.5— 1.8 mm. long, pale, very finely granulate, tuberculate or becoming 54 JOHNSTON papillate towards apex, marked by 6-12 deep more or less sinuous lineate grooves and hence broadly transverse-wrinkled, apex acute, back convex, edges acute, groove closed or narrow; gynobase reaching to about ~ height of nutlets; style just surpassed by nutlets.— Anal. Univ. Chile cxxi. 830 (1908) and FI. Chile v. 235 (1910). Eri- trichium debile Ph. Cat. Pl. Itin. Tarapaca 57 (1891). BOLIVIA. Porost: Paroma, ca. 3800 m. alt., Feb. 25, 1885, F. Philippi (MS, type; G, photo.); Chiguana, 3700 m. alt., March 22, 1921, Asplund 6898 (US). Obviously a close relative of C. diffusa and perhaps only an extreme form of it. The two collections cited agree uncommonly well in gross aspect and technical characters. They come from about 450 km. north of the northernmost known station of C. diffusa and differ from that species in the weak widely branched habit and smaller broader non-tawny fruiting calyces. No particular nutlet seems to be emphasized in this species. Il. Fruiting calyx small, 2-2.5 mm. long; nutlets ca. 2 mm. long. Glomeruliferae............ ws ie ee ae 23. C. glomerulifera. SOUTH AMERICAN BORAGINOIDEAE 55 Fruiting calyx large and coarse, 4-6 mm. long; nutlets 2-3 mm. | Capituliflor : ae. Plant low, with short prostrate or widely spreading densel. ppressed white villous-hispid stems; mature inflores- : ence more or less capitate................ 24. C. capituliflora. Plant erect or ascending, loosely branched; stems appressed- ispid; inflorescence becomin, n _ Plant canescent, abundantly appressed short-hispid on tems and leaves; style very much surpassing the nVels . ccs ccc noe see es ee ee 25. C. longifolia. Plant green, very sparsely hirsute, evidently pustulate; style equalling or shortly surpassing the nutlets. 26. C. spathulata. Plant definitely annual. ruiting calyces armed with stiff hairs, hispid or hirsute. Glomeratae. Nutlets 1.3-2 mm. long; plant green; common, widely dis- persed Species. . 2... nese ec wine ne cs crete nese mee 27. C. glomerata. Nutlets 2.5-3 mm. long; plant canescent; rare local species. oe C. alfalfalis. caer calyces long-villous, lacking stiff hairs. Haplo- Plant slender; cauline leaves 1-1.7 em. long, 2-3 mm. wide; nutlets Verrucoee. 5. ds i cs eee ee 29. C. haplostachya. Plant coarse; cauline leaves 1-2 cm. long, 5-10 mm. wide; Hittite. buperculate oo ca co as et et 30. C. calycotricha. 23. C. glomerulifera (Ph.), comb. nov. Perennial, 3-5 dm. tall; stems tufted, erect or strictly ascending, subsimple or producing slender ascending branches above the middle, more or less abundantly appressed short-hispid; leaves very scabrous, short-hispid, pustulate at least beneath; the lower oblanceolate, crowded, acutish, 3-7 cm. long, 3-5 mm. broad, contracted to a petiole; the middle cauline linear to oblance-linear or lance-oblong, obtusish, broad at the sessile base, ca. 1-3 em. long, 3-4 mm. broad; cleistogamic flowers solitary or glomerate in the leaf-axils all along the stem, the fruiting structures similar to those of the cl gamic fl ;ch gamic fl with corolla 1.5-2 mm. broad, small, white; spikes geminate, bractless, densely flowered, ca. 1 cm. long; fruiting calyx ovate, 2-2.5 mm. long, sessile; mature calyx-lobes lance-linear to lance-oblong, short-hispid, midrib not greatly thickened; ovules 2; nutlets 2, ovate or oblong- ovate, ca. 2 mm. long, granulate, often sparsely tuberculate especially below the middle, apex acute, base truncate or obtuse, back convex or more commonly obtuse with a medial ridge, groove narrow, axial nutlet always present, abaxial one twisted around beside the axial; style equalling or surpassing nutlets, about length of gynobase; gynobase ca. 3 height of nutlets.—Eritrichium glomeruliferum Ph. Anal. Univ. Chile xe. 521 (1895). CHILE. Aconcacua: Juncal, 2400 m. alt., Feb. 25, 1903, Buchtten 144 (G, US, MS). Coaurmmo: Las Mollacas, Cordillera de Ilapel, Jan. 1888, no 56 JOHNSTON collector given (MS, Type; G, photo.). Inpermire: Cordilleras, Cumings 236 ARGENTINA. San Juan: talus in gorge above Bajfios San Crispin, 3500 m. alt., Jan. 1926, Johnston 6106 (G). A very distinct species readily Sa alc from other perennial species by the small size of its flowering parts. 24. C. capitulifiora (Clos) Reiche. Perennial, much depressed; stems few or numerous, decumbent to prostrate, 5-15 cm. long, densely and canescently villous-hispid; leaves oblanceolate, obtuse, thickish, weakly reduced up the stem, densely appressed villous-hispid, abundantly and minutely pustulate beneath; lower leaves rather crowded, 3-5 cm. long, 2-4 mm. broad; cleistogamic flowers in leaf- axils, solitary or glomerate, not numerous; chasmogamic flowers in bractless congested geminate racemes, ca. 1 cm. long, crowded into dense capitate thyrsoid terminal clusters 1-2 em. thick; fruiting calyx oblong, 4.5-6 mm. long, short-pedicellate; mature calyx-lobes oblong to linear, obtuse, erect, velvety with a very dense villous-hispid pu- bescence, midrib not prominent; corolla not conspicuous, 1.5—-2 mm broad, subtubular, white drying brown; ovules usually 2 but occa- sionally 3 or even 4; nutlets 1-2(-4), when 2 with the abaxial one bent around beside the axial one, 2.5-3 mm. long, ca. 1.3-1.5 mm. broad, triangular-ovate, light-colored, granulate, with scattered muri- cations, apex acute, base rounded, back obtuse with a definite medial ridge, sides angles, inner face right-angled bearing an open or closed groove; gynobase ca. } height of nutlets; style surpassing the nutlets. —Anal. Univ. Chile exxi. 822 (1908) and Fl. Chile v. 227 (1910); Hauman, Anal. Soc. Cient. Argentina Ixxxvi. 303 (1918). Eritrichium capituliflorum Clos in Gay, Fl. Chile iv. 467 (1849). Cynoglosso- spermum capituliflorum Kuntze, Rev. Gen. iii. pt. 2, 204 (1898). E. eephalanthum Ph. Anal. Univ. Chile xe. 520 (1895). ARGENTINA. Menpoza: Rio Blanco, 2500-3200 m. alt., Jan. 1908, Hauman 270 (G). San Juan: Los Patos, 3000 m. alt., Ge 533 (G, photo. of TyPE of E. Beate amet ee CHILE. Coqurmso: Ana, F. Philippi (MS); (?) hills, Arqueros, Aug. 836° fee fol ‘Me Aconcagua: Juncal, 2400 m. alt., Feb. 1903, Buchtien 145 hag bei MS). Santraco: Las Co ndes, Jan. 1880, Navarro (MS, mixed wi dimorpha ) The type of E. cephalanthum seems to be lost. Reiche could not find it nor Gall I locate it at Santiago. However, from the descrip- tion, it seems to belong clearly to the present species. It was based upon specimens lacking data. The collection from Arqueros, cited above, I suspect to have been mislabeled. The low hills near the coast are scarcely a likely situation for the present high Andean SOUTH AMERICAN BORAGINOIDEAE 57 species. For further notes on this collection see discussion under C. cynoglossoides. 25. C. longifolia (Ph.) Reiche. Perennial, 1-3 dm. tall; stems several, rather slender, strict, leafy, appressed short-hirsute; leaves oblanceolate, obtuse, gradually reduced up the stem; basal leaves long-petiolate, 7-9 cm. long, 6-8 mm. broad, appressed short-hispid, minutely pustulate; cleistogamic flowers solitary or glomerulate in the leaf-axils particularly below the middle of the stem, inconspicuous; chasmogamic flowers with corollas ca. 3 mm. broad, white, drying brown; spikes bractless, geminate, dense, 1-3 em. long, in a somewhat thyrsoid arrangement; fruiting calyx oblong, subpersistent, 4-5 mm. long; mature calyx-lobes oblong-linear, erect, densely hispid-villous, midrib not prominent; fruit biovulate; nutlets 1-2, when 2 with the abaxial one bent around beside the axial one, 2-2.2 mm. long, 1-1.2 mm. broad, cuneate-oblong, dark or pale, sometimes mottled, granu- late, sparsely tuberculate, apex acute, base obtuse, sides angled, and with thickened edges, back obtuse with a definite medial ridge, inner face right-angled, groove nearly closed though opened at the broad basal forking; gynobase ca. 3 height of nutlets; style much surpassing the nutlets.—Anal. Univ. Chile exxi. 823 (1908) and FI. Chile v. 228 (1910). Eritrichium longifolium Ph. Anal. Univ. Chile xe. 522 (1895). CHILE. Cogurmso: La Polcura, Cordillera de Ilapel, Jan. 1888, collector not given (MS, TypE; G, photo.). Evidently related to C. capituliflora, but certainly distinct, differing in its habit and in the abundance of pubescence. Much more closely related to C. spathulata but a less stiff, more erect plant with more abundant, less rough pubescence on herbage and calyces and having slightly smaller nutlets and a longer style. The known stations for C. longifolia and C. spathulata are very distant. Both are known, however, only from the type-collection. 26. C. spathulata (Ph.) Reiche. Perennial from a strong taproot; stems several, loosely ascending, 1-3 dm. tall, very scabrous, leafy, sparsely appressed short-hirsute; leaves oblanceolate, obtuse or rounded at apex, 2-3.5 cm. long, 6-9 mm. broad, conspicuously pus- tulate, scabrous, sparsely appressed short-hirsute, lower ones petiolate; cleistogamic flowers solitary or glomerulate in the leaf-axils along the middle part of the stems; chasmogamic fl ith corollas 3-3.5 mm. broad, white, drying brownish; spikes bractless, geminate, numerous, 2-5 em. long, becoming loose; fruiting calyx oblong, 4-5 mm. long, subsessile, subpersistent; mature calyx-lobes oblong-linear, erect, obtuse, hispid, midrib weakly prominent; fruit biovulate; nutlets 2 : \|BRAR or ~ i * if 58 JOHNSTON with the abaxial one bent around besides the axial one, 2.5-3 mm. long (the axial one usually the best developed and slightly the darker), ovate-oblong, finely granulate, sparsely tuberculate, apex acute with the very tip blunted, base obtusish, back obtuse with a well developed medial ridge, sides angled with thickened edges, angled ventrally, groove closed except at the broad forking; gynobase ca. } height of nutlets; style equalling height of nutlets or definitely surpassing them. —Anal. Univ. Chile exxi. 823 (1908) and Fl. Chile v. 228 (1910). Eritrichium spathulatum Ph. Anal. Univ. Chile xe. 517 (1895); not E. spathulatum Clarke (1883). CHILE. Cotcuacva: Cordilleras de Popeta, 1881, F. Philippi (MS, TyPE; G, photo.). This species is readily separated from the other coarse perennials by its very sparsely pubescent and quite scabrid herbage. It is known only from the type-collection. 27. C. glomerata Lehm. Annual; plant usually erect, 1-5 dm. tall, more or less hispid or hirsute and frequently finely pubescent as well; stems one or several, shortly branched above or occasionally loosely branched from the base; leaves lanceolate to lance-linear or lance-ovate, 1-4(—5) cm. long, 2-8 mm. broad, from a broad rounded sessile base usually tapered to an acute apex, reduced up the stem and extending into the inflorescence; cleistogamic flowers biovulate, soli- tary or in glomerules or in short bracted spikes produced in the leaf- axils, always present in the upper axils and commonly in all the axils down to the base of the plant, commonly much more abundant than the chasmogamic flowers, corolla very poorly developed and incon- spicuous, fruiting structures as in chasmogamic flowers though aver- aging slightly smaller; chasmogamic flowers usually not abundant and occasionally absent, restricted to and dominating the normal true terminal spikes which are 1-3(—5) em. long, solitary or geminate or even ternate and are bracted only at base; corolla 3-4 mm. broad; fruiting calyx oblong-ovate to oblong, 1.5-3(-4) mm. long; mature calyx-lobes linear to linear-lanceolate, notably unequal, hispid or hirsute; fruit biovulate or rarely quadriovulate; nutlets 2 or rarely 4, 1.3-2 mm. long, tuberculate and coarsely granulate, ovate or oblong- ovate, more or less triquetrous, apex acute, base obtuse, edges de- cidedly angled and usually somewhat thickened, back convex; groove usually open, dilated towards the base where divaricately forked; style from slightly shorter to definitely surpassing the nutlets; gyno- base about 3 height of nutlets.—Lehm. ex F. & M. Ind. Sem. Hort. Petrop. ii. 35 (1836); Don, Gen. Syst. iv. 373 (1838); Greene, Pittonia SOUTH AMERICAN BORAGINOIDEAE 59 i. 110 (1887); Reiche, Anal. Univ. Chile exxi. 817 (1908) and FI. Chile v. 222 (1910). Cryptantha microcarpa F. & M. 1. c.; Don, 1. c.; Greene, l. ec. 111; Reiche, 1. ¢. 818 and |. ¢. 223. Eritrichium cryptanthum A. DC. Prodr. x. 129 (1846); Clos in Gay, Fl. Chile iv. 463 (1849). E. clandestinum A. DC. 1. ¢.; Clos, l. c. Krynitzkia clandestina Gray, Proc. Am. Acad. xx. 274 and 280 (1885). . clandestinum, var. an- gustifolium A. DC. 1. ce. 130; Clos, l. ce. E. clandestinum, var. decum- bens A. DC. 1. ec. 130; Clos, 1. c. E. congestum A. DC. 1. ec. 132; Clos, l. c. 465. Cynoglossum congestum Poepp. in herb. fide A. DC. 1. ¢ 132. Cryptantha congesta Greene, |. c. 111; Reiche, |. c. 816 and I. c. 221. E. asperum Ph. Linnaea xxix. 16 (1857). E. strictum Ph. FI. Atac. 39 (1860) and Viage Des. Atac. 213 (1860); not E. strictum Decne. (1844). Cryptantha microcarpa, var. stricta Reiche, 1. c. 818 and |. c. 223. E. longisetum Ph. Linnaea xxxiii. 189 (1864) and Anal. Univ. Chile xc. 538 (1895). Cryptantha glareosa, var. longiseta Reiche, ]. c. 820 and |. c. 225. E. Vidali Ph. Anal. Univ. Chile xe. 525 (1895). Cryptantha Vidali Reiche, |. c. 817 and 1. c. 222. (?) E. diplasianthum Ph. Anal. Univ. Chile xe. 530 (1895). E. foliosum Ph. Anal. Univ. Chile xe. 533 (1895). C. foliosa Reiche, |. c. 827 and I. c. 232. Cryp- tantha Candolleana Brand in oom Repert. xx. 48 (1924). Cryp- tantha Philippiana Brand, |. c. CHILE. Mavte: poe eg Azocart (MS, tyre of E. foliosum; G, photo.). CotcHaaua: Colchagua, Gay (G, ep py kay of dake, var. angustifolium). OHiaerns: ee Sept Claude-Joseph 434 US). Sant i elector (M antiago, no colle gi ag 2234 or 2881 (us); Nujfioa, Nov. 1922, Claude-Joseph 2107 (US): Renea, Oc no collector given (MS); Cerro de Renea, 1877, no collector given (MS); C rro de re Nov. 1906, Reiche (MS); San Bernardo, Oct. 1877, no BC Seige given (M ae San ae a Oct. 1879, no collector given (MS); Santa ita, Oct. 12, 1879, collector given (MS); San Antonio, Claude-Joseph 1734 and 2869 (US); is Prado, oe Os 1924, Cl laude-Joseph 2806 (US). Vat- PARAISO: near Vifia del Mar, Oct. 19, 1894, Buchtien 145 (MS); sands, near Vina del : 1, 1895, Buchtien (US); along roads, Valparaiso, Oct 9, 1895, Buchtien (US); agers t. 5, Claude-Joseph 3638 and 3639 (U alparaiso, 1879, collector given (MS); paraiso, Cumings 513 (G); Valparai so, 1836, Gaudichaud 1 122 (G, eyoere of peel of i- num, var. de ; Quin coh ca. 3, Werdermann 40 nie 8 2 =e? 4 He MS, 1 photo. a) : acta 1157 G. NY, te gpa io Beta (NY). Aconcacua: Catema, Sept. 1860, no no collector given (MS, TPE of E. longisetum; G, photo.); Islota Tinie, Los Vilos, Nov. "50, 1889, Vidal (MS, tyre of E. Vidali; G, photo.); Zapallar, 1 Johow (IP). Coqummso: Mie ity of Choapa, Oct. 6, 1914, Rose 19218 (US, NY); Talahuen, 1889-90, Geisse (G, MS); Frai Jorge Estan- i rma ie 1 Skottsberg 748 {G); Coquimbo, Nov. 1893, no collector of E. asperum; G, photo.); Coquimbo, Sept. t. 1885, no collector eae ag ger easy eat hap open slope above Aguada Lora, 60 JOHNSTON Cerro Perales, Taltal, Dec. 1925, Johnston 5630 (G); thickets at lower edge of fertile belt, El Rincon near Paposo, Dec. 1925, Johnston 5542 (G); gravelly talus, Aguada Panul, Dec. 1925, see SLAY (G); Mea el Dia az, 1853, Philippi (MS, Type of E. strictum; G, photo.); steep hillside, Tocopilla, Oct. 1925, Johnston 3577 (G). peal g oan rro de Bravos, Oct. = 2 1881, no nec given (MS); San Juan de Pirque, 1914, Baeza (IP); ¢ Nov. 4, 1920, Claude-Joseph 1222 (US); Chile, 1898, Poeppig 193 (G, artes of TYPE of E. congestum). As here taken C. glomerata is extremely variable. While it is possible that future studies will show it as breaking up into several varieties or even close species, the material I have had for study is not sufficiently extended to form a satisfactory basis for the segregation, at this time, of this polymorphous though certainly very natural as- semblage of forms. Thus interpreted the species is readily recognized in its section by its annual habit, its usually very numerous cleisto- gamic flowers and its more or less lanceolate leaves, which from a broad rounded sessile base are usually contracted to an acute apex. In habit of growth, leaf-proportions, pubescence and even in nutlets there is considerable variation. This variation seems to be quite erratic, for I have been unable to detect concomitant variation in un- related characters. The great variation in gross habit is due partly to environment and partly to the varying abundance of cleistogamic flowers. The cleistogamic flowers are always present in considerable numbers, at times to the complete exclusion of the chasmogamic flowers. The extreme latter condition is apparently that represented by the type of C. glomerata. This form has simple bracted spikes and is usually more densely and abundantly branched than the more common form in which the stem is terminated by 2-3 bractless spikes of conspicuous chasmogamic flowers. Eritrichium Vidali, from a coastal islet off the Department of Petorca, is a peculiar form with broad leaves, the uppermost of which are lance-ovate. Eritrichium foliosum, from Constitucion, is a low coastal form with coarse stems and leaves. The type-collection sets a point on the southern limit of the genus in Chile. Eritrichium asperum from Coquimbo is notable for its slightly smallish calyces which seem somewhat more villous than in the plant from further south. Eritrichium a is a small slender plant with few ating. The type of E. premiers is apparently lost, though judging from descriptions it seems to belong clearly to C. glomerata. A very striking form is that plant of the coastal hills of the Bedviace of Antofagasta described as E. strictum. Though widely detached geographically, E. strictum forms part of the lL odws of the fog- SOUTH AMERICAN BORAGINOIDEAE 61 bathed lomas which has a very strong affinity with the flora of central Chile in which the main range of C. glomerata falls. Reiche treated E. strictum as a variety of C. microcarpa, but the characters he enu- merates for it do not hold in light of the new material recently col- lected. While the plant has a fairly characteristic coarse gross habit and a very detached range it does not seem to be decisively separated from C. glomerata. 28. C. alfalfalis (Ph.), comb. nov. Annual, 3-4 dm. tall; stems stiff, subsimple or loosely and ascendingly long-branched, somewhat cinerescent, finely appressed or spreading villous-hispid; basal leaves linear, becoming 6 cm. long; stem-leaves linear or lanceolate, obtusish, firm, 2-3 cm. long, 2-3.5 mm. broad, finely pustulate; cleistogamic flowers very numerous, solitary or glomerate in the leaf-axils or in small-bracted spikes terminating the stems and branches, calyx and fruit similar to those of the chasmogamic flowers; chasmogamic flowers very few, mixed with the cleistogamic flowers in the terminal solitary or geminate spikes; corolla not conspicuous, ca. 2 mm. broad; fruiting calyx ovate or oblong-ovate, 2-3 mm. long, subsessile; mature calyx-lobes narrowly lanceolate, connivent above, hirsute especially on the thickened midrib; flowers biovulate; nutlets 1-2, axial one always present and most persistent, ovate-oblong, somewhat compressed, 2.5-3 mm. long, 1.3-1.5 mm. broad, densely verrucose, apex acute, base truncate, sides acute, back obtuse, groove closed or open and broadly forked below; gynobase 3-3 height of nutlets; style shortly but definitely surpassing the nutlets.—Eritrichium alfalfalis Ph. Anal. Univ. Chile xe. 525 (1895); Reiche, Anal. Univ. Chile exxi. 831 (1908) and Fl. Chile v. 236 (1910). E. rigidum Ph. |. ¢. 529. C. rigida Reiche, |. c. 819 and I. ec. 224. CHILE. Santiaco: Rio Colorado, Jan. 1888, Philippi (MS, Tyre of E. alfalfalis; G, oo Rio Colorado, Jan. 1888, Philippi (MS, Type of E. The specific name suggests that the type of this species came from El Alfalfal, a locality on the Rio Colorado in the mountains about 40 km. east of Santiago. The two collections cited probably came from the same locality and represent merely ecological phases of one species. It is quite possible that the plants concerned represent merely very coarse forms of C. glomerata. The description given 29. C. haplostachya (Ph.), comb. nov. Erect slender annual ca. 1.2 dm. tall, with several well developed ascending branches, ap- 62 JOHNSTON pressed short-hispid; leaves narrowly lanceolate, 1-1.7 em. long, 2-3 mm. broad, sessile by a broad rounded base, scabrous, minutely pustulate, apex acute or acuminate; cleistogamic flowers solitary or glomerulate in all the leaf-axils; chasmogamic flowers in a dense short solitary terminal spike 1-1.5 em. long; corolla ca. 2 mm. broad; fruiting calyx ovate, ca. 1.5 mm. long, subsessile; mature calyx-lobes linear or oblong-linear, densely villous, midrib scarcely developed; flowers biovulate; nutlets 2, opposite one another, firmly affixed, ovate, thick, ca. 1.7 mm. long, 1.5 mm. broad, minutely granulate and coarsely tuberculate, pale, apex acute, base obtuse, back convex or obtuse, sides acute, groove very narrow and broadly forking below; gynobase poorly developed, ca. 3 length of nutlets; style just surpassing the nutlets, short.—Eritrichium haplostachyum Ph. Anal. Univ. Chile xe. 537 (1895). CHILE. Aracama: Piedra colgada, Sept. 1885, no collector given (MS, TyPE; G, photo.). More material of this very distinct species is a decided desideratum. It is known only from the type, which is a small, very slender plant in full fruit. Its nearest relation is found in C. calycotricha of the Coquimbo region. 30. C. calycotricha, sp. nov. Annua ca. 1.5 dm alta basaliter laxeque pauciramosa; ramis ascendentibus delicate adpresseque brevi- hispidis; foliis adpresse hispidis minute pustulatis obtusis, inferioribus linearibus vel anguste oblanceolatis 3-3.5 cm. longis 3-5 mm. latis, mediis oblongis vel ovato-oblongis late affixis 1-2 cm. longis 5-10 mm. latis, superioribus paullo reductis; floribus cleistogamis in axillis foliorum glomeratis et in parte inferiori spicae productis; floribus chas- mogamis albis 2.5-3 mm. latis in spicis geminatis 1-2.5 cm. longis congestis ebracteatis dispositis; calycibus fructiferis ovatis 4-5 mm. longis subsessilibus; lobis calycis linearibus vel oblongo-linearibus vel late lanceolatis obtusis dense molliter villosis fulvescentibus ascen- dentibus cum costa paullo prominenti; fructu biovulato; nuculis 2 anguste ovatis crassis 1.8-2.1 mm. longis brunneis dense tuberculato- granulatis sparse crasseque tuberculatis vel papillatis apice plus minusve acutis basi obtusis margine angulatis dorse obtusis saepe crasse carinatis, sulcis apicem versus clausis infra mediam in areolam triangularem profundam grandem abrupte ampliatis; gynobasi quam nuculae 3 longiori basim versus crassa medium et apicem versus abrupte naive: stylo quam nuculae conspicue longiori. CHILE. Coquimso: Frai J ‘ a ee ee ee Po SOUTH AMERICAN BORAGINOIDEAE 63 A very distinct species apparently most related to C. haplostachya of the Caldera-Copiapé region but widely differing from it in habit. In the Philippi collections it was found determined as E. capituliflorum. Reiche reported that species from Frai Jorge apparently upon the basis of this misdetermination. The plant described and named above is most certainly not closely related to the high Andean C. capituliflora. III. Section Grocarya.—This section is characterized by being amphicarpous, producing ordinary spikes of chasmogamic flowers and at the very base of the stem in the lowermost leaf-axils highly special- ized cleistogamic flowers. The latter are commonly developed just below the surface of the ground. They are always biovulate, strongly compressed and at maturity become acutely ovate in outline and _- closely invested by the tough, much accrescent highly modified calyx. The calyx, commonly, is indehiscent, and has the throat conspicu- ously smaller than the broadly expanded proper tube which tightly invests the fruit. Frequently the mature calyx is strengthened by prominent ribs, these either simple and vertical, or irregularly anas- tomosing and loosely reticulate. These peculiar cleistogamic flowers at maturity become 3-9 mm. long, 2-5 mm. broad and 1-3 mm. thick. In order to distinguish them from the much simpler cleistogamic flowers developed in the section Eucryptantha I have restricted to them, in the present treatment, the term “Cleistogene.” Little is known concerning the early growth of these highly specialized struc- tures. In this paper I have described only their mature fruiting condition. The section is a very natural one and apparently contains some of the most highly evolved members of the genus. With the exception of one high Andean species which occurs just over the Argentine border, the section is entirely Chilean with its center of distribution in the north-central part of the country. It seems to lend itself readily to classification, in fact the principal difficulties in its classification are those concerned with the proper delimitation of C. linearis. Key To SpPEcIEs. Plant a cher ce coarse perennial. Alyssoides...... 31. C. alyssoides, Plant definitely ann . Root fleshy A ye Celie orm. ison gaa Spikes b —— eyeerea -lobes linea: Corolla 5-8 mm. broad; Samiacesiihe flowers 4-ovulate 32. C. involucrata. Corolla 2-5 mm. broad; chasmogamic flowers 2-ovulate 33. C. Volckmanni. Spikes oe calyx-lobes oblong Corolla 2-3 mm. broad; hairs on ealyx straw-colored; set ay in en lochs... .... 6c. 5 34. C. dimorpha. 64 JOHNSTON Corolla 4-5 mm. broad; hairs on calyx bright yellow; ubescence on stems closely appressed..... 35. C. cynoglossoides. Mature calyx 2-2.5(-3) mm. long; nu tlets ca. 2mm. lo ong; cleistogenes rather rohit fe soieitate ai bed. ..36. C. linearis. Mature calyx 3-4.5 mm. long; nutlets 2-3 mm. long; cleistogenes Weakly "ribbed E Beereeoe Pe ee a ae ees 37. C. aprica. Chasmogami late. Corolla ape ncaa 4-8 mm. broad; nutlets 2.5-3 mm. long, homomorphous or practically so, back obtuse. Dolichophyliae rolla 6-8 mm. broad; leaves narrowly linear, ate 1) em. Jong, 1.5-3 mm. broad............ C. dolichophylla. oo grav es i eve de El Ek Chachani, Penn Pennell 13279 ~G, FM rocky slopes, 3700-4000 m. alt ALLECO: margin Rio Renaico, Esperanza, 1915, Baeza Ae pote | between Nacimiento and fae [ 1877, no collector i soa (MS). Concepcion: St. Vicente, 1890, no collector given (MS). Nupue: between Bollen and Coihueco, Oct. 1878, Puga TAS, vasuot I. Pua: “G, photo.). SOUTH AMERICAN BORAGINOIDEAE 81 Mauve: San José, 1861-62, Volckmann (MS). Tauca: Talea, Oct. 1921, Claude-Joseph 1683 (US). CoxncHacua: San Fernando, Sept. 1864, no. collector given (MS). OHtaains: Rancagua, Sept. 8, 1879, Gay (MS); neagua, Oct. 1828, Bertero 442 (G, NY). ered Santiago, Sept. 15, 1921, Claude-Joseph 1 363 (US); San Cristoral, Sept. 1879, no collector “eolibus” San Chris toral, Sept. 1840, Gay 1618 (MS); San a "sd —_ = ‘we 0: Quillota, Oct —_ p Sei i100 sg Valparaiso, 1909, sh (IP). Aconcagua: Uice ae Ge collector given (MS). bo NITE: Maeul, 1917, Base "(P); hile, Oey “G. “NY); Chile, Bridges Py (G). This readily recognized plant of south-central Chile apparently must be accredited to southern Peru. The two collections from the volcano, El Chachani, above Arequipa, seem quite the same as the far removed plant of central Chile. Plagiobothrys tinctorius is related to the North American P. Torreyi Gray, and P. tenellus (Nutt.) Gray, though clearly distinct from both. 4. P. collinus (Ph.), comb. nov. Stems several, ascending to strictly ascending, 3-15 cm. long, branched, usually appressed-villous or hispid-villous; leaves linear to oblance-linear, obtusish, 1-3 cm. long, 2-4 mm. broad, the lowermost clearly opposite; racemes 2-5 cm. long, bracted at base; calyx 1.5-2.5 mm. long, parted into con- nivent lanceolate lobes, fulvescent, subsessile; corolla inconspicuous, 1-1.5 mm. broad; nutlets obliquely ovoid, 1.2-1.5 mm. long, acute, sasinlatocdina with strong sharp ridges, the ventral keel well de- veloped and towards base of nutlet drawn out into a short thick stipe bearing the areola; style about equalling the nutlets in length.— Eritrichum collinum Ph. Linnaea xxix. 17 (1857). Cryptantha collina Reiche, Anal. Univ. Chile exxi. 828 (1908) and Fl. Chile v. 233 (1910). E. inconspicuum Ph. Anal. Univ. Chile xe. 534 (1895). C. incon- spicua Reiche, Anal. Univ. Chile exxi. 820 (1908) and Fl. Chile v. 225 (1910). CHILE. Coquimso: La Serena, Oct. 1878, Philippi (MS, ty inconspicuum; G, photo.); Coquimbo, Sept. 1885, Philippi (MS; G, thot) ); hills, Huanta, ree 1836, Gay 1623 (MS, Type of E. collinum; A very close relative of P. californicus (Gray) Greene, ‘ ae California and adjacent Mexico and perhaps conspecific with it. The Chilean plant is closest to P. californicus, var. gracilis Johnston, and var. fulvescens Johnston, cf. Contr. Gray Herb. Ixviii. 73 (1923), but differs from both in its apparently erect habit and more prominently rugose nutlets. Im size and shape of nutlets, in leaf-form, in inflores- cence and in form and size of the calyx it is closest to the var. fulvescens 82 JOHNSTON though differing conspicuously from it in having softer pubescence. The relationship between the Californian and Chilean plants is very clear and strong, so incontravertable in fact, that some changes in the classification of the North American forms will probably be made. 5. P. armeriifolius (Ph.), comb. nov. Annual, 5.5 cm. tall, erect, short appressed pallid hirsute-villous throughout, herbaceous, from above base producing numerous short strict leafy branches; leaves filiform-linear; strict, 0.5-2 cm. long, 0.8-1 mm. broad; racemes slen- der, densely flowered, 1-1.5 em. long, geminate, ascending, sparsely leafy-bracted; calyx (before anthesis) very elongate, 2.7-3 mm. long, ca. 0.8 mm. thick, strict, divided, the very narrowly linear lobes erect, base abruptly contracted and subsessile; corolla (not quite mature) with tube reaching about 2 height of calyx, the narrow (probably ascending) lobes reaching to about 3 height of calyx, at anthesis the corolla probably distinctly surpassed by the calyx-lobes; fruiting structures unknown.—Eritrichium armertifolium Ph. Anal. Univ. Chile xe. 551 (1895). CHILE. Cotcuacua: San Fernando, Sept. 1864, Philippi (MS, TYPE; G, photo.). _ A curious plant known only from the type specimen which consists of a single small plant, broken off at ground, lacking fruit, and with the flowers only in well developed bud. It was originally described as biennial or perennial and as hard and woody at the base, but that I believe to be incorrect. The plant appears to be an annual which grew in dryish soil. The lower part is no more hard and woody than are similar parts in any of the common annual species of the section Allocarya. I am quite unable to detect any close relationships for P. armeriifolius. Perhaps it is only a freak or extreme ecological form of some well-known species with its identity further masked by its immaturity. More collections of this moana and puzzling plant are greatly desired. 6. P. Germaini (Ph.), comb. nov. Perennial from a cluster of thickened roots; stems usually several, decumbent or ascending, 5-15 em. long, Soacky and densely fine-strigose; leaves numerous, very elongate, linear, 2-5 em. long, 1-2.5 mm. broad, strigose, mostly basal and crowded, those of the stem few, usually the lowest pair op- posite; racemes bractless or very sparsely bracted, tending to become loosely flowered at maturity; mature calyx 3-4 mm. long, base conical and tapering to a pedicel ca. 1 mm. long, lobes strict and densely fulvescent-strigose; corolla 3-7 mm. broad, white; nutlets ovate with a rounded base and acute apex, 1.2-1.5 mm. long, granulate, with a SOUTH AMERICAN BORAGINOIDEAE 83 definite dorsal keel and strong loosely reticulate lateral ridges at least above the middle, below the middle the medial keel and lateral ridges weak or broken up into papillae or murications, scar suprabasal; gynobase pyramidal; style equalling or much surpassing the nutlets.— Eritrichium Germaini Ph. Anal. Univ. Chile xe. 550 (1895). Allo- carya Germaini Reiche, Anal. Univ. Chile exxi. 808 (1908) and FI. Chile v. 214 (1910). CHILE. Matteco: San Ignacio de Pemehue, Cordillera fronting Victoria, 1894-95, Germain (MS, TyPx; G, photo.); Victoria, 1913, Baeza (IP); Victoria, 1914, Stwardo (IP). A remarkably distinct species of uncertain relationship. 7. P. foliosus, sp. nov. Perennis dense. canescenti-strigulosus; caulibus numerosis gracilibus foliosis laxe decumbentibus 6-12 cm. longis e radice gracili verticali orientibus, internodiis saepissime 5-12 mm. longis; foliis numerosis linearibus 1-2.5 cm. longis 0.6—-1 mm latis omnibus (supremis solis exceptis) oppositis; racemis solitariis vel geminatis ebracteatis 2-5 cm. longis; calycibus maturitate strictis, lobis linearibus erectis obtusis ca. 3.5 mm. longis, pedicellis 0.7-1.2 mm. longis; corolla conspicua alba cum limbo 4-6 mm. lato; nuculis ovatis 1.3-1.5 mm. longis irregulariter rugosis carinatis; areola supra- basali obliqua parva; stylo nuculas paullo superante. CHILE. Nustie: Yungay, Feb. 20, 1916, Stuardo (IP, Trpx; G, frag.). A very well marked species perhaps most related to P. Germaini from which it differs in its leafy stems, slender root, and shorter and narrower opposite leaves. 8. P. Kunthii (Walp.) Johnston. Perennial, matted, acaulescent or with stout repent stems, roots fasciculate at nodes; leaves narrowly linear, obtusish, 1—2.5 cm. long, 0.7—1 mm. broad, in a loose basal tuft or clearly opposite along the stem, appressedly and sparsely hispid- villous, margins sparsely ciliate, base dilated and papery; flowers solitary and axillary, in caulescent states in the axils of opposite cauline leaves but in stemless ones springing from the basal tuft of leaves and apparently scapose; calyx cylindrical, sparsely short- villous, base at first narrowly conical but in fruit broad and rounded; calyx-lobes linear or oblance-linear, erect or spreading, becoming 2-4 mm. long; pedicels 3-6 mm. long, becoming stout, usually erect; corolla white with yellowish appendages, limb 4-6 mm. broad, tube 3-3.5 mm. long and 1-1.4 mm. thick; nutlets ovate with a rounded . base and acute apex, 1.4-1.8 mm. long, coarsely reticulate-rugose at least above the middle, frequently spiculate on the ridges, medial keel definite only towards the apex, areola suprabasal; style extremely long, 84 JOHNSTON much surpassing the nutlets——Contr. Gray Herb. Ixviii. 74 (1923). Anchusa Kunthii Walp. Nov. Act. Acad. Caes. Leop. Nat. Cur. xix. suppl. 372 (1843). Allocarya linifolia, var. Kunthii Macbr. Proc. Am. Acad. li. 545 (1916). Antiphytum Walpersii A. DC. Prodr. x. 122 (1846). Eritrichium Walpersii Wedd. Chlor. Andina ii. 90 (1859). PERU. Puno: low pampa on shore of Lake Titicaca, Puno, Oct. 12, 1919, Shepard 4 (G). BOLIVIA. La Paz: margin of Laguna de Pachaujo, between Laripata Sept. 1858, Mandon 383 (NY); wet places, Omasuyos near Achacache, 3950 m. alt., Dec. 1857—April 1858, Mandon 382 (NY); old lake-bed, Pocoata, Feb. 12, 1903, Hill 337 (Kew). This species is known only from the vicinity of Lake Titicaca. It is readily recognized by its very long style and by its large, evidently pedicellate flowers. The original description of A. Kunthii is ex- tremely brief. Since I have seen no authentic material it is possible that I may have the species incorrectly interpreted. However the very distinct species here treated is the only one I know from the Titicaca region with solitary subterminal axillary flowers, as called for in Walpers’s short diagnosis, and so is very probably the species he described. The species as treated here somewhat suggests P.. con- gestus in habit and in pubescence but is probably most related to P. pygmaeus and P. linifolius, particularly to the former. I doubtfully refer to the species a plant collected by Hill (no. 336, Kew) in a damp place near Guaqui at the south end of Lake Titicaca. This plant is glabrescent and has well developed very slender stems, al- though in other characters it is like the plants I have unconditionally referred to P. Kunthii. 9. P. humilis (R. & P.) Johnston. Repent perennial from a tap- root; stems prostrate or frequently with the tips ascending, usually rooted at some of the nodes, glabrescent or sparsely strigose; leaves numerous, all opposite, linear with a rounded apex, 2-6(-11) cm. long, 1.5-3 mm. broad, sparsely strigose to merely ciliate, glabrescent, bases papery and dilated; racemes 2-9 cm. long, usually equalling or sur- passing the subtending leaves, becoming very loosely flowered and somewhat pedunculate, interruptedly bracted; bracts alternately ar- ranged, 0.5-2(-3) em. long; calyx appressed-villulose, canescent with tawny tips, at maturity 2.5-3.5 mm. long, base rounded, pedicels 0.5-1 mm. long, lobes lanceolat d erect; lla white, limb 1-2 mm. broad; nutlets ovate with a rounded base and acute apex, 1-1.5 mm. ng, granulate, irregularly and loosely reticulate-ridged, with a defi- nite medial dorsal keel only above the middle, scar suprabasal; style surpassed by nutlets or at most equalling them.—Contr. Gray Herb. SOUTH. AMERICAN BORAGINOIDEAE 85 Ixviii. 75 (1923). Myosotis humilis R. & P. Fl. Peruv. ii. 5 (1799); m. Asperif. i. 108 (1818). Eritrichium humile A. DC. Prodr. x. 133 (1846) ; Clos in Gay, Fl. Chile iv. 471 (1849); Wedd. Chlor. Andina ii. 88 (1859). Allocarya humilis Greene, Pittonia i. 17 (1887). Cyno- glossospermum humile Kuntze, Rev. Gen. iii. pt. 2, 204 (1898). PERU. Awncasu: dryish brook-bed, Catuc, ca. 24 km. east of ohne 3150 m. ~~ sta arora hey he & Featherstone 2510 (G, FM). Junin ride ta - Pennell 14640 (G, FM); muddy brook-margin, Matucana, 2400 m. April-May 1922, Macbride "k Featherstone 459 (G, FM): 1 Sane Chicla, 3600— 900 m. alt., April 1882, Ball (G, NY); Baiios, Wilkes Exped. (G, NY). Oude: moist slopes, La Raya, sa00-4686 , April 1925, Pennell oe (G, MoquEcua: Carumas near fe vation Ticsani, 4000 m , Feb. 1925 Weberbauer 7325a (G, FM). BOLI A Paz: marshy a La gg 3750 m. alt., Jan. 28, 1919, boxes i MD: near Jungas, 1200 m. alt., Rusby 2581 (N ¥); Undua uavi, 3300 m. alt., Row og “Buehtien 4682 (US): Cheteabare near Quiabaya, 3000 m. alt., Mandon 380 (NY); near Rio Mulluponcu between Laripata and Tani, 3000-3100 oj ae Mandon 379 (G, shee INDEFINITE: Bolivia, pre 1908 (G, NY); Bolivia, Bang 1962 (G, NY, F M). _The type of Myosotis humilis R. & P. came from Pillao, a locality about 50 km. northeast of Hudnuco, Peru. It is described as a hispid annual. As I have seen no authentic material I feel it best to follow usage, largely traceable to Weddell’s Chloris Andina, and accept the name at least tentatively for the widely distributed Peruvian perennial above described. The plant treated here is most related to P. con- gestus but is readily separated by its coarser looser habit, sparser pu- bescence and tawny calyx-lobes. It is usually very quickly recog- nized merely by an examination of the calyx which is very character- istic in having the lobes very tawny at the tips. The species assumes a dwarfed condition exemplified by Buchtien 4682, Pennell 13504, Weberbauer 7325a and Macbride & Featherstone 2510, in which the racemes are glomerate and partially concealed by a tuft of leaves. Weddell probably included such forms in his var. congestum along with the plant which I have treated as P. congestus. These dwarfs forms of P. humilis are quickly distinguished from P. congestus by the very sparsely pubescent or glabrescent leaves and, of course, tawny calyx-lobes. 10. P. congestus (Wedd.) Johnston. Perennial from a taproot, repent; stems a prostrate, rooting at the ge much branch- ed, sparsely villous or gl all opposite, linear to oblance-linear, 6-20 but commonly 10-15 mth, long, 1-2 mm. broad, somewhat cinerescent, appressed hispid-villous, apex rounded, 86 JOHNSTON with somewhat dilated papery ciliate bases; racemes usually glomer- ate in the upper axils and commonly surpassed by the adjacent leaves, frequently elongating and becoming 1-2(—3) cm. long, leafy-bracted throughout, the bracts 5-10 mm. long and alternate; calyx appressed- villulose, cinerescent or very slightly flavescent, at maturity 2.5-3 mm. long, base rounded, pedicels 0.5-1 mm. long; calyx-lobes linear or linear-lanceolate, erect; corolla white, limb 1-2.3 mm. broad; nutlets ovate with a rounded base and acute apex, 0.9-1.2 mm. long, granu- late, irregularly and loosely ridged, with the medial dorsal keel definite only above the middle, scar suprabasal; style equalling the height of the nutlets or definitely surpassed by them.—Contr. Gray Herb. Ixvili. 75 (1923). Eritrichium humile, var. congestum Wedd. Chlor. Andina ii. 88 (1859). PERU. Puno: Chuquibambilla, gravelly banks of riverlet on puna, 3850-3900 m. alt., April 1925, Pennell 13397 (G, FM); Occa ghey Prov. Huancané, 3700 m Wk: Dec. 19 919, el age 107 (G, NY, US). Mogquvecva: GYM. near Volcano’ Ticsani, 4000 m alt., Feb. 1925, i hickoter 73825 BOLIVIA. La Paz: on puna, La Paz, 4100 m. alt., Feb. 1910, Buchtien 43 (G, US, FM); on puna, La Pa az, 4100 m. alt. , Jan . 25, 1907, Buchtien 43 (US); Chuquiaquillo n ear La Paz, April 1857, Mandon 381 (G); Omapusa near Ac 4000 tm Jan. atases, 1859, Mandon 381 (NY); Copa- cabana: cg ‘D4 1903, Hill : aa (Kew). Or = ‘between Oruro and Cocha- bamba, ecg im. alt., March 17, 1892, Kuntze (NY). Tariya: Escayache near Tarija, 3 m. alt., Jan. 30, 1904, Fiebrig. 3019 (G, US). ARGENTINA. TucaMan: La Cienega, Sierra de Tucaman, Jan. 1874, Lorentz & Hieronymus 639 (US, FM). This species is very closely related to P. humilis, but has a more southerly distribution and is smaller and more slender throughout. Its foliage is rather persistently pubescent and cinereous, the racemes are more or less glomerate and the calyx-lobes are not noticeably tawny at the tips. As originally treated by Weddell the present plant was apparently confused with the dwarf states of P. humilis. The original description, however, calls for plants differing from P. humilis in having congested racemes and stems with more numerous and much shorter branches. This applies completely to the plant here treated as P. congestus It seems not improbable that the collections of Meyen treated by Walpers, Nov. Act. Acad. Caes. Leop. Nat. Cur. xix. suppl. 371 (1843), as Amsinckia humifusa also snd to the present species. However, since the specific name used by Walpers was obviously saree from | one of Poeppig’s i Baas, and since Walpers’s n is quite ambiguous applying equally well to various oa of south central Chile as to the Bolivian P. congestus, it seems SOUTH AMERICAN BORAGINOIDEAE 87 best to reserve Walpers’s name for the Chilean species including Poeppig’s collection number 55 which is cited by Walpers. Just what this species is I am not certain. DeCandolle, Prodr. x. 133, footnote (1846), thought it might be P. procumbens 11. P. pygmaeus (HBK.) Johnston. Perennial, repent; stems prostrate, slender, rooting at the nodes, sparsely strigose, sometimes loosely branched but more often densely so and congested, internodes usually short; leaves narrowly oblance-linear, 0.8-1.5 em. long, 0.8-1.2 mm. broad, all opposite, very sparsely hispid-strigose and frequently so only on the margins, apex obtuse, gradually contracted towards the narrowly connate expanded papery base; flowers solitary in the axils of the opposite cauline leaves, subsessile and strict; calyx sparsely appressed-villous, with erect or loosely spreading lanceolate lobes be- coming 2-3 mm. long, subsessile or with very short obscure pedicels 0.3-0.9 mm. long; corolla white, limb 2.5-3.3 mm. broad, tube 1-1.4 mm. long and 0.8 mm. thick; nutlets ovate, with a rounded base and acute apex, 1.2-1.6 mm. long, roughened with rounded irregularly reticulate ridges which are frequently obscure or absent. below the middle, medial dorsal keel definite only towards the apex, areola supra- basal; style decidedly shorter than the nutlets.—Contr. Gray Herb. Ixviii. 74 (1923). Anchusa pygmaca HBK. Nov. Gen. et Sp. iii. 92 (1818). Eritrichium pygmaewm Wedd. Chlor. Andina ii. 89 (1859). Lithospermum alpinum R. & S. Syst. iv. 742 (1819). epee Losa: Chuquiribamba, Nov. 16, 1876, André 4442 ek FM). Azuay: vicinity of Cumbre, Sept. 24 24, 1978, Ros e, Pac 22947 (G, NY). Inperinite: Ecuadorian Andes, Ff sett 5309 (G). The type of this species was collected at 4100 m. alt. near the summit of Antisana, Prov. Pichincha, Ecuador. It is probably most related to P. linifolius but is smaller in all parts and is more slender and more compact in habit. 12. P. linifolius (Lehm.) Johnston. Perennial; stems trailing, rooting at the nodes, slender with the internodes usually equalling or surpassing the leaves in length, very loosely branched, not congested into a mat; leaves narrowly oblanceolate or linear-oblanceolate, 1-2.5 cm. long, 1.5-3 mm. broad, all opposite, acute, very sparsely hispid- strigose, usually somewhat ciliate with strict hairs, noticeably con- tracted towards the dilated papery narrowly connate base; flowers solitary in the axils of the opposite cauline leaves; calyx sparsely ap- pressed-villous, with erect or spreading lance-linear lobes, becoming 2.5-4 mm. long; pedicels becoming evident, 1—1.5 mm. long, spreading; corolla white, 4-6 mm. broad, tube 2-2.3 mm. long and 1.2-1.3 mm. 88 JOHNSTON thick; nutlets ovate with a rounded base and acute apex, 1.5-1.9 mm. long, roughened with rounded irregularly reticulate ridges which occasionally are obscure below the middle, medial dorsal keel usually definite only towards the apex, areola suprabasal; style shorter than the nutlets.—Johnston, Contr. Gray Herb. Ixviii. 74 (1923). Anchusa linifolia Lehm. Asperif. i. 215 (1818). Antiphytum linifolium DC. Prodr. x. 121 (1846). Eritrichium linifolium Wedd. Chlor. ii. 89 (1859). Krynitzkia linifolia Gray, Proc. Am. Acad. xx. 266 (1885). Allocarya linifolia Macbr. Proc. Am. Acad. li. 545 (1916). Anchusa oppositifolia HBK. Nov. Gen. et Sp. iii. 91, t. 200 (1818). ECUADOR. InpeFrIniTE: Quitensian Andes, 1855, Couthouy (G). ; dtr NariNo: Meneses, April 29, 1876, André 2906 (G, NY, This species was based upon collections made in southwestern Colombia, Prov. Narifio, between Pasto and the Ecuadorean border. Weddell reports it from El Quindio apparently upon the basis of a collection by Triana. Though certainly distinct, the species seems to be most related to P. linifolius of Ecuador. 13. P. Macbridei, sp. nov. Annuus; caulibus gracilibus pluribus prostratis vel laxe ascendentibus 3-7 cm. longis adpresse breviterque villosis; foliis linearibus 2-3 cm. longis 1-2 mm. latis acutiusculis basem versus paullo dilatatis breviter sparseque hispido-villosis in- ferioribus oppositis; racemis laxe floratis ubique bracteatis; bracteis alternis linearibus 0.5-2 cm. longis; calycibus adpresse villosis ma- turitate 2-2.5 mm. longis, lobis lanceolatis érectis; corolla inconspicua subtubulari ca. 1 mm. lata; nuculis ovatis plus minusve angulatis ca. 1.5 mm. longis irregulariter tuberculatis vel papillatis obscurissime dorso carinatis; areola suprabasali grandi triangulari concava vel plus minusve excavata. PERU. Lia: in short grass, Viso, 2700 m. alt., May 1922, Macbride & Featherstone 599 (FM, rypx; G, 1soryPe). Clearly related to the Chilean P. procumbens and perhaps only a form of it. More collections of this plant are needed before its exact status can be finally decided. In lacking any transverse ridges and being merely tuberculate or papillate its nutlets are readily distin- guished from those of P. procumbens. The size and general form of the nutlets, however, very much suggest that species as does also the . large more or less excavated scar. 14. P. procumbens (Colla) Gray. Annual; stems usually several, prostrate or laxly ascending, strigose or appressed short hispid-villous, 5-25 em. long, usually rebranched; leaves linear or spathulate- or SOUTH AMERICAN BORAGINOIDEAE 89 oblance-linear, 1-3.5 cm. long, 0.8-2.3 mm. broad, obtusish, lower- most opposite; racemes elongate, leafy-bracted; bracts alternate, in- terrupted, 1-1.5 cm. long; calyx rather densely appressed-villous, somewhat tawny, at maturity 2-4 mm. long with linear or lanceolate erect or loosely ascending lobes, shortly pedicellate; corolla inconspicu- ous, 1.5-2.5 mm. broad; nutlets ovate, more or less angulate, 1.3—1.7 mm. long, reticulate-rugose and granulate-tuberculate, with the roughenings prominent, frequently (especially on the somewhat heter- omorphous axial one) the ridges partially or completely replaced or sometimes surmounted by murications papillae or even subulate fre- quently glochidiate appendages, medial dorsal keel always definite at least above the middle, margins frequently ridged and hence somewhat angulate, ventral face with a large deltoid more or less ex- cavated suprabasal scar lying usually below the level of the well developed ventral keel; axial nutlet frequently differentiated being usually dulled with minute pubescence-like spicules and having the back more papillate and frequently provided with glochidiate ap- pendages; style short, surpassed by the nutlets; gynobase hemispheric- al or pyramidal.—Proc. Am. Acad. xx. 283 (1885); Johnston, Contr. Gray Herb. Ixviii. 80 (1923). Myosotis procumbens Colla, Mem. Acad. Torino xxxviii. 130 (1834). Eritrichiwm procumbens DC. Prodr. x. 133 (1846); Clos in Gay, Fl. Chile iv. 470 (1849); Ph. Anal. Univ. Chile xc. 543 (1895). Allocarya procumbens Greene, Pittonia i. 17 (1887); Reiche, Anal. Univ. Chile exxi. 807 (1908) and FI. Chile v. 212 (1910). E. tenuicaule Ph. Linnaea xxix. 18 (1857). A. tenuicaulis Macbr. Proc. Am. Acad. li. 544 (1916). (?) E. illapelinum Ph. Anal. Univ. Chile xe. 548 (1895). ARGENTINA. Cuasvut: Rio Palena, March wd 1900, ir orn (G). CHILE. »LCHAGUA: Malloa ae 20, sane a : ‘ , no or given (MS, TYPE E. ilapelinum; G, photo.); Cerrillos, t. Ovalle, 1917, Baeza (IP). InpEFinire: no locality given, Bertero 445 (G); Quillota and Rancagua, Bertero 1 i159 and 443 (NY}. This is a very common and very frequently collected plant of central 90 JOHNSTON Chile and apparently the predominating one of the section Allocarya in the Santiago-Valparaiso region. Very clearly it is the Myosotis procumbens of Colla. It has been the plant most commonly associ- ated with Ruiz & Pavon’s Lithospermum muricatum, though in most cases with some doubt. That latter species is said to have been based upon material from Concepcion, which is quite south of the known range of P. procumbens in Chile and whence I have seen no material that is satisfactorily covered by Ruiz & Pavon’s description. Litho- spermum muricatum is described as having muricate nutlets which is not the case in P. procumbens. Being quite uncertain as to the iden- tity of L. muricatum R. & P., I have been forced to list it among the poorly understood and unrecognized species. Plagiobothrys procumbens is closely related to P. Greenei and related forms of California. It is a less robust plant, with small lanceolate calyx-lobes and somewhat smaller nutlets than in P. Greenei, but is quite similar and probably conspecific with P. Piperi Johnston, cf. Contr. Gray Herb. Ixviii. 75 (1923). The Chilean plants vary in the roughenings of the nutlets. There are two common forms which apparently grow in the same places. In one form the nutlets are alike, all of them being reticulate-rugose. In the other form the axial nutlet is more or less differentiated, usually being duller in color and having rather prominent frequently more or less glochidiate papillae or subulate appendages, which either replace or surmount the reticulate ridging. Practically all the specimens I have seen fall into one or the other of these two forms. These forms most decidedly do not have distinct ranges. In a few cases I have noted that the early flowers on a plant had the axial nutlet differen- tiated, whereas the later flowers produced only homomorphous nutlets Another very much rarer variation is represented by Baeza’s col- lection from the Province of Coquimbo. This has nutlets slightly smaller but otherwise quite indistinguishable from those particular forms of P. Greenei, sensu lat., which Piper, Contr. U. S. Nat. Herb, xxii. 82 and 89 (1920), treated as Allocarya Eastwoodae and A. Greenci. In the Chilean form mentioned the nutlets are all alike, 1.5 mm. long, dull and are all armed with abundantly glochidiate subulate-append- ages. The scar is very deeply excavated. More of this peculiar form is needed so that its status can be determined. 15. P. polycaulis (Ph.), comb. nov. Annual; stems 5-15 cm. tall, one to several, loosely ascending to erect, strigose; leaves linear - ob: i r, 2-3 cm. long, 1-2 mm. broad, strigose; racemes in- terruptedly bracteate or bracted only towards base, becoming loosely towered; calyx usually densely appressed hispid-villous, commonly SOUTH AMERICAN BORAGINOIDEAE 91 fulvescent, becoming 2.5 mm. long with erect or strictly ascending oblong-lanceolate or broadly linear lobes; pedicels short; corolla small, 2-2.5 mm. broad; nutlets ovate often broadly so, ca. 1 mm. long, low reticulate-rugose, keeled dorsally, rounded laterally; scar suprabasal, small, deltoid, solid, on or slightly above level of the ventral keel; gynobase broadly pyramidal; style surpassed by nutlets.—Eritrichium polycaule Ph. Anal. Univ. Chile xe. 542 (1895). EE. delicatulum Ph., l. c. 544. E. flavicans Ph. |. c. 544. E. wes oe Ph.1..o. 547; E. bracteatum Ph. |. c. 548. E. vernum Ph. |. ¢ CHILE. Va.prvia: Valdivia, meg * Philippi (MS, Tyre of £. vernum; G, photo.) Bro-Bro: Coigu e, 1915, O oa (IP). Concepcion: San Vicente, Nov. 1887, Philippi (Mor ie vane 1G.» Nuste: Chillan, 1869, Man. Ant. de Solis (MS, type of E£. delicatulum; G, photo.); ‘‘ Nuble ?, col. Puga?’ TYPE of E. graminifolium; G, Oo Curico: Prov Curico, meget idal (MS. tyPE of E. bracteatum; G, photo.). Santiaco: (?) between Colina and Batuco, Sept. 1899, Reiche (MS) Digs auenigae specimen "without data (MS, Type of E. polycaule; G, photo This plant apparently replaces P. procumbens in south-central Chile. It is clearly related to that species but is distinguished by its smaller nutlets which do not have cristate dorsal margins and have a small solid sear seldom sunk below the level of the ventral keel. The inflorescence is rather less abundantly bracteate and is frequently practically bractless. 16. P. calandrinioides (Ph.), comb. nov. Annual; stems several, prostrate or loosely ascending, 5-18 em. long, strigose; leaves oblong- linear to narrowly linear, obtusish, 1-5 cm. long, 1-2.5 mm. broad; racemes loose or dense, interruptedly pauci- to multi-bracteate; calyx strigose, usually somewhat tawny above, becoming 2-3.5 mm. long, with erect or strictly ascending linear-oblong or lanceolate lobes; pedicels obscure, 0.5-1 mm. long; corolla 1-3 mm. broad, white; nut- lets oblong-ovate, 1.5-2.3 mm. long, compressed, dark and somewhat lossy, irregularly rugose with the broad ridges usually low and sinu- ately transverse, granulose-tuberculate, medial dorsal keel obscure, ventral face broadly angled; scar suprabasal, deltoid to cuneate, about flush with crest of the well developed ventral keel, somewhat oblique; style much surpassed by the nutlets——Eritrichium calandri- nioides Ph. Anal. Univ. Chile xc. 541 (1895). E. nubigenum Ph. ex. Meigen in Engler, Bot. Jahrb. xvii. 267 (1893), nomen. Allocarya sessiliflora, var. nubigena Ph. ex Reiche, Anal. Univ. Chile exxi. 806 (1908) and FI. Chile v. 212 (1910). E. albiflorum of Griseb. Abh. K. ' Ges. Wiss. Goettingen vi. 131 (1854), as to Lechler’s plant, not synonymy. P. Lechleri Johnston, Contr. Gray Herb. Ixviii. 79 (1923). 92 JOHNSTON o Negr Oct. 0, “1914, * Fisher 135 “G, US, FM). Sanra Gas. arid subsaline places near Rio San a Cruz, Feb. 19, 188 2; gl ni (G); Isla Pavon, Rio Santa Cruz, Feb. 9, Ts82, Spegazzini (G); wet places seta Rio Deseado, Jan. 1899, Ameghi ino (G); Patagonia, lat. 50° “53°, Moreno & Tonini 529 and an atin ered collection (NY). This species appears to be the common Patagonian member of the section Allocarya. It has been much confused and has been misde- termined as E. humile, E. uliginosum and E. procumbens. With the exception of a collection of P. procumbens from the Cordilleras of Chabut, P. calandrinioides is the only species of its section that I have seen from Patagonia or Fuegia. It is very readily recognized by its large oblong-ovate nutlets which are compressed and are marked by broadly spaced wide, low, more or less sinuous transverse ridges. Its nearest relative is P. oppositifolius. The range of P. calandrinioides is interesting and perhaps worthy of some comment. The plant, though seeming to be primarily Pata- nian, occurs on the western slope of the Cordilleras above Santiago. Although the Cordilleras are highest in this district a number of herbaceous Boraginaceae, commonly of low altitudes, appear to cross the continental divide there. The most striking examples of these are Cryptantha globulifera, Amsinckia tessellata and Plagiobothrys ver- TUCOSUS 17. P. oppositifolius (Ph.), comb. nov. Annual; stems several, prostrate, 5-15 cm. long, subsimple or loosely branched, loosely ap- pressed short hispid-villous; leaves broadly linear, 1-2 cm. long, 1.5-3 mm. broad, obtusish, sparsely appressed hispid-villous; racemes bract- ed nearly throughout, becoming loosely flowered and 3-5 em. long; calyx densely appressed hispid-villous, becoming 2-3 cm. long with lanceolate or oblong-linear erect or strictly spreading lobes; pedicels ming 1 mm. long; corolla 2-3 mm. broad; nutlets narrowly to broadly ovate, 1.5-1.8 mm. long, thick, not strongly compressed, ir- regularly rugose with the ridges usually broad and low and frequently absent or obscure particularly below the middle, medial dorsal keel and usually rounded and obscure below the middle, ventral face prominently angled; scar almost basal, oblique, small and solid; gyno- broadly pryamidal; a much surpassed by the nutlets.—Eri- trichium oppositifolium Ph. Anal. Univ. Chile xc. 542 (1895). Allo- carya i ie ale Reiche, Ana. Univ. Chile exxi. 807 (1908) and FI. SOUTH AMERICAN BORAGINOIDEAE 93 Chile v. 212 (1910). EE. cinereum Ph. 1. ¢. 545. A. cinerea Reiche, l.c. 808 and 1. c. 213. E. limoniwm Ph. |. c. 546. CHILE. Matteco: Ercilla, Araucania, Nov. 1887, Philippi (MS, TYPE of E. limonium; G, photo.); Araucania, Nov. 1887, Philippi (MS, TyrE of cinereum; G, photo.). CautTin: Temuco, 1915, Baeza (IP). INDEFINITE: without data (MS, Tree of E. oppositifolium; G, photo.). A close relative of P. calandrinioides differing in its weakly com- pressed nutlets and more southerly Chilean range. 18. P. pulchellus (Ph.), comb. nov. Annual; stems several, prostrate or decumbent, 1-1.5 dm. long, branched, appressed short- villous; leaves linear, 1-3 cm. long, 1.5-2.5 mm. broad, very sparsely strigose, margins sparsely short-ciliate and pustulate; racemes 2-5 em. long, usually geminate, bractless; calyx densely appressed short- villous, fulvescent especially at tips, becoming 2-3 mm. long, with oblong-lanceolate erect or spreading lobes; pedicels ca. 1 mm. long; corolla 4-5 mm. broad; nutlets ovate, 1.3-1.5 mm. long, obscurely an finely granulate, sparsely papillate, occasionally obscurely transverse- rugose towards the base, medial dorsal keel very sharp above the middle; scar suprabasal, deltoid, concave, clearly below the level of the strong ventral keel; gynobase globose-pyramidal; style just sur- passed by nutlets.—Eritrichium pulchellum Ph. Anal. Univ. Chile xe. 545 (1895). CHILE. Ma.teco: [Traiguen?] Araucania, Nov. 1887, Philippi (MS, TYPE; G, photo.). I was unable to locate in the Philippi Herbarium any material given as from Traiguen, the locality cited by Philippi when he described the species. The species is represented at Santiago by two plants of the same collection which have been both labeled and determined by Philippi. These have only the following data: “Eritrichium pul- chellum Ph., Araucania, Nov. 1887.” During Nov. 1887 Philippi was collecting in the province of Malleco, cf. Gartenflora xxxviii. 88-90 (1889). Since Traiguen is in that province perhaps the locality was supplied from memory in an attempt to give amore precise source for the specimen than merely the vague indefinite regional name, “ Araucania.”’ As defined by Philippi his species is an aggregate. The two plants in the herbarium at Santiago represent different species, one being the plant here treated as P. pulchellus, the other apparently a phase of P. corymbosus. The original description is very ambiguous, though in the majority of points it seems to fit the plant I have described above especially so where the nutlets are described as “ mui arrugadas” and the hairs of the calyx as “amarillentos.” 94 JOHNSTON The species as here defined is most related to P. corymbosus. In having papillate rather than rugose nutlets, however, the plant is so different that it seems best treated as specifically distinct. Although the relationships are patent I have seen no suggestion of the occurrence of transitional forms. 19. P. uliginosus (Ph:), comb. nov. Annual; stems usually solitary, weak and slender but apparently erect, becoming 2—4 dm. tall, sparsely and obscurely villous-strigose, branches few and strictly ascending; leaves linear or lance-linear, 1-5 cm. long, 1.5-2.5 mm. broad, obtusish, very sparsely villous-strigose, mostly opposite; racemes usually geminate, bractless, 3-15 cm. long, becoming elon- gated and loosely flowered; calyx densely appressed tawny-pubescent, becoming 2-3 mm. long, with the narrowly lanceolate lobes erect to spreading; pedicels usually slender, becoming 1-3 mm. long and some- what spreading; corolla conspicuous, white, 4-6 mm. broad; nutlets ovate, ca. 1.7 mm. long, pale, granulate, reticulate-rugose with the rugae narrow and sharp, medial dorsal keel definite, ventral keel strong, scar nearly basal; gynobase hemispheric-pyramidal; style reaching to tip of nutlets.—Eritrichium uliginosum Ph. Anal. Univ. Chile xliii. 519 (1873). Allocarya uliginosa Greene, Pittonia i. 14 (1887); Reiche, Anal. Univ. Chile exxi. 807 (1908) and FI. Chile v. 213 (1910). Cynoglossospermum uliginosum Kuntze, Rev. Gen. iii. pt. 2, 204 (1898). (?) E. tenuifolium, var. pulchellum Ph. Anal. Univ. Chile xe. 547 (1895). (?) A. tenurfolia, var. pulchella Reiche, |. c. 805 and I. ¢. 211. CHILE. Nuste: Chillan, Reed (G); “Prov. de Chillan,” Dec. 1869, Philippi (MS, tyre of E. uliginosum; G, photo.). LypEerrnire: (?) collection without data (MS, type of E. tenuifolium var. pulchellum; G, photo.). When E. uliginosum was described the type specimens were given as from the province of Colchagua. This was probably a slip of the pen, for the only specimens in the herbarium at Santiago covered by the original description are those labeled, “ Eritrichium uliginosum Ph. Prov. de Chillan, Decemb. 1869.” The species is closely related to P. corymbosus but differs in its more sharply ridged nutlets, and in its more slender, erect and taller habit. It is positively known only from Nuble, which is some distance north of the range of its relative. The type of E. tenuifolium, var. pulchellum is doubtfully associated with P. uliginosus. It agrees with the present species in the sculp- flowers, but differs from the Chillan material in its low (only 5-7 cm. tall) stoutish habit and conspicuously rufous pubescence. The source of the type material is not known. SOUTH AMERICAN BORAGINOIDEAE 95 20. P. pedicellaris (Ph.) Johnston. Annual; stems solitary or few, trailing, subsimple, 5-10.cm. long, finely and closely short-strigose; leaves linear, 3-5 em. long, 1-1.5mm. broad, very gradually contracted towards the acutish apex, finely and closely short-strigose; racemes not much if at all surpassing the leaves, 1-3 em. long, leafy-bracted at the base, very loosely flowered; pedicels slender, those of the lower flowers elongate and spreading, becoming 8 mm. long; calyx appressed short-villous, fulvescent, becoming 3 mm. long, with erect slender linear lobes; corolla conspicuous, ca. 4 mm. broad; nutlets ovate, ca. 1.3 mm. long, reticulate-rugose, with the ridges narrow but not promi- nent, medial dorsal and ventral keels fairly well developed; scar just above the base of the nutlet, small, oblique; gynobase low, broadly pyramidal; style elongate, surpassing the nutlets—Contr. Gray Herb. Ixviii. 75 (1923). Eritrichium pedicellare Ph. Anal. Univ. Chile xc. 549 (1895). Allocarya pedicellaris Reiche, Anal. Univ. Chile exxi. 809 (1908) and FI. Chile v. 214 (1910). E. tenuifolium, var. longipes Ph. Anal. Univ. Chile xliii. 518 (1873). A. tenuifolia, var. longipes Reiche, |. c. 805 and lI. e. 211. CHILE. Inperinite: Pinales de la cordillera de ise arene . —” (MS, tyres of E. pedicellare and E. tenuifolium var. longipes; ji Philippi’s species and variety cited above are alk based on specimens of the same collection and are unquestionably synonymous. The exact type-station is somewhere in the Cordillera de Nahuelbuta, the range of mountains whose crest is the boundary between the province of Arauco and adjacent Bio-Bio and Malleco. The species is closely related to P. corymbosus. It appears to differ, however, in its shorter subsimple very leafy stem, racemes which are not projected above the leaves, and very elongate lower pedicels. Before the species can be definitely accepted, however, more material should be studied and some idea gained as to the constancy of the characters here given. 21. P. corymbosus (R. & P.), comb. nov. Annual; stems usually several, decumbent, branched, 5-20 cm. long, finely strigose; leaves linear to filiform-linear, 1-5 em. long, 2—-2.5 mm. broad, mostly opposite, finely and sparsely strigose; racemes elongating, slender, bractless, fre- quently geminate, becoming 4-6 cm. long; mature calyx 2-3 mm. long, usually very tawny with a rather dense appressed pubescence, with the lance-linear lobes erect or ascending; pedicels 0.5-2 mm. long, usually strict and stiffish; corolla evident to conspicuous, white or cream- colored, 4-6 mm. broad; nutlets ovate, 1-1.3 mm. long, finely granu- late, irregularly somewhat reticulate-rugose with the ridges low and rather obscure, dorsal keel obscure or definite only towards apex, 96 JOHNSTON ventral keel definite; scar suprabasal, oblong, solid; gynobase pyra- midal; style shortly surpassing the nutlets—Myosotis corymbosa R. & P. Fl. Peruv. ii. 5 (1799); Lehm. Asperif. i. 82 (1818); Reiche, Anal. Univ. Chile exxi. 834 (1908) and Fl. Chile v. 239 (1910). Cryptantha corymbosa Johnston, Contr. Gray Herb. Ixviii. 54 (1923), Cynoglos- sum sessiliflorum Poepp. in herb. Eritrichium sessiliflorum DC. Prodr. x. 133 (1846); Clos in Gay, FI. Chile iv. 470 (1849). Allocarya sessilifolia Greene, Pittonia i. 17 (1887). A. sessiliflora Reiche, |. ¢ 806 and |. c. 211. E. humile, var. capitatum Clos in Gay, |. c. 471; Gray, Proc. Am. Acad. xx. 267 (1885). E. tenuifolium Schlecht. ex sched.; Ph. Anal. Univ. Chile xliii. 519 (1873), nomen; Ph. 1. e. xe. 546 (1895). Krynitzkia tenuifolia Gray, Proc. Am. Acad. xx. 267 (1885). A. tenuifolia Greene, Erythea iii. 57 wee Hoahe ce. 805 and |. ¢. 210. P. tenuifolius Johnston, |. ¢. HILE. Luanquruve: playa del puerto de Nahuelhuapi, collector not given, no. 37 (MS). Vativia: Valdivia, Oct. 1857, Philippi ( MS); Canelos, collector not ee no. 295 (MS); Rafe ‘Jan. 1887, Otto (MS): shore of Lago Soniget Jan. 60. Phitzpps (MS); Arique, Nov. 1854, Lechler 255 (MS); i Nov. 1850, .Lechler 255 (G); streambank, Valdivia, 1899, Tackhice (US); Valdivia, 1896, Saag (US); Valdivia, 1862, Bridges 785 m. a 3 ; H 140 t., 1924, Hollerma ); Panguipulli, 1924, Claude-Joseph 2693 (US) LLECO: Collipulli, 1915, Baeza ( )5 [? Tr: ] Araucania, Nov. 1887, Phil - (MS). Bro-Bro ei 97 A very variable species that os study may cause to be further enlarged so as to include P. pulchellus, P. uliginosus and P. pedicellaris. As here defined, however, the species includes large-flowered annuals with a sparse usually reddish-brown pubescence, short pedicels and nutlets with low irregularly reticulate ridges. It is one of the more southerly ranging of the Chilean species and appears to be quite common within the area of its dispersal. Tam accepting Plagiobothrys corymbosus as the proper name for the present species since Dr. A. Brand, who has seen authentic material, writes me that Myosotis corymbosa R. & P. is an Allocarya apparently of this relationship. The original description of M. corymbosa gives, “corolla calyce triplo major.’”’ The species described above is the only large-flowered one growing in southern Chile that seems to fit the original description. The type is said to have come from Con- cepcion. I refer E. sessiliflorum here after a study of a photograph of the type. The type specimen seems to have evident corollas and the characteristic dark dense pubescence on the calyces. Poeppig, who collected the type, explored in the province of Concepcion and SOUTH AMERICAN BORAGINOIDEAE 97 Bio-Bio and hence within the known range of the species as here defined 22. P. pratensis (Ph.), comb. nov. Annual or perhaps persistent; stems several, prostrate or decumbent, 1-2 dm. long, densely strigose; leaves 1-2.5 cm. long, oblong-spathulate to linear, obtuse, finely strigose; racemes bractless, 2-3 cm. long, densely flowered; calyx densely silky-strigose, becoming 2-3 mm. long with erect lance-oblong or broadly linear obtuse lobes; pedicels 1-2 mm. long; corolla 3-4 mm. broad; nutlets broadly ovate, 1.5 mm. long, finely granulate, rough- ened with low rather inconspicuous reticulate ridges, dorsal keel rounded but definite to below the middle; areola almost basal, small, oblique, clearly much below the level of the strong ventral keel; style surpassed by nutlets or about reaching to their tips.—Eritrichum pratense Ph. Linnaea xxxili. 192 (1864); Reiche, Anal. Univ. Chile exxi. 810 (1908) and FI. Chile v. 215 (1910). CHILE. Luanquinve: between Roble and Pilmaiquen, Dept. Osorno, Jan. 1861, Philippi (MS, TypE; G, photo.). Both Philippi and Reiche have described this species as perennial. In the Museo Nacional at Santiago there are two plants of the original collection, both determined as E. pratense by Philippi. One of the plants is unquestionably annual, the other has a thickened root, but is otherwise similar. It is evident that Philippi’s and Reiche’s state- ments were based upon a study of the plant with the thickened root. This specimen exhibits the upper 1.5 cm. of the root which is about 4 mm. thick and bears at its crown what appears to be vestiges of old stem-bases. I suspect that the plant has been browsed upon and that its root-development is not at all typical of the species. DovuBTFUL oR EXCLUDED SPECIES. AMSINCKIA HUMIFUSA [Poepp.] Walp. Nov. Act. Acad. Caes. Leop.-Car. xix. suppl. 371 (1843); A. DC. Prodr. x. 133, note (1846). Cynoglossum humifusum Poepp. in sched. ex Walp. 1. ¢. 372; Brand, Pflanzenr. iv. Fam. 252, i. 150 (1921). Benthamia humifusa Druce, Rep. Bot. Exch. Cl. Brit. Isl. iv. 298 (1916).—This plant is evidently a member of the section Allocarya. Its aor however, is too vague for recognition. See discussion under P. congestus. hrys mesembryanthemoides (Speg.) pchectieteg Contr. ps Herb. Ixviii. 79 (1923) and I. c. xxv. 42 (1925). Eritrichium mesem- bryant. ides Speg. Anal. Soc. Cient. Argentina lili. 136 (1902).—Not a species of the Boraginaceae. This appears to be a member of the Portulacaceae and perhaps a CALANDRINIA. 98 JOHNSTON PLaciopoTurys Muricatus (R. & P.) Johnston, Contr. Gray Herb. Ixviii. 79 (1923). Lithospermum muricatum R. & P. Fl. Peruv. ii. 4 (1799); Lehm. Asperif. ii. 327 (1818). Eritrichium muricatum A. DC. Prodr. x. 132 (1846); Ph. Anal. Univ. Chile xc. 540 (1895). Allocarya muricata Reiche, Anal. Univ. Chile exxi. 809 (1908) and FI. Chile v. 215 (1910).—The type of L. muricatum is said to have come from Concepcion, Chile. It is described as a prostrate annual herb. with opposite leaves, small inconspicuous corollas and muricate nutlets. Possibly it is the same as P. polycaulis which has been collected near Concepcion. That species is an inconspicuously flowered plant but its nutlets are certainly not muricate. 12. Amsinckia Lehm. Calyx cut to base into erect lanceolate or oblong lobes. Corolla tubular or salverform; tube cylindrical, glabrous, unappendaged ; lobes obtuse. Style filiform, included; stigma capitate, emarginate. Ovules 4. Cotyledons 2-parted. Nutlets 4, erect, angulate-ovoid, smooth or rough, unmargined, strongly keeled ventrally; areola infra- medial, small, carunculate. Gynobase frustate, about half the height of the nutlets—Annual herbs. Leaves alternate, linear to ovate, usually veinless. Racemes usually bractless——Del. Sem. Hort. Hamburg 7 (1831). Benthamia Lindl. Nat. Syst. 241 (1830), nomen. A genus of considerable technical difficulty which centers in western North America and has only two South American representatives. Key To Specigs. gigpecsSoemeem a 3-3.5 mm. long; corolla 6.5-7.5 mm. long; Niitia = Aree See : oe re Cae. cea ais foes i ieee fe 1. A. tessellata. el se, 2-2.8 mm. long; corolla m usually slender m. long; plants Pi ad eae ew ees ee sa bea Ce ke 2. A. hispida. . Amsinckia tessellata Gray. Rather coarse plants, 8-20 cm. on with loosely or strictly ascending stems, shaggy-hirsute with the younger parts also somewhat hispid-villous; leaves firm, obtuse, ap- pressed-hispid, usually densely pustulate; lower leaves somewhat crowded, oblanceolate, 4-6 em. long, 7-9 mm. broad; middle cauline leaves gradually reduced, oblong to oblong-ovate; racemes usually short and densely flowered, 1-3 cm. long, solitary, bractless or bracted only y appressed 1 tawny-hirsute as well as short a pressed white-villous, t y 6-8 mm. long with erect coarse ase linear obtuse lobes; pedicels ca. 1 mm. long; corolla yellowish, 6.5- SOUTH AMERICAN BORAGINOIDEAE 99 7.5 mm. long, subtubular, limb ca. 2 mm. broad, throat weakly flaring, tube ca. 1 mm. thick; nutlets triangular, angulate, 3-3.5 mm. long, tessellate with flattened very crowded tuberculations, frequently with short transverse rugae along the medial line, the back very broadly obtuse, almost flat, with a weakly developed medial ridge.— Proc. Am. Acad. x. 54 (1875); Macbr. Contr. Gray Herb. xlix. 6 (1917). ARGENTINA. Cuasour: along Rio Carren-leufti, March 1, 1900, Spegaz- 2U is "CHILE. Sant1aco: Las Arafias silver mines, Nov. 1861, Philippi (MS). This species, frequent in the deserts of southwestern United States and now first reported from South America, was long ago known and distinguished by Spegazzini, Anal. Soc. Cient. Argentina lili. 137 (1902), who treated it as A. pseudolycopsicoides under the supposition ported the plant from Rio Carren-leufti and Lago Nahuel-huapi. Al- though apparently indistinguishable from some of the low spreading forms from North America, the two South American collections studied are not so tall nor so erect as the common forms of the species in the deserts of United States. 2. A. hispida (R. & P.) Johnston. Erect annual, 2-7 dm. tall, simple or loosely branched, conspicuously shaggy-hirsute and with the upper parts frequently finely villous-hispid as well; leaves hirsute, acute, gradually reduced up the stem; lower leaves usually crowded, oblanceolate, 5-15(-18) cm. long, 5-15(-24) mm. broad; cauline leaves linear to lanceolate, usually strictly ascending; racemes solitary or geminate, bractless or bracted only at base, much elongating, be- coming very loosely flowered and 1-2.5 dm. long; calyx somewhat tawny, hirsute and more or less hispid-villous, becoming 4-6 mm. long with acuminate-linear or lance-linear lobes; pedicels ascending, 1-3 mm. long; nutlets triangular-ovate or triangular, angulate, incurved, 2-2.8 mm. long, granulate-tuberculate, commonly with strong ir- regular transverse crests, back definitely obtuse, the medial ridge frequently longitudinally cristate—Johnston, Contr. Gray Herb. Ixxiii. 75 (1924). Lithospermum hispidum R. & P. Fl. Peruv. i. 5 (1799); Lehm. Asperif. ii. 328 (1818). A. angustifolia Lehm. Del. Sem. Hort. Hamburg 7 (1831) and Linnaea vi. Litt. 74 (1831); F. & M. Ind. Sem. Hort Petrop. ii. 26 (1835); DC. Prodr. x. 118 (1846); Clos in Gay, Fl. Chile iv. 473 (1849); Reiche, Anal. Univ. Chile cxxi. 833 (1908) and FI. Chile v. 238 (1910); Johnston, Contr. Gray Herb. Ixx. 44 (1924). Benthamia angustifolia Druce, List Brit. Pl. 103 (1908). A. parviflora Bernh. Del. Sem. Hort. Erf. (1833) and Linnaea x. Litt. 100 JOHNSTON 73 (1836). L. calycinum Moris, “Enum. Sem. Hort. (1831)” and Mem. Acad. Torino xxxvii. 98, t. 22 (1833). L. chilense Colla, Mem. Acad. Torino xxxviil. 127, t. 40 (1834). A. angustifolia, var. pseudo- lycopsioides Clos in Gay, FI. Chile iv. 473 (1849). A. pseudolycopsi- coides Speg. Anal. Soc. Cient. Argentina liii. 137 (1902). Eritrichium pachnophilum Wedd. Chlor. Andina ii. 87 (1859). A. basistaminea Cesati, Atti Acc. Sc. Nap. ser. 5, vii. 14 (1873). B. basistaminea Druce, Rep. Exch. Cl. Brit. Isl. iv. 298 (1916). E. Mandonii Ball, Jour. Linn. Soc. Bot. xxii. 51 (1885). A. angustifolia, var. microcarpa Speg. Anal. Soc. Cient. Argentina liii. 136 (1902). Cryptantha Spegaz- zintt Johnston, Contr. Gray Herb. Ixviii. 54 (1923). Plagiobothrys aurantiacus Brand in Fedde, Repert. xx. 46 (1924). ARGENTINA. Santa Sad Bajo of sil Argentina, 1907-08, Fur- long 120 (G, NY); Patagonia, lat. 50°-53°, Moreno & Tonini 106, 309 and 6381 (NY). Cuasor: om Sabo “shoe Feb - zint (G); no locality given, 1898, Kowbowky ( EGRO vicinity of General Roca, 2. . alt., Sept. 25, 1914, Fischer 81 (G, US, ; dry eerily: a7 at junction of Rio Neuquen and Rio Negro, Jan. 21, 1898, pegazzini CHILE. Me EGALLANES Negro, Nov. 1867, Cunningham (G); Packet Harbor, Nov. roca 1133 Qs), Oun Harbor, Nov. 1852, Lechler 1133 (MS); Fuegia, Feb. 1879, (MS). : Chillan, Man. Ant. de Solis (M OHigains: Sash a, ‘July iss, lio not given en (MS); Rancagua, fae 379 (G). Santraco: Santi o, Aug. 1830, Gay 1628 (MS); ae te: Nov. 1860, no collector given (MS); "San ntiago, et. 1858, no collector given (MS 3 ntiago, Noy. 1918, Claude-Joseph 706 (US); Santiago, Gay 48 G ghee of : an P); Quinta Normal, Nov. 1921, Baeza (IP); Talagante, 1916, J Baeza (IP). Vatparaiso: Limache, Sgro (US); hills, Quillota, 1829, ; i 2 RA PON a aie, Warierinane 761 (G). InpeEFIniTE: costa, Non, rp 1920, Cidudedenek 1210 (US); no e las la: unas, March, D’Orbigny 1448 (G, frag. TYPE of Reine). ines azna, 3900 m. alt., Apr. 3, 1921, Asplund 5893 Be). La Paz: General Campero, 4200 m. alt., March 4, 1921, splur Casevays ater Sorata, 2700-2900 m. alt., ’Mandon 378 (G, PERU. Cusco: aoe 3500-3700 m. alt., Herrera 1006a (US). (G : fe yaya ae Rio Blanco, 4500 m. a a , March, 1923, Macbride Pee . ‘ yon slope, Ri co, m. — May 1922, Featherstone 729 (G FM), Chie alt, (Apa 1882, 7: , 3600-3900 Ball (NY). -Huanuco: edge of alf: & Featherstone 2288 (G, FM). alfe-fild, Lata, 2100°m, SOUTH AMERICAN BORAGINOIDEAE 101 ECUADOR. CaNar: borg of Cafiar Sarg ta, sid Rose 22728 (G, NY). Picurycua: plains of Chillagallo, Sept y (G, fragment). INDEFINITE: without data but probably eect see 662 (N wer Considering the extent of its range, this species is quite constant and shows very little of that capacity for bewildering variation which is so characteristic of its immediate relatives in North America. In habit and corolla-structures and -form, the plant is not unusually variable. The most striking variations are those of nutlet-size. The majority of the specimens studied had nutlets about 2.5 mm, long. In Asplund’s Bolivian collections, and in that cited from extreme northern Chile, the nutlets are slightly larger, becoming 2.8 mm. long. On the other hand in the material sent me by Spegazzini from the confluence of the Rio Negro and Rio Neuquen as representing A. an- gustifolia, var. microcarpa, the nutlets are about 2 mm. long. While worthy of note these variations in size of nutlets scarcely merit no- menclatorial recognition. 13. Lappula Moench. Calyx 5-parted into spreading lanceolate lobes. Corolla with a rather short tube; lobes rounded, ascending, imbricate; throat closed by intruded appendages; stamens affixed in the tube, included; fila- ments slender, short; anthers oblong, obtuse. Style short surmount- ing the subulate-columnar gynobase, commonly surpassing the mature nutlets; stigma subcapitate. Ovules 4. Nutlets 4, erect, smooth or verrucose, narrowly but firmly attached to the gynobase along the length of the well developed ventral keel, back angulate or margined by a single or double row of prickles which by confluence frequently form a wing-like or cupulate border—Annual or rarely perennial herbs. Leaves alternate, usually narrow, firm and veinless. Flowers small, blue or white, on usually erect pedicels or rarely subsessile, in bracted racemes.—Meth. 416 (1794); Johnston, Contr. Gray Herb. Ixxiii. 60 (1924). Echinospermum Sw. ex Lehm. Asperif. i. 113 (1818). Staurina Nutt. Jour. Acad. Philad. i. 182 (1848). A very difficult genus, the South American forms of which are very imperfectly known. The species are mainly critical ones, being largely separated by technical characters of the fruit, such as size and form of the nutlets, their surface and the characters of the dorsal-margin. Key To Species. Nutlet-margin armed with a row of stout flattened glochidiate mibuinge Wien: 6 2 eee 1. L. Redowskii. Nutlet-margin bearing an up-turned cupulate wing............. 2. L. texa 102 JOHNSTON 1. Lappula Redowskii (Hornem.) Greene. Annual; stems erect, 1.5-3 dm. tall, ascendingly long-branched from the base, appressed villous-hispid; leaves appressed villous-hispid becoming somewhat glabrate, lower ones oblanceolate or spathulate, cauline ones linear or oblong; corolla small, white, 1-2 mm. broad; nutlets 2-3 mm. long, homomorphous, verrucose or rugose-verrucose, margined dorsally with a single row of compressed subulate glochidiate appendages.— Pittonia ii. 182 (1891). Myosotis Redowskii Hornem. Hort. Hafn. i. 174 (1813). Echinospermum patagonicum Speg. Anal. Soc. Cient. Argentina lin. 79 (1902). L. patagonica Macl. Rev. Fl. Patag. 214 (1915) ng ee es Santa Cruz: dry places near Puerto de San Julian, Feb. 3, Peg ree (G). CHapuT: deserts near omy near Rio Chabut, Feb” 1900, Spegazzinit (G). Rio Necro: vicin y of General Roca, Fischer 12h (G, US, FN). MeEnpoza: vicinity of pn pad Jan. 1908, Carette 256 (G). The material above cited is quite variable and I suspect that several varieties may eventually be distinguished among the Argentine forms of the species. The material available to me for study is hopelessly inadequate for a satisfactory understanding of these forms. 2. L. texana (Scheele) Britt. Annual; stems erect, 1—1.5 dm. tall, appressed hispid-villous; leaves appressed hispid-villous, lower ones spathulate, the upper ones linear or oblong; corolla small, white, 1- 1.5 mm. broad; nutlets homomorphous, more or less verrucose, with a broad toothed up-turned and hence deeply cupulate margin, the marginal teeth triangular-subulate and glochidiate—Mem. Torr. Bot. Cl. v. 273 (1894); Johnston, Contr. Gray Herb. Ixx. 49 (1924) ARGENTINA. Santa Cruz: dry places, Puerto de San Julian, Feb. * 1882, Spegazzini (G). The above cited collection seems to be indistinguishable from L. texana, var. columbiana (Nels.) Johnston, |. c., of northwestern United States. It was sent me mixed with specimens of L. Redowskii. 14. Hackelia Opiz. Calyx cut to the base into spreading ovate to oblong or lanceolate lobes. Corolla with short or elongate tube; lobes rounded, imbricate; throat with trapeziform intruded appendages; stamens included, af- \ fixed at middle of tube; filaments slender, short; anthers oblong to elliptical Style slender, scarcely if at all surpassing the nutlets; stigma capitate. Ovules4. Nutlets 4, erect, ovate, affixed ventrally to the pyramidal gynobase by a broad medial areola, margin with SOUTH AMERICAN BORAGINOIDEAE 103 subulate glochidiate appendages which are frequently confluent at the base, back smooth or with glochidiate appendages.—Biennial or perennial or rarely annual herbs. Leaves alternate, broad and veiny. Flowers in naked or inconspicuously bracted racemes paniculately . disposed. Pedicels slender, recurving in fruit.—Opiz in Bercht. FI. Boehm. ii. pt. 2, 146 (1839); Johnston, Contr. Gray Herb. Ixviii. 43 (1923). A genus centering in western North America with outlying species in South America, and Eurasia. The South American forms are very poorly understood 1. Hackelia revoluta (R. & P.) Johnston. Biennial or perennial; stems 4-8 dm. tall, erect, one to several, usually loosely branched above, appressed hispid-villous being usually antrorsely so above and retrorsely so below the middle; leaves acute, the upper face scabrous, appressed short-hispid with bulbous- or pustulate-based hairs, the lower surface paler, sparsely pubescent to velutinous, the bases de- current forming obscure angles to the stems; lower leaves with lance- olate to elliptical blades, 5-15 cm. long, 1-5 em. broad, narrowed below into a petiole 5-15 cm. long; middle cauline leaves narrowed to a sessile base, gradually reduced up the stem; racemes terminal and from upper axils; calyx cut into acute, ovate or ovate-lanceolate lobes; corolla blue or white, the limb plane or shallowly convex, 3-7 mm. broad; nutlets somewhat heteromorphous, 2.5-4 mm. long, some of them armed with long slender glochidiate appendages, others with only a few short glochidiate appendages.—Contr. Gray Herb Ixviit. 45 Sie Cynog’ ossum revolutum R. & P. Fl. Peruv. ii. 6 (1799); Asperif. i i. 147 (1818). Lappula revoluta Brand in Fedde, mae xiv. 148 (1915). C. ovatifolium Griseb. Abhandl. Ges. Wis. Goettingen xxiv. 271 (1879). JL. revoluta, f. ovatifolia Brand, |. ¢. C. Fiebrigii Krause in Engler, Bot. Jahrb. xxxvii. 634 (1906). L. revoluta, f. Fiebrigit Brand, 1 ¢ C. parviflorum Krause, |. c. C. an- dicolum Krause |. c. 635. L. costaricensis Brand in Fedde, Repert. xviii. 310 (1922). H. costaricensis Johnston, |. c. 46 L. guatema- lensis Brand in Fedde, Repert. xviii. 311 (1922). COLOMBIA. CunprnaMARCA: moist rocky canyon, Rio San Francisco above i ae 2700-2800 m. alt., Pennell 1939 (NY); meadow on sabana, Sibate, 2600-2650 m. alt., Ponnali 2473 (NY); wet forest slope on mount south of Sibate 3000 m. alt. ., Pennell 2448 (G, ; ueron, mountains east of Bogoté, Holton 530 (G). Touma: forest, Old Quindio Trail between the summit and La Ceja, 3100-3500 m. alt., Killip & Hazen 9515 (G, NY); forest below Paramo de Ruiz, 3100-3500 m. a. Pennell 3102 (G, NY). ALDAS: wayside on hill, Pinares above Salen m. alt., — 9183 os Cauca: paramo, Las Delicias, ree of Popayan, 2800-— alt., , Lehmann BT T. 673 (NY). Inperrinire: San Juan, Gaines, 104 JOHNSTON Have 8-10, 1876, André 2268 (NY); without locality, Ariste-Joseph A-856 ECUADOR. InveEFINiTE: without locality, Spruce 5554 (G) and Jameson RU. Cusco: Cusco, 3 au ay m. alt., March 1925, Herrera 525 (Us, Cusco, 3000-3600 m. , July 1923, Herrera (US). BOLIVIA. CxuqQuisaca: *Cilderillo, 3300 m. alt., Fiebrig 2476 and 3174 ° ARGENTINA. Caramarca: La Ollada, 3100 m. alt., Jan. 10, 1916, Jérgensen 1077 (G); El Cordallo, Jan. 1, 1917, Forqerion 1077 (US). The present treatment of the South American species of Hackelva is quite unsatisfactory and must be subjected to severe revision when more material becomes available for study. I feel quite sure that at least two South American species will then be recognized. The ma- terial I have cited from Colombia and Ecuador differs from that from Peru, Bolivia and Argentina in being commonly more robust and more evidently hairy, and in having tomentose growing parts, usually vil- lulose faucal appendages and larger firmer leaves that turn dark in drying. This plant of northwestern South America also occurs in the mountains of Central America whence I have studied material from Panama (Pittier 3106), Costa Rica (Pittier 7537, Dodge 3429), Guate- mala (Seler 3144, Heyde & Lux 3043) and Mexico (Ghiesbreght 115). The Guatemalan collections mentioned are the basis of Lappula guate- malensis Brand. Judging from descriptions I assume that L. cos- taricensis Brand, from San José, Costa Rica, is a small-flowered form of this same species. Two collections from Mt. Orizaba, Mexico (Botteri 179 and Seaton 22) are quite like the Central American plants but differ in their very small corollas. Since the type of Cynoglossum or Lappula mexicana came from a low mountain just to the east of Orizaba it is possible that it is the same as the plants from Orizaba and hence only a small-flowered phase of the plant of Central and northern South America. In that case the Mexican plant from Hidalgo (Coulter 1056), Mexico (Pringle 5259, 9318; Bourgeau 881 and Harshberger 130) and Oaxaca (Pringle 4832; N elson 1045) passing as mexicana is without a name. The material I have from Peru, Bolivia and Argentina is quite in- adequate as a basis of worth-while opinions as to the status of the forms and the several names applied to them. The types of Cyno- glossum revolutum R. & P.and C. parviflorum Krause, came from central Peru in the western part of the Department of Junin. The type of C. andicolum Krause was obtained in northern Peru in the Depart- ment of Cajamarca. The genus Hackelia is also represented in the _ Andes of the Department of Lima whence it has been reported by Brand. The material I have cited from the Department of Cusco SOUTH AMERICAN BORAGINOIDEAE 105 I assume, on phytogeographical grounds, to be similar to that from the departments of Lima and Junin and these latter hence different from the plant of Colombia and Ecuador. To my knowledge the genus has been reported from only two local- ities in Bolivia. Perkins, Engler Jahrb. xlix. 214 (1912), cites a col- lection from the Department of La Paz. Fiebrig made numerous col- lections at Calderillo in southern Bolivia, one of these being the type of C. Fiebrigii Krause. The collection by Fiebrig seems clearly con- specific with the material that I cite from Cusco. From Argentina Brand cites collections from the provinces of La Rioja and Tucaman. The type of C. ovatifolium Griseb. came from Tucaman. The collections I cite from Catamarca are very slender and have lanceolate leaves and small flowers and perhaps are not referable to Grisebach’s species. 15. Selkirkia Hemsley. Calyx cut to base. Corolla with tube short-cylindrical; lobes rounded, imbricate; throat with broad-trapeziform intruded append- ages; stamens included; filaments slender, filiform, short, attached at middle of tube; anthers oblong, about equalling the filaments. Style slender, scarcely surpassing the nutlets; stigma capitate. Ovules 4. Nutlets 4, winged-margined, with subulate glochidiate appendages dorsally, attached ventrally by a broad medial areola to a pyramidal gynobase.—Shrub. Leaves alternate, firm. Flowers in naked ra- cemes more or less corymbosely disposed. Pedicels slender, erect in fruit.—Bot. Challenger i. pt. 3, 47 (1884). A monotypic genus very closely related to Hackelia and perhaps best treated as a subgenus of it. It is separated from Hackelia by no structural characters, differing only in its fruticose habit, firmer foli- age, erect fruiting pedicels and more or less corymbose inflorescence. Certainly Selkirkia is less distinct from Hackelia than are the species of Eritrichium which Brand, Fedde’s Repert. xxii. 104 (1925), has re- cently, I believe improperly, referred to the genus. On the Juan Fernandez Islands, as on other oceanic islands, various genera, pre- vailingly herbaceous on the continent, have developed species that are decided shrubs. No one gives generic rank to the woody insular forms of Chenopodium, Plantago and Erigeron on the Juan Fernandez Islands and I question the advisability of doing so with the present insular shrubby manifestation of Hackelia. To those who doubt the close relationship of Hackelia and Selkirkia I commend a study and comparison of the plate of Selkirkia given by Hemsley, |. c., and that of Hackelia floribunda given by Hooker, Fl. Bor. Am. ii. t. 164 (1838). 106 JOHNSTON 1. Selkirkia Berteri (Colla) Hemsley. Shrub 1-2 m. tall; branches densely canescent-strigose, becoming glabrate, leafy on only the outermost decimeter, older parts roughened by leaf-scars; leaves ob- lanceolate or obovate-oblanceolate, somewhat acuminate, contracted towards the base into a winged petiole ca. 5 mm. long, the blade 4-9 cm. long, 1-3 cm. broad, paler beneath, very sparsely strigose, becom- ing glabrous in age but minutely roughened by the persistent tuber- cular bases of the hairs, midrib conspicuous, veins very obscure; in- florescence a corymbose panicle, bracted only in the lower part, 6-10 em. tall and about as thick, branches with slightly spreading pubes- cence; calyx strigose, the lobes ovate, imbricate, acute, ca. 1 mm. long in flower and almost 2 mm. long in fruit; pedicels slender, ascending, 5-10 mm. long; corolla white with a spreading limb 8-9 mm. long; corolla-tube 2.5-3 mm. long, 2.5 mm. thick, glabrous inside; throat open; lobes orbicular, a little longer than the tube; nutlets ca. 4 mm. long, the body compressed ovate with a broad coarsely and irregularly lobed margin, dorsally bearing several coarse compressed subulate appendages which are sparsely and weakly glochidiate at tip, surface glabrous or very sparsely strigose, anterior face with a large flat tri- angular scar; style shortly but clearly surpassing the body of the nutlets.—Bot. Challenger i. pt. 3, 48, t. 57 (1884); Ph. Anal. Univ. Chile xe. 554 (1895); Johow, Estud. Fl. Juan Fernandez 84 (1896); Reiche, Anal. Univ. Chile exxi. 250 (1907) and Fl. Chile v. 208 (1910); Skottsb. Nat. Hist. Juan Fernandez ii. 163 (1921); Brand, Pflanzenr. iv. Fam. 252, i. 16 and 163 (1921); Johnston, Contr. Gray Herb. Ixx. 45 (1924). Cynoglossum Berteri Colla, Mem. Acad. Torino xxxviii. 132, t. 43 (1834); DC. Prodr. x. 153 (1846); Clos in Gay, Fl. Chile iv. 478 (1849 LE: pepiass de Villagra, Masatierra, Skottsberg 132 (US); Juan CHI Fernandez, Downton (G); Juan Fernandez, Oct. 1872, Reed (MS); Juan Fer- nandez, Nov. 1864, Philippi (MS); Masatierra, 1892, Johow (iP). The plant is known only from Masatierra Island of the Juan Fer- nandez group. IV. Tree CyNoGLossEAk. Nutlets ascending or divergent or rarely even inverted or suberect, straight or slightly bent, appendaged or verrucose or smooth, usually : areola lateral to enh near the radicle, unmargined, not tr yramidal or columnar; style entire; stigma 3 , capitate; corolla blue or hie —Johnston, Contr. Gray Herb. Ixxiii. 69 (1924); Brand, Pflanzenr. iv. Fam. 252, i. 2 (1921). SOUTH AMERICAN BORAGINOIDEAE 107 Key To GENERA. Nutlets with glochidiate appendages; perennials or biennials with alternate leaves, lower leaves stalked; inflorescence bractless or bracted only at the base..........-..--- 16. Cynoglossum, Nutlets with uncinately tipped appendages; annuals with the lower leaves opposite and unstalked; inflorescence bracted throumhout., 0... 2. ee eee en wee wae cee ene seuk trae 17. Pectocarya. 16. Cynoglossum [Tournef.| L. Calyx cut to beyond the middle, somewhat accrescent at maturity and spreading or reflexed. Corolla cylindrical or funnelform; lobes broad, spreading, imbricate; throat with trapeziform oblong or sub- lunate appendages; stamens included, affixed in the tube; filaments short; anthers oblong or elliptical. Ovules 4. Nutlets 4, depressed- ovoid or orbicular, glochidiate, ascending or divaricate, back flat or convex, frequently with an elevated margin, attached by a small or large medial to apical scar to a convex or pyramidal gynobase and frequently with a free subulate prolongation decurrent on the short entire style.—Biennial or perennial or rarely annual. Leaves alter- nate, the basal ones long-stalked. Racemes elongating, usually ebracteate, rarely bracted at the base.—Sp. Pl. 134 (1753); Gen. Pl. 65 (1754). A cosmopolitan genus of about 50 species of which about 8 are in- digenous to America. Key To SPECIES. appendages; co ca. 1 cm. broad, conspicuously veiny; upper leaves amplexicaul...........-.------- 1. C. creticum. Nutlets 2.5-3 long, smooth between the glochidiate n ; corolla ca. 0.5 em. broad, very obscurely pper leaves not amplexicaul.........---. . C. zeylanicum. Attachment-scar not pro apically nor decurrent on the ies. Nutlets very widely spreading or divaricate, attachment- bapical and small; mature pedicels 1 em. long or co ne ep 3. C. limense. all nerved. Nutlets 2-2.6 mm. long; corolla 7-12 mm. broad; leaves usually sparsely villous-strigose above, larger ones 6-12 em. long; plant fruticulose below; Chile... .4. C. paniculatum. Nutlets ca. 6 mm. long; corolla 5-6 mm. broad; leaves glabrous above except near margins, larger ones 10- 30 em. long; plant herbaceous to base; Colombia and a ik ors coe te PS ee ee 5. C. Trianaeum. 108 JOHNSTON 1. Cynoglossum creticum Mill. Biennial; stems coarse, erect, solitary, simple, 3-7 dm. tall, densely pilose; leaves pilose, somewhat canescent; lower leaves oblanceolate, 10-18 cm. long, 2-4 cm. broad, gradually contracted below into winged stalks, obtusish; middle and upper leaves gradually reduced up the stem, lanceolate or oblong, with broad rounded sessile amplexicaul bases; racemes solitary or gemin- ate, 5-20 cm. long, disposed tually intoa!] few-branched panicle, bractless or bracted only at the base; calyx appressed-pilose, cut to near the base into broadly lanceolate or oblong lobes, 3-4 mm. long at anthesis, at maturity widely spreading and 5-9 mm. long; pedicels slender, 6-10 mm. long, recurving; corolla violet, evidently veined, broadly funnelform, 7-9 mm. long, limb 9-11 mm. broad; nutlets ovate-orbicular, depressed, 4—5.5 mm. long, unmargined, covered with glochidiate appendages and papillae, divaricate, the attachment-scar small, circular with a linear prolongation decurrent on the style; style 3-4 mm. long.—Gard. Dict. ed. 8 (1768); Brand, Pflanzenr. iv. Fam. 252, i. 129 (1921); Johnston, Contr. Gray Herb. Ixx. 33 (1924). C. pictum Soland. in Aiton, Hort. Kew. i. 179 (1789); Reiche, Anal. Univ. Chile exxi. 248 (1907) and Fl. Chile v. 206 (1910). C. molle Ph. Linnaea xxix. 18 (1857); Ph. Anal. Univ. Chile xe. 552 (1895). ARGENTINA. Buenos Arres: Cerro Redondo, Sierra Baya, Nov. 11, 1924, Castellanos (G). s ILE. Vaxprvia: Cuesta de Sote, collector not given (MS). ARauca: Canete, 1909, Tzschalran (IP); Contulmo, 1909, Johow (IP). COoNCEPCION: Tomé, Nov. 1855, Germain (MS, Tyre of C. molle; G, photo.); i 1924, Looser (Looser Herb.). Nusue: Chillan, Dec. 1869, Reed (G); between Chillan and Bafios, Jan. 1877, Philippi (MS); Nuble, 1878, Puga (M8); Valle del Renegada, Jan. 1877, Philippi (MS). Tauca: Talca Oct. 1921, Claude-J oseph 1685 (US). Santiago: San Juan i San José de Maipo, 1911, Johow (IP); Puente Alto, 1923, Baeza (IP); El Volean, 1925, Looser ser Herb.); 25, Looser (Loo Herb.); Santiago, Nov. 1924, Claude-Joseph 2868 (US). A native of the Mediterranean region well established and a pest in the fields of central Chile. It is also to be reported from eastern Argentina, where it has apparently appeared only recently for I can find no report of the occurrence of an introduced Cynoglossum in that country. In Chile, where it is well known under the name “ trupa,” it was first mistaken for an indigenous plant and described as C. molle by Philippi. Some of the material I have examined has been mistaken for C. officinale L., another species of the Old World, which is to be expected as an introduction in South America. It is similar to C. creticum but differs in having the back of the nutlets depressed, mar- gined, and lacking any papillae between the glochidiate appendages, also in having the corolla obscurely rather than conspicuously veined and the upper cauline leaves not amplexicaul. SOUTH AMERICAN BORAGINOIDEAE 109 2. CG. zeylanicum Thunb. Biennial; stéms erect, simple or loosely branched above, 3-8 dm. tall, strigose (towards the base retrorsely so); leaves evidently veiny, densely strigose, usually somewhat ciner- escent; lower leaves oblanceolate, 1—2 dm. long, 2~4 cm. broad, gradually contracted to a narrowly winged stalk; middle cauline leaves ellip- tic or lance-elliptic, sessile by a rounded or deltoid base, 4-6 cm. long, 1-2 cm. broad, reduced up the stem; racemes geminate, disposed in a loose panicle, bractless, unilateral, 5-15 em. long; calyx densely strigose, cut below the middle into obtuse ovate or deltoid lobes, at anthesis ca. 2 mm. long, slightly accrescent at maturity becoming 3-3.5 mm. long and widely spreading; pedicels 2-5 mm. long, recurving; corolla blue, broadly funnelform, 4-5 mm. long, limb concave and ca. 5 mm. broad; nutlets ovate, compressed, 2.5-3 mm. long, unmargined divaricate; attachment-scar narrowly ovate, subapical with a linear prolongation decurrent on the style; style 2-3 mm. long.—Thunb. ex. Lehm. Neue Schrift. Nat. Gess. Halle iii. pt. 2, 20 (1817) and Asperif. i. 116 (1818); Brand, Pflanzenr. iv. Fam. 252, i. 134 (1921); Johnston, Contr. Gray Herb. Ixx. 33 (1924). C. furcatum Wall. in Roxb. FI. Ind. ii. 6 (1824). CHILE. Vauprvia: Valdivia, 1899, Buchtien (US). Inperinire: ‘San Juan der Muhle,” Feb. 1891, Philippi (Ms). A species of southeastern Asia introduced in southern Chile. Brand, l. c., cites “ Anchusa zeylanica Vahl” of Hornemann, Enum. Hafn. 3 (1807) and Hort. Hafn. i. 176 (1813), as the name-bringing synonym for C. zeylanicum. Lehmann, however, in publishing Cynoglossum zeylanicum (in the synonymy of Echinospermum zeylanicum Lehm.) ’ definitely excludes the references to Hornemann’s works just cited, ef. Asperif. i. 116 and 217 (1818). The notes by Jacquin, Ecol. Pl. i. 47 and 152 (1811-16) and Lehmann, |. c., would indicate that the plant Hornemann had in mind was a species of Bothriospermum and not a Cynoglossum. In any case “ Anchusa zeylanica Vahl” of Horne- mann merits no recognition since it is a nomen subnudum and was - published only incidentally in the discussion of another species. 3. C. limense Willd. Perennial (?); stems several, lax, strigose, simple or sometimes erectly branched, 3-6 dm. tall; leaves veined, lanceolate to lance-oblong, acute, 5-12 em. long, 9-30 mm. broad, pale below and strigose, green above and finely strigose; upper leaves sessile by a triangular or rounded base; lower leaves larger and gradu- ally contracted into a winged stalk; racemes geminate and solitary, 5-10 em. long, unilateral, bractless; calyx 3 mm. long at anthesis, campanulate, cut to about middle into rounded lobes, in fruit ac- 110 JOHNSTON crescent with oblong to lance-linear stellately spreading lobes 5-7 mm. long; pedicels recurved, slender, 3-10 mm. long; corolla 7-10 mm. long, pale blue becoming cream-colored or brownish in drying, limb concave 7-10 mm. broad, lobes obovate, tube 2-3 mm. long and broadly cylindrical, throat with intruded trapeziform appendages that have decurrent hairy bases; stamens attached at middle of tube with the anthers oblong ca. 1.3 mm. long and longer than the filaments; nutlets compressed, ovate-orbicular, unrimmed, 2.5-3 mm. long, di- varicate or slightly ascending, attachment-scar small circular and subapical; style 2-3 mm. long.—Sp. PI. i. 762 (1798); Lehm. Asperif. i. 177 (1818); Brand, Pflanzenr. iv. Fam. 252, i. 142 (1921); Johnston, Contr. Gray Herb. Ixx. 33 (1924) and 1. ec. lxxiii. 77 (1924). C. decur- rens, var. limense A. DC. Prodr. x. 153 (1846). C. decurrens R. & P Fl. Peruv. ii. 6 (1799); Lehm. |. c. 142; H. & A. Bot. Beechey Voy. 37 (1830); A. DC. 1. ¢.; Clos in Gay, Fl. Chile iv. 479 (1849); Reiche, Anal. Univ. Chile exxi. 249 (1907) and Fl. Chile v. 207 asain C. alatum Mol. Sagg. Chile ed. 2, 280 (1810). CHILE. Concepcion: vicinity of Concepcion, 1855, Germain (G). Legge woods near Arique, Nev; Lechler 321 ( Ms); Niebla, collector not 8S). Luanquinve: shore of Rio Puelo, carrera ‘del arraco, fan. 8, a ee Fellez (MS). InperFrinire: “Concepcion, Valdivia,’’ collector The name C. limensis Willd. is based upon a plate and description given by Feuillée, Jour. Obs. Phys. ii. 765, t. 49 (1714). According to Feuillée he found the plant at Ilo, a minor part on the desert coast of southernmost Peru. No true species of Cynoglossum is known from Peru or northern Chile and it seems more than probable that Feuillée obtained his specimens at Concepcion, where he found many of the plants that he reports, and that through some confusion of data he incorrectly attributed the plant to Peru. Certainly the original plate and description apply very well to the plant that has been collected at Concepcion and that which has been current as C. decurrens R. & I doubtfully refer to C. limense two peculiar collections I saw in the herbaria of the Museo Nacional and Instituto Pedagégico at Santiago. - Both these collections are from the vicinity of Concepcion, Chile, one collected in Nov. 1855 at Tomé by Germain, and the other at Cerro Caracol in 1921 by Baeza. Both differ from the collections which I have unreservedly referred to C. limense in having spreading pu- rather than appressed, and in having smaller more acute elscitabbe: Germain’s collection is most peculiar since it has the corollas only 3-4 mm. long, the calyces scarcely accrescent and only 3.5 mm. long in fruit, the mature pedicels only 1-3 mm. long and the SOUTH AMERICAN BORAGINOIDEAE lll ripe nutlets ca. 2.5mm.long. Baeza’s collection, however, has corollas and fruit of typical C. limense. It is clear that these forms do not represent typical forms of the species but more material must be available for study before it would be wise to attempt to define the forms or give them nomenclatorial recognition. 4. C. paniculatum H. & A. A weak perennial, in age becoming somewhat fruticulose below and producing a few erect subsimple branches 2-8 dm. long, strigose or merely appressed hispid-villous; leaves oblanceolate, acute, 5-12 cm. long, 1-2 em. broad, gradually and but little reduced up the stem, contracted gradually towards the point of attachment, usually veinless, margins sparsely ciliate, lower surface pale and strigose, the upper sparsely short villous-strigose, lower leaves with a short narrow petiole, the upper sessile by a subulate base; inflorescence a very loose open panicle, 1-3 dm. long, leafy- bracted towards the base; calyx 1.5-2 mm. long, densely strigose, cut to below middle into deltoid or short-lanceolate lobes, scarcely ac- crescent in age and strongly reflexed; pedicels slender, divaricate, straight, 0.8-3 cm. long; corolla funnelform, ca. 5 mm. long, light blue; limb concave, 5—6 mm. broad, lobes obovate and 2.5-3 mm. long; tube cylindrical, 1.5 mm. long, about equalling the calyx; faucal ap- pendages trapeziform, pubescent; stamens affixed at middle of tube; anthers elliptic, ca. 0.5 mm. long, about twice as long as the filaments; nutlets orbicular-ovate, ascending, 2-2.6 mm. long, unmargined, at- tachment-scar occupying most of the ventral face; gynobase pyram- idal; style 1.5-2 mm. long.—Bot. Beechey Voy. 37 (1830); DC. Prodr. x. 153 (1846); Clos in Gay, Fl. Chile iv. 478 (1849); Reiche, Anal. Univ. Chile exxi. 248 (1907) and Fl. Chile v. 206 (1910); Brand, Pflanzenr. iv. Fam. 252, i. 137 (1921); Johnston, Contr. Gray Herb. Ixx. 33 (1924). C. Azocarti Ph. Anal. Univ. Chile Ixv. 62 (1884) and l. ce. xe. 553 (1895). C. paniculatum, var. Azocarti Reiche, lc. C. siete: f. Azocarti Brand, 1.c. C. paniculatum, f. Philippianum ses le CEPCION: tog phioda deo Philippi M phe patil se 1880-1, cart (MS, Type of ©. Azocarti; G, ph to ; Tat A: Talea, ehnaas: Joseph 2071 (US). Vawuparaiso: Cura Oct. tag Philipyi (MS); Rianne. (lek 100s, Cliedeeecph 9670 (US); La Palma near Quillota : 580 1161 (G). gg se Dhcahea 9 le April 1877, Reed “G). A> very distinct species which though apparently uncommon is widely distributed over south-central Chile. It is best known from the coastal hills of the province of Valparaiso. In addition to the localities I.have listed, Brand, 1. c., cites collections of Poeppig from 112 JOHNSTON Concon and Chorrillos. The type came from Concepcion. In de- scribing the species Hooker & Arnott mention a collection made by Macrae at “ Murillo Bay in Peru” — perhaps to an unimportant locality in the department of Arequipa. I do not believe that C. paniculatum grows in Peru and canis rae the record is the result of a misdetermination or, very much more likely, confused locality-data. The plant from Constitucion named C. Azocarti by Philippi is worthy of no nomenclatorial recognition. The leaves in the type-specimen are very loosely though strongly revolute. This condition is clearly the result of insufficient pressure during the preparation of the speci- mens. Though the revolute leaves appear very narrow, in fact their real breadth is not unusually small for the species. 5. C. Trianaeum Wedd. Perennial or biennial; stems several, erect, simple, 3-8 dm. tall, glabrous below, villous-strigose above; leaves acute or acuminate, contracted gradually towards the base, oblanceolate, firm, conspicuously bicolored, beneath pale and abun- dantly hispid-strigose, above shiny and (margins excepted) glabrous; lower leaves 1-2 dm. long, contracted to an alate stalk 1-2.5 cm. long and 4-5 mm. broad; upper leaves contracted to a narrow sessile base, gradually reduced up the stem; base of leaves ciliate, slightly broad- ened, somewhat clasping and decurrent forming angles on the stem; racemes solitary, terminal, becoming very loose and 1.5 dm. long, bractless above but near base leafy-bracted; calyx at anthesis broadly campanulate, ca. 4 mm. long, cut to about the middle into broad _ teeth, densely strigose; calyx in fruit reflexed, slightly accrescent, surpassed by the nutlets; pedicels slender, ascending, 1.5-3 cm. long, usually nodding at maturity; corolla blue to violet, limb spreading, 7-12 mm. broad, lobes obovate; tube shorter than the lobes, broadly cylindrical, 2-4 mm. long; appendages trapeziform, pubescent; stamens attached at middle of tube, filaments short about equalling the length of the oblong anthers; nutlets ovate, compressed, ca. 6 mm. long, unmargined, ascending; attachment-scar very large, ovate, occupying most of the ventral face of the nutlet; gynobase pyramidal; style ca. 2 mm. long.—Chlor. Andina ii. 90 (1859); Oliver in Hook. Icones Pl. xxv. t. 2458 (1896); Brand, Pflanzenr. iv. Fam. 252, i. 136 (1921); Johnston, Contr. Gray Herb. Ixx. 33 (1924). COLOMBIA or ECUADOR: “‘Casapamba?”, André 2952 (NY). A montane plant of western Colombia and Ecuador. The type came from the cordilleras of the province of Antioquia, Colombia, but _ apparently most of the subsequent collections have been made at the southern end of the Cordillera tral in the provinces of Narifio and SOUTH AMERICAN BORAGINOIDEAE 113 Cauca. The species is a very distinct one and is widely separated from its congeners in Mexico, California and Chile. 17. Pectocarya DC. Calyx cut to base, accrescent, divaricate at maturity. Corolla tubular; lobes ascending, imbricate, small; throat with small intruded appendages; stamens included, short, attached at middle of tube; filaments short; anthers oblong or elliptical. Style very short; stigma capitate. Ovules 4. Nutlets 4, obovate to linear, divaricate, paired, usually margined, pubescent with uncinate hairs, apically affixed to a low very broadly pyramidal gynobase.—Slender spreading annual herbs with the lower leaves opposite and linear. Flowers in leafy- bracted very elongate false racemes which constitute the bulk of the plant.—DC. in Meisner, Gen. i. 279; ii. 188 (1840); Brand, Pflanzenr. iv. Fam. 252, i. 94 (1921). Ktenospermum Lehm. Del. Sem. Hort. Hamb. 17 (1837), nom. nudum; (Ctenospermum) Post & Kuntze, Lex. 152 (1904). Gruvelia A. DC. Prodr. x. 119 (1846). A very distinct genus of west American herbs. In his recent treatment of the genus Brand, |. ¢., recognizes only 3 species, viz., P. setosa Gray of North America aid P. pusilla (A. DC.) Gray and P. lateriflora (Lam.) DC. of both North and South America. I am convinced, however, that what he has called P. lateriflora is an aggre- gate of 7 recognizable entities which have technical fruit-characters, distinct ranges and characteristic habits of growth. Iam treating the 4 of these forms which occur in South America as species, and believe that the North American forms should be treated likewise. Key TO SPECIES. Nutlets equally divergent, much compressed, obviously angular an mboidal in outline.......---..-.++pseorere settee 1. P. pusilla. Nutlets ace nt in es thickish, not obviously angular nor rhomboid outline. and sides of neitinte armed with coarse subterete subu- late appendages tipped with uncinate hairs............ 2. P. anomala. Back of nutlets smooth or at most papillate, margin with a bai eae or pectinately cut wing-margin bearing inate Fruit. pennmae g that produced at base of with reflexed subpersiste nt peti onal i - di f the middle and up Sarena: eda ee Saye 3. P. dimorpha. Fruit of one sort, that enna at the base of the plant similar to that on the w Nutlets linear, 2.5-3 mm. long, sides parallel........ 4. P. gracilis. Nutlets obovate or obovate-oblong. 114 JOHNSTON Nutlet-body 2-2.5 mm. long, with medial dorsal ridge. 5. P. lateriflora. Nutlet-body 1.5-2 mm. long, lacking medial eter I ere res ei ae ws ees . P. boliviana. 1. Pectocarya pusilla (A. DC.) Gray. Plant small and slender, strigose, with erect or ascending simple stems; leaves linear or oblong- linear, obtuse, 0.5-2 cm. long, 1-2 mm. broad, sparsely strigose; flowers all chasmogamic; corolla short salverform with ascending lobes; nutlets 4 or fewer, rhomboidal or rhomboid-obovate, homo- morphous, radially arranged, 1.8-2.3 mm. long, 1.5-2 mm. broad; margin acute but wingless, ciliate with uncinate hairs.—Proc. Am. Acad. xii. 81 (1876); Reiche, Anal. Univ. Chile exxi. 247 (1907) and Fl. Chile v. 205 (1910); Brand, Pflanzenr. iv. Fam. 252, i. 96, fig. 1g (1921); Johnston, Contr. Gray Herb. Ixx. 38 (1924). Gruvelia pusilla A. DC. Prodr. x. 119 (1846); Clos in Gay, Fl. Chile iv. 482, t. 52 bis, fig. 3 (1849). P. chilensis, var. californica Torr. Pacif. R. R. Rep. iv 124 (1857); Johnston, |. c. Ixxv. 43 (1925). P. pusilla, var. Magilintts Brand, l. ¢ CHILE. be ih pet dry slope, {? Punta de] La Playa Ancha, Aug. 1830, Bertero 212 (G, NY). 7 This species although based upon material from central Chile seems to be very rare there. Brand, |. c., reports a collection made at Rancagua by Philippi. There are, however, no collections of the species in the Philippi Herbarium in the Museo Nacional nor in the herbarium of the Instituto Pedagégico at Santiago. In fact with the exception of the Philippi collection cited by Brand, the species is not known to have been collected in Chile for nearly a century. It is, however, well known from central and northern California where it has been frequently collected. 2. P. anomala Johnston. Stems slender, finely strigose, not nu- merous, erect or ascending; leaves filiform-linear, 1-3 em. long, 0.7-1 mm. broad, appressed short hispid-villous; flowers all chasmogamic; _— short salverform with ascending lobes; nutlets densely pu- the pubescence short soft and uncinate; body of nutlet ob- idee. 1.5-1.7 mm. long, ca. 0.7 mm. broad, without a proper *k and upper sides | 0.6-1 mm. long that are terminated with esa tipped bristles. — Contr. Gray Herb. ixx. 35 bss n south slope of El Chachani, 2400 m. ; at, March 192, a Hinchey 41", mire) elly soil G5, Ponnall ste vi — above Arequipa SOUTH AMERICAN BORAGINOIDEAE £15 An extremely well marked species known only from the vicinity of Arequipa, Peru. It seems to be most related to P. lateriflora, but is quickly separated from that species by the lack of a definite margin on its nutlets and by the peculiar subterete subulate appendages that cover them. 3. P. dimorpha (Johnston), comb, nov. Stems slender but stif- fish, prostrate or widely ascending, numerous, finely and closely ol aa somewhat canescent; leaves filiform-linear, 1-3 cm. long, 0.5-— . broad, canescent, finely strigose and occasionally short- hispid; saedls short-salverform with ascending lobes; fruits dimorphic, those of the cleistogamic flowers (borne at the base of the stems) all reflexed and with narrow margins; nutlets of the chasmogamic flowers (borne along the stems) with body oblong or narrowly obovate-oblong, ca. 2 mm. long, 0.8-1.1 mm. broad, tending to be heteromorphous with the abaxial one or one of each opposing pair tending to be less broadly margined, slightly smaller, more pubescent and more per- sistent than the others; margins of the nutlets of the normal fruits very much reduced or more characteristically well developed, ascend- ing or erect and frequently much incurved, pectinately or lacerately cleft or divided or even dentate with the lobes triangular to subulate and terminated by uncinately tipped bristles——P. gracilis, var. di- morpha Johnston, Contr. Gray Herb. Ixx. 37 (1924). CHILE. Aracama: Vallenar, Oct. 1914, Rose 19331 (US); Rio de Laguna one east of Vallenar, 2140 m alt., Johnston 891 (G); Valle de Carrizal, a Peacung, not given (Ms); Yerba buena, 1885, R. Godoi de Collao (M 3); Bard urrias, 1885, Yale (MS, G); Atacam ha een [? fe vba Geisse 59 (NY); Quebra da de Chanchoquin north ch of Reoenepns, #8 tat velly bench in hills north of Copiapé6, 600 m. alt., ree F025 (G); Copiapé, Sept. 1885, San Roman (MS); Desert ot pa [Caldera- Copiapé capaci Mey 1282 (G, TypE; NY, tsoTyPeE). ANTOFAGASTA: a small colony on ry dry rocky hillside at western end of Llano —— Dept. of Taltal, Watson S656 (G). The present plant is perhaps most related to P. gracilis, but is quickly etcaoubel by its stiffer stems, shorter more heteromorphous nutlets and presence of cleistogamous flowers at base of stem. Its range is north of that of P. gracilis, covering the western part of the province of Atacama and extreme southern Antofagasta. It is a very distinct species and certainly worthy of recognition. 4. P. gracilis (R. & P.) Johnston. Stems very slender, numerous, prostrate or widely ascending, finely and closely strigose, canescent; leaves filiform-linear or narrowly oblance-linear, 1-3(-4) cm. long, 0.6-1.5(-1.8) mm. broad, numerous, closely fine-strigose; flowers all chasmogamic; corolla short-salverform with ascending lobes; nutlets 116 JOHNSTON homomorphous or only obscurely heteromorphous; nutlet-body linear- oblong or somewhat spathulate-oblong, 2.5-3 mm. long, 0.6-1 mm broad, with the erect or strictly ascending margin pectinately parted or divided into subulate or narrowly triangular teeth each terminated by an uncinately tipped bristle—Contr. Gray Herb. Ixx. 36 (1924). Myosotis gracilis R. & P. Fl. Peruv. ii. 5 (1799). Echinospermum gracile Lehm. Asperif. i. 129 (1818). Rochelia gracilis R. & S. Syst. iv. 111 (1819). Cynoglossum lineare R. & P. 1. c. 6. P. linearis DC. Prodr. x. 120 (1846). P. chilensis DC. 1. c.; Clos 3 in Gay, Fl. Chile iv. 480, t. 52 bis, fig. 2 (1849). P. gracilis, var. genuina Johnston, I. ¢. P. lateriflora of Reiche, Anal. Univ. Chile exxi. 247 (1907) and FI. Chile v. 205 (1910), in largest part. ARGENTINA. as Cruz: fields near Rio Gallegos, March 11, 1882, Spegazzini (G). Rio Necro: vicinity of General Roca, 250-360 m alt., Sept. 28, 1914, “Fischer 123 (G, US, FM); fields — Rio Negro, Nov. 19, 1904, Berg (G). ecm Frutillo, Oct. 1913, Sanzi CHILE. OHi«ee ee Yims er 1878, no nolkecioe given (MS). SANTIAGO! a Cine dos ti 684, 1496 and 2241 (US); Santiago, Sept. 1900, Hastings 56 (US); Santiago, Sept. 1840, Gay 1616 (MS); Cerro San Cristoral, 2 no ; Sa oral collector gi (MS); so San Cristobal, Oct. 1924, 4, Mont ? (G); Cerro San Cristobal, Aug. 1925, Looser (G); San Cristobal, 1913 start 1917; Baeza ap): San Bernardo t. 1877, no collector gi ven (MS); Santa Rita, Oct. 1879, no ahem given (MS); Chacabue Sept. or 1864, no collector given (MS); Tiltil, 1911, Horst Jat Navia, 1914, ee (IP). Vauranaiso: Valparaiso, bag ot 30, pets ien (US); : Valparaiso, 1856, Harvey (G); Concon, Poeppig (NY). Aco . Bet F elipe, Aug. 1921, _CladesFoseph 1351 (US); Zapillar, 1912, Ford: (IP); Palquico, Oct. Rose 19189 in pt. (Ui Se ew ee . , Oc Es cP 1914, eae 19269 US); ig Oct. e, 1914, 1885 ACAMA a Grande eas st of Vallenar, 2140 m. alt. gest vs 6283 (G 2. TT neeee NITE: eal not given, Aug. 9, 1917, Skatisber ted (G); loeality not given, Bridges (G); locality not given, Bertero 2 1 (G, NY). The species concerned here is the common one of central Chile. It is certainly Cynoglossum lineare R. & P., the type of which came from Santiago (Sancti Jacobi), where the present species is common. The type of Myosotis gracilis R. & P. is said to have come from Con- { . Although Reiche gives the range of the plant as south to arts of Bio-Bio, I have not seen any collections from south In Argentina the species seems to be uncommon. Brand and Dusén report it from Rawson, Chabut. Spegazzini cites four localities in Santa Cruz, viz. Rio Gallegos, Rio Santa Cruz, San Julian and Rio Deseado. A Pectocarya cited by Brand from La Incrucijada, La 3 Bio, x may represent the present species but from phytogeographical tions would seem to be more probably P. boliviana. SOUTH AMERICAN BORAGINOIDEAE 117 Outside its range in central Chile and southern Argentina the species also occurs in southern California. It has a definite character and is quickly recognized by its very elongate linear nutlets. The subulate toothing of the nutlets in this species, more than in any other, merits the adjective “ pectinate.” 5. P. lateriflora (Lam.) DC. Stems ascending or erect, not very numerous, stiffish, finely appressed hispid-villous or strigose; leaves filiform-linear to narrowly lance-linear or linear-lanceolate, 1-5 mm. long, 1-2.5 mm. broad, usually definitely contracted towards both ends, appressed hispid-villous and usually somewhat canescent; flowers all chasmogamic; corolla short-salverform with ascending lobes; nutlets tending to be heteromorphous with the abaxial one usually more pubescent and roughened than the other ones which commonly are smooth and glabrate or sparsely pubescent; nutlet-body obovate or obovate-oblong, 2-2.5 mm. long, 1-1.8 mm. broad, with a definite medial dorsal ridge, margin becoming well developed spread- ing, pectinately and lacerately cleft or divided into subulate or nar- rowly triangular teeth terminated by uncinately tipped bristles.— Prodr. x. 120 (1846); Brand, Pflanzenr. iv. 252, i. 95 (1921); Johnston, Contr. Gray Herb. lxx. 35 (1924) and I. ec. lxxiii. 76 (1924). La net sum lateriflorum Lam. Encye. ii. 239 (1786) and Planch. t. 92, fig. 2 (1791); Lehm. Asperif. i. 166 (1818). Mattia lateriflorum Don, Gen Syst. iv. 310 (1838). Rindera lateriflora R. & S. [Syst. iv. 762 Asi9) ex Steud. Nom. Bot. ed. 2, ii. 460 (1841). C. pilosum R. & P. F Peruv. ii. 6, t. 111b (1799); Lehm. |. c. 167. M. pilosa Don, l. c. b pilosa R. & S. {l. c.]ex Steud. l.c. (?) Ktenospermum linifolium Lehm. Linnaea xii. Litt. 84 (1838), nom. nudum. Lima: open rocky hill, Amancaes, 200-400 m. ae 14806 G EMD; open sandy slo Atocongo, 300-500 m. alt, wn Tayo (G, FM); open grassy knolls, Canta, 2800-2900 m val “Pennell 14583 (G, FM); Obrajillo, Wilkes Exped. (G, US); loose stony soil, Rio Blanco, 3600 m. alt., Macbride & Fe . : ge roc illsi Weber (FM). A bac Johnston 3534 (G); sandy slopes, desert hills, Mfallendo, Hitchcock 22382 ve). OQUEGUA: open mixed formation, eats. 2200-2300 m. alt., We 7396 (FM). Characterized by its obovate definitely margined nutlets and com- paratively coarse stems and leaves. It is a plant of the lomas and desert margins of western Peru. Brand reports a collection of Spruce’s, no. 5309, from the Andes of Ecuador. Whether this ss collection represents the species, as here defined, I do not know 118 JOHNSTON 6. P. boliviana (Johnston), comb. nov. Stem slender, decumbent or ascending, finely and closely strigose, canescent; leaves linear- filiform to linear, 1-3 em. long, 0.8-1.5 mm. broad, strigose, canes- cent; flowers all chasmogamic; corolla short-salverform with ascending lobes; nutlets heteromorphous with the abaxial one usually slightly the most roughened and pubescent; nutlet-body obovate-oblong, without a medial ridge, 1.5-2 mm. long, the spreading margin pec- tinately divided into numerous ‘arstepenid or more commonly few distant subulate teeth which are tipped by uncinate bristles.—P. gracilis, var. boliviana Johnston, Contr. Gray Herb. Ixx. 37 (1924) and 1. c. Ixxiii. 76 (1924). oe Porosr: Chiguana, 3700 m. alt., Asplund 3897 (US, bie, a, 3700 m. alt., Asplund 5896 (US). “Ononos oe 3900 m Asp pS (US La Paz: General Campero, 4200 m. alt., Pte This species is closely related to P. lateriflora of western Peru, but is a smaller and more slender plant of the puna. Its nutlets are smaller, transversely rugose, lack a definite medial dorsal ridge and have a margin that is commonly much dissected and frequently represented only by several distant subulate teeth. Brand, Pflanzenr. iv. Fam. 252, i. 95 (1921), reports P. lateriflora from Incrucijada in La Rioja, Argentina, upon the basis of a collection by Hieronymus & Niederlein (no. 466). This collection may represent P. boliviana. Reprinted from Ruopora, Vol. 30, February—May, 1928. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY LX XIX I. Geocaulon, a new Genus of the Santalaceae..........--..-+-- 21 Il. The American and eastern Asiatic Beckmannia............--- 24 III. The eastern American Variety of Polystichum Braunwi......... 28 IV. The American Representatives of Asplenium Ruta-muraria. . 37 V. The eastern American Occurrence of Athyrium alpestre»....... VI. Primula § Farinosae in America... ..----- +++ +00 - eee e eres 59 By M. L. Fernatp Dates OF IssvE Pages 21 to 3D, plate 150... <.-..- 6: e+ vee eee eee re eee 9 March, 1928 af to 49, plates 161 to 168.......---.+- eee ee eter teers 10 April, 1928 ff 59 CO Bs RR ae 15 May, 1928 ey i so ee we ee ieee ee 13 June, 1928 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY,—NO. LXXIX. M. L. FERNALD I. GEOCAULON, A NEW GENUS OF THE SANTALACEAE Ever since I first met Comandra livida Richardson, in 1894, I have had a strong conviction that it is out of place in the genus Comandra Nutt. Many times subsequently in the field, as I have gathered the juicy scarlet false-drupes from the filiform axillary peduncles or have examined the bronze or green, often unisexual, flowers, I have made a mental protest against the inclusion of this plant of the moss and damp humus in the same genus with Comandra umbellata (L.) Nutt. or with C. elegans (Rochel) Reichenb. f. Finally, in August, 1923, with the intention of settling the question, an abundant series of inflorescences was collected on the Shickshock Mountains; but not until the present time have they been closely studied. It now becomes quite clear that in essentially all its characters C. livida departs from all the species of Comandra proper: C. umbellata (L.) Nutt., C. Richardsiana Fernald, C. pallida A.DC., C. elegans (Rochel) Reichenb. f., C. cali- fornica Eastw., and perhaps others. In true Comandra the stoutish creeping or sprawling superficial or subterranean stems or rootstocks are covered with a loose and freely exfoliating corky to papery whitish-brown cortex; in C. livida the very slender and cord-like reddish to dark-brown subterranean stems have a tight and smooth cortex. In true Comandra the inflorescences are terminal panicles or corymbs of numerous small umbels, each umbel subtended by a tardily deciduous or persistent foliaceous involucel; in C. livida the 1-3 simple mostly 3-flowered umbels are borne on 22 Rhodora [FEBRUARY filiform peduncles from the axils of the middle leaves, and the brown, scarious involucres are caducous. The flowers of true Comandra are perfect, the calyx-tube free from the summit of the ovary and in fruit slightly prolonged as a neck; and the ascending turbinate limb consists of petaloid whitish lobes. The flowers of C. livida are andro- dioeceous, the central 1 (rarely 2) perfect, the lateral mostly staminate and promptly dropping after anthesis, or sometimes all the flowers staminate; the calyx-tube is completely adnate to the ovary, not prolonged above it, and the rotate limb consists of bronze or green herbaceous lobes. In true Comandra the elongate disk is shallowly lobed, its lobes much shorter than the filaments, and it reaches the summit of the prolonged tube; in C. livida the salverform disk arises from the base of the throat and its long lobes about equal the filaments. In true Comandra the style is filiform and prolonged; in C. livida conical and very short. Finally, the fruit of true Comandra is a dry nut with only the lower half or two-thirds of the coriaceous calyx- tube adherent to it, the upper portion forming a free neck below the erect lobes; in C. livida the fruit is a scarlet and juicy false-drupe, with the succulent calyx-tube completely surrounding the nut. Differing in every fundamental character (of rootstock, flowering habit, involucre, fertility of flowers, shape and texture of perianth, adnation of tube, position and lobing of disk, form and length of style, and, above all, in the very different fruit), and restricted to the Hud- sonian to arctic-alpine regions of North America, Comandra livida does not seem to be congeneric with the species of true Comandra, plants of more southern range in America, with an isolated species in southeastern Europe. | In many characters Comandra livida is much closer to the monotypic Nestronia (or Darbya) of the southeastern United States. Like staminate Nestronia wmbellula Raf., Comandra livida has the simple few-flowered umbels on filiform axillary peduncles, borne always below the terminal leaves, the involucral bracts caducous (in all material of Nestronia examined completely absent), and the calyx herbaceous and with spreading limb. The pistillate or perfect flower of Nestronia, like that of Comandra livida, has the ovary quite inferior and ripening into a false-drupe with completely adnate calyx-tube. _ But in many characters Comandra livida departs from Nestronia. _ The latter is a branching shrub, with opposite leaves; C. livida at =~ suffruticose, with simple herbaceous flowering stems and 1928] Fernald,—Geocaulon, a new Genus 23 alternate leaves. Nestronia is dioecious or polygamo-dioecious, the pistillate flowers solitary (not umbellate) and with the limb a mere crown; C. livida with the perfect flower central in the umbel with the staminate and with a spreading limb. In Nestronia the calyx-lobes are normally 4, in C. livida 5; in Nestronia the calyx-tube of the staminate flower is elongate and lined by the elongate, shallowly lobed or undulate disk; in C. livida short-campanulate or salverform, with the lobes of the disk prolonged. It is thus clear that in some of its technical characters Comandra livida is nearer Nestronia than it is to Comandra; yet it is not satis- factorily placed with either. It is, therefore, here proposed as Geocaulon (Santalaceae), gen. nov. Flores andro-dioici, centrales pl quel phroditi, laterales masculi. Calyces herbacei, priores et masculi inferne turbinato, feminei campanulati; lobis ovato-acutis, aeneis vel viridibus, patentibus persistentibusque. Fasciculus pilorum e basi interna lobi cujusve ad antheram tendens ubique adhaerens. Stamina lobis opposita, filamento subulato, anthera ovoidea, bilo- culari. Discus epigynus hypocrateriformis, lobis elongatis filamenta subaequantibus. Ovarium inferum, tubo calycis inclusum. . Stylus conicus, brevis. Stigma capitellatum. Drupa ovoideo-globosa coc- 1 pedunculata 3 (2-4)-flora; bracteis involucri scariosis brunneis caducis; floribus masculis caducis. (I"j, earth, and xaog, stalk, from the long, The following, selected from many specimens, are characteristic. Laprapor: Makkovik, Stecker, no. 99; Hopedale, Bowdoin College Exped. no. 245; Square Island, 1864, B. P. Mann; Chateau Bay, Bowdoin College Exped. no. 75; Forteau, Fernald & Wiegand, no. 3284. NEWFOUNDLAND: Burnt Cape, Fernald, Wiegand, Pease, Long, Gris- com, Gilbert & Hotchkiss, no. 28,113; Mistaken Cove, Wiegand, Pease, Long & Hotchkiss, no. 28,112; Brig Bay, Fernald, Long & Dunbar, 24 Rhodora [FEBRUARY no. 26,610; Deer Pond, Wiegand, Gilbert & Hotchkiss, no. 28,114; Port Saunders, Fernald & Wiegand, no. 3283; Bay of Islands, Eames «& Godfrey, no. 6037; Sandy Lake, Fernald & Wiegand, no. 3282; Grand Falls, Fernald, Wiegand, Bartram & Darlington, nos. 5327, 5328; Tilt venir Fernald & Wiegand, no. 5331; St. John’s, Robinson & Schrenk, no. 152. QueEBeEc: Archipel du Petit Mécatina, St. John, o. 90,397; Ile a ‘la Proie, Archipel de Mingan, Victorin & Rolland, nos. 18,506, 22,055; Baie Sean laire, Anticosti, Victorin, no. 4270; Seven Islands. C. SB. inson, no. 731; éninsule, Baie de Gaspé, Victorin et al. no. 17,331; Tabletop Mts., Gaspé Co., Fernald, Dodge & Smith, no. 25,713; New Carlisle, 1902, W Miams & Fernald; Lac Saint-Jean, Victorin, no. 15,754; Tadousac, 1892, Kennedy; Riviére du Loup, Victorin, no. 134; St. Alexandre de Se adele 1880, Pringle; Black Lake, F. ernald & Jackson, no. 12,077. MaGDALEN IsLAN aa ee Island, St. John, no. 1851. New BRUNSWICK: St. John, r. James. Maine: Mt. Katahdin, 1900, Fernald; Mt. Silk Franklin Co., 1894, Fernald, 1902, Kainalion: Mt. Abraham, Kn owlton, no. 606; Wes t Quoddy Head, Lubec, Fernald, no. 1715; Roque Bluffs, 1913, ‘Raouton. New Hampsuire: Mt. Clinton, Eggleston, no. 2399, Pease, no. 12,303; Imp Mt., Pease, no. 16,766; . Ingalls, A. H. Moore, no. 4096, Pease, no. 11,210. VERMONT: Mt. Bag ines Pringle et al. Ontario: Pic River, Loring; Anvil Timagami Region, Anderson & Anderson, wr nase, Isle Royale, 1849, Whitney; Keweenaw Co. = 63, Robbins; Marquette Island, 1913, W. H. Manning. MuinNEsoTA: border of Lake of the Woods, snath n ‘49°, Richardson. Manirosa: Churchill, J. M. Macoun, no. 79,398; Taka Manitoba, 1881, Macoun. Sas- KATCHEWAN. 1857-8, Bourgeau. ALBERTA Rocky Mts., Drum- nor Banff, rows, #5 etal. BritisH Coren: Macleod’ s Lake, lat. acoun, no. 1559; Revelstoke, Shaw, no. 31. YuKon: Dawson, East nos. ni, 491. Axaska: Lake Iliamna Region, Gorman, no. 1 Pilea lividum was beautifully illustrated, as Comandra livida, in Hooker, Flora Boreali-Americana, t. clxxix B. Although the accom- panying description indicates no difference between the central and lateral flowers, the artist noted the difference and showed the central perfect flower much larger than the others. II. THE AMERICAN AND EASTERN ASIATIC BECKMANNIA Ir has often been pointed out that the plant of northwestern North and northeastern Asia which has passed as Beckmannia erucaeformis (L.) Host is not identical with the true B. erucaeformis (Phalaris erucaeformis L.) of southeastern Europe and southwestern Asia; but not until Hultén’s recent scholarly publication upon the 1928] Fernald,—American and Asiatic Beckmannia 25 flora of Kamtchatka! have the two plants been adequately con- trasted. Hultén shows that there are two quite distinct species of Beckmannia2 True B. erucaeformis of Italy, Greece, Hungary, southern Russia and Asia Minor is a perennial with bulbous or “incrassated’’ bases, simple or but slightly compound and very short panicle-branches closely appressed to the rachis: spikelets mostly 2-flowered and strongly rounded; the glumes nearly semicircular, coriaceous and with a rounded keel and very narrow membranaceous border; the thick lemma short-mucronate or very short-pointed, scabrous and beset with long hairs along the median nerve. The plant of northeastern Asia and northwestern America, on the other hand, has the culms with soft bases, usually described as annual, commonly coarser and taller; the leaves broader; the panicle much more branched, with the principal branches longer and somewhat spreading; spikelets larger, mostly with 1 perfect and 1 aborted floret, pyriform; glumes rounded-triangular, broadest toward the apex, thin, strongly carinate, with a whitish scarious border; lemma thin, long- mucronate, glabrous or only slightly short-pilose on the dorsal nerve. For this plant of eastern Asia Hultén proposes the name Beckmannia baicalensis (W. Kuznetzow) Hultén,* based upon B. erucaeformis, var. baicalensis W. Kuznetzow (1913). Hultén gives a careful discussion of the two plants and makes a good case for the specific segregation of B. baicalensis; but, unfor- tunately, he overlooked at least two available names which were published earlier than Kuznetzow’s in 1913. In order clearly to understand these names it is well to start with the original Phalaris erucaeformis L.4 Linnaeus did not differentiate the two and certainly included them both under P. erucaeformis which had its “ Habitat in Sibiria, Russia, Europa australi.” After a brief original diagnosis, “PHALARIS panicula lineari secunda, calycibus bifloris,” which certainly applies to the plant from “Russia, Europa australi,” he gave three citations: “Dactylis spicis numerosis -alternis culmo adpressis longitudine internodiorum, calycibus bifloris. Roy. lugdb. 1Eric Hultén, Flora of Kamtchatka and the Adjacent Islands, i,—Kungl. Svenska Vetenskapsakad. Handl. Ser. 3, v. no. 1 (1927). 2In some of our American manuals we are told that the gen “In us was named honor of Johann Beckmann, . . . teacher of Natural History at St. Petersburg.”’ The original statement of Host, however, was: ‘*Hoc proprii generis gramen in honorem Cl, viri Joannis Beckmanni, in Universitate Goettingensi Oeconomiae ruralis Proft nominavi’—Host, Ic. Gram. Austr. iii. 6 (1805). * Hultén, 1. c. 119 (1927). 4L. Sp. Pl. 4. 55 (1753). 26 Rhodora [FEBRUARY 57. Gmel. sibir. I. p. 130. t. 29” and “Gramen palustre, locustis erucaeformibus. Barr. rar. 1158. t. 2.” Royen’s description and his citation of the description and plate of Barreliero show that he had the Italian plant and although Barreliero said of the plant “ Annum est,” ' he certainly had the plant of Italy, and the Linnean specific name was obviously derived from his Gramen palustre, locustis Erucae- formibus, a name given from the resemblance of the green spikes to caterpillars on the branches of a tree. Linnaeus’s reference to Gmelin’s Siberian plant alone belongs to Beckmannia baicalensvs. Gmelin’s plate is thoroughly typical for the latter species. With Linnaeus’s own diagnosis and the plants of Royen and of Barreliero coinciding, and with the specific name clearly derived from Barreliero’s account of the Italian plant, there can be no doubt that the name B. erucaeformis should be restricted to the plant of southeastern Europe and Asia Minor. The species is well illustrated in Host, Ie. Gram. Austr. iii. t. 206 (1805); Nees, Gen. i.t. 40 (1843) and Reichen- bach, Ic. Fl. Germ. i. t. elxxi (1850). That the wide-ranging northern plant is not identical with the Mediterranean species seems to have been first recognized by Steudel in 1846. Receiving the plant from Japan, Steudel, on account of the single perfect flower, the disarticulation of the spikelets below the glumes and the superficial resemblance of the plant to Panicum bri- zoides Jacq. and others with the inflorescence of Echinochloa, described it as Panicum Syzigachne.2 Later, however, recognizing that his Japanese species belonged to Beckmannia, he reduced it to a variety of B. erucaeformis, “Variat: statura et omnibus partibus majori- bus. Panicum Syzigachne. Steud. Flora 1846. p. 19. Japon.’ * In 1880 Dr. George Thurber, treating the grasses of California and retaining our plant as B. e fe , made the note: “ In all —— specimens we have thus far aoeaat the upper floret is wanting.” Thurber’s observation was soon again emphasized by Vasey, ee in diagnosing Beckmannia, described the spikelets as “consisting of two perfect flowers, the terminal one only fertile, (or in American specimens the lower flower is suppressed); and at the same time Vasey called the American plant B. erucaeformis, var. uniflorus Seribn.,° but without diagnosis further than by implication from the 1! Barr. Plantae per Galliam, Hispaniam et Italiam Observatae, 105 (1714). ? Steud. Flora, xxix. 19 (1846). Steud. Syn. 1928] : Fernald,—American and Asiatic Beckmannia 27 preceding note. In 1896 B. erucaeformis uniflorus Scribn. was used by Beal! for the American plant which was well described and illus- trated by drawings of the spikelet supplied by Scribner; and in the 6th edition of Gray’s Manual, Watson & Coulter took up B. erucae- formis, var. uniflora, a name dropped by Hitchcock from the 7th edition, where the plant is treated unequivocally as B. erucaeformis, although the genus, based upon a 2-flowered species, was by Hitchcock now defined as having “Spikelets 1-flowered in our species.” In his later work the American and Asiatic plant is still maintained by Hitchcock as B. erucaeformis (without var. uniflora cited as a syno- nym); but the facts of the case are covered by the statement under the generic description: “Spikelets 1-flowered, rarely 2-flowered’’; ? and under the specific description: “ The European form has 2-flowered spikelets.” 2 In 1913, W. Kuznetzow again named the Asiatic and American plant, his material coming from Transbaikalia, whence his new name B. erucaeformis, var. baicalensis,’ upon which Hultén based the specific combination B. baicalensis. In taking up Kuznetzow’s varietal name as the specific name for a plant which extends half way around the northern hemisphere Hultén was apologetic: “ Although his variety name is not very suitable for the species, I have thought it most correct to retain it” (p. 121). Unfortunately, however, - correct following of nomenclatorial rules forces us to use for the species a name in some ways less desirable than B. baicalensis. The correct name of the plant seems to be : BrecKMANN1A Syzigachne (Steud.), n. comb. Panicum Syzigachne Steud. Flora, xxix. 19 (1846). B. erucaeformis, var. Steud. Syn. Pl. Gram. 15 (1854). B. erucaeformis, var. uniflorus Scribn. in Vasey, Descr. Cat. Grasses U. S. 8 (1885), name only; in Beal, Grasses N. A. ii. 428, fig. 77 (1896). B. erucaeformis, var. baicalensis W. Kuznet- zow, Bull. Angev. Bot. vi. No. 9: 584 (1913). B. batcalensis (W. Kuznetzow) Hultén, Kungl. Svenska Vetenskapsakad. Handl. Ser. 3, v. no. 1: 119 (1927). Beckmannia Syzigachne, as previously stated, was well illustrated by Gmelin. Scribner’s figures of the spikelet of his B. erucaeformis, var. uniflorus are excellent, and Mrs. Chase’s illustration (as B. erucae- formis) in Hitchcock, Gen. Grasses U. S. fig. 108 is very characteristic. 1 Beal, Grasses, N. A. ii. 428, fig. 77 (1896). 2 Hitchc. Gen. Grasses U. S. 180, 182 (1920). 2 W. Kuznetzow, Bull. Angev. Bot. vi. No. 9: 584 (1913). 28 Rhodora [FEBRUARY III. THE EASTERN AMERICAN VARIETY OF POLYSTICHUM (Plate 159) Tue plant which has been passing in eastern America as Polystichum Braunii (Spenner) Fée abounds in rich woodlands and glades or on shaded talus and rock-slides of northern and western Newfoundland, Anticosti Island and the Gaspé Peninsula of Quebec. Westward and southward it becomes more local, extending west to Algoma District, Ontario and south (at low altitudes) to Cape Breton Island and Col- chester and Kings Counties, Nova Scotia and Charlotte County, New Brunswick, and (chiefly in the mountains) to Aroostook, northern Penobscot, northern Piscataquis, northern Somerset and Franklin Counties, Maine, Carroll and Grafton Counties, New Hampshire, Berkshire County, Massachusetts, Ulster and Delaware Counties, New York, Sullivan County, Pennsylvania and Keweenaw County, Michigan. It is thus completely isolated from P. Braunii of Europe, which is there a plant of decidedly southern range: localized in upland woods of the Caucasus and southern Russia, and from the Transyl- vanian and Croatian Alps to the Maritime Alps and the Pyrenees, extending northward in the mountains to southern Germany, where (according to Milde) it is abundant in some regions of Silesia. North of southern Germany it is found at two stations in southernmost Sweden and at scattered stations in southern Norway, thence extend- ing to a point (Brénn6d) about midway along the western coast of Norway. It is not in arctic Europe, nor does it reach the British Isles, Iceland and Greenland; and the plant of eastern America does not extend north of the St. Lawrence basin, being decidedly a Canadian type. Outside Europe, P. Braunii is known in Eurasia only in northeastern Asia, whence it apparently extends across into southern Alaska. The Japanese plant is at least varietally distinct, var. japonicum Christ; and in Alaska there is a related plant, P. alas- kense Maxon, which in its more attenuate pinnae and more tapering sete of the pinnules closely simulates the Kamtchatkan P. Braunit, ar. kamtschaticum C. Chr. & Hultén, Kungl. Vet. Akad. Hand. v. no. 1: 38, t. 2 (1927), but its fronds are simply bipinnate, those of var. kamtschaticum tripinnatifid. European authors are in the habit of citing the Hawaiian Islands as also having P. Braunii, but the Hawaiian plants, P. haleakalense Brack. and P. Hillebrandit Carruth., are thoroughly distinct in many obvious characters. The Alaskan 1928] Fernald,—American Polystichum Braunii 29 material which may belong to P. Braunii is fragmentary and its identity, therefore, not readily settled, and at least some of the plants of eastern Asia are more like the eastern American than the European plant. The plant of the Canadian forest of southeastern Canada and the northeastern states, unknown on the Labrador Peninsula or in Greenland, has obviously been long isolated from the European plant, unknown in arctic Europe, the British Isles and Iceland. As a general rule, flowering plants with such ranges would be found to have quite definite characters of flower and fruit and would stand as good species: such pairs as Carez loliacea L. (Eu., e. Asia and n.w. Am.) and C. trisperma Dew. (Atl. N.A.), C. alba Scop. (Eu., e. Asia) and C. eburnea Boott (Atl. N. A.), Luzula pilosa (L.) Willd. (Eu. and w. Asia) and L. saltuensis Fern. (Atl. N.A.), Hepatica nobilis Schreb. (Eu.) and H. americana (DC.) Ker. (Atl. N.A.), ete., ete.; but, although showing recognizable, though slight, differences in the scales of the stipe and rachis, in the texture of the frond and the toothing of the pinnules, the American and European plants have essentially identical spores. P. Braunii of Europe is generally described as having herbaceous fronds: “Blatter . .. weich ... nicht tiberwinternd”—Luerssen; “Textur krautig’’—Christ; “Frondes . . seulement membraneuses” —Rouy; but in the plant of eastern America they are subcoriaceous. Consequently, in the European plant the venation of the pinnules is distinctly seen under low magnification; in the American ordinarily it is rather obscure. The largest scales of the stipe-bases in the European plant are rather firm and are prolonged into bristle-tips 5-8 mm. long; in the eastern American they are much thinner and shorter- pointed (the bristle-tip 1-4 mm. long). Under considerable magni- fication the median cells of the large scales of the European plant appear elongate-linear, with heavy cell-walls and extremely narrow lumina; while in the eastern American plant the thinner scales show shorter-linear to oblong cells (mostly 100-300 « long) with very thin walls and broad lumina, a difference similar to that separating the European Thelypteris spinulosa, var. dilatata (Hoffm.) St. John and the American var. americana (Fischer) Weatherby. In the European plant the back of the rachis (especially the lower half) bears innumer- able retrorse soft acicular scales which are far more numerous the lanceolate scales; in the eastern American plant the relation is 30 Rhodora [FEBRUARY reversed, the lanceolate scales being more abundant than in the European, the acicular scales fewer and shorter. In the European plant the terminal (and often the marginal) bristles of the pinnules are 1-2 mm. long; in the plant of eastern America 0.5-1.3 mm. long. Only a few good indusia of the European plant have been available for study, consequently it has not been possible to make a satisfactory comparison of this organ in the two plants; but, as stated, the spores present no appreciable difference unless it is that the American are minutely larger. With essentially identical spores and outline of frond, pinnae and pinnules, the two plants illustrate the conservatism of the ferns. They have obviously been long isolated but their segregation has proceeded only far enough to affect the superficial vegetative char- acters. These, however, are sufficient to make it clarifying to dis- tinguish the eastern American (and northern Chinese) plant as a geographic variety; and since the first American record of the plant was based upon its discovery by Frederick Pursh in the Green Moun- tains of Vermont in 1807 it is appropriate that it be called Porysticuum Braunu (Spenner) Fée, var. Purshii, n. var., forma typica recedit frondibus subcoriaceis; stipitis squamis scariosis latis- oblongis vel linearibus 100-300 y. longis lumine lato parietibus tenuibus pallidis; rhacheos squamis lanceolatis fibrillosisque illis pluribus; pinnularum aristis terminalibus 0.5-1.3 mm. lon is.—Cool woods, glades and shaded talus, northern and western Newfoundland, Anti- costi Island and Gaspé Peninsula, Quebec to Algoma District, Ontario, south to Cape Breton and Colchester and Kings Counties, Nova Scotia, York and Charlotte Counties, New Brunswick, Aroos- took, northern Penobscot, northern Piscataquis, northern Somerset and Franklin Counties, Maine, Carroll and Grafton Counties, New Hampshire, Berkshire County, Massachusetts, Ulster and Delaware Counties, New York, Sullivan County, Pennsylvania and Keweenaw County, Michigan; southward chiefly at altitudes from 300-1529 m. (1000-5000 feet); also northern China and Sachalin Island. TYPE: Smuggler’s Notch, Vermont, August 9, 1877, C. E. Faxon (in Gray Herb.), distributed as Aspidium aculeatum, var. Braunii. This is the plant which has regularly passed in eastern America as Aspidium aculeatum, A. aculeatum, var. Braunii, Polystichum acu- leatum, var. Braunii and Polystichum Braunii. oh EXPLANATION OF PuiaTE 159 _ Potysticaum Braunu, var. Pursuu. A small p from Cole- brook, New Hampshire, Pease, no. 10,387. wren? a Denk ai ¥. ae (To be continued) Rhodora Plate 159 pe # - . bp es fi «< rif Fy ay “EN ~ ‘B54 % id az. Herbarium of A. 8. Pea a dud ve ton Naw Kaajnad Bornmicat OWS Moy, 1982 4 (> Or Oe 46 X47 aves anies OF ARTES HEAR PHASE eo. M. Me pe rioaa oF COS PotysticHumM BRAUNU, VAR. Pursell. X CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY,—NO. LXXIX. (Continued from page 30.) IV. THE AMERICAN REPRESENTATIVES OF ASPLENIUM RUTA-MURARIA Ir has been repeatedly remarked by American students of ferns that, whereas Asplenium Ruta-muraria L. of Europe acquired its name from its common habitat, the plant of North America is rarely, if ever, found on brick- or stone-walls. European botanists are essentially unanimous as to the habitat of their plant. Thus, more than a century ago, the great student of the ferns, Schkuhr, stated it: “Hab. ad muros et in fissuris rupium”;! and later authors give the following statements. Moore: “on rocks and on ruins, and in abun- dance on old walls” ;2 Hooker: “Common on rocks and old walls and buildings” ;3 Lowe: “ Delighting to grow on old walls, both brick and stone, and more especially on the northern side, and preferring to mount up the walls as high as possible” 4 Britten: “The name is anaptone, . . . although not confined to walls, it is essentially a wall plant”;> Christ: “In bewohnten Europa ist die Art von dem anstehenden Fels und den gréssern Blécken auf die Mauern, sowohl die rohen Trockenmauern der Strassenborde und Kulturterrassen, als auf die iiltern, mit Mértel verbundenen Mauern gewandert, sodass wohl im ganzen solche kiinstliche Standorte vorwiegen; daher 1 Schkuhr, Krypt. Gewiichse, 75 (1809); rupium misprinted “pupium,”’ 2 Moore, Nat. Printed Brit. Fern, oct. ed. ii. 121 (1860). 1). ‘ Lowe, Our Native Ferns, ii. 222 (1867). 5 Britten, Eur. Ferns, 111 (1881). 38 Rhodora [Marcu der Name”? Druery: “common on old a bridges, etc. almost invariably rooted in old mortar.”? Contrasted with: dnd typical statements of the European habitat we have in America, such statements as the following. Torrey: “Limestone rocks, nas in shady situations: rare”’;? D. C. Eaton: “clefts of calcareous oc . . .. but not seen on walls in America” ;* Clute: “It loves ie ere nooks on dry cliffs”;' Waters: “Tt grows on sunny, exposed cliffs, or boulders”; Stuenen: “Seams, pockets, and ledges of calcareous rock: usually exposed to sun or in partial shade.’” From these characteristic statements it would appear that the plant of eastern America is a conservative and comparatively rare fern, that of Europe an aggressive and common type. The ranges of the Eurasian and the American plants are also strikingly contrasted. Asplenium Ruta-muraria, according to the late James Britten, is “found in most parts of the United Kingdom [Great Britain and Ireland]”; thence “It extends throughout Europe, from Norway to Spain and Portugal, Italy, the Mediterranean Islands, Greece, and Turkey. It is found both in North and South Africa, and in many parts of Asia, from the Ural Mountains to Thibet and Kashmir,”® ete. Christensen, however, cuts out the South African occurrence and says merely: “ Europa. Asia septentr.— Himalaya. U. S. A. orient.”;? but he has subsequently cited it from southwestern China near the Thibetan border." Ledebour'! does not cite it from east of Lake Baikal and it is apparently not found in extreme eastern Asia.“ The American plant occurs from Vermont to southern Ontario, south to Alabama and Arkansas; 7. e. it has an Alleghenian range. Long experience has shown that the Alleghenian flora is to a great extent made up of ancient species, with their nearest affinities in ie or adjacent continental eastern Asia, rather than in western hrist, Die Varietaten und Verwandten des Asplenium Ruta muraria L., Hedwigia, xii. i (1903). * Druery, Brit. Ferns, 76 (1910). * Torr. Fl. N. Y. ii. 492 (1 spe , Our 8 Waters, eke U 157 (1903). + 3 1928] Fernald,—American Asplenium Ruta-muraria 39 Eurasia. Consequently, it would be most unusual for a rare or conservative plant, which in America is unknown north of Vermont and southern Ontario, to be common in Europe but rare or quite unknown in eastern Asia. Furthermore, it is quite contrary to our ordinary experience, that a local and non-aggressive Alleghenian species should in Europe be a common plant, there taking possession of the artificial walls built by man: roadside- and garden-walls, the mortar of brick or stone buildings, bridges, and even church-steeples. There are plenty (more than a thousand) of cases of youthful and aggressive European plants introduced into temperate eastern America and quickly overrunning the region; but, with the exception of the calcicolous aquatic, Elodea canadensis, conservative plants of temper- ate eastern America have rarely, if ever, shown pioneering tendencies in Europe. From these considerations it would be surprising if the common European Asplenium Ruta-muraria and the rare fern of Alleghenian America are really identical. Nevertheless, I can find little indica- tion in literature that they have even been suspected of being different. Most students of the ferns, Hooker, Moore, Mettenius, Luerssen, Milde, D. C. Eaton, Underwood, Slosson, Christensen and many others, seem to have regarded them as strictly identical; and some have even made positive assertions of the identity. Thus, Torrey, in 1843, said of the American fern: “In all respects similar to the Euro- pean plant, except that it is usually smaller”; and Christ,? making a detailed study of the variations of the species, recognized no less than 17 varieties and subvarieties in Europe and identified with 4 of these European varieties the plants of eastern America. Very rarely a mild suggestion of a difference has been voiced. Thus, Michaux made the “Obs. Europaeo multo minus”; Pursh, similarly said: “The American plant is very small, scarcely ever exceeding an inch or two in height”; and Torrey’s comment to the same effect has just been quoted. But since Torrey, apparently no one has com- mented even on the difference of size. When the two series are carefully measured, however, the accuracy of the observations of Michaux, Pursh and Torrey becomes quite evident. I have studied 154 plants of the European series, with an 1 Torr. 1. ¢. 492 (1843). 2 Christ, 1. c. (1903). 3 Michx, Fl. Bor.-Am. ii. 266 (1803). ‘ Pursh, Fl. Am. Sept. ii, 667 (1814). 40 Rhodora [Marcu average of 10 or 12 fronds each, and 290 plants of the American series, with an average of 6 or 8 fronds each. Measurements of stipe and frond and count of segments give the following results: the stipes of the European plant range from 1.8-15 cm. long, with an average length of 7 cm., of the American series 1-7.5 cm. with an average length of 4.5 cm.; the fronds of the European plant range from 1.5-7.5 cm. in length, with an average of 4 cm., those of the American plant from 1-6.5, with an average of 3.3. cm.; the largest fronds of each collection show in the European 7-60 segments (average 33) in the American 3-42 (av. 15). The difference of size noted by Michaux, Pursh and Torrey is, thus, very real. But the differences are not limited to size. In European Asplenium Ruta-muraria the teeth of the segments are commonly bordered by a whitish cartilaginous rim which is frequently prolonged into a minute tip, and the mature sori become so confluent that their boundaries are quite obscured. In the American plant the teeth of the segments are coarser than is usual in the European and the cartilaginous border is only slightly, if at all, developed; and the mature sori are rarely completely confluent. is In the European plant the chaffy tips of the rhizomes and the chaffy bases of the stipes project above the root-bearing region of the short and thick rhizome and are easily seen under a low-power lens; in the American the tips of the rhizome and the chaffy stipe-bases are hidden among the rootlets of the elongate rhizome and can be seen only by dissecting away rootlets and marcescent old stipe-bases. This really important difference of the rhizomes has been beautifully shown, but apparently quite unappreciated, in some of the best illustrations. It is well displayed in such European illustrations as those of Schkuhr,! Lowe,? Britten? and Druery* and in the American plates of John Robinson,® Williamson,’ D. C. Eaton’ and Tilton.’ When the basal chaff is examined it will be found that the scales of the European Asplenium Ruta-muraria are 3-6 mm. long, made up of a lattice-like frame with whitish lumina, the broadest part of the scale with usually 6-9 rows of cells, the cell-walls much thinner than erns, ii. t. xlviii. (1869). ? Britten, l. c. figs. on pp. 111 (copied from Gerarde’s Herbal) and 112 (1881). 4 Druery, Brit. Ferns and Vars. t. v. (1910). s 7D. C. Eaton, 1. c. t. xv. fig. 1 (1879). 8 Tilton, Fern Lover’s Comp. 89 (1922). 1928] Fernald,—American Asplenium Ruta-muraria 41 the broad lumina; while the basal scales of the American plants are shorter and firmer, 1.5-4 mm. long, with fewer rows of cells, the cell- walls as thick as the lumina. In the European plant the stipe usually bears slender curving trichomes or linear-filiform scales 1-3 mm. long; in the American series the stipe is quite naked or only rarely with a few short trichomes. The indusia of the two plants are so similar as to present no quickly recognized differences, but, as already noted, the sori of the American are rarely as confluent as in the European. The spores are of practi- cally the same size; but, whereas those of the European plant are marked with coarse irregular ridges, the spores of the American have a much finer sculpturing. From this analysis it is apparent that Eurasian Asplentum Ruta- muraria is in its behavior and its technical details quite different from the Alleghenian plant which for a century and a quarter has been confused with it. It is, therefore, proper to separate the Alleghenian species as em. longa subcoriacea; segmentis 3-42 (medio 15) cuneato-rhombeis grosse dentatis, dentibus deltoideis vel oblongis obtusis vel subacutis margine plerumque vix hyalino; soris distinctis vel subconfluentibus; ris minute rugulosis.—A. Ruta-muraria of Am. authors, not L.— Caleareous cliffs and ledges, rare and local, Vermont to southern Ontario, south to Alabama and Arkansas. m ma pec Cliff, July 14, 1887, Faxon; dry limestone ledge, Milton, August 6, Boott; Mt. Philo, Charlotte, September 9, 1881, Faxon; Snake Mt., Addison, June 31, 1896, Eggleston; Pittsford, Slosson; Clarendon Gorge, August 27, 1897, G. H. Ross; H Kennedy; Manchester, M. A. Day, no. 253; cliffs, North Pownal, July 25, 1898, Eggleston; August 1, 1898, Churchill; crevices of ledges, usually in half-shade, The Cliff, North Pownal, June 22, 1901, F. G. 42 Rhodora [Marcu Floyd; with Camptosorus, Gregor Rocks, North Pownal, August 12, 1902, Blanchard. Massacuusetts: Mt. Toby, Sunderland, August 10, 1871, L. B. Tuckerman, and later collections by Bishop, Jesup et al; rocky hill, Stockbridge, June 27, 1900, Ralph Hoffmann; shaded limestone ledges, Sheffield, July 6, 1906, Bissell (ryPE in Gray Herb.); lime ledge, Sheffield, June 1; and September 30, 1919, Churchill. Conne_ECTICUT: lime ledges, Canaan, August 14, 1912 and October 28, 1916, F. G. Floyd; shaded limestone ledges, Salisbury, September 22, 1903, Bissell; ledges, Miles Mt., Salisbury, October 19, 1916, F. G. F loyd; calcareous rocks by Hous satonic River , Kent, Austin & Eames, no. 4059; Bull’s Falls, above Gaylordsville, July 28, 1889, I. Holden; dry calcareous rocks, Gaylordsville, Austin «& Eames, no. 8308; moist shaded limestone ledge by coyaeaeenag fie er, New Milford, A. E. Blewitt, no. 1205; dry limestone on shaded bank of Housatonic, New Milford, Austin & Eames, no. ee NES York: Poughkeepsie, Van Gieson; Little Falls, Vasey; Jamesville, July, 1885, Hulst. New ERSEY: Vicinity of Newton, Dowell, no. 4816; limestone rocks, Sussex Co., 1867, Austin. PENNSYLVANIA: Wilkesbarre, 1867, H. eh oe Easton, Traill Green; Bushkill Creek, Northampton Co. limestone cliffs, Centre Co., Tuttle & Rothrock. VIRGINIA: ‘at 365— 460 m. (1200-1500 ft.), Mt. Crawford, Rockingham Co., Heller, no. 785; Natural Bridge, May 30, 1891, Churchill, and at later dates by Kennedy, T. O. Fuller et al; Wythe Co., June 7, 1872, A. H. Curtiss. ae P. O. Schallert. Onrarto: north end of Manitoulin Island, Georgian Bay, Scott in herb. Geol. Surv. Can. no. 66,415. MICHIGAN: without locality, State Collection. Inp1ana: limestone, Jefferson Co., 1872, J. Hussey. TENNESSEE: on rocks along Tennessee R., Knox- ville, Ruth, no. 556; Holston River, below Concord, July, 1872, W. Faxon ; Post Oak ‘Springs, Roane Co., Pollard & Maxon, no. 412. Missounrt: calcareous rocks, near Lott’ s Mills, Perry Co., November 15, 1886, 6, C. H. Demetrio; wet mossy rocks, Shannon Co., Bush, no. 436; on boulders i in woods, Montier, Bush, no. 2795; shaded limestone rocks, near Montier, E. J. Palme er, no. 19,343. _ Iniusrrations (as A. Ruta-muraria): J. Robi inson, Ferns in their Homes and Ours, t. xvi. Pe moe Ferns of Ky. t. xix. Nad Rede’ Etchings, t. xxiv. (1879); D. C. Eaton, Ferns N. A. . t. xv. fig. 1 (1879); Tilton, Fern Lover’s Comp. 89 (1922). European Asplenium Ruta-muraria is so variable that many varie- ties have been proposed, Milde recognizing 10, Christ 17. A. erypto- lepis, on the other hand, is a tolerably constant plant. The fronds of young or small individuals have fewer and more dilated segments than in the older or larger plants but such variation is found in the 1928] Fernald,—American Asplenium Ruta-muraria 43 individual colonies and is of no taxonomic importance. In one area, however, the limestone cliffs of Clifton Gorge, of the Little Miami River in Greene County, Ohio, the plant has so far departed from typical A. cryptolepis that it may well be designated as a geographic variety. The six different collections before me are constant in having lanceolate segments with long-attenuate tips and incised margins. This extreme may be called A. cRYPTOLEPIS, var. Ohionis, n. var., segmentis lanceolatis incisis apice longe attenuatis.—Outo: Springfield, 7. G. Lea; Clifton, Greene County, 1873, J. Y. Bergen, Jr.; limestone rocks, Yellow Springs, Greene Co., August 3, 1902, June 15, 1905, L. 5S. Ho kins (tyre in Gray Herb.); on Niagara limestone, Clifton Gorge, Little Miami River, August 22, 1920, Clara G. Mark. Of this Ohio variety, Mr. L. S. Hopkins wrote, in 1907: “The single station known for it is in Greene County, and.at the present time numbers approximately one hundred plants whose short stiff rootstocks find solid footing in the cracks and crevices of a western exposure of Niagara Limestone. No other fern with which I am acquainted is so exacting in its choice of ahome. This station has been under my observation for several years, and quite a number of plants have been collected from it. Just around a corner and not over five feet from the place where strong, vigorous plants grow, not a single plant has been found on a southern exposure of the same material and ecological conditions.” —L. S. Hopkins, Fern. Bull. xv. 6 (1907). That the plants of var. ohionis have been more numerous is indi- cated by the collections of Lea and of Bergen (two different collec- tions) and by the fact that material of it has been widely distributed to different herbaria. Such a specimen, collected at Clifton in 1877 by G. E. J. Spencer, was the basis of the erroneous identification by Christ! of the Ohio plant with the European 4. Ruta-muraria, var. lanceolum Christ—the European plant, as shown in Christ’s plate, having the numerous (42-54) segments of A. Ruta-muraria, with their teeth few and small or obsolete, A. cryptolepis, var. ohionis having the few (5-28) segments and other characters of A. cryptolepis, but with elongate tips and incised margins. 1 Christ, Hedwigia, xlii. 171 (1903) 44 Rhodora [Marcu V. THE EASTERN AMERICAN Aceh gsenea tea OF ATHYRIUM ALPESTR (Plates 161-168) In 1917 Dr. F. K. Butters pointed out! that the plants which had long passed in America as Athyrium alpestre (Hoppe) Rylands or as Phegopteris alpestris (Hoppe) Mett. differ from the Eurasian type and he called our plant A. alpestre, var. americanum; and a year later Maxon, emphasizing, besides the characters noted by Butters, some habital points of the plant of western North America, elevated it to specific rank as A. americanum (Butters) Maxon? and designated as the type (which Butters had neglected to do) a plant from the Ikirk Mountains. Still later, Christensen, who certainly knows the Eurasian plant, after discussing the Eurasian range of A. alpestre said: “In America the closely allied form A. alpestre americanum Butters.”* The treatments recently published give us, then, the choice of calling the American plant either A. alpestre, an endemic American variety, an endemic American species or a “closely allied form,” and it is with the hope of throwing some clearing light upon the dilemma that the present notes and the accompanying photo- graphs are presented. In eastern America Athyrium alpestre or its representatives are as yet known from only two regions: the gulches of Bard Harbor Hill, one of the Highlands of St. John, in northwestern Newfoundland, and the northern and northeastern slopes of the Table-top Range in Gaspé County, Quebec. On my four different visits to these areas I have during the first days studied the plants with enthusiasm and have put up, altogether, hundreds of sheets of specimens; but in each case, the plants soon proving to be abundant, my enthusiasm for them has gradually waned. The plant of Gaspé had been included by Butters and by Maxon with the cordilleran var. americanum or A. americanum; consequently, when the Newfoundland material was labeled it was assumed to be the same and it was later recorded* without question as var. americanum. Comparison now shows, however, that the Newfoundland and Gaspé plants are not quite identical, nor are either of them satisfactorily placed with the cordil- leran plant. The plant of Eurasia has decidedly “leafy” fronds, ‘ Butters, Ruopora, xix. 204 (1917). * Maxon, Am. Fern. Journ. viii. 120 (1918). *C. Christens. in Hultén, Fl. Kamtch. and Adj, Isl. i, 40 (1927). * Fernald, Ruopora, xxviii. 117, 148 (1926). 1928] Fernald,—Eastern American Athyrium alpestre 45 ordinarily elliptic-lanceolate to ovate-lanceolate (pl. 161), one-fourth to one-half as broad as long, but in extreme specimens from open habitats (pl. 162) they are narrowly elliptic-oblong. Their larger pinnae are spreading or obliquely ascending, from oblong-lanceolate to narrowly deltoid and pinnatifid into broad-based oblong ultimate lobes; the sori are median and when well developed 0.75-1.4 mm. across and seemingly without any indusium. According to Butters, “Careful dissection shows, however, that the receptacle of the sorus is slightly elongated along the vein, and under a compound micro- scope it is almost always possible to find a vestigial indusium in the same position that the indusium holds in forms of A. Filix-femina with round sori. This indusium . . . often consists merely of two or three hairs joined together side by side at the base. ig Athyrium alpestre, var. americanum Butters was described as having the segments of the fronds narrower and more distant than in the European plant, the ultimate ones linear; sori only 0.5—-0.7 mm. across, submarginal and partly covered by the reflexed marginal teeth; vestigial indusium wanting. Butters added that, “ As certain inter- mediate forms occur, it is the opinion of the author that this plant should be considered a distinct geographical variety rather than a species.”” Maxon, on the other hand, subsequently said “ An examina- tion of the very ample material in the National Herbarium, however, including some of the numbers cited by Butters as intermediate in leaf cutting, reveals no specimens which are truly intermediate in this or other respects . . . The American material is essentially uniform in all respects save size, and since it differs constantly from the European plant it should rank as a distinct species, A um americanum. The invariable absence of reduced indusia, which might be regarded as an inconsequential point if the plants were otherwise like the European, is a substantiating character of some worth; but disregarding this feature, the plant is different enough in gross characters to warrant separation. The conspicuously narrow, oblique, widely separated segments give it a strict, singularly skeleton- like aspect widely different from that of the leafy European plant, which has the segments spreading and more broadly attached; and the oblique, elongate-deltoid pinnae (with the basal pinnules often greatly produced) are equally at variance from the spreading, oblong- acuminate pinnae of the European species.” Another character of the plant of northwestern North America not emphasized by Butters 46 Rhodora [Marcu nor by Maxon is the almost consistently linear-oblong to lance- oblong outline of the frond which certainly averages much narrower in proportion to its length than does the European frond. Reduced specimens of the European plant (pl. 162) have fronds very similar in outline to some of var. americanum (pl. 163) but well-developed European plants (pl. 161), have an elliptic- to ovate-lanceolate out- line. If the extreme plant of western North America (pl. 164) were alone to be considered, Athyrium americanum would upon some characters stand well apart; but when we take into account the Gaspé material, which both Butters and Maxon identify with A. americanum, and the Newfoundland material, the specific separation of the American from the European plant becomes more difficult. Such typical Newfoundland specimens as shown in pl. 165 and 166 (a small plant and a large frond from the same station) have essentially the outline, pinnae and degree of dissection of the two extremes of the European plant shown (pl. 161 and 162) and they also have the sori large and submedian as in the European plant. It does not seem possible, then, to separate them specifically or even varietally from the European material. It should be noted, furthermore, that their pinnae and those of the two European specimens illustrated are a bit more oblique than and quite as “elongate-deltoid” as in the type of A. americanum (pl. 167) specially selected by Maxon. Inci- atlas one of the stated characters of A. americanum was that sorus is “protected by a reflexed tooth of the pinnule.” This cee (which seems to be more ecological than morphological) i is obvious in most American specimens but in some extra-American material (for instance Davidsson’s Iceland plant, pl. 162) the teeth are even more vigorously reflexed than in the type of A. americanum; and many continental European specimens show some reflexing of the teeth. As a differential character it has no value. The —- plant (pl. 168) differs at once from the European and Newfo material seen in its tripinnate or, in large plants, almost quadripinnate fronds, with the ultimate strongly toothed segments linear or linear-lanceolate and remote; and its submarginal sori are only 0.3-0.8 mm. across. Its fronds have the elliptic- lanceolate to ovate-lanceolate outline of the European and New- foundland plant, rather than the linear-oblong to narrowly lance- oblong outline of A. americanum. The Gaspé plant is, therefore, 1928] Fernald,—Eastern American Athyrium alpestre 47 intermediate between that of Europe and Newfoundland and that of the cordilleran region; the European and Newfoundland plant at one extreme, the cordilleran at the other extreme of a series.! Of the characters which have been ascribed to Athyrium americanum as distinctive the reputedly more oblique and elongate-deltoid pinnae are no more oblique nor deltoid in the type of A. americanum than in some characteristic specimens of the European plant, while the recurving of the marginal teeth is more conspicuous in some European specimens than in many of the American. The characteristically narrow outline of A. americanum appears in several sheets of European material, while the Gaspé plant, which both Butters and Maxon have considered identical with the cordilleran, has, when well devel- oped, as broad an outline as the most extreme European plant. There remain to distinguish the continental American plants the narrower and more distant segments with smaller and mostly sub- marginal sori and the absence of the minute vestigial indusium which is often found, after sufficient search, in the European. But some specimens of the European plant before me have sori as small as in ~ the American, the Gaspé plant has fronds as broad as in the broad- est European, and at one of its stations (Southwest Gulch, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 27,245) the fronds of the Newfoundland plant are unusually divided and thus make a strong approach to those of the Gaspé plant. In view of these facts I am forced to the conclusion, originally reached by Butters and recently subscribed to by Christensen, that A. alpestre, var. americanum is a geographic variety rather than a species; but that the plant of Gaspé is not var. americanum, but is a second variety standing morphologically as well as geographically midway between typical A. alpestre and its var. americanum. These conclusions may be summarized in the following brief synopsis. 1 This geographic phenomenon, the Newfoundland variant of a circumpolar series e European than to the plant of continental eastern America, is frequent and I have specially noted it in the case of Cypripedium parvifiorum, var planipetalum Fern. Ruopora, xxviii. 168 (1926) and Habenaria viridis, var. inter- a Fern. 1. c. 173 (1926). ‘The Cypripedium of nortt d western Newfoundland is so similar to the Eurasian species that, at the time of describing the Newfoundland plant, I suggested that C. Calceolus of Eurasia and C. parvifiorum of continental North America might be extremes of one circumpolar type, with the Newfoundland plant bridging the morphological gap which separates them. Similarly, the northern Newfoundland (and northwest Am variety of Habenaria viridis is exactly erican) intermediate between typical H. viridis of Europe and its var. North America, It to add Athyrium alpestre to the A is interesting, therefore, showing this European and Newfoundland relationship. 48 Rhodora [Marcu Fronds pete or somewhat tripinnatifid, elliptic-lanceolate to sig ee eolate — lance-oblong), mostly one- pies to ehalf as broad as long; ‘pixnnles gx ag inane olate, sais gre broad- faded oblong ultimate lobes mostly approximate: sori Smear or submedian, the larger one 5 Sorta Pit ue alpestre, var. typicum. Fronds tripinnatifid, tripinnate or pearly. quadripinnate} ulti- ate segments of the pinnules lin r linear-lanceolate and — pet chiefly nimcratel vainly 3-0. Fronds elliptic-lanceolate to ovate-lanceolate, from three arly one-half as broad as long, tripinnate to arly pe Ne es Var. gaspensis. Fronds Phdarcoblong to narrowly lance-oblong, from one- - epaarsicgny as broad as long, tripinnatifid to ral eee ec a ee aa Nee a Var. americanum. ATHYRIUM ALPESTRE (Hoppe) Rylands, var. typicum. Aspidium alpestre Hoppe, Neue Taschenbuch, 216 (1805). Pse a aspestre (Hoppe) Newm. Phytologist, iv. 370 (1851). Phegopteris alpestris (Hoppe) Mett. Fil. Hort. Lips. 83 (1856). Athyr. alpestre mile south of Deer Pond, August 20, 1925, Wiegand, Gilbert & Hotch- kiss, nos. 27,212, 27,243—see pl. 165 and 166; turfy and mossy quart- zite rocks along Man’s Humbug Brook, Highlands of St. John, August 21, 1925, Fernald & Long, no. 27,244: wet quartzite rocks and gravel along brook, Southwest Gulch, northeast of summit of r Hill, August 22, 1925, Fernald, Wiegand, mde Gilbert & Hotchkiss, 27,245; these stations all at altitudes from 460 m (850-1 , the specimens all distributed as var. americanum and so recorded ras Fernald, Ruopora, xxviii. 117, 148 (1926). Superficially Athyrium alpestre, var. typicum so strongly resembles A. Filiz-femina (Willd.) Presl as readily to be mistaken for it; in fact, the two were both included by Linnaeus in his Polypodium rhaeticum. It is at once distinguished by its almost complete lack of an indusium (sometimes represented by a minute vestige), on which account the plant has found a place in Polypodium and Phegop- teris and as a distinct genus, Pseudathyrium. Its fronds are firmer than in A. Filiz-femina, almost coriaceous, and its spores are blackish and reticulated. Fronds of the European plant are shown in pl. 161 and 162, of the Newfoundland plant in pl. 165 and 166, the photo- graphs most kindly made by Professor J. F. Collins. __ Var. gaspense, n » frondibus elliptico-lanceolatis vel Mone lanceolatis 2-7 dm. Dee 0.7-2.5 dm. latis tripinnatis vel subquadri 1928] Fernald,—Eastern American Athyrium alpestre 49 pinnatis, segmentis ultimis linearibus vel lineari-lanceolatis plerumque distantibus; soris 0.3-0.8 mm. diametro submarginalibus.—North- eastern region of the Table-top Range, Gaspé County, QuEBEc: crevices of granitic rock, altitude 750-1050 m., easterly and northerly re es of Table-top Mountain, August 9, 1906, Fernald & Collins, 151 (small plants of exposed situation); forming extensive areas in (pe ee of alpine brooks, easterly and northerly slopes of Table- top Mountain, August 9, 1906, Fernal ollins, no. 15la (large plants up to 9 dm _ high)—nos. 151 and 15la distributed as Phegopteris alpestris and cited by Butters as A. alpestre, var. americanum; alpine and subalpine meadows and hivakailion at about 1100 m. alt., north- eastern slope of Mt. Dunraven, August 2, 1923, Fernald, Dodge & Smith, no. 25,384 (Type in Gray Herb.); brooksides and meadows at about 975 m. alt., above the cascades, head of Gorge of Northeast Branch of Riviére Ste. Anne des Monts, August 5, 1923, Fernald, Dodge & Smith, no. 25,385; subalpine meadows on eastern base (alt. about 900 m.) of Mt. Au Clair, August 10, 1923, Fernald & Smith, no. 25,386; nos. 25,384-25,386 distributed as var. americanum The hel iia photographed by Professor Collins, is thpieted in pl. 1 Var grees Butters, a a 204 (1917), excluding plan f Quebec. A. americanum (But ers) Maxon Fern Journ. viii. 120 (1918), where a pe i aoe is designated (Heacock, 554).—Alaska to Colorado, Nevada and Calif om type-specimen is illustrated in pl. 167; ore ioe in pl. 163 and 164 EXPLANATION OF PLATEs 161 To 168 (Photographs by J. F. Collins) 161, ATHYRIUM ALPESTRE aor Switzerland; 162, from eee tg A. ALPESTRE, var. AMERICANUM fro Co., California a, 0. 9538; 164, from Tuolumne Meadows, . California. Smiley, joa 810. 165 and 166, . ALPESTRE from Newfoundland, Wiegand, Gilbert ele sg ng 167, TYPE of A. — var. — ANUM, from Selkirk M ” Beitigh pg 3 Heacock, no. 168, TYPE of A. ALPESTRE, var. rife from pé Co., Quebec, Fernald, Dodge "& "Smith, no. 25,384. (To be continued.) Plate 161 Rhodora ATHYRIUM ALPESTRE FROM SWITZERLAND, X 1g Rhodora Plate 162 Olafur Davidsson. bxee bhi Amy es te . i? 4 ATHYRIUM ALPESTRE FROM IcELAND, X *% Plate 163 Rhodora PLANTS OF THER PACIFIC ATHRYIUM ALPESTRE, VAR, AMERICANUM FROM CautrForniA, X 4 Rhodora Plate 164 ~ a ~ . os 7 ~ od ate . Atbegportem alpcste' epee MEA. Erte COA Bores! Flora of the Serra Newada CALIFORNIA me S/d Pk ge fh wars ih (fas fxo beat my £ on we ne befor t perinnt ee Sta dmts. o 1% ra atin ee Fave ATHRYIUM ALPE3TRE, VAR. AMERICANUM FROM CALIFORNIA, X Rhodora Plate 165 ATHYRIUM ALPESTRE FROM NEWFOUNDLAND, X “4 Rhodora Plate 166 x _ ATHYRIUM ALPESTRE FROM NEWFOUNDLAND, Plate 167 Rhodora AMERICANUM, X 14 Type or ATHYRIUM ALPESTRE, VAR. Rhodora Plate 168 af sorties Type or ATHYRIUM ALPESTRE, VAR. GASPENSE, X 73 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY,—NO. LXXIX (Continued from page 49.) VI. PRIMULA § FARINOSAE IN AMERICA (Plate 169) Tue genus Primula, only slightly represented in America, but one SF ee 8 eet et 60 Rhodora [APRIL classification. In.America we know little of the complications which the student of the Eurasian flora must consider in untangling the species; but in the § Farinosae we have a slight illustration of these difficulties. This is best shown by the fact that the European P. farinosa L., which is apparently not found in America, has for more than a century held an undisputed, though changeable, position in our flora; and, although perfectly distinct species with natural geographic ranges, such as P. mistassinica Michx. (1803), P. decipiens Duby (1844), P. incana Jones (1895) and P. specuicola Rydb. (1913), have from time to time been set off, their authors have often stated that their new segregates were being distinguished from P. farinosa, which they believed to occur elsewhere in America. Primula farinosa seems first to have been listed as a member of the American flora in 1813 when Muhlenberg! cited it as a Canadian plant with white corolla. Just what he referred to is not clear unless it were P. mistassinica, forma leucantha. P. farinosa was more definitely admitted to our flora by Nuttall? in 1818, as growing “On the calcareous gravelly shores of the islands of Lake Huron; and . . . in the outlet of Lake Michigan,” Nuttall’s eat ‘belns really a species intermediate between P. farinosa and P mistassinica. In 1822 Torrey again identified the plant of the shores of Lake Huron with P. farinosa, saying: “On a careful comparison of the American plant with specimens of P. farinosa, from Germany and Norway, I can find no difference except that the leaves are more toothed than crenate in the former”; and, on account of the leaves Torrey* later called the plant of Lakes Huron and Michigan P. farinosa B. americana but wrongly identified with it P. pusilla Goldie,’ this time saying: “ Professor Hooker, however, thinks the P. pusilla of Goldie to be very distinct from P. farinosa; though there can be no doubt that it is the plant described above.” Goldie’s P. pusilla, however, as shown by his description, illustration and locality (near Quebec), was a common broad-leaved phase of P. mistassinica, with which species it has subsequently been generally united; and it is not conspecific with the plant described by Torrey. Gradually in America the conviction became firmly established, that P. farinosa has farinose lower leaf-surfaces and calyx, P. mistas- _ } Muhl. Cat. 19 (1813). * Nutt, Gen. 4: 119 (1818). ' *Torr. Am. Journ. Sci. iv. 1822 ‘Torr. Fl. No. and Mid. U. U. 8. i. 213 (1824), 322, t. xi, fig. 2 (1822). 1928] Fernald,—Primula § Farinosae in America 61 sinica green and efarinose; and with this highly inconstant vegetative character as the leading difference the two are distinguished in recent manuals. In the Synoptical Flora Gray! so maintained P. farinosa for the farinose plants of the Gulf of St. Lawrence region (Labrador to Nova Scotia and eastern Maine), Lake Superior, the Rocky Mountains and southern South America (four quite distinct species) and merged them without question with the very different Eurasian plant; while, merely because of its efarinose quality, he placed under P. mistassinica the otherwise quite distinct P. stricta Hornem. of the arctic regions. In 1907, still clinging to the tradition that all plants with pronounced mealiness on the foliage or calyx are P. farinosa but not satisfied that the species was a unit, I suggested? the division of the American plant into typical P. farinosa and three varieties; var. americana Torr. (a species endemic on the shores of the upper Great Lakes), var. macropoda Fernald (a mixed series, but primarily based on the plant centering about the Gulf of St. Lawrence) and var. incana (Jones) Fernald, based on P. incana of the Rocky Mountains and Great Plains. But the most reactionary treatment is that of Pax & Knuth? in Das Pflanzenreich, where not only are P. mistassinica and P. decipiens (“P. magellanica’’) mingled with P. farinosa, but where P. farinosa, subsp. eufarinosa, var. genuina Pax is made to include, of course, true P. farinosa of Europe, but also the Rocky Mountain and Great Plain P. incana and (by citation of specimens) the species of the Gulf of St. Lawrence which I have called P. farinosa, var. macropoda, but which is really quite distinct from both the European and the Rocky Mountain species. Only a casual examination of typical European P. farinosa, the Great Plain and Rocky Mountain P. incana and the eastern American P. fari- nosa, var. macropoda is needed to show how artificial is a classification which makes the three species quite identical (subsp. eufarinosa, var. genuina!) : European P. farinosa, a very slender plant with the subulate involucral bracts 4-7 mm. long, mostly only 14-14 as long as the filiform pedicels, the mature (fruiting) strongly ribbed calyx 4-6 mm. long, the corolla-lobes 4-6 mm. broad; P. incana coarser, with the lanceolate to linear-oblong flat involucral bracts 0.5-1 cm. long, mostly nearly equaling to exceeding the short and stout flowering pedicels, the mature calyx 8-10 mm.. long and only obscurely ribbed, 1 Gray, Syn. Fl. ii. pt. 1: 58 (1878). 2 Reopora, ix. 15, 16 (1907). 2 Pax & Knuth in Engler, Pflanzenr. iv’. Primulaceae (1905). 62 Rhodora [APRIL the corolla-lobes 2-3 mm. broad; P. farinosa, var. macropoda, a coarse plant with subulate involucral bracts 0.5-1.4 cm. long and with stout pedicels, the mature almost ribless calyx 5.5-11 mm. long, the corolla-lobes 3.5-5.5 mm. broad. Or again, the essentially Canadian Primula mistassinica, which superficially more closely simulates P. farinosa than do the other species, but which has consistently different seeds, is a very delicate and slender plant with subulate involucral bracts only 2-6 mm. long, elongate filiform pedicels, mature calyx 3-6 mm. long and only 2-3.5 mm. in diameter, and tiny smooth or obscurely reticulated seeds rarely 0.5 mm. long (European P. farinosa with coarser and con- spicuously pebbled or reticulated seeds); but the Magellanic species (included by Pax & Knuth along with the Canadian P. mistassinica under P. farinosa) is stout (scapes up to 4 mm. thick), with flat lanceolate involucral bracts 6-10 mm. long, umbel almost capitate owing to the abbreviated stout pedicels, mature calyx about 1 cm. long and 5 mm. in diameter, and with the largest seeds of the section, fully 1 mm. long and covered with long and conspicuous murications. That as plants P. mistassinica and the coarse plant of southern South America are not conspecific is perfectly apparent and their inclusion along with P. incana and others in P. farinosa by Pax & Knuth is due to the reliance by those authors upon single key-characters, rather than upon the sum-total of characters which really mark the different species but which, throughout their work, they largely ignore. The result of reliance upon single, and almost exclusively vegeta- tive, characters is inevitably either under-classification, such as has just been illustrated, or over-classification and the complete segrega- tion in a treatment of plants which are really closely related or even of individuals of a single species. The sections of Primula defined by Pax and Knuth well illustrate this difficulty. Thus, the only char- acter given by them in Das Pflanzenreich for the § Minutissimae (3 species of Thibet and the Himalaya) i is ff Species stoloniferae, ” the sections by “Bracteae involucrales basi gibbosae vel siseide detihastne,* yet P. farinosa and four others in the section 1928] Fernald,—Primula § Farinosae in America 63 are promptly grouped together because they have “Bracteae basi vix gibbosae.” Based upon single characters, such as are above illustrated, the sections and keys to species of Das Pflanzenreich are perplexing and contradictory in the extreme; and it is impossible to avoid the conviction that a vast reduction and reorganization will be necessary before Primula attains a natural classification. The Farinosae in America are certainly real entities and when their different characters are closely studied they resolve themselves very satisfactorily into species with definite combinations of characters, especially of flowers, fruits and seeds, and with clearly defined and natural geographic ranges. As a result of an intensive study of this small group extending over several weeks the following treatment of the American species is proposed as at least some advance over the treatments heretofore available, although, on account of poor material, it is still necessary to treat two or three plants in a tentative and, therefore, unsatisfactory manner.! Besides the material in the Gray Herbarium and the herbarium of the New England Botanical Club, I have had the great advantage of examining the splendid series of specimens in the National Herba- , species of § Nivales ey therefore, not included in the following synopsis is not nay age a ts earliest specific name and should have its nomen- clature clarified. he gest treated by Gray in the Synoptical Flora as P. nivalis Pallas. "subseauently, it hast been n repeatedly shown that the plants of the ering Sea regi q f central and western Asia; but the telat nevea relied upon by Pax and by Greene to separate the Bering Sea plant into two species do not hold. Thus Pax, following Greene, distinguished P. eximia Greene from P. p neal hoe phy ) er by ‘‘Corollae lobi . . . acuti, non aaeguorsuat ed lobi e w only in size of parts the two are apparently not —- species, but they are well marked varieties and Mrs. Busch so treats them. She fails, however, to take up the earliest specific name, not apparently because it . of difficult pronunciation, but he consistently retains the oldest name of whatever rank and consequently — fi wie pumila (Ledeb.) Pax. By the International Rules the f the two P, ype Kjellm., var. pumila (Ledeb.), n. comb. 7. nivalis, 7. pumila Ledeb., Fl. Ross. iii. 10 “aera. P. nivalis i Syn. FL. N. A. ii. pt. 1: 59 (1878), in Fur Oc. iii. pany : Greene, 1. c. 251, 260 (1897); J. M. Macoun, l. c. 569, t. xciii (1899). P. arctica Koid- Zumi, The Bot. Mag. (Japan), xxv. 216 (1911). P. eget var. arctica (Koi ) E. Busch, Fl. Sib. et Orient. Extr. Cem. 65: 75, fig. A (1926 64 Rhodora [APRIL rium of Canada (cited as “Can.’’) kindly placed at my disposal by Dr. Malte, and the Primulas in the herbaria of the University of Minnesota (cited as “Minn.’’) generously loaned by Dr. Butters, and of the University of Pennsylvania submitted by Mr. Fogg. Key To AMERICAN SPECIES OF PRIMULA § FARINOSAE a. Bracts of involucre subulate, lanceolate or linear-oblong above the dilated base, tapering gradually to the tip; their bases either gibbous, rounded or tapering, rarely much prolonged petioled or merely narrow: base, farinose or dlarincee: : mature shorter than to at most twice as long as the calyx, thick- cylindric to giinectd-ovond, 2-5 mm. in diameter: seeds dark-brown or fulvous, smooth or rough—c. De Comparatively ms t plants: scape 0.6-4 mm. in di- ameter jus the involucre, excluding the umbel 0.1. 1-4.5 dm high: inyobsstal bracts 0.3-1.4 cm. long: pedicels 0. 4-1 mm. Fig diameter: mature calyx 3.8-11 mm. long, 2.5-6 mm. in diameter at summit of tube: anthers and Kets a ertopped by the corolla-tube, not xserted from the throat of Phe shrivelled : cap- mm. in diameter: atte muricate or tinetly Teticulated (ef. no. 6), 0.5- ong—d. d. Lobes of mature calyx obtuse to pwns but not subu- ne ‘tinted et Leaves green beneath, very rarely a little farinose, subentire or obscurely undulate-dentate, 0.5-4 em. long, 0.2-1.5 em. broad: mature calyx 3.8- 6 mm. long, efarinose or only scantily farinose; the lobes about half as long as the tube: limb of corolla 5-8 mm. broad; lo’ vidi oblong or narrowly cuneate, 1-3 pag broad, shallowly emarginate, the segments a B Abas lon we ee is 1. P. stricta. Leaves strongly farinose (rarely efa ae beneath, mostly dentate, re 13 cm. long, 0.3-3 m. broad: mature cal 1 mm. long, is farinose; lobes about equaling the tube: limb of corolla 9-13 mm. broad; aie broadly ob- cordate, 3.5-5. . bro nts orth Wee Oe 2. P. laurentiana. involucre lanceolate to Pca flat falightly 2 nara only on )- Calyx copio arinose; the oblong obtuse or rarely acutish a shorter a: the tube: corolla-tube tly exceeding the calyx; limb € 10 mm. broad, lilac, wi oblong to eee mm. lobes 2-3 the) only a exceeding the calyx: seeds 0.5-0.7 merceery Ben ge ens Soh anes ” 3. P. incana. 1928] Fernald,—Primula § Farinosae in America 65 Calyx efarinose or only sparingly farinose; the ob- long-lanceolate acute to obtuse lobes oenias the tube: corolla-tube shorter than equaling the ee its white (rarely ilae-tinged) limb 1.2-2 em. broad, lobes 3.5-6 m bemee sitaie ‘ distinetly Satna the sate: ‘ ; bovoid...... 4. P. decipiens. d. Lobes of hatiie calyx sharply acuminate to sec oar sp hy en ase be ae mm. long; its limb 6-10 P, specuicola. wa ae a ee ee, Se le ee OOS Re i eee oe e. Comparatively slender: scapes 0.3-1.8 mm. in phe just below the involucre, excluding the umbel 0.1—2.5 xg involucral bracts 2-6 mm. long: pedicels 0. 2-0. 5 mm. in diameter: mature calyx mm. long, 2-3.5 m reer in diameter at summit of tube: stigma o tops of anthers exserted nage: the — of the aoivel. led corolla: capsule 2-3 mm. in diameter: s smooth or only pie f tetioulathd prs wae ‘reticulated in no. 6), 0.3-0 m. long—f. Most of the ean merely cuneate at base or narro owed gradually to the broad subpetiolar base, with tie airs of. feet: involucral sgtsicrn sa sacca P| gibbous at base—g. a. Seeds stronaly angulate and truncated, oe beneath P. intercedens. g. Seeds ro ea nearly smooth or lay specie copiously fa farinose a somewhat rhom- c: pedicels and calyx farinose.....-.------ 7. P. ajanensis. joaues oven, very rarely atta oblanceolate to pe pedicels efarinose: calyx bee IPE OL eg ue sie pare ye Ceca = ae . mistassinica. Most a the ave with petioles longer than the rather abruptly dilated rhombic, ovate or cu neate-obovate nce Ae the blades with few (1-7) pairs of teeth con- fined to the upper half or Byngise Corolla-limb 7-10 mm. broad; its emarginate to ob- a sige wee gy involucral — > eed ee ela Hee ee P. parvifolia. Corolle L 22 em. broad; its deeply perish 5-8 mm. broad: involucral bracts usually ceate gibbons OC hee ee ee 10. P. borealis. m iain uae or, rarely slightly dentate, dis- tnety slender petioted arinose: limb of corolla 5-9 mm ; aa fee its cuneate lobes distinctly shorter than the tube, 1.6-4 mm. broad, cleft a third or half | : mature capsules ender-cylindric, tapering at summit, becoming 2-3 times as long as the yx, 7-13 mm. prnaag 1.8-2.1 mm. in diameter: seeds : e-brown to stramineous, smooth......--...- 11. P. egaliksensis. - Bracts of solute phious or narrowly obovate, obtuse 0: abruptly c seeps + tp; their bases often prolonged i in narrow saccate & les 1-1.5 mm. long: leaves slender, en . P. sibirica. 66 Rhodora [APRIL 1. P. srricra Hornem. Leaves green, or only sparingly farinose beneath, oblanceolate to narrowly obovate, gsi to obscurely undulate- dentate, 0.5-4 em. long, 0.2-1.5 cm. broad: scape 1.5-30 em. high, rather strict and stout, 1-2 mm. in diameter ee the inflorescence, green or purplish and efarinose: involucral bracts lance-subulate, pedicels erect or nearly so, in anthesis from shorter than to twice as long as the bracts: calyx urceolate-campanulate, efarinose, mn —— 8. 36 mm. long, 3.5-6 mm. in diameter at summit - tube; the lobes oblong to narrowly deltoid, obtuse to acute, about half as Sem as the tube: teehee lilac or violet; the tube distinctly digee the limb 5-8 mm. broad; lobes oblong to narrowl y cuneate, 1-3 mm. broad, shallowly bet any the sha 0.2-1 mm. long; capsule ellipsoid only slightly exceeding the calyx, 3-4 mm. in diameter: seeds more or less angulate, dark-brown or fulvous, 0.5-0.8 mm. long, Spicy reticulated. —Hornem. in FI. Dan. viii. fase. 24, t. mecelxx 1810); Duby i “ 880) ; ; ; a Pax & Knuth in Engler, Pflanzenr. iv™’. 86 (1905); J. M. Macoun & Holm, Rep. Can. Arct. Exped. 1913-18, v. pt. A. t. xi. fig. 6 (1921); FE. Busch, FI. Sib. et Orient. Extr. iv. Cem. 65:38 (1926). P. farinosa, 8. stricta (Hornem.) Wahlenb. Fl. Lapp. 60 (1812). P. sshd ag niana Lehm. Monogr. Prim. 55, t. 4 (1817); Hook. FI. Am. il. 120 (1838), in small part only. P. Pre ro Picakes ex x W. Nyl. & Saelan, Herb. Mus. fond (1859) 32, acc. to Pax & Knuth P. mistassinica Gray, Syn. Fl. N. A. ii. pt. 1: 58 (1878) in part, not Michx. (1803). P. farinosa, var. mistassinica Pax, Engler’s Jahrb. x. 200 (1889) in part, not P. mistassinica Michx. (1803). P. farinosa, var. anges es in Engler, Pflanzenr. iv’. 84 (1905) in part, not P. stricta, var. groenlandica Warming | (1886). and subarctic Eurasia, Greenland and North America; with us south to northern Labrador, northwestern Quebec, northern Ontario and Alberta. The following American specimens have been examined. GREENLAND: without locality, ex Lehmann; Umenak, Rink; Ata, lat. 70° 16’, kngnet 6, 1921, A. E. Porsild; Qeqerlatik Najarsuit, lat. ugust 3, 1911, M. P. & A. E. Por sild; Kiigsinerssuaq and Ata, lat. 70° i7’, aly 11, 1923, M. P. Porsild; Itivdleq-Fjord, Quingua, lat. , July 6, 1926, M. P. Porsild; the Porsild specimens all distributed as P. mistassinica. LaBrapor: moist banks, Nachvak, R. Bell, no. 15,829 (Can.), as P. mistassinica; Rama, A. Stecker, no. 78, as P. farinosa. — : Ungava (“ ierthern eet 1884, L. M. Turner, July, 1897, A LP, Low, no. 24,529 (Can.), as P. egalik- sensis; River Kovik, lat. 61° 59’ , Hudson ‘Straits, Low, no. 23,025 (Can.), as P. farinosa; Richmond Gulf, June 28, 1890, Spreadborough, ‘no. 14,421, June 12, 1899, Low, no. 63,242 (Can.), bed us P sibirica; 1928] Fernald,—Primula § Farinosae in America 67 north of Cape Jones,! James Bay, A. P. Low, no. 63,244, as P. stricta, altered to P. sibirica; damp banks, South Twin Island, James Bay, J. M. Macoun, no. 15,831 (Can.), as P. farinosa. ONTARIO: “ growing below high-water mark,’’ west coast of Hudson Bay, lat. 56°, August, 1886, J. M. Macoun, no. 15,850 (Can.), as P. mistassinica or P. sibirica; Cape Henrietta Maria, Spreadborough, no. 62,555 (Can.); mouth of Ekwan River, James Bay, Dowling, no. 34,526 (Can.), as P. farinosa; The Beacon, mouth of Moose River, Spreadborough, no. 62,554 (Can.), as P. sibirica. Manrrosa: Churchill, lat. 58° 50’, J. M. Macoun, no. 79,388, as P. stricta or P. farinosa; Churchill River, oa Coronation Gulf, Port Epworth, Cox & O’ Neill, no. 581; Bernard Harbour, Frits Johansen, no. 347; Great Bear River, Elizabeth Taylor, no. 87; shore, Great Slave Lake, R. Bell, no. 23,151 (Can.). Yukon: near mouth of Lewis River, Gorman, no. 1052 (Can.). ALBERTA: ocky Mountains, Drummond, as P. farinosa or P. scotica; head of Pabocton Trail, S. Brown, no. 1107, as P. borealis. The Greenland and American plants cited seem to me quite insepar- able from Scandinavian material of typical P. stricta, although the tiny plants from the Arctic coast of Mackenzie might be set off as var. jacutensis E. Busch; they agree closely with Mrs. Busch’s descrip- tion and figure of the plant of northeastern Siberia, but seem more like dwarfed arctic extremes than a true variety. Pax & Knuth exclude P. stricta from America and cite all Green- land and Labrador material under P. farinosa, var. groenlandica, which they base upon P. stricta, var. groenlandica Warming, Svensk. Vet. Akad. Hand. xii. Afd. iii. No. 2: 21, fig. 7, A-D (1886). Lange, Conspect. Fl. Groenl. 260 (1887), however, maintained both P. stricta and P. stricta, var. groenlandica Warm. in the Greenland flora, remarking that the latter is a “Forma intermedia inter P. strictam et P. egaliksensem.” Surely, the figures of var. groenlandica published by Warming are of a plant scarcely, if at all, separable from P. egaliksensis. The distinctly petioled leaf with abruptly dilated entire blade, the large and plane involucral bracts, short flowering pedicels, comparatively slender calyx with narrow lobes, only slightly exserted corolla-tube with deeply-notched lobes are all characters of P. egaliksensis and Warming’s figures are readily matched in that species, but not in P. stricta. In fact, Warming’s 1 The material is labeled in the hand of the late J. M. Macoun “North of Cape Jones, Hudson Strait’; but Low’s report for 1899, when it was collected, explicitly refers Bay ’’—See Low, Geol. Surv. Can. to ‘‘Cape Jones at the entrance to James Ann. Rep. n. s. xii, 144A (1902). N 68 Rhodora [APRIL figures are not appreciably unlike the original plate of the white- flowered P. egaliksensis Wormsk. in Hornem. FI. Dan. ix. fasc. 26: t. mdxi. (1816), except that P. stricta, var. groenlandica has purple corollas. Just such a plant occurs across boreal America, in northern Newfoundland, northern Quebec, Alberta, British Columbia and Alaska; and of the abundant series which I have collected and studied in Newfoundland (six numbers, representing all stages from young flowers to mature fruit) the smaller specimens exactly match Warm- ing’s original figures of P. stricta, var. groenlandica and, better still, the four individuals in the Gray Herbarium of his type collection, gathered at Itivnek-Elvens in the Holstensborg District of Greenland on July 13, 1884. The Newfoundland plant with violet corollas, watched closely in the field, can be separated from typical white- flowered P. egaliksensis only by its intense color. Its fruit, collected at a specially marked station, is quite like that of P. egaliksensis, except that the capsules are deeper-colored: the slenderly cylindric capsules 2-3 times as long as the calyx and only 1.5-1.8 mm. thick (the capsules of P. stricta ellipsoid, only slightly exceeding the calyx and 3-4 mm. in diameter; the capsules of P. farinosa likewise ellipsoid, only slightly exserted and thick). Furthermore, in both P. farinosa and P. stricta the dark-brown seeds are obviously muricate or reti- culated, in P. egaliksensis the stramineous or pale-brown seeds smooth or at most obscurely reticulated; and the seeds of P. stricta, var. groenlandica are like those of P. egaliksensis. It is now very clear, then, that P. stricta, var. groenlandica Warming belongs neither to P. stricta with which he placed it not to P. farinosa to which it was transferred by Pax & Knuth, but that it is a variation of P, egaliksensis, as Lange has already suggested. It is, furthermore, clear that most of the material cited by Pax & Knuth under P. farinosa, var. landica really belongs to P. stricta. 2. P. laurentiana, Fernald, nom. nov. (Plate 169). Leaves farinose (rarely efarinose) beneath, obl lat tulate or narrowly rhombic- ovate, mostly petioled and dentate, 1-13 cm. long, 0.3-3 cm. broad: scape 0.1-4.5 dm. high, strict and stout, 0.6—3 mm. in diameter below the inflorescence, often farinose at summit: involucral bracts lance-subu- late or strongly involute, usually strongly saccate or gibbous at base, 0.5-1.4 cm. long: umbel 1-17-flowered: pedicels erect or strongly r, m practically wanting to 5 cm. long, comparatively stout (up. to 1 mm. thick): calyx urceolate-campanulate, usually _ fartnose, in maturity 5.5-11 mm. long and 3-6 mm. in diameter at ‘summit of tube; the lobes lanceolate, oblong or narrowly deltoid, 1928] Fernald,—Primula § Farinosae in America 69 obtuse to acute, about equaling the tube: corolla lilac; the tube but slightly exserted; the limb 9-13 mm. broad; lobes broadly obcordate, 3.5-5.5 mm. broad, with segments 1.5-3 mm. long: style and anthers not exserted from the yellow throat: capsule ellipsoid, from slightly exserted to twice as long as the calyx, 2.5-5 mm. in diameter; its valves splitting into linear halves 1.5-2 mm. wide: seeds angulate, 0.5-0.8 mm. long, conspicuously reticulated: flowers with the fragrance of Narcissus Jonquilla; roots musky.—P. pusilla Sweet, Brit. Fl. Gard. ser. 2, i. t. 5 (1831), not Goldie (1822). P. farinosa var. macropoda Fernald, Ruopora, ix. 16 (1907) mostly, including the type-specimen, not P. macropoda Craib, Notes Roy. Bot. Gard. Edinb, xi, 176 (1919). P. scotica Hook. Fl. Bor.-Am. il. 120 (1838), not Hook. in Curt. FI. Lond. iv. t. 133 (1821). P. farinosa, var. genuina Pax in Engler, Bot. Jahrb. x. 199 (1889) as to eastern American citations. nur incana St. John, Can. Dept. Mines. o. 126: 104 (1922); Fernald, Ruopora, xxviii. 224 (1926); not var. incana (Jones) Fernald Ruopora, ix. 16 7). P. farinosa of eastern Am. authors, not L Nova Scotia and eastern and north-central Maine. The following, selected from many specimens, are characteristic. LABRADOR: banks of Naskaupi River, about 18 miles from mouth, Wetmore, no. 103,037; Indian Harbor, lat. 54° 27’, Ralph Robinson, no. 102; Battle Harbor, Bowdoin College Exped. no. 104; Barge Point, July 17, 1913, W. E. Ekblaw; Forteau, 1870, S. R. Butler. NEWFOUNDLAND: turfy limestone barrens, Burnt Cape, } Griscom, Gilbert & Hotchkiss, no. 28,905; gravelly limestone shore, stone ledges, St. Barbe, Fernald, Long & Dunbar, no. 26,956; talus of calcareous sandstone escarpments, no. 28,912; calcareous rocks and talus, Port Saunders Harbor, Fernald & Wiegand, no. 3889; conglomerate limestone, etc., Cow Head, Fernald & Wiegand, nos. 3885, 3887; boggy spots on bas crests, Twillingate, Fernald, Wiegand & Bartram, no. 6068; dry sea-cliffs, i ald. bare spots, Quesxc: limestone and calcareous sandst Fernald & Wiegand, nos. 3888, 3890, Fern : 28,914; grassy shore, Wapitagun, July 14, 1927, H. F. Lewis Can.); rocky shore, Goynish, St. John, no. 90,675; limestone headland, 1 'The original labels read ‘ Labrador,’’ but by recent decision of the Privy Council _ Blanc Sablon is transferred to Quebec. 70 Rhodora [APRIL Pointe-aux-Esquimaux, Mingan, St. John, no. 90,674; plages cal- caires de la petite riviére, Pointe-aux-Esquimaux, Victorin& Rolland, o. 18,485; sur les caleaires du rivage, le 4 Marteau, Mingan, Vintorss & Rolland, no. 18,565; ae ie rivages calcaires, Grande Ile, Mingan, Victorin & Rolland, no. 21,832; Baie Ellis, Anticosti, Violorin, no. 4188; cegthegg agiloalcie, Rividre Jupiter, Anticosti, Victorin & olla md, 25,139; Salt Lake, Anticosti, J. Macoun, no. 15, (Can.); Bohas entire conglomerate (calcareous) senschifis) eearor venture Island, Fernald & Collins, no. 1148, Victorin et al, 17,642, 17,644; cliffs and ledges, Percé, J. M. Macoun, no. 68,949 (Can.), sting, Fernald & Pease, nos. 5319, 5320, 5434, 5435, 5554, Fernald & Collins, nos. 1147, 1149; sur les pre ne Anse A I’Indien, Victorin, Rolland, Brunel & Rousseau, no. 643; limestone -cliffs, Cape Rosier, Frits Johansen, no. 103,287 (Gan calcareous sea-cliffs, Christie, Fernald & Pease, no. 25,233; sea-cliffs, Tourelle, Griscom, Mackenzie & Smith, no. 25,981; calcareous sea-cliffs, Sabo Hughes, Fernald & Pease, no. 25,232; wooded banks of the St. Lawrence, Matane, August 1 F. Forbes; shaded calcareous cliffs, Bic, Fernald & Collins, nos. 243 (TYPE in rei Herb.), 1146; wet shore of the St. Lawrence, Temiscouata, Augu pe Pringle; Pointe a Persil, Riviére du Loup, Victorin, no. MacpaLEN Te ices Entry Island, June 23, 1861, Hyatt, Verrill . ‘Shaler Nova Scotia: dripping diffe. Bixter’s Harboor, July 10, Lop F. G. Floyd; cliffs and ledges, Morden, W. H. sesieteton no. 644 (Can.); crests of ogee cliffs by Bay of Fundy, near > peaelse Bissell, Bean, White & Linder, no. 22,234; turfy crests and slopes of exposed head- sey Markland (Cape Forchu), Fernald, Bartram, Long & Fassett, no. 24,327; Chebogue Point, John Macoun, no. 81,152. MAtneE: Houlton, 1880, 1881, Kate F arbishs foot of Mt. ‘Kinens Moosehead Lake, August, 1866, AR&@ C. E. Smith; north side of Mt. Kineo, September 21, 1887, G. G. enrger dees Libby Islands, Machias- port, Cushman & Sanford, no. P. laurentiana is the plant ae originally as P. farinosa, var. macropoda. account of P. macropoda Craib it is necessary to assign anew name. It isa coarser plant than the European P. farinosa to which it has always been referred, either as identical or as a geo- graphic variety. The dwarf northern extremes (from southern Labrador and northern Newfoundland) simulate P. farinosa and have often been mistaken for it, but in its typical development, P. laurentiana is taller and stouter; the bracts of the involucre longer, 0.5-1.4 em. long (in European P. farinosa 4-7 mm. long); the pedicels stouter; the calyx commonly much more farinose, urceolate-cam- panulate, in fruit 5.5-11 mm. long and 3-6 mm. in diameter (the efarinose to but slightly farinose calyx of P. farinosa more turbinate, 1928] Fernald,—Primula § Farinosae in America 71 in fruit only 4-6 mm. long and 2.5-4 mm. in diameter); the capsules larger and the seeds with more conspicuous reticulation. Plate 169 is from a photograph taken at the type-station by Professor J. F. Collins. In view of the pronounced selection of calcareous habitats by P. laurentiana it is worth noting that Contejean classified the European P. farinosa as one of the “Calcifuges presque indifférentes, cependant plus nombreuses sur les sols privés de calcaire;”! and that Warming indicates? P. farinosa as a typical oxylophyte, listing it along with Vaccinium uliginosum and V. Oxycoccus. the other hand, Tansley makes Primula farinosa in Great Britain distinctly calcicol- ous, saying, “ Actaea spicata and Primula farinosa also seem to have found the siliceous soils of the Leeds and Halifax district an effectual barrier against a southern extension of their range” ;° and he definitely lists it as one of the characteristic plants which “On the Pennines, for example, . . . occur on the swamps of the limestone hills.” In P. laurentiana the pedicels are commonly elongate, but at the northern part of its range they may be very short or almost wanting. These specimens with abbreviated pedicels have been confused with the Great Plain and Rocky Mountain P. incana, but they differ from that species in their involucral bracts, broad corolla-lobes, and other characters which show them to be merely dwarfed states of P. lauren- tiana. Other plants of P. lawrentiana with unusually small calyces and capsules have sometimes been identified with the plant of the Great Lakes which Torrey described as P. farinosa, var. americana; but the Great Lake material, though often quite farinose (and thus strongly simulating P. farinosa and P. laurentiana) has a technical character which allies it as much to P. mistassinica: the capitate stigma or the tops of the anthers protruding from the throat of the shrivelling corolla. The plants of Newfoundland and Quebec which have been misidentified with P. farinosa, var. americana are clearly only attenuate forms of P. laurentiana. The plants of Mt. Kineo, Maine are extreme cases of this attenuation; the leaves being remark- ably thin and delicate, the scapes unusually slender, and the small calyces with unusually thin and sharp lobes; but these seem to be only slight ecological modifications, presumably due to the habitat, 1 Contejean, Influence du Terrain sur la Végétation, Ann. Sci. Nat. sér. 6. ii. 300 ee weesiala Oecology of Pl. ed. Groom & Balfour, 193 (1909). ? Tansley, Types of British Vegetation, 157 (1911). 72 Rhodora [APRIL at the foot of a north-facing precipice rather than in more exposed and better illuminated spots, such as the plant usually selects. Ordinarily P. laurentiana, like P. farinosa and P. incana, has the lower surface of the leaf strongly farinose or whitened with waxy particles. When the specimens have been dried over extreme heat or when immersed in alcoholic solutions in poisoning, the wax is often removed and the leaves have a deceptive post-mortem green- ness. Occasionally, however, considerable colonies of P. lauwrentiana are found with absolutely green and efarinose foliage; and, unless their characters of calyx, corolla, capsule and seeds are carefully examined, they are likely to be misidentified either as P. stricta, which is high-northern, or with P. mistassinica, which abounds through much of the range of P. laurentiana but which has more slender scapes and pedicels, calyx and capsules, shorter bracts and smaller and only obscurely pebbled seeds. The green form of P. laurentiana is apparently parallel with P. farinosa, var. denudata Koch of Europe. It is merely a minor form, but as a striking varia- tion, may appropriately be designated . LAURENTIANA, forma chl 1 n.f., foliis subtus efarinosis.— NEWFOUNDLAND: Cape Norman, Wiegand & Long, no. 28,909; Sacred Island, Wiegand, Gilbert & Hotchkiss, no. 28,908; Sandy (or Poverty) Cove, Fernald, Long & Dunbar, no. 26,955; Capstan Point, Flower Cove, F ernald, Long & Dunbar, no. 26,957; Yankee Point, Wiegand & Hotchkiss, no. 28,904; turfy limestone barrens, “Hs Pe ninsula, oye 27, 1925, Fernald, Wiegand, Long, Gilbert & Hotchkiss, no. 28,913 (ryPE i in Gray Herb.); Bard eae ghd Fernald & Long, no. 28,911. Quesec: Anticosti Island, Pursh; lets Perroquets, ingan, Victorin & Rolland, no. 18,421; wet mossy swale, Puffin Island, St. Mary Islands, July 27, 1927, H. F. Lewis (Can.). Matne: base of Kineo Cliff, Moosehead ‘Lake, July 24, 1866, August, 1867, A. H. ae yl Smith (Penn.), June 8, 1878, F. S. Bu nker 3. P. incana Jones. Leaves strongly (rarely only slightly) farinose beneath, tiohe Roa neat or spatulate, without petioles or in attenuated plants with winged petioles, obtuse, shallowly denticulate, 1.5-8 em. bei 5 5-2 em. broad: scape 0.5-4.5 dm. high, strict and inal 1 hes mm. in diameter below the crete , farinose at summit: e-ca’ _ maturity 8-10 mm. long and 4—5 mm. in diameter; the oblong ob _ or rarely acutish lobes shorter than the tube: corolla lilac; the tube 1928] Fernald,—Primula § Farinosae in America 73 slightly exceeding the calyx; the limb 6-10 mm. broad, with oblong to cuneate-obovate lobes 2-3 mm. broad; stigma and anthers not exserted from the yellow throat: capsule ellipsoid, only slightly exceeding the calyx: seeds strongly angled, 0.5-0.7 mm. long, conspicuously reticulated. —Proc. Cal. Acad. ser. 2, v. 706 (1895). P. farinosa, var. Gray, Proc. Acad. Nat. Sci. Phila., 1863: 70 (1863). P. dealbata Engelm. in Gray, |. ¢. 1863) as synonym. P. americana Rydb. Bull. Torr. Bot. Cl. xxviii. 500 (1901). P. farinosa, var. genuina Pax, in Engler’s Bot. Jahrb. x. 199 (1889) in part, as to Rocky Mt. material. nosa, subsp. eufarinosa, var. genuina Pax & Knuth in Engler, Pics: iv*57, 83 (1905), as to synonymy in part and citation of Colorado material. P. farinosa, var. incana (Jones) Fernald, Ruopora, ix. 16 (1907). P. farinosa, var. macropoda Fernald, Ruopora, ix. 16 (1907) in small part (as to plants of Saskatchewan, Athabasca and Mackenzie). [Illustrated as P. farinosa in Clements & Clements, Rocky Mt. Flowers, t. 16, fig. 1 (1914) and by MeCalla, Wild Fl. W. Can. 37 (1920) _—Meadows, bogs and damp places, Mackenzie to Colorado and Utah. Mackenzie: Great Slave Lake, Richardson. SASKATCHEWAN: without locality, 1858, Bourgeau; Carlton House, Richardson; bank of Saskatchewan R., Prince Al rt, J. Macoun, n cCalla, no, 2590 (Can. e rH. H. Gaetz, no. 7476 (Can. ); boggy ground West Fork, Water Coulée, near Rosedale, Moodie, no. 942; bank w River, Calgary, Malte & Watson, no. 118,331 (Can.); Elbow y Fas J. Macoun, no. 24,528 (Can.); Rocky Bar, 20 mi. south of Jackson, Merrill & Wilcox, no. 1039; wet ik Adams Ranch, Jacksons Hole, Merrill & Wilcox, no. 990; low m near North Pilot Butte, Merrill & Wilcox, no. 749; meadow 20 m mi. west of Big Piney, Sublette a tees & "Payson, no . 2648. CoLo- RADO: Rocky Mts., lat. 39° 1862, Hall & pie asi no. 378; South Park, E. L. Hughes, no. a Grape Creek, Custer Co., July 2, 1888, Demetrio; Gunnison, Baker, no. 361. Jones’s original material was from cold bogs at the head of Sevier River, Utah, I have not seen the type but the description clearly 74 Rhodora [APRIL belongs to the characteristic plant of the Rocky Mountain and Great Plain area. The name P. dealbata Engelm. (1863), having been published merely as a synonym, cannot displace the validly published, but later P. incana Jones (1895). In its smaller extremes P. incana closely simulates P. stricta, but when well-developed it is quite distinct in its large and copiously farinose leaves, flat and rather broad bracts, and longer and broader strongly farinose calyx. In its subcapitate inflorescence and plane bracts P. incana is nearer related to P. decipiens of southern South America (P. magel- lanica of authors) than to other members of the Farinosae. This relationship of the Magellanic and Rocky Mountain plants, long familiar in other groups, was clearly recognized by Asa Gray, who, however, failed to detect the characters which separate the two. In enumerating the Rocky Mountain plants of Hall & Harbour, Gray said: “378. PRIMULA FARINOSA L., var. foliis sessilibus; umbella capitata; calyce cylindraceo tubum corollae subaequante. P. dealbata, Engelm. in litt. But it exactly accords with the eftchand figure of P. farinosa, var. Magellanica of Hooker’s Flora Antarctica (P. decipiens, Duby), and with my Antarctic specimens, a that the calyx is perhaps a little longer, and the corolla bluish-purple. t is interesting thus to connect the Antarctic with tg northern fori, by specimens from the Rocky Mountains in about lat. 40°. As stated, Gray overlooked some very real characters: P. decipiens (P. farinosa, var. magellanica) with longer calyx-lobes, the corolla with shorter tube and with much larger white (rarely lilac-tinged) limb (1.2-2 em. broad) with lobes twice as broad, capsule much longer, and seeds the largest in the section (1 mm. long), rounded- obovoid and conspicuously muricate; but the affinity of P. incana is, nevertheless, with P. decipiens rather than with P. farinosa. 4. P. pecrprens Duby. Leaves spatulate to narrowly ovate, obo- vate or rhombic, tapering to a sessile base or broad petiole, serrulate- ong, 0.7-2 cm. broad, efarinose or farinose beneath: scape 0.3-5 dm. high, stout, up to 4 mm. in diameter below the umbel, f. at summit: involucral bracts flat, lanceolate, 6-10 mm. long, broadly gibbous at base: wmbel subca capitate, few- to many-flow vered: pacers ga stout, a oe sascbaemesate in of . 2-10 mm.: calyx in maturity se, ) Gray, Bree end. Nat. Sci. Phila. (1863) 70. 1928] Fernald,—Primula § Farinosae in America 75 calyx: seeds dark-brown, 1 mm. long, rounded-obovoid, conspicuously vcate.—Mém. Fam. Prim.—Mém. Soc. Phys. d’Hist. Nat. Genéve, x. 46, t. ii. fig. 1 (1844); Duby in DC. Prodr. viii. 44 (1844). P. magellanica Hook. Fl. Ant. ii. t. exx. (1847); Skottsberg, Bot. ( i . Phan. 93 (1910); Vallentin & Cotton, Ill. Fl. Pl. and Ferns Falk. Isl. t. 41 (1921); not P. magellanica Lehm. (1817). P. farinosa Gay, FI. Chil. iv. 367 (1849), not L.— Falkland Islands and from Fuegia northward along the Andes of Chili and Argentina to lat. 38° (acc. to Skottsberg). There is no doubt as to the application of the name Primula decipiens. Duby clearly illustrated the plant and his description explicitly gives the diagnostic characters: “ foliis . . . serrulatis Cieke ek 8 _ invol. . . . calyces subaequantis foliolis elongato-lanceolato-linearibus . . ., calycis . - - - laciniis obtusis r . . lobis late obcordatis as ors aa Flores subsessiles etiam in planta fructifera; corollae majores quam in P. farinosa.” Lehmann’s P. magellanica, to which all subsequent authors except Duby have referred the Falkland and Patagonian plant, differed in many striking characters from the plant with subcapitate white flowers which Duby correctly set off as P. decipiens. Lehmann’s description called for “Folia . . . dentato-crenata ” and his figure of a large plant shows teeth less salient than in large plants of P. decipiens. P. magellanica had “Involucri foliola : unguicularia,” but the bracts of P. decipiens are flat and hardly unguiculate. The umbel of P. magellanica was described: “ Pedicelli unciales : Calyx . . . lacinis ovatis, acutis, Corolla®. . . carnea: . . . laciniis cuneiformi- 1 In the text Hooker treats the plant as a variety, but the caption of the plate mula Magellanica. 76 Rhodora bus”; and the plate accurately coincides with the description, showing flowering pedicels twice as long as the involucre, ovate acute calyx- teeth only half as long as the tube, and narrowly cuneate corolla- lobes only 3-3.5 mm. wide. P. decipiens, however, has pedicels nearly obsolete or in fruit only a few millimeters long, the calyx- lobes oblong and obtuse to only subacute and equaling the tube; the white (rarely lilac-tinged) corolla with cuneate-obovate lobes much broader (3.5-6 mm. broad). All authors except Duby have, as said, consistently treated the plant of subantarctic South America as P. magellanica or as P. farinosa, var. magellanica (or sometimes as P. farinosa), but either Lehmann’s plant did not come from the Straits of Magellan, as he supposed, or else he had a very rare species which has escaped subsequent collectors. It is to be noted that he received the plant indirectly and it is probable that there was some error as to its geographic origin: “Pulchra haecce et nova species Parisiis a Dom. de Jussieu absque nomine, sed cum nota: hab. ad fretum Magellanicum, communicata mecum est. In nullo alio herbario eam vidi; neque minus, qui plantam detexerit, cognitum habeo.” At any rate, unless P. decipiens is far more variable than the six collections before me and the descriptions or plates of such authors as Duby, Hooker, Decaisne, Skottsberg and Vallentin & Cotton indicate, it is wiser not to take up for it the name P. magellanica Lehm. The occurrence of P. decipiens in southern South America, sepa- rated by about 78 degrees of latitude from its nearest ally in Colorado and Utah, has naturally attracted comment. Thus Hooker, failing to note the characters of the involucre and the very distinct seeds and peat agra reducing the plant to varietal rank under P. farinosa, said: “One argument which militates against the commor origin of the individuals from the opposite hemispheres, must not be overlooked; it is the absence of the plant, and, indeed, of the whole genus, in any part of the Andes [i. e. the Cordillera] south of 39° north lat.; a circumstance which makes it very difficult to account for its appearance in the two opposite temperate zones, if all the individuals of both hemispheres are supposed to have sprung from one parent. Gray’s comment in 1863, when he identified the Great Plain and ‘Rocky Mountain plant with the Magellanic, has been quoted in the discussion of P. incana. Franchet (1889) and Macloskie (1905) 1 Hook. Fl. Ant. ii. 337 (1847). 1928] Benke,—Some Illinois Asters 77 have reiterated the fact, but have added nothing to its interpretation. It is at least noteworthy that the Magellanic plant is closest related, apparently, to P. incana of the northern Cordillera and Great Plains (Utah and Colorado north to Mackenzie) and that the smallest extremes of the latter are separated only with difficulty from the cir- cumpolar arctic and subarctic P. stricta. In eastern North America, centering on the unglaciated areas about the Gulf of St. Lawrence, dwarf extremes of P. laurentiana, superficially so strongly resemble P. incana that they have been mistaken for it, and in its efarinose form P. laurentiana is separated from the arctic P. stricta only with difficulty. Similarly, P. farinosa of temperate Eurasia is often difficult to distinguish from P. stricta and under their treatment of the latter species Pax & Knuth specially say: “Species haec valde affinis Pr. farinosae et forsan melius pro ejus varietate habenda.’ From this line of evidence it may well be that the arctic P. stricta is the progenitor from which have been derived P. farinosa and other species of Eurasia, P. lawrentiana of the Gulf of St. Lawrence region, P. intercedens of the upper Great Lakes, P. incana of the North American cordillera and still farther isolated, P. decipiens of the southern American cordillera.' s of descent from livin ic species have been suggested in Fernald. Persistence of riots in Unglaciated Soe. of Boreal America, M A xv, (1925); and numero’ es Primula stricta, with endemic repre- ntatives along the North American cordillera or about the Gulf of St. Lawrence, have eccie or allies; in Puccinellia, Draba, Sazifraga, Empet , Euphrasia, Antennaria, Agoseris, Taraxacum, etc. while others, like Carex ac C, Macloviana, Plantago juncoides, etc., show little, if any, differentiation, (To be continued.) CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY.—NO. LXXIX. (Continued from page 78.) em. long, 0.7—2 cm. broad, farinose to efarinose beneath: scape 1-1.6 dm. high, 1.5-2 mm. in diameter, efarinose: involucral bracts lance- attenuate or -involute, 4-10 mm. long, dorsally thickened but scarcely gibbous at base: umbel 6-12-flowered: pedicels strongly ascending, in anthesis 0.5-1 cm., in fruit sometimes elongating to 4 cm. long: calyx farinose, campanulate, becoming turbinate, in maturity 6-9 mm. long mm. in diameter; the lance-attenuate to -subulate lobes about equaling the tube: corolla-tube yellowish, 8-10 mm. long, its violet limb 10 mm. broad, with the narrowly cuneate lobes ginate: stamens inserted near the middle of the tube; the anthers and stigma not no. type was from “loose soil, under overhanging cliffs in the alcove-like heads of the canyons, characteristic of the limestone bluffs of San Juan River,” near Blufis. P. specuicola differs from P. farinosa and from all other American species which have been called P. farinosa in the very sharp calyx- lobes, long corolla-tube and short capsule. I have seen no good flowers but Rydberg’s description of the corolla-limb as “dark 86 Rhodora [May violet”? and Miss Eastwood’s characterization of it as “crimson,” suggest much deeper coloring than in the other Farinosae. Neither of these authors mentions a yellow eye such as occurs in most of the species. The long corolla-tube, deep color of the limb, shape of fruiting calyx, scarcely gibbous bracts, and thin sinuate-dentate leaves suggest that P. specuicola may be nearer related to P. Rusbyt Greene than to the Farinosae; but P. Rusbyi has much larger flowers, the leaves with vadraneice: or callous teeth and the caudex bearing marcescent brown shea 6. P. intercedens, nom. nov. Leaves yellow-farinose iit some- times green, firm, berate, spatulate or oblanceolate, 1-7 cm. long, cuneate at base but scarcely petioled, crenate-dentate or ibontibe with the margin commonly revolute: scape 0.4—2.5 dm. high, filiform to stoutish, 0.5-1.8 mm. in diameter and sometimes slightly farinose at summit: bracts of the involucre lance-attenuate, involute or subulate at tip, 3-6 mm. long, slightly dilated and thickened but scarcely gibbous at base: umbel 1-10-flowered: the stiffly fastigiate filiform pedicels at first only 1-10 mm. long, in fruit lengthening to 0.6—3 cm .: calyx turbinate-campanulate, commonly farinose, in wathoass 3.5-4 mm. long, in fruit 4-6 mm. long, cleft nearly to the middle into lanceolate to narrowly ovate acute to obtuse lobes: corolla lilac; the slender tube exserted from the yellow throat, at least of shrivelled corollas: capsule cylindric, once-and-a-half the le sngth of the calyx, 2-3 mm. in diameter; its valves splitting into linear halves 0.5-1 mm. wide: seeds ones te and truncated, strongly rugose or reticulated, 0.5-0.7 mm. long.—P. farinosa Nutt. war i. 119 (1818); Torr. Am. Jour: Sci. iv. 59 “(1822); not L. (1753). P. farinosa, B. americana Torr. Fl. No. Mid. U.S. i. 213 (1824), excluding syn _P. pusilla Goldie; Fernald, Ruopora, ix. 16 (1907); not P. americana Rydb. (1901). P. pusilla Hook. Bot. Mag. Ivii. t. 3020 oe), not Goldie (1822).—Shores of the upper Great Lakes, southwestern Ontario, northern Michigan and northeastern : t 2, 1865, A. A. wet rocks, Isle Royale, July, 1889, Sandberg, (Minn.); north shore of Thunder Bay , Lake Huron, Alpena Co., July 3, 1895, C. F. Wheeler; = of Thunder Bay near Alpena, July 5, 1895, C. F. Wheeler; on near Lake Huron yom in damp sand, near Alpena, June 24, i812 ©. K. Dodye, no. 3 i ture near cedars and near shore of Straits of Mackinac, t Tenace, May 14, ne W. H. Manning; Keweenaw Peninsula, Robbins. Minnesota: Two Harbors, Lake 1928] Fernald,—Primula § Farinosae in America 87 Co., June, 1891, Sandberg (Minn.), June 1891, E. P. Sheldon, no. 4400; damp basic eruptic rocks, Susie Island, Cook Co., September, 1927, Butters (Minn.); Agate Bay, L. H. Bailey, no. 477 (Minn.); Grand Marais, Cook Co., August 4, 1875, 7. S. Roberts (Minn.), August 17, 1906, H. L. Lyon, no. 924 (Minn.). Nuttall’s P. farinosa was from “calcareous gravelly shores of the islands of Lake Huron; around Michilimakinak, Bois Blanc, and St. Helena, in the outlet of Lake Michigan: abundant.” Torrey’s P. farinosa, 8. americana, described in detail by him, came from “shores of Lake Huron, Lake Michigan, &c. Douglass and Nuttall.” It is the plant of upper Michigan and the adjacent region of Ontario and Minnesota which has passed as P. farinosa. Unlike true P. farinosa, P. intercedens has the involucral bracts without basal auricles, the umbels few-flowered (in P. farinosa often with 20-25 flowers), the tube of the fruiting calyx only obscurely (in P. farinosa strongly) nerved, and the anthers borne higher on the tube, near the throat and slightly protruding from the shrivelled corolla. In the latter character and in the protruding of the style in the long- styled form P. intercedens is closer to P. mistassinica than to Ps farinosa, P. stricta, P. laurentiana and P. incana; but it has the strongly angled and conspicuously rugose seeds as in these species, in this important character as well as in the ordinarily heavy development of wax on most of the leaves at once departing from typical P. mistas- einica. In perfectly characteristic fruiting material P. intercedens seems to stand quite apart, but as yet I have been unable to find floral differences; and, in view of the fact that P. mistassinica, with its rounded and less rugose seeds, may be sometimes farinose, the difficulty is increased. P. intercedens is here proposed without full confidence of its specific value; but it seems most likely that, in the upper Great Lake region, where it occurs in the same areas as typical P. mistassinica, the two have become much crossed. Elsewhere across the continent, from Newfoundland to Yukon and British Columbia, P. mistassinica is very constant in its seed-characters. It does not seem probable, then, that P. intercedens (P. farinosa, 8. americana Torr.) is to be considered a variation of P. mistassinica in which the seeds have taken on (or retained) the characteristics of the coarser P. stricta, P. laurentiana and P. incana. The true status of the plant can be worked out only by the botanists of the upper Great Lake region. 88 Rhodora [May P. aJANENSIS E. Busch. Leaves wie yellow-farinose beneath, ated or rhombic-spatulate, 0.6-2 cm. long, 0.5-1.1 em. broad, narrowed to a winged or broad and spa poke the blade dentate: scape slender, 0.3-1 dm. high, farinose at summit: involucral bracts subulate-lanceolate, 3-5 mm. long, gradually dilated and thickened but hardly saccate at base: umbel 2-12-flowered: pedicels filiform, farinose, 5-10 mm. long, i in fruit sometimes up to 2 or 3 em. long: calyx cam- panulate, in maturity 4-6 mm. long and about 3 mm. in diameter, cleft about half its ‘length into lanceolate to oblong lobes: — violet; the tube only slightly exceeding the calyx; the limb 1-1.2 ¢ broad, with obcordate lobes about 5 mm. broad: stigma or summits ‘of anthers slightly exserted from the throat: capsule cylindric, about twice the length of the calyx, 2.5 mm. in diameter: seeds not seen.— FI. Sib. et Orient. Extr. iv. Cem. 65: 34 (1926).—Eastern Siberia and Alaska. ha caeds dampish moorlands, Nunivak Island, August 8, 1891, J. M. Macoun, as P. stricta or P. borealis. Primula ajanensis was based on material from Ochotsk Province in eastern Siberia, but Mrs. Busch’s clear description and beautiful illustration of it show it to be the plant twice collected by J. M. Ma- coun on the cruise of the British Bering Sea Commission, once on Nunivak Island (above cited) and later at Plover Bay, Siberia (dis- tributed as P. borealis 8. P. mistasstntca Michx. Leaves efarinose or rarely a little fari- nose, oblanceolate, spatulate or cuneate-obovate, sessile or narrowed to short hed din petioles, 0.5-7 em. long, 0.2-1. 6 cm. 1. broad, many of | 1-10-flowered: pedicels ieee 0.2-0.5 mm. in diameter, much adic the bracts, in maturi = 5-3.5 em. long: calyx slenderly campanulate, 3-6 mm. long, ray. in diameter, efarinose or rarely farinose, cleft to the middle into one soli oblong or narrowly ovate obtuse to acute lobes: corolla-tube yellow, exserted; the pale-pink, lilac or bluish-purple (or white in forma LEUCANT THA) limb 0.8-2 cm. , its obovate or cuneate-obcordate or emarginate lobes 3-6 mm. broad; eye orange or yellow (except in var. NOVEBORACENSIS) : stigma or tops of anthers protruding from the throat of the shrivelled corolla: capsule subcylindrie, once-and-a-half to twice as long as the yx, 2-3 mm. in diameter; its valves splitting into linear halves 0.5— 0.9 mm. wide: seeds 0.4-0.6 mm. long, rounded, nearly smooth or obscurely linear-reticulated. A very variable species, with three strongly marked variations 1928} Fernald,—Primula § Farinosae in America 89 sie! wh nang 0.8-2 em. broad, with conspicuous orange or Ofls fiieb pink, lilac or bidishi-putple.. occ Var. typica, SAWDUA WHO. icra ee a eee eee Forma leucantha. Limb of prs 0.8-1 em. broad; without conspicuous yellow a r. noveboracensis. Var. typica. P. mastassinica Michx. FI. Bor. han: i. 124 (1803); Lehm. Monogr. Prim. 63, t. 7 (1817); Hook. Bot t. Mag. lvii. t. 2973 (1830); Duby in DC. Prodr. viii. 43 (1844); Gray, Syn. FI. N. A. ii. pt. 1: 58 (1878) in part. P. pusilla Goldie, Edinb. Phil. Journ. vi. 322, t. 11, fig. 2 (1822). P. Hornemanniana Hook. Fl. Bor.-Am. ii. 120 7) (1838), in part, not Lehm. (181 P. farinosa, var. mistassinica (Michx.) Pax, Engler’s Bot. Jahrb. x. 200 (1889) i in part. P. farinosa, subsp. mistassinica (Michx. Pax & Knuth i in Engler, A erie iv?37, .) Engler’s Bot. Jahrb. xix. 396 (18 94). P. Maccalliana Wiss, a Torr. Bot. Cl. xxvii. 389 (1900); Brown & Schaefer, Alp. FI. Rocky Mts. 227, t. lxiii. (1907); Rydb. Fl. Rocky Mts. 647 aan: — Southern Labrador and N catbondianil to southern Yukon, south to central Maine, northern Vermont, northern Michigan, northern Wis- consin, northern Minnesota, southern Alberta and rickets British Colum tei The following, selected from about 1 umbers, are representative. LABRADOR: grassy banks, Lake rs apace Hamil- ton River, Ad. P. Low, no. 6035 (Can.). N&wFOUNDLAND: bare wet peat and depressions in aia bog-barrens, Rock Marsh, Flower Cove, Fernald, Long & Dunbar, no. 26,958; a, and turfy pockets in lime- stone fied Brig Bay, Fernald, Long & Dunbar, no. 26,959; borders of ponds on the limestone tableiaiid. Table Mt., Port-a-Port Bay, Fernald & ton, John, no. 10,860; by Grand Codroy Riv. er, July 10, 1912, J. R. Lunt. Quesec: Eskimo Point, June, 1909, C. W. Townsend, St. John, no. 90,676 (Can.); Anticosti, ‘June 20, 1861, Hyatt, Shaler & Verrill; sur les talus calcaires, Rivitre Vaureal, Anticosti, Victorin, Rolland & Louis-Marie, no. 21,831; rivage du Lac Salé, Anticosti, Victorin & Rolland, no. 25,107; by springs, caleareous summit and northerly slopes of Percé Mt., "Percé, Fernald & Collins, no. io endroits sourceux, au pied de Montagne Sainte-Anne, Percé, Victori Rolland, Brunel td Recto: no. 17,645; calcareous cliffs, facing ak alt. 900-1125 m., Table-top Mts., Fernald & Collins, oe 707, 708; pray tie oe oi, alt. 300 m., Lac Pleureuse, F ernald wens is & ledges by Escuminac Brook, Grand Cascapedia River, July 13-15, 1905, Williams, Collins & F. ‘ernald; Riviére des Goelands, Michauz, TYPE in herb. Mus. d’Hist. Nat. Paris (“ad lacus Mistassins ns Canadam a = fretum Hudsonis aie FI. ant Ripe Sy — Cinels ds, "not L seis does not enumerate it from Lake Mistassini. Riviére speci aes ap was 90 Rhodora [May ec, May 21, 1863, ex herb. Univ. Laval; rochers humides, Ile Dorcas’ V islosin: no. 16,122. New Brunswick: moist ground, Connor, Madawaska, Pease, no. 2248; Grand Falls, Malte & Watson, no. 118,333 (Can.); crevices of calcareous rock, Gorge of Aroostook River, July 17, 1902, ee. Collins & F ernald > wet rocks, Wood- stock, J. Macoun, no. 22,545 (Can.); Kennebecasis River, July, 1880, Hay; rocky shore of St. John — near mouth, 1884, Hay Nova ScoTIA: wet boggy banks, Truro, un, no. 15,841 (Can.); aco Upper Stewiacke, Miss A. L. Archibald, no. 69,371 (Can.). MAINE: gravelly shores, St. Francis, Fernald, no. 2427; seepy - elly shores, Fort Kent, Fernald, no. 2428; ledgy banks of Aroostook Riv er, Masardis, September 8, 1897, F ernald; sandy river a Fort Fair- field, Fernald, no. 82; Houlton, 1881, Kate Furbish; limestone rocks, a of Piscataquis River, Foxcroft, May 1 1894, Mary LF pene crevices of ledges along river, Dover, Mi 12, 1895, May 21, 1896, G. B. Fernald; springy limy ‘gravel beach, Township i ix, Range 1; Somerset Co., St. John & Nichols, no. 2444. VERMONT: wet calcare- ous cliffs, Willoughby Mountain, Pringle et al; rocky ledges, Mt. An- nance, Willoughby Lake, September 20, 1855, Wm. Boott; roadside, Willoughby, June 12, 1896, “ay 3 21, 1905, Kennedy: 4th of July Slide, Willoughby, June 12, 190 5, Kennedy; bog near Willoughby Lake, June 1, 1909, Churchill. Onrarto: Lake Nipigon, J. Macoun, no. the name used by. Michaux for a large river emptying from Lake Mistassini into Hudson Bay. ‘Le 4 Socaiaats {1792] . . . A 10" 14 entré dans le Lac Mistassin. . . . Le 5 fait environ 8 4 10 lieues et diné sur la rive des Goelands a 16 lieues de dis- mula mistassinica was not seen b; Michaux on jac the tied of Like os seems: but in the report of his collections from Lake Mistassini, Rupert River and Rupert House, Geol. and Nat. Hist. Surv. Can. n.s. i. 36-44D (1886) the late J. M. Macoun recorded. sea segrg ale the Lake not from of the Vake Mistassini"—Macoun, Cat. Can. PL The organized portion of __ the National Herbarium a hata A886). Se a reall 1928] Fernald,—Primula § Farinosae in America 91 County, August 17, 1901, M. A. Barber. Wisconsin: sand beach near Rowley’s Bay, Door County, Pease, no. 18,003; moist ledges by St. Croix River, St. Croix Co., aad no. 4238 (Minn.). Mrn- NEsoTA: Encampment, Lake Co., Sheldon, no. 4768 (Minn.); moist rocks, Grand Marais, Rosendahl ‘é Butters, no. 4643 (Minn.); Two oe Sandberg (Minn.); Knife River, Sheldon, no. 4948 (Minn.). Manirosa: Mile 256, Hudson Bay Ry., July 12, 1917, J. H. Emerton; Glare atce Lake, July 2, 1917, J. H. Emerton; Sewell, J. Macoun, no. 12,726 (Can.). SASKATCHEWAN: without locality, 1857-8, Bour- geau; north side of Lake Athabasca, J. W. Tyrrell, no. 15,840 Can.). iamnaeh: Mackenzie River, Richardson; west shore of Great Bear Lake, J. M. Bell, no. 22,956 (Can.); southwest and north shore of Great Slave Lake, C. F. Howe, no. 1019 (Can.). ALBERTA: Red Deer River, J. Macoun, no. 15,842 (Can.); Rocky Mountains, Drummond, Burke; oben: of North Fork and North Branch of Saskatchewan, S. Brown, no. 919; Bear Creek pod No. 1, S. Brown nff, S. n ee no. 1327, June 4, 1904, E. M. floyd (Penn.); Devil’s Head Lake, Sanson, no. 22,140 (C an.); Laggan, J. Macoun, no. 68,723 (Can.). British CoLUMBIA: near Vida, gsi 1896, C. E. Cummings; river flats, Yoho Valley, alt. 5000 ft., Butters & Rosendahl, no. 1434; Torrent Fan, Emerald Lake, Heacock, no. 14; Six-Mile Creek, May 18, 1905, Edith M. Farr (Penn.). YUKON: Frances River, lat. 61°, Dawson, no. 15,844 (Can.). Forma LEUCANTHA Fernald, Ruopora, xxi. 148 (1919).—With the typical form, often more abundant. NEWFOUNDLAND: Glenwood, Fernald & Wiegand, no. 6072; Grand Falls, Fernald, Wiegand, Bar- tram & Darlington, no. 6070; Millerton Junction, Fernald & Wiegand, no. 6071; Grand Lake, 1906, Owen Bryant; Savage Cove, Fernald, ie Pease, Long & Gilbert, o. 28,890; Middle Arm, Bay of — W & Griscom, no. 10,405; Birchy Cove (Curling), Fernald & Wiegand, no. 3881; Table Mountain, Port-d-Port Bay, F ernald . “- Sher no. 10,861. QUEBEC: Eskimo Pointy ‘Saguenay Co., Jun W. Townsend; Percé, J. M. Mac o. 68,948. VE cee Willughby, May 24, 1904, Kennedy. Var. noveborace , Var. nov., foliis obovatis vel late oblanceolatis; corollae limbo 0.8-1 cm. lato, fauce vix flava.—P. mistassinica Torr. Fl. N. Y. ii. 7 (1843); Paine, N. Y. State Cat. 18 Sth Ann. Rep. 1 (1865); Clinton, N. Y. State Mus. 24th Rep. 101 (1872); Prentiss, Bull. Torr. Bot. Cl. iv. 15 (1873); Dudley, Cayuga Fi. 60 (1886) ; Rowlee, eo Torr. Bot. Cl. xx. 69 (1893); Peck, N. Y. State Mus 46th Rep. 48 (1893); House, Wild Fl. N, Y. ii. 211, t. 159B (1918); House, Annot. List Ferns and Fl. Pl. N. Y. 588 (1924); Wiegand & Eames, Fl. Cayuga L. Basin, 338 (1926); not Michx. —Cold or wet 92 Rhodora [May cliffs and ravines northwestern and central New York, southern Ontario, southern Michigan and northern Illinois. New York: “Penn Yan,” Sartwell (ace. to House “the specimen in the Sartwell herbarium is labelled ‘Hammondsport’”’); Annsville, Taberg, June, 1846, Knieskern & Vasey; Taberg, Paine; wet rocks, Taberg, ane 23, 1920, House, no. 6911 (Can.); cliffs of Fish Creek, Oneida Co., Paine; Crooked Lake, Sartwell; dripping cliff, south side of ravine ear ——— Falls, Fall Creek, Ithaca, May 27, 1915, A. J. E ames, no. 4804 (TYPE in Gray Herb.); wet limy cliffs, Taughannock Ravine pe vicinity, Ulysses, MacDaniels, no. 3020; other stations cited in Portage and on Salmon River, Oswego Co. ONTARIO: en May 31, 1907, J. Macoun, no. 88,032 (Can.). Micuican: Clifto June, 1883, F. E. Wood; rocks, Grand Ledge, 10 miles west of State College, May 3, 1890, C. P. W heeler; face of wet rocks in moss, bank of G 10 miles west of State College, May 10, 1890, C. F. W heeler. Pdr de limestone cliffs, Jo Davies County, June 15, 1891, H. S. Pepoon (Minn.) Var. noveboracensis has smaller flowers than most of the more northern and typical P. mistassinica, but occasional northern plants have them as small. In its lack of a well marked orange eye it is most striking, true P. mistassinica having the throat bordered by a brilliant yellow or orange ring (as in most of the § Farinosae). The leaves of var. noveboracensis are somewhat distinctive but similar foliage occasionally occurs in the more northern type with larger corolla and yellow eye. I have sought in vain for any character upon which to separate P. Maccalliana from P. mistassinica, but can find absolutely nothing to separate it from the variable plant of eastern Canada, northern New England and Newfoundland. In describing it, Wiegand said: “From the eastern P. Mistassinica it differs quite markedly in its stouter habit, broader and more farinose leaves, short-pedicelled flowers, larger acute-lobed calyx, and large bluish, not pink or flesh- colored, corolla with tube scarcely longer than the calyx, and with a very prominent yellow eye.” In interpreting the description of P. Maccalliana it should be borne in mind that Wiegand knew as P. mistassinica. the var. nossboracenets, which is a very small extreme without “prominent yellow eye.” As for the other characters, an abundant series of specimens shows that P. mistassinica may have in the same general region either obtuse or acute or even attenuate calyx-lobes, the leaves efarinose or slightly farinose, dentate or entire, and the pedicels and corolla-tube very variable in length. The blue color emphasized by Wiegand is not a constant character. 1928] Fernald,—Primula § Farinosae in America 93 Professor Butters, who has collected the plant at Banff and elsewhere in the Canadian Rockies, assures me that the fresh flowers are pink or lilac-purple and that he has never seen any decided blue in them except as a post-mortem character in dried specimens. The late Stewardson Brown, who also knew P. Maccalliana in the field, says, “Flowers . . . pale pink or bluish” and Mrs. Schiffer’s photo- graph of it looks quite like eastern P. mistassinica. In this connec- tion it is not inappropriate to quote the description by the late John A. Paine of the station of P. mistassinica, var. noveboracensis on Fish Creek, Oneida Co., New York:— “Near dripping water the plants grow most abundantly and largest, often eight or ten inches in height, and bearing a cluster of as many flowers. These vary in color, from pure white, through different shades of pink, to deep blue. The leaves also vary in form, from round obovate to oblong lanceolate; often entire, commonly more or less toothed; usually smooth beneath, but frequently white mealy. A whole cliff-side scattered over with these variegated Primroses is one of the loveliest sights in all our flora.”! If P. Maccalliana, originally separated from the plant of New York by “bluish, not pink or flesh-colored, corolla” may have, as described by those who know it in the field, “Flowers . . . pale pink” and if the New York plant may have them “from pure white, through different shades of pink, to deep blue,” it is obvious that color of the corolla in these cases is no safer as a specific criterion than it is in most other plants! P. Maccalliana more often has the leaves entire or subentire and angnty. farinose, but plenty of western specimens have th f: leaves usly dentate. Conversely, many eastern plants have some or all of the leaves subentire to entire and occasionally eastern plants show them more strongly farinose ay any from the Rocky Mountains. 9. P. parvirotia Duby. Leaves efarinose, cuneate-obovate, spatu- late or rhombic, 0.4-3 em. long, 1.5-7 mm. broad, the lower scarcely petioled, the upper with winged petioles, the margin denticulate: scape se 0.4-1.8 dm. high, efarinose: bracts of involucre lance- or linear- 2-5 mm. long, dilated and thickened but not saccate umpbel 2-0 domes: pedicels filiform, fig mm. long: calyx cam- 3-5 mm. long, cleft 1 Paine, N. Y. State Cat. 18th Ann. Rep. 105—Reprint, 53 (1865). 94 Rhodora [May cylindrical, up to once-and-a-half the length of the calyx, 2 mm. in Duby in DC. Prodr. viii. 42 (1844). P. mistassinica Cham. & Schlechtd. Linnaea, i. 213 (1826), in large part; Eastwood, Bot. Gaz. Xxxili, 212 (1902); not Michx. (1803). P. tenuis Small, Bull. Torr. Bot. Cl. xxv. 320 (1898). P. borealis, var. parvifolia (Duby) Pax in Engler & Prantl, Pflanzenr. iv’. 81 (1905). P. Chamissonis E. Busch, FI. Sib. et Orient. Extr. iv. Cem. 65: 28 (1926).—Alaska and adjacent Siberia. ALasKA: several stations cited by Chamisso & Schlechtendal and by Mrs. Busch; Cape Nome, 1900, Blaisdell, no. 129, as P. mistassinica. Duby’s P. parvifolia, described “ad fretum Beering (Fisch.! in herb. cl. Kunth)”, is treated by Pax as a small-leaved variety of P. borealis Duby; while Mrs. Busch, in her remarkably exhaustive treatment of the Siberian Primulaceae, makes no mention of it, even asasynonym. In view of the reduction by Pax of the wholly distinct P. Loczii Kanitz of central Asia to P. borealis (see discussion under the latter) it need not be surprising that Pax should also include in P. borealis the very distinct P. parvifolia. Surely, Duby, who published both P. parvifolia and P. borealis, was too discerning a specialist to separate them if they differ only by a millimeter or two in the length of the leaves and by the slightly different outline of the latter. Duby indicated other differences, notably: in P. parvi- folia “ corollae lobis emarginatis,” in P. borealis “corollae . . lobis obcordatis semi-bifidis.”” The plant of the Bering Straits region and Alaska, described in detail and beautifully illustrated by Mrs. Busch as P. Chamissonis, well matches the description of Duby as does the Blaisdell material from Cape Nome. Should it eventually prove that I am in error in identifying these with P. parvifolia, the name P. tenuis Small (1898) should be used for the plant here described. efarinose, 2.5-10 em. high: bracts - involucre siubulaté: often unguiculate, 2-6 mm. long, dilated and usually slightly saccate-gibb e: umbel 1-12-flowered: pedi ) long calyx turbinate-campanulate, efarinose or barely atineee, somewhat gibbous at base, in maturity 5-6 mm. long, cleft pence’ its — into lance-ovate acute to obtuse segments: corolla lilac; the tube equaling the calyx or slightly epsicted: the limb 1.2-2 em. broad, its obcordate lobes 5-8 mm. broad: stigma or tops of anthers pro- truding from the throat of the wilted corolla: capsule cylindric, only 1928] Fernald,—Primula § Farinosae in America 95 slightly exserted: seeds less than 0.5 mm. long, smoothish.—Mém. am. Prim.,—Mém. Soc. Phys. d’Hist. Nat. Genéve, x. t. ii. fig. 2 (1844) and in DC. Prodr. viii. 43 (1844); Ledeb. Fl. Ross. iii. 15 (1847-49); Pax & Knuth in Engler, Pflanzenr. iv’. 81 (1905), excl. var. parvifolia; Ostenfeld, Vidensk, Selsk. Skrift, i. Math.-Naturv. KI. no. 8: 60 (1910); J. M. Macoun & Holm. Rep. Can. Arct. Exped. 1913-18, v. pt. A, 184A, t. xi. fig. 4 (1921); E. Busch, FI. Sib. et Orient. Extr. iv. Cem. 65: 31 (1926). P. mistassinica Cham. & Schlecht. Linnaea, i. 213-215 (1826), in part; and other authors on the flora of Bering Sea region; not Michx. (1803). P. sibirica, var. borealis (Duby) Kurtz, Engl. Bot. Jahrb. xix. 397, 468 (1894). P. borealis, var. Loczii Pax in Engler & Prantl, Pflanzenr. iv™’. 81 (1905), not P. Loczyi Kanitz (1891).—Aretie Mackenzie and Yukon, arctic and western Alaska and adjacent Siberia. The following American specimens have been examined: MacKenzie: Warren Point, Stringer, no. 62,251 (Can.); Cape Bathurst, Frits Johansen, no. 533; Pillage Point, Arctic coast, July 19, 1924, W. H. B. Hoare (Can.); mouth of Moose River, Mackenzie Delta, July 18, 1924, Hoare (Can.). Yukon: Herschell Island, Stringer, no. 14,423 (Can.), Frits Johansen, no. 265 (Can.). AtaskKa: Collinson Point, Camden Bay, Frits Johansen, nos. 50, 119, 119a (Can.); west of Konganovik, Camden Bay, F rits Johansen, no. 56; Point Hope, 1884, Cruise of the Corwin; Kotzebue Sound, Fischer, Mertens; Schismaroff Inlet, Chamisso, as P. mistas- sinica; Koyuk River, collector unknown; St. Michaels, Norton und, Bannister on Western Union Extens. Telegr. l. Ex 1865-66, as P. stricta; St. Michaels, 1881, J. Muir. Simmons Phytogeogr. Arct. Am. Archipel. 119 (1913) cites specimens from Banks Lanp. Primula borealis is a variable species, but not so variable as one would infer from various published accounts of it. From its immedi- ate allies it is distinguished, when in good condition, by the unusually large corolla with. very deeply obcordate and broad lobes. From the first it has been often misinterpreted as small-flowered. Duby described it as growing “in Americae occidentali-borealis sinubus Schischmarfe}ff et Sancti Laurentii (Cham.! Bunge!)” and his illustra- tion showed corollas barely 1 cm. broad but with the very character- istic deep notching of the lobes. The material of Chamisso’s plant from “Sinus Schischmareff” in the Gray Herbarium has the flowers badly shrunken and shriveled so that they give no adequate impres- sion of the size of the fully expanded corolla. Mertens’s material, originally labeled as coming from “Sinus St. Laurentii,” but changed in pen to “Sinus Kotzebue,” also has badly shriveled flowers as does further material from Kotzebue Sound collected by Fischer. I have not seen the Bunge material from “Sancti Laurentii” (whether from . 96 Rhodora [May St. Lawrence Island, Alaska, or from St. Lawrence Bay, Siberia is not clear); but it is evident that these old collections which were available to Duby and his immediate successors failed to display the large corolla which is seen on some of the more modern specimens. Consequently we find Gray, in the Synoptical Flora of North America, describing P. borealis with “lobes of the purple corolla oblong, barely 2 lines long, deeply notched” and reducing to it P. egaliksensvs, which has the smallest flowers of any American species, instead of the largest flowers. The specimens in the Gray Herbarium studied by Gray show that he had only poorly prepared material of P. borealis and that he confused with it both the typical white-flowered P. egaliksensis and its violet-flowered form. Similarly, Mrs. Busch, working with the collections at Leningrad, describes the corolla “lobis 3-5 mm. lg., loco latissimo 3-5 mm. It.”; but Pax & Knuth, more accurately say “limbus 12-14 mm diametiens.”’ The fully expanded flowers of well-prepared specimens before me are 1.2-2 cm. broad and I have seen no well-pressed flowers with corolla-lobes less than 5 mm. broad at summit, while in the largest specimens they are 8 mm. broad. The largest extreme of P. borealis, with leaves from 1-4 cm. long, has been identified by Pax with P. Loczyi Kanitz; and making up his P. borealis, var. Loczii of a collection from Bering Strait and the plant described by Kanitz, he gets the extraordinary range, “ Alaska: t. Lorenz . Kansu: am Chaji-san . . . ” Kansu, in high central Ank would hardly be expected to share a species or variety in so technical a group as Primula with the shores of Bering Strait, fully 5500 km. to the northeast; and a glance at the description and the detailed illustration of P. Loczyi' at once shows that they are not closely related. To be sure, Kanitz spoke of his species as “ Ex affinitate P. borealis Duby.” P. borealis, however, has the leaves long-petioled, with the 1-7 pairs of teeth confined to the upper half of the rhombic blade, P. Loczyi has the cuneate-obovate leaves scarcely petioled and dentate nearly to base; the involucral bracts of P. borealis are lance-subulate and gibbous at base; in P. linear, flat and tapering at base (consequently not of § Farinosae); the calyx of P. borealis is gibbous at base and cleft to the middle, the ) Kanitz Agost, dvénytani Gyiijtések Eredményei Grof Széchenyi Béla Kele- tazsiai viabot (s77=1890),—Plantaran in Expeditione Speculatoria Comitis Béla nig entrali C cor ena oan 36, t. IIT. # {ihol): Ranehoesl Getesias. Belen (Bot.) ii. 713, t. 3, figs. 1-2 (1898). 1928] Fernald,—Primula § Farinosae in America 97 calyx of P. Loczyi non-gibbous and cleft about one-third its length; the anthers of the short-styled form of P. borealis are inserted at and slightly protrude from the throat; in P. Loczyi they are borne about midway on the corolla-tube. That P. Loczyi is neither a variety of P. borealis nor closely related to that species should be apparent. 11. P. EGauiKsensis Wormskj. Leaves thin, efarinose, ovate, oblong, obovate or spatulate, entire or obscurely undulate (very rarely obscurely dentate), slender-petioled, 0.5-5 em. long, 0.2-1.5 em. wide, the petiole often equaling to exceeding the blade: scapes strict, efari- nose, pale-green, 0.12.4 dm. high, 0.5-2.3 mm. in diameter at summit: involucral bracts green, lanceolate or lance-oblong to lance-subulate, flat to involute, dilated and gibbous-saccate at base, 2.5-7 mm. long: umbel 1-9-flowered: pedicels strict, in anthesis from about equaling to thrice Monogr. Prim. 64, t. vii. (1817); Pax, Engler’s Bot. Jahrb. x. 198 (1889); Pax & Knuth in Engler & Prantl, Pflanzenr. iv’. 77 (1905). P. sibirica Hook. FI. Bor.-Am. ii. 121 (1838), in part; J. M. Macoun & Holm, Rep. Can. Arct. Exped. v. Pt. A. 18A, t. xi. fig. 5 (1921); r. Labrador to Alaska, south to northern Newfoundland, eastern Quebec, shores of James Bay, Alberta and British Columbia.—GREENLAND: Igaliko (type-locality), July 27, 1888, Rosenvinge, August 6, 1925, A. E. & M. P. Porsild. Laprad Exped. 1891, no. 103; Waghorne, no. 15,849 (Can.); damp mossy rocks, Battle Harbor, August, 1911, C. S. Williamson (Penn.). NEWFOUND LAND: mossy brooksides and damp turfy slopes, Sacred Island, Fernald & Long, no. 28,899; mossy and turfy trap cliffs and talus, Anse aux Sauvages, Fernald, iegand & Long, no. 28,900; g tundra, Schooner (or Brandy) Island, Pease & Long, no. 28,903; moist turfy or peaty depressions in limestone barren, Cook Poi 98 Rhodora [May & Gilbert, no. 28,894; springy swales and turfy seis Boat Harbor, Fernald, Wiegand «& Long, no. 28,895; swale near mouth of brook, Watts Bight, Pease, Griscom, Gilbert & Hotchkiss, no. 28 ,897; gravelly barren east of Big Brook, Pease, Griscom, Gilbert lotchkiss, no. 28,896; wet limestone barrens on the Highlands northeast of Big Brook, Fernald, Wiegand & Hotchkiss, no. 28,902; peaty and turfy limestone barrens west of Big Brook, Long & Gilbert, no. 28,893; bog, Flower Cove, July 12, 1920, on E. Priest, September 2, 1923, A. G. Huntsman; swales and wet peaty limestone barrens, Capstan Point, Flower Cove, Fernald, gee & Dunbar, no. 26,960; springy swale Long, Gilbert & H. si no. 28, 898. QUEBEC: Ungava Ba ay, 1884, L. M. Turner; along Noyava River, near Fort Chimo, Spreadborough, no. 14,420 (Can.); gravelly patch at head of rocky shore, Middle St. Mary Island, bined 22 and 23, 1927, H. F. Lewis (Can.); mossy bank, island off Pointe au Maurier, Charnay, St. John, no. 90,677 (Can.); sur les ae at Pointe-aux-Ammo onites, Mingan, Victorin & Rolland, no. 18,418; sur les plages calcaires, Ile & la Vache Marine, Mingan, Victorin & "Rolland, no. 18,419; sur les corniches calcaires, Ile Téte 4 la Baleine, Mingan, Victorin & Rolland, no. 21 -aectah sur le 25,109; Richmond Gulf, Spreadborough, no. 34,449 (can. as P. stricta; a of Cape Jones, Hudson Bay, éw, no. 63,243 (Can.), as P. stricta altered to P. sibiric dance banks, sea a James fea J. M. prec no. 15,847, as P. mistassinica.. ONTAR Frits Johansen, no. 369 (Can.), as P. sibirica; west shore of Great Be J. M. Bell, no. 22,955 (Can.), as P. sibirica. YUKON: moist tops pers near mouth of Lewes River, Gorman, no. 1052 in part (Can.), mixed with P. stricta. ALASKA: wet meadows, Yakutat, 4 a Walker, no. 1052; Chignik Bay, hindex Peninsula, June 19, 1874, er. Forma violacea, nom. nov. Whole plant darker: scapes and pedicels suffused with violet or purple-black: bracts and _ calices at yi nor or spotted, often blackish: corolla violet or deep-lilac. sibirica Hoo r.-Am. ii. 121 ( 838), in part, not Jacq. ily P. borealis Gray, Syn. Fl. N. A. ii. pt. 1: 58 (1878), in ha’ not Duby (1844). P. stricta, v nsk. Vet. Akad. Fl. Groenl. 260 (1887). P. farinosa, var. groenlandica (arming Pax, in Engler, Pflanzenr. iv’. 84 (1905), in small part 1928] Fernald,—Primula § Farinosae in America 99 y P. sibirica, var. arctica Fernald, RHopora, xxviii. 98, 99, 105, 224 (1926), not Pax (1905). P. farinosa, subsp. P. groenlandica (Warming) W. W. Smith & Forrest, Notes Roy. Bot. Gard. Edinb. xv. no. Ixxvi. 24 (1928). P. groenlandica (Warm.) W. W. Smith & Forrest, |. c. 49 (1928).—Throughout the range of the species, and sometimes more abundant. GREENLAND: Itivnek, Holsteinsborg Distr., July 13, 1884, Warming & Holm, (duplicate TYPE of P. stricta, var. groenlandica); mixed with typical P. egaliksensis, Igaliko, August 6, 1925, A. E. & M. P. Porsild, NEWFOUNDLAND: wet hollows in limestone gravel-barrens, Cook Point, Fernald & Gilbert, no. 28,917, Fernald, Gilbert & Hotchkiss, no. 28,919; wet bank on limestone barrens, Cape Norman, Wiegand, Griscom, Gilbert & Hotchkiss, no. 28,918; swale near mouth of brook, Watts Bight, Pease, Griscom, Gilbert & Hotchkiss, no. 28,901; peaty limestone barrens back of Big Brook, Fernald & Long, no. 28,915; sandy and gravelly margin of Big Brook, Fernald, Wiegand, Long & Gilbert, no. 28,916. Quesec: River Kovik, Hudson Straits, Low, no. 23,026 (Can.); Little Whale River, Low, no. 63,249 (Can.). Auperta: North Fork of the Saskatchewan, 6 miles above North Branch, S. Brown, no. 1024; Upper Maligne Valley, S. Brown, no. 1145; Maligne Lake, S. Brown, no. 1184; all as P. borealis; damp places, Kicking Horse Pass, Dawson, no. 15,826 (Can.). Bririsa COLUMBIA: shore of Atlin Lake, Guilliam, no. 101,403 (Can.). ALASKA: swamps, Popoff Island, Shumagin Islands, June 28, 1872, M. W. Harrington. The name forma violacea is here used for P. stricta, var. groen- landica Warming (discussed in detail under P. stricta, p. 67). Forma violacea is most probably the true philogenetic color-form of P. egaliksensis, the white-flowered plant, originally described, being a presumable albino, such as we get in the normally lilac-flowered P. mistassinica; and, treated as a mere color-form, the name forma violacea is more appropriate to it than would be the name originally given by Warming. The plant, changing its rank from variety to forma, does not necessarily have to retain an inappropriate name. W. W. Smith & Forest, in their Sections of the Genus Primula (Notes Roy. Bot. Gard. Edinb. xvi. no. Ixxvi. 24), which comes to hand while this is going through the press, treat P. egaliksensis (using the later and incorrect spelling P. egalliccensis) as a species but under P. farinosa enumerate among the subspecies (listed as binomials’ in violation of general botanical usage and of the Inter- national Rules of Botanical Nomenclature) a “P. groenlandica, Warming.” In the index, however, they definitely give “P. groen- landica” as a “Species.” Warming, of course, had no such species, nor did he have a subspecies of P. farinosa. He published P. stricta 100 Rhodora [May var. groenlandica, and, as the specimens show, his plant was a violet- flowered P. egaliksensis P. egaliksensis has many times been confused with P. sibirica, to which it is more closely allied than to other species. Like P. sibirica it has long-petioled, mostly entire, thin and dilated leaves, broad and often flat involucral bracts, copiously glandular-ciliate calyx-lobes, deeply notched corolla-lobes, and very slender capsules; but in P. sibirica the involucral bracts are usually longer and more dilated upward (often slightly obovate) and they have more pro- longed and narrower auricles; the calyx is longer (5-8 mm. long); the corolla is much larger (1-1.8 cm. broad) and with broader lobes; and the summits of the stamens or (in the long-styled form) the stigma slightly protrude from the throat. P. sibirica is a plant of arctic and alpine Eurasia, reaching America only in the high mountains of Alaska and Yukon; P. egaliksensis is strictly American, occurring mostly between latitudes 50° and 60° N., though sometimes reaching the Arctic Coast, and found only slightly north of 60° in southern Greenland. 12. P. srprrica Jacq. Leaves green, efarinose, slender-petioled, 1-7 em. long; the thin blade suborbicular, ovate or elliptic, rounded at summit, entire, or obsoletely dentate, abruptly contracted at base, shorter than to equaling the petiole, 0.5-2 cm. broad: scape slender, 0.45-1.5 dm. (-3.8 dm. acc. to Busch) high, efarinose: involucral bracts oblong or narrowly obovate, 4-11 mm. long, obtuse or abruptly contracted at tip, often ee at base into narrow saccate auricles 1-1.5 mm. long: umbel 1-4 (-8)-flowered: ipedivele filiform, 0.7-4 cm. long: calyx efarinose, wig i! campanulate, in maturity 5-8 mm. long, cleft one-third its length into oblong-ovate fulvous-ciliolate tee corolla-tube abe or less exserted, very slender; the lilac limb 1-1.8 em. broad, with obcordate or deeply emarginate broadly cuneate or ia lobes: stigma or summits of anthers slightly exserted from the throat: capsule slenderly cylindric, somewhat narrowed at summit, from altgndly exserted to twice the length of the calyx, about 1.5 mm. cw ter: seeds not seen.—Misc. Austr. i. 161 (1778); Lehm. ye Pre 60, t. 5 (1817); Hook. Bot. Mag. lix. t. 3167 (1832) and Ist 3445 (1839) Bo. Dg DC. Prodr. viii. 43 (1844); Gray, Syn. Fl. N. A. ii. pt. 1: 38 (1878), in small part only; Pax in Engler’s Bete Jahrb. x. 197 (1889); Pax & Knuth in Engler, Pflanzenr. iv”’. 76, 77, incl. vars. (1905); E. Boeh. Fi. Sib. et Orient. Extr. iv. aan: 65: 61 1 (1926). P. finmarchica Jacq. sc. Fore i. 160 (1778). P. norvegica Retz. Fl. Scand. Prodr. 55 (1795). P. intermedia Ledeb. Mém. Acad. Pétersb. v. 519 aes For further synonymy see Pax & Knuth, |. c. and E. Busch, |. e—Arctic Europe, Ural 1928] Fernald,—Primula § Farinosae in America 101 Mountains, arctic and central Asia; Alaska and Yukon. ALAsKa: Kuskokwim Valley, 1884, Weinmann; cited by Pax & Knuth from Lynn Canal. Yuxon: Lake Tahko, near head of Yukon River, June 24, 1883, Schwatka, no. 81. The citation by Hooker, Fl. Bor.-Am. ii. 121 (1838) of P. sibirica from “Barren country between lat. 60° and 69°, . . . Dr Richardson” can hardly be accepted as unquestionably belonging here. Hooker said “I have carefully compared Dr. Richardson’s specimen with authentic ones of the European P. Norvegica, the Asiatic P. Sibirica, and the P. Egaliccensis from Hornemann himself, and there can be no question about the propriety of referring them all to one and the same species.” P. egaliksensis, however, is now known to be quite distinct from P. sibirica and specimens before me show it from several stations between the west shore of Hudson Bay (Churchill, lat. 58° 51’) and Alaska. It is most probable, then, that Hooker’s P. sibirica was P. egaliksensis; for both the white- and purple- or violet-flowered plants of the latter have been repeatedly misidentified as P. sibirica. The identity of the violet-flowered P. egaliksensis is discussed in the comments upon P. stricta, var. groenlandica Warm. on p. 67. The citations “ Pr. integrifolia Oeder in Fl. Danica t. 188 (1767)” and “Pr. rotundifolia Pallas, Reise Prov. russ. Reiches III. (1776) 223” are given by Pax & Knuth in the synonymy; and the first is made the basis of the varietal combination, P. sbirica, “Var. B. integri- folia (Oeder) Pax.” Both these names would, thus, appear to have been published earlier than P. sibirica, but examination of the original publications at once shows that they cannot be taken up to displace it. Oeder did not publish the binomial P. integrifolia. Accompanying a recognizable plate of P. sibirica was the text: “Tap, CLXXXVII. iis carnosis integerrimis. Hall. Helv. 485. ; : Primula, integrifolia, fee “ ‘mis glabris oblongis, calycibus tubulosis b inn. pl. 205. . 102 Rhodora [May plant with P. integrifolia L. and certainly was not proposing a new species. Incidentally his “ Primula, integrifolia, foliis integerrimis” was the third of a series of names and would technically be “ published in synonymy”; and, furthermore, it is not clear that Oeder was at that time following binomial nomenclature. The emphasis by typography on certain adjectives was not uniformly confined to a single “specific” name. Thus, in the same fascicle we find “Camp- anula alpina linifolia caerulea C. B. Prodr. 34” with the synonym “Campanulae, rotundifoliae, varietas. Linn. Sp. pl. 232” (Tab. elxxxix); “ Arthritica hiberna. Gesn.” . with the synonym “Primula, veris, acaulis, foliis dentatis selansae’® (Tab. exciv) and “ Medica echinata minima. J. B. . . ” with the synonym “ Medicago polymorpha minima. Uiin: Sp. pl.” (Tab. cexi); while in some cases (Tabs. ccxiv, ccxxxii, ccxxxiii) no roman type was used. It is, therefore, very clear that Oeder had in these early fascicles of Flora Danica no clear-cut intention to use binomials exclusively and such as crept in in synonymy were merely “accidental” binomials. At any rate, the name Primula integrifolia, already validly used by Linnaeus for another species could not be used for a later species and it is unfortunate that in making Oeder’s misidentification of Primula integrifolia L. the basis of a variety of P. sibirica Pax should have excluded perfectly valid varietal combinations. The first unequivocal varietal name for var. integrifolia Pax (1905) seems to be var. kashmiriana Hook. f. Bot. Mag. evi. t. 6493 (1880). The name Primula rotundifolia Pallas has no nomenclatorial status. It was used casually and without diagnosis in the running text of his eee and combinations based upon it should be excluded: such a as P. sibirica, var. rotundifolia (Pall.) Pax in Engler’s Bot. Jahrb: x. 197 (1889). IpENTITIES oF NUMBERED EXSICCATAE CITED 1, P. stricta 8, P. mistassinica 2, P. laurentiana ; . forma leucantha _ a forma chlorophylla 8b, P. i var. noveboracensis 3, P. incana 9, P. parvifolia 4, P. decipiens 10, P. borealis 5, P. specuicola 11, P. egaliksensis SF. intercedens r. s forma violacea Clea P. ajanensis 12, P. sibirica 0.1, 69,371 (8 | «Bitter Ee, ae ne, 477). a6 , nase M., no. 23,955 (11); 22,956 (8). 1928] Fernald,—Primula § Farinosae in America 103 BELL, RosBert, no. 15,829 (1), BISSELL BEAN, py & LINDER, no. 22,234 (2). 9 (3). Bowpbo1n CoLLEGE EXPED., n 3 (11); 104 (2). spusirt Me no. 13 (8); 900 (8); et (8); 1024 (11a); 1107 (1); 1145 (11a); 1184 PAA de & RosENDAUBL, nos. 1327 and 1434 (8). Carrngs, C. E., no. 89 722 (1 ): CHAMBERLAIN, no. 2 CoLuLins, FERNALD & PrASE, nos. 5319, 5320, 5434, 5435, 5554 (2). 581 (1 i CusHMAN & SANFORD, no. (2). Dawson, no. a 826 (11a); IB oy (8). Dopeg, C. K., no. 3 (6). Dowsixa, no. 34 525 (11 aS. 34,526 (1). AM Eastwoop, no. 68 ®. FERNALD, na "82, 2427, Nese (6). FERNALD, ARTRAM SETT, no. 24,327 (2). FaRNAaLp , CoLLins, ve 246 (2): "707" and 708 “)s 1145 (8); 1147, 1148 and 114 , no Wie, ae & meee” ~ 25,983 FERNALD & Gunes no. 28, 894 (11); % 17 Re FERNALD, GILBERT & ow hal no. AE a). Fernaip, Griscom & GILBERT, no. 28, 302 Fernatp & Lone, no. 28,899 (11); 28,911 (aad: 28,912 (2); 28,915 (11a). FERNALD, Lone & DUNBAR 0. 26,955 (2a); 26 26,956 (2); 26,957 (2a); 26,958 and 26,959 (8); 26,960, 26,001 and 26,902 (11). ERNALD, LONG & Foaa, n 0. 375 (2 pa & Be tous me 25,332 and 3 (8 . "10, TO Br (8). LD OHN, no. a), FERNALD & WIEGAND, no. 3881 (Sa); 3885-3890 (2); 6071 and 6072 (8a). FERNALD, WIEGAND ARTRAM, no. 6068 FERNALD, WIEGAND, BarTRAM & DaRLI NGTON, no. 6070 (8a). FERNALD, ae & ect no. OTCHKISS, no. 28,902 (1 1). > RNALD, Weasak D& roy 28, 895 a ee eree ,900 (11); 28,914 (2). ERNALD gaat: Lone ILBERT, no. Fenwatp, Wiecanp, Lona, GILBERT & Paine eee * 28,898 (11); 28,913 meine, Wiecanp, Pease, Lone, GriscoM, Grtpert & HorcHKiss, no. GUILLIAM, no. 101,403 (11a). A . L., no. 45 (3). : JOHANSEN, ese 38 (10); 56 56 (10); 119 (10); 265 (10); 347 (1); 369 (1); Low, A. P., Ae (11a); 24,529 (1); 63,242 (1); 63,243 (11); Lone & Gripert, no. 28,893 (11 104 Rhodora [May Lyon, H. L., no. 924 (6). aie cia 3020 (8b (8b). Mackenzie & Gniscom , no. 10,402 (2); 10,405 (8a). Macoun, J., no. 5313 (3); WL 776 (2); 12, ee (2); 12,742 (2); 12, 726 (8); 15,830 (2); 1 833 (2); 15,837 (2); 15,83 8 (2); 15, 841 (8); 15, 842 (8); 15,846 (6 Bey: 8); 22,545 aN 24 528 (3); 54,260 (6); 68,725 (8); 81,152 @); 88,032. (8a); 101, 401 (3). Macoun, J. M., 0. 15, 831 (1); 15,847 (11); 15,850 (1); 68,948 (Sa); 68,949 (2); “79, 388 (1); 79, 390 (11 Matte & Watso N, no. 118, 331 (3), 118,333 (8). MERRILL & Witcox, no s. 749 9, 990 and 1039 (ay. NELSON, nos. 1871 and ue Pd aF Payson & PAYSON, nos. and 2648 (3). PEASE, nos. 2248, 17,971 i 18,003 (8 PEASE, ag RISCOM, Gitpert & Horcuxiss, nos. 28,896 and 28,897 (11); 28,901 PEASE Sik Lobe: no. 28,903 (11); 28,6 907 (2). Pease, Lone & GILBERT, no. ‘ioe (8a). 82 ROsENDAHL, no. 4238. (8). RosENDAHL Oni nos. 1327, 1434 and 4643 (8). (8 SHeEtpon, E. P., o. 4400 ( (6); 4768 (8); 4948 (8). Sp PREADBOROUGH, ar 14,420 (11); 14,421 (1); 19,852 (3); 34,449 (11); 62,554 (1); 62,555 (1); 6 62,556 (8). 1CKER Sr, Joun, si 90,674 (2); 90,675 (2); 90,676 (8); 90,677 (11). JOH 444 (8). Sr. Jou NicHoLs, no nic, nos. a aa TAYLOR, gy 38 (3); o (1); 110 (3). TYRRELL, no. pda nos. ary ai 4188 (2); 16, sy (8). Vicrorin & Roiianp, nos. 18,418 and 18,419 (11); 18,421 (2a); 18,465 a 18,485 (2); 21 "330 (ia), 21, 832 (2); 25,107 (8); 25,109 (ani 35,139 om cTorRIN, RoLtuaNnp, B & Rousseau, nos. 17,642 and 17,643 (2); 17. 645 (8). ICTORIN, ROLL s-MaRIg, no. 21,831 (8). Wacnonne, no. 5 Boost ):15 8 ,849 (11). Wiecanp, GinBert & Horcuxiss, no. 28,908 (2a); 28,918 (11a). Wiecanp & Horcuxiss, no. 28,904 (2a). Wiscanp & Longe, no. 28,909 (2a). EXPLANATION OF PuiaTe 169. LAURENTIANA X 34. Photograph taken by pater d- F. Collins at pig sina type-locality, Bic, Quebec. g Rhodora Plate 169 PRIMULA LAURENTIANA X 34 CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY LXXX {SSUED JUN 26 1928 RECORDS PRELIMINARY TO A GENERAL TREATMENT OF THE EUPATORIEAE—VII By B. L. Roprnson PUBLISHED BY THE GRAY HERBARIUM OF HARVARD UNIVERSITY CAMBRIDGE, MASS., U. S. A. 1928 RECORDS PRELIMINARY TO A GENERAL TREATMENT OF THE EUPATORIEAE—VII. SINCE publishing the preceding paper in this series, the writer has had access to many sources of further information concerning the Eupatorieae. During the summer of 1927 he visited, chiefly for the study of this group, the great herbaria in Zurich, Geneva (both at the Botanic Garden and at the University of that city), Paris, Berlin, South Kensington, and Kew, receiving at all these places permission to examine and photograph specimens without limit and to make dissections where needful to their precise identification. He was also permitted to borrow from these herbaria much undetermined material for more extended study on his return to Cambridge. He wishes here to record his gratitude for the many favors shown him and to express his sincere thanks to the directors of the herbaria at the botanical centers mentioned. As in former papers of the series the source of the specimens studied is indicated in the text by abbreviations to show the herbaria to which they belong (K. for Kew, BM. for the British Museum of Natural History, Par. for the Museum of Natural History at Paris, Brl. for the Botanical Museum at Berlin-Dahlem, Gen. for the Conservatory and Botanic Garden at Geneva, Univ. Gen. for the University of Geneva, ete.). Besides material seen on his journey or subsequently lent by European establishments, the writer has seen much material from American sources. Through many loans, gifts, or exchanges his work has been constantly aided, as for many years, by the New York Botanical Garden, the United States National Herbarium, the Academy of Natural Sciences at Philadelphia, the Field Museum, Missouri Botanical Garden, California Academy and University of California. : : He would furthermore express particular gratitude for aid received from several South American sources. Thus, Professor Miguel Lillo, Director of the Museum of Natural History at the University of Tucumfn, has most kindly sent to the Gray Herbarium an extended and very discriminatingly selected series of specimens to illustrate the Eupatorieae of the Argentine Republic; and Professor L. R. Parodi of the Faculty of Agronomy and Veterinary Science at Buenos Aires has contributed material of much interest from different parts of the Argentine Republic. These sendings, together with an admir- ably prepared and carefully identified series of specimens recently acquired from Mr. S. Venturi, are doing much to clarify points 4 ROBINSON hitherto obscure through lack of adequate Argentine material. From Peru some specimens possessing high interest and including several novelties have been received for study (through the Field Museum) from Professor A. Weberbauer and (directly or through the United States National Museum) from Professor F. L. Herrera of the Uni- versity at Cuzco. Through a loan from the University of Minnesota it has been possi- ble to examine some further numbers of the admirable Brazilian collection made some years ago by the late Professor E. W. D. Hol- way. A notable collection, secured in western Jalisco by Mrs. Ines Mexia while on an expedition encouraged and aided by the University of California, has yielded several Eupatorieae of high interest, in- cluding a new genus of marked character and several species of great rarity rediscovered after many years. The most noteworthy single source of the material treated in the following pages, however, has been the collection made by Messrs. Ellsworth P. Killip and Albert C. Smith on their recent journey to northern and northeastern Colombia—an expedition undertaken and exceptionally well carried out by these collectors under the auspices of the New York Botanical Garden, the United States National Museum, the Arnold Arboretum and the Gray Herbarium. The Eupatorieae alone of this collection amounted to 218 numbers, representing 82 different plants, including 10 new species, 11 new varieties and forms, 1 new hybrid and several specimens showing important extensions of range. The following diagnoses and notes are designed to give published record to such new or otherwise notable plants as it has thus far been possible to identify from the extensive and varied material indicated above. PIQUERIA GALIOIDES DC. Prod. v. 105 (1836); Robinson, Proc. Am. Acad. xlii. 11 (1906); also Contrib. Gray Herb. Ixxvii. 4 (1926). Another specimen of this rarely collected Peruvian plant has been recently examined by the writer, namely Prof. Weberbauer’s no. _ 5379 in the herbarium of the Botanical Museum at Berlin. The Species was originally based by De Candolle upon fragmentary material, showing only the inflorescence and a few of the uppermost leaves, which were described by him thus “foliis linearibus integerri- mis subenerviis” and “ Folia 8-10 lin. longa, vix lineam lata.” Better material now at hand, both from the recent collection of Dr. Pennell (mentioned i in Contrib. Gray Herb. Ixxvii. 4) and Prof. Weberbauer’s : ieee sete ree. shows that the fully developed cauline leaves RECORDS OF THE EUPATORIEAE,—VII 5 become 3 em. long, have a distinct midnerve, and are often provided on one or both sides with from one to three salient sharp teeth. De Candolle stated that the heads were 3-flowered, but dissection of several shows that they vary from 2-flowered to 4-flowered on the same individual. There seems no reasonable doubt that the material secured by Pennell and by Weberbauer is conspecific with that long ago obtained by Haenke. Prof. Weberbauer’s admirable field notes add several details to the knowledge of this slightly known plant. They are as. follows: “Tracht: Strauch, 14 m. hoch. Bliitenfarbe: weiss. Fundort: westliche Andenhinge zwischen 13° und 14° siidl. Breite, iiber dem Hafen Pisco: Pampano. Standort: steiniger Boden mit ganz vereinzelten Pflanzen (hauptsichlich einjihrige Krauter, regengriine Striiucher und Cacteen). 1200-1300 m., Mai 1910.” Mexianthus, gen. nov. Compositarum. Capitula homogama dis- coidea uniflora, flosculis hermaphroditis. Involucrum cylindratum, squamis paucis inaequalibus valde imbricatis scariosis subhyalinis. Corolla regularis tubulosa, faucibus turbinato-campanulatis, limbi dentibus 5 brevibus recurvatis. Antherae oblongae leviter connatae apice omnino exappendiculatae basi rotundatae vel obscure cordatae. Styli rami subfiliformes apicem versus leviter clavatim incrassati. Achaenia graciliter obovoideo-subprismatica 5-costata. Pappus e squamis 5-7 oblongis laciniatis obtusis basin versus plus minusve connatis compositus.—Suffrutex, foliis alternis. Capitula parva uniflora sessilia in glomerulos globosos capitiformes paniculatos dis- posita. Species hucusque cognita unica mexicana. M. mexicanus, spec. nov., erectus 6 dm. vel ultra altitudine aspectu glaber; caule basi lignoso tereti juventate obscure puberulo tardius glabrato; foliis alternis petiolatis utroque acuminatis serrato- dentatis supra basin pinnatim 5-nervatis utrinque viridibus supra parce hirtis subtus subglabris membranaceis 8-12 cm. longis 4-7 cm. latis; petiolo gracili 1-2 cm. longo apicem versus a laxa paniculata; glomerulis graciliter pedunculatis globosis ca. 15-40- capitulatis 7-8 mm. diametro; capitulis arcte sessilibus unifloris; involucri squamis 3(—5) inaequalibus obovato-oblongis valde imbri- catis tenuibus scariosis apice rotundatis glabris; corolla alba glabra; tubo proprio ca. 0.5 mm. longo; faucibus turbinato-campanulatis ca. 0.9 mm. longis; dentibus limbi 0.3-0.4 mm. longis; achaentis nigris 2-2.5 mm. longis apicem versus paullo hispidulis; papp! squamis ca. 0.5 mm. longis—Mextco: Jalisco: sandy soil near stream, Arroyo del Chorillo, Quimixto, alt. 60 m., Dec. 1, 1926, Ynes Mezxta, no. (typz, Gr., U. S.). : This notable plant curiously simulates in habit and inflorescence 6 ROBINSON the anomalous Eupatorium monanthum Sch.-Bip., which also has 1-flowered heads, sessile in spherical glomerules, but on closer examina- tion the plant here described proves entirely distinct, not only specific- ally but as to genus, for it has a scale-pappus entirely different from the fine bristle-pappus of Eupatorium. Furthermore, its anthers are strictly exappendiculate at the apex, showing thus the characteristic feature of the Piquerinae. In this subtribe the plant falls with the West Indian Phania, the Mexican Ageratella, and the new Peruvian genus 7rychinolepis, described below. From all these, it is amply distinguished by its compound heads, single floret, and scarious few-scaled involucre, as well as in its more ample foliage, and in many minor details. It is a pleasure to dedicate this remarkable plant to its discoverer, Mrs. Ynes Mexia, whose courageous exploration of little known portions of the Sierra Madre has brought to light many plants new to science or otherwise of special interest. Trychinolepis, gen. nov. Compositarum. Capitula homogama tubulifiora pauciflora. Involucrum cylindratum; squamis paucis sub- aequalibus angustis vix imbricatis. Receptaculum planum nudum. Corollae aequales regulares, tubo proprio fauces vix ampliatas subaequante, limbo brevi 5-fido, lobis anguste deltoideis recurvatis. Antherae vix connatae apice exappendiculatae basi integrae. Styli rami longi filiformes maturitate valde recurvati apicem versus subabrupte clavato-incrassati et nigrescentes. Achaenia 5-angulata. compositus.—Frutex sublaevis. Folia opposita dentata. Capitula parva numerosissima in thyrsum ovoideum disposita. Floscula 4-5. Species hucusque cognita unica aliquantum xerophytica andina peruviana. Nomen e toebyrvoc, laceratus, et Aex/c, squama, derivatum. T. Hoppii, spec. nov., fruticosa obscure sparseque puberula; caule tereti pallido juventate cum globulis resinosis sparsis ornato maturit- ate laevissimo; foliis breviter petiolatis rhombeis acutis 3-nervatis argute serrato-dentatis (basi cuneata excepta) firmiter membranacels utrinque viridibus juventate cum globulis resinosis adspersis maturit- ate breviter obscureque hirtellis 3.5-5 cm. longis 1-2 cm. latis saepe in axillis proliferis; thyrso terminali ca. 8 cm. alto et crasso; capitulis 4-5-floris 7 mm. longis aliis sessilibus aliis breviter pedicellatis; involucri squamis 4-5 lineari-oblongis acutis stramineo-viridibus paullo granulatis; corollis ca. 3 mm. longis; tubo proprio 1.5 mm. pallido; faucibus subaequilongis purpureis granulatis; dentibus limbi albis ; in angulis hirtellis; pappi squamis ca. 5 irregu- lariter laciniatis 1-1.2 mm. longis corollam vix dimidio aequantibus.— RECORDS OF THE EUPATORIEAE,—VII 7 Peru: on rocks and sand, Arequipa, alt. 2500-2800 m., July 1925, Werner Hopp, no. 28 (Type, in herb. Bot. Mus. Brl., phot. and slight fragm. Gr.). This interesting and rather perplexing plant was found by the writer among some still undetermined specimens at the Botanical Museum in Berlin and it was kindly lent to him for more detailed examination in Cambridge. It has a very close habital resemblance to several species of Ophryosporus and like that genus has the anthers unappendaged at the summit. However, the pappus is very different, being composed of scales instead of hairs and being much shorter relatively to the corolla than is the case in Ophryosporus. Trychino- lepis has a pappus exceedingly like that of the small West Indian genus Phania, but the habit is quite different. The heads of Phania are many-flowered and disposed in a rather open inflorescence, while in Trychinolepis they are 4-5-flowered and borne in a dense thyrse. In Phania the style-tips are less thickened and are not in the same degree nigrescent. In Phania the corolla is white or nearly so; in Trychinolepis the throat is rather deep purple. These differences (together with the wide geographic separation) make one hesitate to refer to Phania the Peruvian xerophyte here in question. OxyYLoBuUs GLANDULIFERUS (Sch.-Bip.) Gray. The writer, Proc. Am. Acad. xlix. 486 (1913), has already drawn attention to the anomal- ous history of this species and shown that its type was Liebmann’s no. 238, since that was the material from which it was first described and published by Hemsley. Though Schultz-Bipontinus had (in herb.) applied to the same specimen the name Ageratum glanduliferum var. albiflorum, his manuscript name has of course no validity. The white- flowered form (since it was obviously the basis of the specific descrip- tion) becomes ipso facto the typical form of the species. An occurrence of this species is western Venezuela has long been known from the Sierra Nevada, Merida, Funck & Schlim, no. 1155 (K., Par., Gen., fragm. Gr.). From the collectors’ notes on the label it is learned that here also the flowers were white. It is now possible to report the occurrence of the species in Colombia, where it appears in two color forms, as follows: — Forma a. typicus, corollis (saltim limbo) albis—Ageratum glan- duliferum B. albiflorum Sch.-Bip. ex Gray, Proc. Am. Acad. xv. 26 (1879), without char. Ozylobus glanduliferus var. albiflorus Klatt, Leopoldina, xx. 75 (1884).—Cotompra: Dept. Santander: Péramo de Romeral, alt. 3800-4100 m., Jan. 29-30, 1927, Killip & Smith, no. 18,526 (Gr.). Also southern Mexico and eastern Venezuela. orma b. coloratus, forma nova, corollis coloratis; tubo purpureo; 8 ROBINSON dentibus limbi roseis—CotomBia: Dept. Santander: Paramo de Santurban, near Vetas, alt. 3950-4160 m., Jan. 17, 1927, Killip & Smith, no. 17,519 (Gr.). FLEISCHMANNIA SCHAFFNERI Gray, Proc. Am. Acad. xvi. 101 (1880). This species was founded upon a specimen collected near San’ Francisco in the valley of San Luis Potosi more than half a century ago. No subsequent record of the species is known to the writer. It seems well worth while, therefore, to mention its redis- covery by Mrs. Mexia at a somewhat distant locality in another Mexican state, as follows: JaLisco: an erect herb, in crevices of rocks by a streamside, Real Alto; trail to Arroyo del Jaguay, Sierra Madre Occidental, alt. 2500 m., Feb. 25, 1927, Ynes Mewia, no. 1752 (Gr., US.) Stevia jaliscensis, spec. nov., fruticosa erecta verisimiliter 1-2 m. altitudine; caule juventate tomentello maturitate prope basin usque ad 1 cm. crasso; ligno duro albo; cortice firmo atrogriseo longitudina- liter rugoso; foliis oppositis petiolatis ovato-lanceolatis obtusis vel acutis crenatis vel subirregulariter serrulatis basin versus integris abruptius cuneatis supra puberulis subtus griseo-tomentellis penni- veniis 5-10 cm. longis 1.5-3.5 em. latis; petiolo 12-25 mm. longo; corymbo composito — convexo 8-22 cm. diametro; capitulis 12-14 mm. longis in cymulas 3-capitulatas dispositis; capitulo terminali sessili; ait lateralibus distincte pedicellatis; involucri squamis lineari-oblongis abrupte acutatis dorso patenter puberulis vel tomentellis; capillis albis attenuatis eglandularibus; corollis albis vel albidis; achaeniis uniformibus cum coronula subintegra scariosa ca. 0.4 mm. alta coronatis—Mexico: Jalisco: hills near Guadalajara, alt. about 1525 m., Palmer (’86) 224 (Gr., Par.), 673 -. ., Par., K.); Pringle, nos. 1773 (rrp, Gr., isoTypEs, Gen., Par., K.), 9984 (Gr., K. A plant foonierdy identified and several times distributed as S. subpubescens Lag., but differing in its longer petioles, somewhat ooser inflorescence, relatively broader and often obtuse leaves, and subtomentellous phyllaries. 8. mitopoda, spec. nov., herbacea verisimiliter annua erecta — diffuse ramosa 4 dm. vel siltra altitudine gracillima; caule tenui tereti pallide brunneo obscure puberulo subglabrato leviter flexuoso; foliis oppositis vel suboppositis vel in inflorescentia alternis, inferiori- _ bus ut videtur caducis, superioribus lanceolatis subsessilibus basin apicemque versus attenuatis integris lateraliter serratis a basi triner- __vatis tenuiter membranaceis utrinque viridibus glabris obscurissime - _ Pertceiates 2-3.5 em. longis 6-9 mm. latis; panicula diffusissima RECORDS OF THE EUPATORIEAE,—VII 9 obscure albo-puberula non glandulifera; pedicellis filiformibus pleris- que I-1.8 cm. longis leviter curvatis; capitulis saepe nutantibus 5-6 mm. longis 5-floris; involucri squamis anguste lanceolatis acutis tenuibus viridibus minute albido-puberulis, pilis non glanduliferis; corollis albis bilabiatis; lobis 3 limbi exterioribus patentibus lanceo- lato-oblongis 1—1.4 mm. longis, lobis 2 interioribus suberectis deltoideo- lanceolatis 0.6-0.9 mm. longis; faucibus ca. 2 mm. longis; tubo proprio ca. 0.4 mm. longo; achaeniis nigris 2.5 mm. longis in angulis sursum scabratis, omnibus cum cupula denticulata 0.2 mm. alta coronatis; achaenio uno biaristato, duobus uniaristatis, duobus exaristatis; aristis violascentibus tenuibus minute barbellatis ca. 2.8 mm. longis.—Mexico: on stony slopes, Rio Balsas, Guerrero, 20 Oct. 1904, C. & E. Seler, no. 4286 (typE, Bot. Mus. Berlin, phot. and fragm. Gr.). This highly characteristic species is obviously most closely related to S. Aschenborniana Sch.-Bip., which, however, has somewhat larger heads, deep purple corolla-throat, copious gland-tipped puberulence, and usually four of its achenes triaristate. From the also related S. micrantha Lag., S. mitopoda differs conspicuously in its much more diffuse’ inflorescence and larger bilabiate corolla. S. MONARDAEFOLIA HBK. Nov. Gen. et Spec. iv. 147 (1820). With this species there may be doubtfully associated a much more slender plant, which produces runners at the base and which has smaller and less pubescent leaves. In all the more important techni- cal characters of inflorescence, florets, achenes, etc., it is in essential However, the considerable at least a varietal recognition. The two varieties may be dis- thus: tinguished Var. a. typica, erecta 6-12 dm. alta, turionibus ut videtur nullis; foliis maximis 4-12 cm. longis subtus in superficie et praecipue in nervis venisque griseo-pubescentibus.—S. monardaefolia HBK. I. ce. Var.? 8. repens, var. nov., multo gracilior 3-4 dm. alta; basi distincte repens et stolonifera; Solis 3-4 cm. longis utrinque viridibus subtus in nervis venisque parce —Mexico: Jalisco: erect, common, Real Alto, Arroyo de las Ciiebs, alt. 2500 m., Ynes Mexia, no. 1610 (Gr., US.). Possibly a distinct species but marked chiefly by vegetative traits of doubtful constancy. S. Nelsonii, spec. nov., fruticosa erecta vel adscendens veri- — 1 m. altitudine in foliis involucrisque subvernicosa; caule et 1 primo griseo-tomentellis tardius glabratis; erectis atropurpureis ioliis lanceolato-oblongis serratis utroque acutis vel acuminatis 10 ROBINSON Gray Herb.). A shrubby thickish-leaved species near S. phlebophylla Gray, but having narrower more lanceolate leaves (with more widely spreading veins and less prominulent veinlets) and subglabrous in- volucre. S. (§ Podocephalae) orizabensis, spec. nov., perennis gracilis verisimiliter subherbacea; ramis teretibus purpurascentibus juventate inconspicue puberulis tardius glabratis; internodiis 6-7 cm. longis; foliis oppositis (eis inflorescentiae plerisque alternis), infimis rhombeo- ovatis graciliter petiolatis, superioribus oblongo-lanceolatis sessilibus sed basin versus angustatis quasi alato-petiolatis, omnibus integerri- mis apice rotundatis vel saepius acutis membranaceis utrinque viridi- bus et glabris 4-8.5 em. longis 1-3.5 cm. latis multo supra basin 3-nervatis; inflorescentia laxa paniculata cum foliis milto reductis bracteiformibus sessilibus elliptico-lanceolatis ornata 1.5 dm. alta et crassa; pedicellis gracilibus minute glandulari-puberulis plerisque 4-6 mm. longis; capitulis saepe nutantibus ca. 1 cm. longis 5-floris; involucri squamis 5 lineari-lanceolatis acutis viridibus minute glandu- lari-puberulis; corollae limbo albo, dentibus 2-nervatis ca. 2 mm. longis et 1.2 mm. latis obtusis; faucibus subcylindratis purpurascenti- bus ca. 3 mm. longis; tubo proprio paullo contracto ca. 2 mm. longo; achaeniis 3.5 mm. longis similibus omnibus cum cupula scariosa paullo denticulata vix 2 mm. alta coronatis—S. lava Robinson & Seaton, Proc. Am. Acad. xxviii. 107 (1893), in part, solely as to pl. of Bourgeau, but not as to type.—Mexico: region of Orizaba, Bot- tert, no. 1125 (BM., K., Gen.); Escamilla, Bourgeau, no. 3331 (TYPE, Gr.; K., Gen.). species in many respects close to S. micradenia, but having its leaves more evenly distributed on the stem, the upper internodes less elongated, the upper (relatively narrower) leaves subsessile, and the inflo hee more copiously bracteate. S. Puncrata. The status and confused application of this bino- mial merit at least a brief discussion. The two uses of this name have been as follows: No.1. S. puncrata (Ort.) Pers. Syn. ii. 403 (1807), based upon _— Ageratum punctatum Ort. Dec. 37 (1797). RECORDS OF THE EUPATORIEAE,—VII 11 No. 2. S. puncratra (Jacq.) Sch.-Bip. Linnaea, xxv. 286 (1853), based on Ageratum punctatum Jacq. Hort. Schon. iii. 28, t. 300 (1798). No. 1. This application of the name, though resting upon the earlier synonym, has been pretty generally neglected. It was listed by Steudel in the first edition of his Nomenclator, but in the second edition was reduced to the synonymy of S. eupatoria (Spreng.) Willd. Spec. Pl. iii. 1775 (1804), a species based on Mustelia eupatoria Spreng. Nachtr. 1, Bot. Gart. Halle, 28 (1801) & Trans. Linn. Soc. vi. 152, t. 13 (1802). Steudel’s reduction has been followed by Hemsley, Biol. Cent.-Am. Bot. ii. 85 (1881) and by the Index Kewen- sis. Of course, if this supposed identity between Ageratum punctatum Ort. and Mustelia eupatoria Spreng. could be confirmed Ortega’s earlier specific name would have to be taken up under Stevia, but the writer has found no means of substantiating the identity in question and several traits are mentioned in Ortega’s original descrip- tion which render it doubtful, indeed rather unlikely. From his all too few specific characters it appears that his plant had linear- lanceolate carinate leaves toothed at the apex and narrowed at the base—traits which would agree well enough with many specimens of the variable S. serrata Cav., but are by no means appropriate to S. eupatoria (Spreng.) Willd. in which the leaves are narrowl oblong, flat (see illustration of Sprengel and note of Schultz “ foliis semper planis”), subentire, and not at all narrowed to the rather abruptly contracted base. It is true, Persoon, when transferring Ageratum — Ort. to Stevia may well have had some material of S. oria before him, but his work is very sketchy and probably Seiad. for it will be noticed that while attempting to separate Ageratum punctatum Jacq. from the earlier A. punctatum Ort., he cites Jacquin’s descrip- tion and plate in the treatment of each of his segregates. From the facts stated it will be evident that there are no pres- ent grounds for replacing S. eupatoria (Spreng.) Willd. by S. pune- tata (Ort.) Pers. It would seem more probable that the latter is best placed as a doubtful synonym of S. serrata Cav. No. 2. This application of the binomial, Stevia punctata, has been scarcely employed except by Schultz and the plant so designated by him (namely the one with narrowly oblanceolate to linear leaves which are conspicuously toothed, gradually narrowed to the base, and in nature often carinate) has in general = sce properly passed under the name S. serrata Cav. Ic. iv. 33, t. 355 (1797). Just why Schultz displaced the earlier S. serrata ee a the later specific name of Jacquin is not clear. It is true, Schultz was interpreting 12 ROBINSON the species somewhat more broadly than either Jacquin or Cavanilles, but this of course constitutes no ground for changing its name. From the above it will be seen that no. 2 falls definitely into the synonymy of S. serrata Cav., while no. 1 may be more doubtfully disposed of in the same manner. Symphyopappus compressus (Gardn.), comb. nov. Eupatorium compressum Gardn. in Hook. Lond. Jour. Bot. v. 473 (1846); Walp. Rep. vi. 709 (1847). Symphyopappus polystachyus Bak. in Mart. Fl. Bras. vi. pt. 2, 368, t. 98 (1876), as to plant described and figured but not as to name-bringing synonym, viz. Eupatorium poly- stachyum DC. which is E. crenulatum Spreng. DeCandolle’s descrip- tion as well as the name of his E. polystachyum, calling for somewhat spicate or racemelike branches of the inflorescence should, even without further evidence, have prevented Baker from associating this name with the corymbiferous plant described and figured in the Flora Brasiliensis. While recently at the Museum of Natural History in Paris the writer was able to examine and photograph an excellent sheet of the type of De Candolle’s E. polystachyum, namely “Mus. imp. Bras. sub. n. 488” and found it precisely E. crenulatum Spreng., the species treated and illustrated by Baker, l. c. 321, t. 84, as E. dendroides Spreng. See Hieron. in Engl. Bot. Jahrb. xxii. 776 (1897). Baker must have been misled by the likeness of the foliage in which there is in fact a fair degree of similarity between these otherwise exceedingly different plants. After discarding the name polystachyum, which from its type is quite inapplicable, it appears that the next name applied to the corymbiferous plant was Eupatorium compressum Gardn. The plant is somewhat transi- tional between Eupatorium and the segregate genus Symphyopappus, but apparently must be classified with the latter if the genera are kept apart. abrupte contractam versus integris utrinque viridibus supra secundum nervos venasque maximas brevissime hirtello-puberulis a puncto supra basin 3-nervatis subtus paullo pallidioribus textura firmiter membranaceis 10-14 cm. longis 8-9 em. latis venulis subdisjunctim __ Ppellucidis; petiolo 2-3 cm. longo apicem versus plus minusve alato; _ corymbo oppositirameo 3 dm. diametro; ramis curvato-adscendenti- _ bus elongatis incurvo-puberulis; bracteis ovatis vel spatulatis vel RECORDS OF THE EUPATORIEAE,—VII 13 (supremis) lanceolato-subulatis multo reductis; capitulis ca. 25- floris 8 mm. longis 3.5-4 mm. diametro arcte sessilibus in glomerulos subsphaericos dispositis; involucri squamis ca. 32 ca. 4—5-seriatim imbricatis albidis glabriusculis eroso-ciliolatis apice rotundatis dorso 3-5-nervatis; corollis cylindratis gracilibus 3.3 mm. longis limbum (vix 0.2 mm. longum) versus subcaeruleo-purpureis; achaeniis 2.5 mm. longis in angulis obscure hirtellis; pappi setis ca. 40 tenuiter capillaribus albis—CoLompB1a: Dept. Santander: in thickets between Narifio and El Tambor, alt. 150-300 m., Dec. 4, 1926, Killip & Smith, no. 14,967 (rypE, in Gray Herb.). This species is transitional between §§ Subimbricata and Cylindro- cephala. It shows probable affinity to E. acuminatum HBK. and perhaps to E. morifolium Mill. Its leaves are thicker and more sparingly and coarsely toothed than in the former and the heads more densely glomerate than in the latter. From both, the species here proposed differs in its much more numerously flowered heads. . It is named in honor of one of its discoverers, Mr. Albert C. Smith, who accompanied and effectively aided Mr. Killip upon his recent exploration of northern and eastern Colombia. E. (§ Supmmpricata) ARBUTIFOLIUM Benth. PI. Hartw. 199 (1845). This Colombian species, thus far known only from the Department of Cauca, can now be reported from the Departments of Santander and Norte de Santander, where collected by Messrs. Killip & Smith. The recently obtained material falls into two pretty well marked varieties, as follows: Var. «. typicum, ramis ramulisque inflorescentiae patenter copioseque ferrugineo-pubescentibus; _pilis articulatis.—E. arbuti- folium Benth. 1. c—Coromsta: Dept. Cauca: in woods near Huambia, Hartweg, no. 1100 (K., phot. Gr.); “San Jose,” San Antonio, bushy forest, Cordillera Occidental, alt. 2400-2700 m., June 28, 1922, Pennell & Killip, no. 7423 (Gr.). Norte de Santander: edge of woods on eastern slope of Péramo del Hatico, between Toledo and Pamplona, alt. 2900 m., Mar. 12, 1927, Killip & Smith, no. 20,599 (Gr.), the last a luxuriant form with pubescence and all important characters as in the type, but with somewhat larger leaves, some of them as much as 17 cm. long and 6 cm. wide. : : Var. 8. sublaeve, var. nov., ramis inflorescentiae glabris vel subglabris; ramulis breviter inconspicueque puberulis; aliter var. typico simillimum.—Co.omBta: Dept. Santander: forest near Charta, alt. 2000 m., Killip & Smith, nos. 18,870 (Gr.), 18,871 (TYPE, in Gray Herb.); swamp near Charta, Killip & Smith, no. 18,906 (Gr.); open bank near La Baja, alt. 2200-2600 m., Jan. 19-28, 1927, Killip & 14 ROBINSON Smith, no. 17,129 (Gr.); wooded banks of the Rio de la Baja, below Baja, alt. 2200-2300 m., Killip & Smith, no. 18,288 (Gr.); Mesa de los Santos, alt. 1500 m., Dec. 11-15, 1927, Killip & Smith, nos. 15,188 (Gr.), 15,201 (Gr.). According to field notes of the collectors this smoothish variety sometimes appears as a shrub, 1.3-4 m. in height, sometimes as an “herbaceous vine.”’ Its floral parts are sometimes found to be all white, in other cases the corolla-tube is said to be greenish and the limb white to light pink. The specimens appear to be quite identical as to essentials. E. BpaccHaroiwEs HBK., Nov. Gen. et Spec. iv. 132 (1820). This highly characteristic species exhibits two well marked varieties, which may be distinguished thus: Var. a. typicum, foliis utrinque glaberrimis.—Synon. and distrib. as stated by the writer in Proc. Am. Acad. liv. 287 (1918). Var. 8. araneotectum, var. nov., foliis supra glabris subtus dense pallideque araneoso-lanatis; pilis implexis ut videtur viscidis sed neque glandulari-capitatis nec conspicue articulatis; aliter var. typico simillimum.—Cotomstia: Dept. Santander: in wood on western slope of Péramo Rico, alt. 3000-3600 m., Jan. 15-19, 1927, Killip & Smith, no. 17,230 (Gr.). According to the excellent field-notes of the collectors, this is a viscid shrub 2-3 m. high, with corolla proxim- ally greenish and distally pinkish, with deep pink style-branches. As evidence of intergradation of the two varieties is still lacking, their status is subject to some doubt. However, in all respects other than that of pubescence they appear identical and though the pubes- cence character is striking it seems not to justify specific separation. E. baccharoides var. araneotectum X crassiceps. A plant, which from its intermediate nature seems likely to be a hybrid of this parentage, was collected in CoLomsia: Dept. Santander: on the western slope of Paéramo Rico, alt. 3600 m., Jan. 15-19, 1927, by Messrs. Killip & Smith, no. 17,715 (Gr.). It has the same nodding or recurved inflorescence highly characteristic of both its supposed parents, but has heads intermediate both in size and in the number of phyllaries and florets, and has leaves which are neither glabrous as in E. crassiceps nor densely arachnoid-lanate as in E. baccharoides var. araneotectum, but covered (on the lower surface) with a sparse indument composed of hairs of precisely the same nature as those in the Meginaed just mentioned. The inflorescence has many more heads than in E. crassiceps but is much less dense than in E. bac- charotdes var. araneotectum. It will be noted, furthermore, that both of the supposed parents, as well as the putative hybrid, were collected RECORDS OF THE EUPATORIEAE,—VII 15 in similar paramo conditions in a region traversed by the expedition of Messrs. Killip & Smith between January 15 and 19. Evidences of hybridity have been rarely found in the genus Eupatorium, but this case appears probable and a hybrid status seems the best disposition of the intermediate at least until it can be known from more extensive material. Perhaps the best evidence lies in the fact that the inter- mediate seems to exhibit no trait, however slight, which is not found in one or both of supposed parents, though these themselves differ rather widely especially in the size of the heads. E. borinquense (Britton), comb. nov. Osmia borinquensis Brit- ton, Sci. Surv. Porto Rico and Virgin Ids. vi. 288 (1925). Through the kindness of Dr. Britton the writer has had an opportunity to examine the best available (though somewhat immature) material of this “straggling vine-like shrub” and agrees with Dr. Britton that it represents a species not previously described. , However, prolonged study has not revealed any satisfactory boundary between Eupator- ium and its attempted segregate -Osmia. Therefore it seems best to place this interesting novelty in the older and more comprehensive genus. E. Bucheri, nom. nov. Through an unfortunate and much re- gretted error in copying, this name was published as E. Ruckeri by the writer, Contrib, Gray Herb. Ixxvii. 35 (1926). As will be readily understood it was the intention to name the plant for its collector. Unfortunately his name, somewhat obscurely recorded on a manu- script label, was mistakenly inferred to be Rucker, but (as is now learned from Dr. N. L. Britton) is in fact Bucher. An orthographic correction of this nature is certainly permissible under Art. 57 of the International Rules of Botanical Nomenclature. E. (§ Subimbricata) carnosifolium, spec. nov., suffruticosum 2 m. altum; caule tereti robusto medulloso 6 mm. vel ultra diametro pallide brunneo minute papilloso-puberulo et sordescenti; internodiis textura carneis sed omnino non coriaceis a puncto 1-3 cm. supra basin 3(-5)-nervatis aliter penniveniis; petiolo 1.5-5 cm. longo; inflorescentia terminali 7-20 cm. diametro ad anthesin saepe sub- densa sed tardius laxiuscula planiuscula; pedicellis 6-15 mm. longis sordide puberulis; capitulis 50-60-floris ca. 1 cm. altis et 1.2 cm. latis; involucri campanulati squamis ca. 25 subtriseriatim imbricatis, extimis ovatis acutiusculis ca. 3 mm. longis et 1.4 mm. latis dorso 16 ROBINSON sordide puberulis, intermediis oblongo-lanceolatis subacutis vel acutis 6 mm. longis et 2 mm. latis etiam dorso puberulis, intimis lanceolatis acutis 3-nervatis ciliolatis aliter subglabris ca. 7 mm. longis et 1.2 mm. latis; corollis carneo-roseis tubulatis 4.8 mm. longis summum versus paullo ampliatis; limbo brevi 5-dentato; achaeniis nigris 3mm. longis glabris; pappi setis ca. 45 flavescenti- vel brunnes- centi-albis 4.7 mm. longis——Campuloclinium crenatum Schlecht. ex Mart. Herb. Fl. Bras. 345 (1837), not E. crenatum DC., nor Gomez. Conoclinium mikanioides Sch.-Bip. in sched. pl. Riedel, acc. to Bak. in Mart. FI. Bras. vi. pt. 2, 306 (1876) in synon., not FE. mikanioides Chapm.—Braziu: Bahia: wooded coastal marshes near [hleos, Martius Herb. Bras. no. 668 (tTyPE, in Gray Herb., isorypE, Mo.); southern part of the province, Blanchet, nos. 3124 (Univ. Gen.), 3124A (Gen.). Ceara: Serra de Maranguape, Ule, no. 9119 (Brl.). Brazil without indicated locality: Luschnath, no. 78 (Brl.), Riedel (Gr.). This plant, long known as a recognizable entity, has had no valid designation. The specific name crenatum, given to it with brief diagnosis by Schlechtendal under the now obsolete generic name Campuloclinium, cannot be applied under Eupatorium because antedated by homonyms. The plant is one of several diverse elements placed by Baker (I. c.) in E. sordescens DC., a species which as to its type was exceedingly different, being of § Eximbricata and having much smaller about 20-flowered heads and more deltoid-ovate membranaceous leaves. From E. Lundianum (E. Schlechtendalit -DC., Mikania resinosa Spreng., Conoclinium resinosum Sch.-Bip.) the ink here under discussion differs in its much larger leaves, longer petioles, more pointed outer phyllaries and shorter (sordid rather than rufescent) puberulence. E. (§ Cy~inprocePHaLa) CHRISTIEANUM Bak. in Mart. Fl. Bras. vi. 298 (1876). While visiting the Kew Herbarium in the summer of - 1927, the writer was able to examine the type of this little known species of southern Brazil and to identify with it the puzzling plant of Paraguay and northern Argentina, which he had some months before described as E. laevigatulum, Contrib. Gray Herb. Ixxvii. 21 (1926), after it had by others been incorrectly referred to E. gracili- florum DC. and to E. Orbignyanum Klatt. The reduction of E. laevigatulum to E. Christicanum automatically extends the range of the latter to Paraguay and northern Argentina. ____E. § CytryprocerHaa) coLumB1ANuUM Heer, Mém. Soc. neuchat. a rasy Nat. v. 421 (1913). This Colombian species, said by its author a : the Departments of Antioquia and Tolima and y several times collected about Bogoté, can now be RECORDS OF THE EUPATORIEAE,—VII 17 extended to the Departments of Santander and Norte de Santander, where it shows marked variation as follows Var. a. typicum, capitulis 9-12-floris; foliis crassiusculis supra valde rugoso-bullatis subtus dense sordideque piloso-pubescentibus.— For details of occurrence see Robinson, Proc. Am. Acad. liv. 274 (1918) Var. 8. umbratile, var. nov., capitulis 5~-7-floris; foliis tenuibus planis vel vix subrugosis brevissime inconspicueque in nervis venisque puberulis.—Cotomsia: Dept. Santander: in dense forests, Quebrada de Pais, north of La Baja, alt., about 3200 m., Killip & Smith, nos. 18,814 (TyPE, in Gray Herb.), 18,796 (Gr.). Compared alone with the typical variety these numbers might seem to warrant separation as an independent species, but a transi- tional form is shown by another collection in the same series, namely: Dept. Norte de Santander: eastern slope of Paéramo de Santurban, toward Mutiscua, alt. 3600-3900 m., Feb. 20, 1927, Killip & Smith, no. 19,604 (Gr.). E. CosTaRIcENSE O. Ktze., Tig Gen. i. 337 (1891). So far as known to the writer this species, collected by Dr. Kuntze at Angos- tura, Costarica, in 1874 and described by him in 1891, has never been’ subsequently recorded. This seems rather strange in considera- tion of the large amount of field work since accomplished in Costarica —exploration far exceeding in extent and care the hurried collections of Kuntze. It was therefore a matter of much interest to the author to find an unnamed Ecuadorian specimen of the species in some unindentified material at the Botanical Museum in Berlin. The sheet, kindly lent for more careful examination at Cambridge, was there compared with the type of E. costaricense, which was on loan from the New York Botanical Garden. The Ecuadorian specimen was collected in tropical vegetation of the Province Esmeralda along the Cachalte River in August, 1904, by the late Father Sodiro, being no. 14 of a small series of Compositae sent to Prof. Hieronymus for identification in 1909. Comparison of the specimens of Kuntze and of Sodiro discloses close correspondence even to trifling char- acteristics. There can be no doubt whatever that the two are speci- fically identical. The geographic jump from Costarica to Ecuador, while striking, is not to be regarded as anomalous. A number of plants are now known which are common to the ¢ordilleras of these two countries. Father Sodiro’s discovery yields another instance of this floral interrelationship and adds another species to the ever lengthening list of South American Eupatoriums. . § Subimbricata) crassiceps, spec. nov., fruticosum gracile 18 ROBINSON 1.3-4 m. altum glabrum glutinosum; caule firmiter lignoso fusco- brunneo tereti albo-medulloso 5 mm. vel ultra diametro; ramis oppo- sitis patenter adscendentibus; internodiis plerisque 5-20 mm. longis; foliis oppositis breviter petiolatis oblongis obtusis vel subacutis serrulatis basi cuneatis coriaceis penniveniis utrinque glabris sub- lucidis supra planis subtus reticulato-venulosis 3.5-6 cm. longis 12-22 mm. latis; petiolo 2-4 mm. longo lato atrorubro; inflorescentiis ramos terminantibus curvato-nutantibus vel subpendulis plerumque 3-5-capitatis; pedicellis fusco-brunneis vel -purpureis viscosissimis 2-3.6 cm. longis; capitulis pro genere maximis 1.5-2 cm. longis e crassis multiflosculosis; involucri squamis ca. 35 valde inaequalibus ca. 4-seriatim imbricatis tenuibus viridibus vel purpureis viscosis multistriatulis margine eroso-serrulatis, extimis ovatis, ceteris ovato- oblongis vel oblongis; corollis ca. 8 mm. longis pallide roseis; tubo proprio 1.5-2 mm. longo; faucibus paullo ampliatis ca. 5 mm. longis; limbi dentibus 5 ovatis acutiusculis 0.5 mm. longis; achaeniis nigris sursum praesertim in angulis hispidulis 5.5 mm. longis; pappi setis ‘ca. 35-40 inaequalibus. albis acutis sublaevibus.—CotomBta: Dept. Santander: edge of Péramo de Santurbdn, near Vetas, alt. 3500-4000 m., Killip & Smith, no. 17,583 (typr, in Gray Herb.); Péramo de Romeral, alt. 3800-4100 m., Jan. 29-30, 1927, Killip & Smith, no. 18,534 (Gr.). A notable species with exceptionally large heads in nutant or recurved inflorescences similar to those of the much smaller-headed E. baccharoides HBK. E. (§ Suprmericata) pEsectuMm. Robinson, Contrib. Gray Herb. Ixxvii. 12 (1926). This species was based on a rather fragmentary specimen (in the Museum of Natural History at Paris) collected by d’Orbigny at” Miraflor, “Bolivie.”” At the time of description no Bolivian place of this name was found on any map or in any atlas of Bolivia available to the writer, and he expressed some doubt whether the locality intended might not have been Miraflores in Argentina. Recently Dr. I. M. Johnston has brought to the writer’s attention the fact that a small place in Bolivia called Miraflor is mentioned in the general narrative and shown on the maps of d’Or- bigny’s explorations. It is situated slightly to the northeast of Potosi. This establishes the type-locality, which had previously been in doubt. The matter is of interest just now as the species has recently been rediscovered at a second and rather remote station ____ as follows: Peru: Dept. Cuzco: in the Sacsahuaman hills, alt. 3600 m., duly, 1927, F. L. Herrera, no. 1577 (Gr.). The Peruvian material, _ though collected some 800-900 km. to the northward, is closely olivian. RECORDS OF THE EUPATORIEAE,—VII 19 E. (§ Supmericata) pripymMum Klatt, Ann. Naturhist. Hofmus. Wien, ix. 356 (1894). Material apparently conspecific with this highly characteristic Bolivian species has recently been received by the author from the Province of Tucumdén in the Argentine Republic. The specimens in question formed a part of an exceedingly interesting sending most kindly forwarded by Professor M. Lillo. While agreeing with the original Bolivian material in all significant traits, the Argen- tinian plant differs in the nature of its pubescence and in consequence the species, as now known, may be varietally divided as follows: Var. a. typicum, ramis juvenilibus et pedicellis et involucri squamis exterioribus puberulis vel subtomentellis cum pilis brevibus acutis eglandularibus.—E. didymum Klatt, 1. e—Botnrvia: La Paz: Prov. Larecaja: Moyabayam near Sorata, Mandon, no. 262 (Gr.). Var. 8. glandulitectum, var. nov., ramis juvenilibus et pedicellis et involucri squamis extimis breviter puberulis cum pilis brevissimis minutissime glandulari-capitulatis.—ARGENTINA: Prov. Tucuman: Cuesta de San Javier, Aug. 17, 1913, Castillon, no. 2960 (Gr.); in same locality, among rocky cliffs, alt. 1100 m., Aug. 1914, ex herb. M. Tallo, no. 13,153 (Gr.). E. (§ Cylindrocephala) farinosum, spec. nov., fruticosum erect- um vel reclinatum ramosissimum; caule tereti albo-medulloso juven- pubescentibus (pilis densis gracilibus) et sparse cum glandulis globosis adspersis subtus multo pallidioribus in nervis pubescentibus in superficie densissime cum glandulis globosis sessilibus albidis ornatis ea. 3(-4.5) cm. longis et 1.5(-3) em. Jatis; petiolo 2-10 mm. longo tomentello; corymbis multis convexis subdensis saepe paniculatim dispositis; pedicellis filiformibus ad 5 mm. longis; capitulis ca. 23- floris 9-10 mm. longis 6 mm. diametro; involucri subcylindrici squamis ca. 20 obtusis albidis viridi-striatis ciliolatis, extimis ovatis ca. 2 mm. longis et 1.3 mm. latis, intermediis oblongis ca. 5 mm. longis et 1.5 mm. latis, intimis oblanceolato-linearibus ca. 7 mm. longis et 1 mm. latis; corollis albis vel purpurascentibus a basi ad limbum gradatim ampliatis non in tubum et fauces distinctas divisis ca. 4 mm. longis; dentibus limbi patenter recurvatis deltoideis ca. 0.4 mm. longis; achaeniis gracilibus ca. 4 mm. longis deorsum de centibus in angulis minute scabratis aliter glabris; pappi setis ca. 35 tenuiter capillaribus vix scabratis albis—VENEZUELA: erect bush 20 ROBINSON with white or purplish flowers, Loma de Moron, vicinity of Valera, State of Trujillo, Nov. 17, 1922, Prof. H. Pittier, no. 10,719 (TYPE, Gr., IsOTYPEs, Par., etc.); also as a shrubby trailing less branched and larger leaved white-flowered plant in forest at Dividive, State of Trujillo, no. 10,824 (Gr., Par., etc.). Locally called estoraque and pesebrito and used as floral decoration. This species in many respects rather closely simulates forms of the widely distributed and variable E. odoratum L. but is readily distinguishable by the dense (meal- like) granulation on the lower leaf-surface. Var. asetosum, var. nov., ramulis et costa foliorum asetosis sed breviter molliterque pubescentibus.—Co.omsra: Dept. Santander: shrub, 2.4~3 m. high, in woods, alt. 3300-3600 m., western slope of Paramo Rico, Jan. 15-19, 1927, Killip & Smith, no. 17,769 (Gr.)- os Subimbricata) Herrerae, spec. nov., herbaceum perenne ___ Yel fruticosum aspectu glabrum sed obsolete granulatum vel puberu- RECORDS OF THE EUPATORIEAE,—VII 21 lum plus minusve glutinosum; caule subhexagonali medulloso usque ad 8-10 mm. diametro; internodiis 1-9 cm. longis; ramis patenter adscendentibus; foliis oppositis breviter petiolatis oblongo-lanceolatis attenuatis argute serrato-dentatis basi cuneatis et integris concoloribus utrinque laete viridibus subglabris a basi penniveniis 6-14 cm. longis 2-5 em. latis dense puncticulatis, venulis utrinque immersis; petiolo ca. 5 mm. longo crassiusculo angustissime marginato; corymbis densis fastigiatim ramosissimis multicapitulatis rotundatis folia suprema vix superantibus minute pulverulentis; capitulis 4—6-floris ca. 1 em. longis subsessilibus vel arete sessilibus binis vel trinis ad apices inflorescentiae ramulorum gracilium flexuosorum dispositis; involucri squamis ca. 8 multo inaequalibus brevissime ciliolatis dors apicem versus saepe cum glandulis globosis minutis instructis et plus minusve vernicosis, extimis deltoideo-lanceolatis acutis 2.5 mm. longis, intimis lineari-oblongis obtusiusculis ca. 5 mm. longis; argute angulatis deorsum attenuatis minute glandulari-atomiferis ca. 2.6-3 mm. longis; pappi setis ca. 30 inaequalibus flavescenti-albis vix scabratis.—Prru: Dept. Cuzco: Urubamba Valley, alt. 3000 m., Aug. 1927, Prof. F. L. Herrera, no. 1593 (rye, in Gray Herb.). To this species may also be referred several specimens, probably all from the Department of Cuzco, 1839-40, Claude Gay (Par., phot. and fragm. Gr.). Among the species previously described this approaches in habit E. drepanoides Robinson, which however has more slender and relatively longer petioles, distinctly 3-nerved leaves and a much longer somewhat arachnoid ciliation of its phyllaries. E. Herrerae also approaches E. coelocaule Robinson, but that has a terete hollow stem, longer and more slender petioles, somewhat differently disposed E. (§ Subimbricata) heteroneurum (Ernst), comb. nov. Ina Acad. liv. 339) prepared some ten years ago, doubt was expressed as to the precise identity of Critonia heteroneura Ernst, Flora, Ivii. 210 (1874). As no authentic material had been available for . nation and as the species appeared from description to agree in 1 more essential characters with the. variable and widely distributed Eupatorium morifolium Mill. (also known to occur In northern oe ROBINSON Venezuela), it was provisionally referred to that species. The subsequent opportunity to examine authentic material of Dr. Ernst’s plant and to study what is certainly conspecific material from Colom- bia shows that it differs from E. morifolium in having smaller only 4-5-flowered closely sessile heads instead of 8-16-flowered perceptibly pedicellate heads, and also in having more decidedly angulate branches and somewhat thinner, brighter green, and more finely pellucid- punctate leaves. It thus becomes possible to recognize it as a distinct species. Happily Dr. Ernst’s specific name, not having been other- wise applied in Eupatorium, may be retained in the new generic position. The material recently received is as follows: CoLOMBIA: Dept. Santander: shrub 2.4-3 m., in edge of forest, Rio Suraté valley, above Surata, alt. 2000-2300 m., Jan. 5-6, 1927, Killip & Smith, no. 16,691 (Gr.). Dept. Norte de Santander: tree 5-6 m. high, in woods along stream, Loso and vicinity (north of Toledo), alt. 2200-2400 m., Killip & Smith, no. 20,435 (Gr.). These new stations not only add another Eupatorium to the flora of Colombia but extend the known range of this species westward from the region of Caracas. E. (§§ Conociintum) nuMILE (Benth.) Hieron. Beih. Bot. Centralbl. xill. 68 (1902); Robinson, Contrib. Gray Herb. Ixxvii. 19 (1926). Hitherto this rare and little known species has been collected only on the mountains near Bogoté, first by Hartweg and much later by the Princess Therese of Bavaria. Its range may now be extended to Dept: Santander and Dept. Norte de Santander and furthermore three color-forms may be indicated as follows: Forma a. typicum, corollae limbo et styli ramis albis.—Cono- clinium humile Benth. Pl. Hartw. 199 (1845), so far as can be judged from examination of isotype in herb. N. Y. Bot. Gard.—CoLoMBIA: Cordillera de los Andes, near Bogoté, Hartweg, no. 1102 (K., N. Y.). Dept. Santander: Péramo de las Puentes, above la Baja, alt. 3500- 3700 m., Jan. 25, 1927, Killip & Smith, no. 18,233 (Gr.). Forma b. caeruleum, forma nova, corollae limbo et styli ramis pulchre coloratis caeruleis; aliter formae typicae simillimum.— Cotomsia: Dept. Santander: Péramo de las Puentes, above la Baja, alt. 3500-3700 m., Jan. 25, 1927, Killip & Smith, no. 18,190 (Gr.). Dept. Norte de Santander: Péramo de Romeral, alt. 3800-4200 m., Jan. 29-30, 1927, Killip & Smith, nos. 18,564 (rypx, in Gray Herb.), 18,664 (Gr.). Forma c. erubescens, forma nova, corollae limbo et styli ramis ___ Ppallide roseis—Cotomsta: Dept. Norte de Santander: between _ Mutiscua and Pamplona, alt. 3400 m., Feb. 23, 1927, Killip & Smith, 2 - no. 19,731 (Gr.). RECORDS OF THE EUPATORIEAE,—VII 23 E. (§ CyLINDROCEPHALA) IVAEFoLIUM L. Syst. ed. 10, 1205 (1759), as tuaefolium. To this species, though with some doubt, the writer would refer an extraordinary plant collected in northern Argentina vy . Pedro Jérgensen. In it the involucre, instead of being subcylindric and 4—5-seriate, becomes in most of the heads elongate- turbinate and very attenuate at the base, the phyllaries being greatly increased in number and, as they extend down the pedicels, gradually reduced in size. This very striking departure from normal EL. ivaefolium would certainly justify specific separation were it certain that it did not arise from some teratological modification. The latter seems probable for two reasons. In the first place, except for the downward prolongation of the involucre, the plant corres- ponds in all particulars to southern forms of the widely distributed and somewhat variable E. ivaefolium, and in the second place the involucres on Mr. Jérgensen’s plant (of which several specimens in different herbaria have been examined) are not uniform. few of them are subcylindric and about 5-seriate as in the typical E. ivaefolium, while the others are greatly altered even on the same individual. This variability on the same plant renders the involucral character here of doubtful value and makes it probable that it is due to some teratological influence. However, the form is so striking that it is bound to be separated to a certain extent from the typical form of the species. It may accordingly be given status as follows: Var. perturbinatum, var. nov., involucri (capitulorum plerorum- que) deorsum attenuati perturbinati usque ad 13 mm. longi squamis multiseriatim imbricatis ARGENTINA: Formosa: Guayalee, Jérgen- sen, no. 3351 (Gr.). E. (§ Subimbricata) jaraguense, spec. nov., fruticosum 6-12 dm. altum; caulibus erectis subteretibus purpureo-brunneis albo- medullosis apicem versus ramosis; pubescentia duplici e pilis aliis densis rectis brevissimis patentibus glandulari-capitatis aliis sparsis multo longioribus attenuatis curvatis eglandusis; internodiis plerisque 14 cm. longis; foliis alternis subsessilibus vel breviter petiolatis rhombeo-ovatis acutis basi cuneata excepta argute serratis 4-6 cm. longis 1-2.5 cm. latis membranaceis utrinque breviter pubescentibus subtus paullo pallidioribus supra basin subtrinervatis; corymbis terminalibus sublaxis compositis basi foliaceo-bracteatis; pedicellis saepe 2-3 cm. longis adscendentibus rectis; capitulis ca. 45-floris 14 mm. altis 11 mm. diametro; involucri campanulati squamis lanceolatis valde inaequalibus ca. 4-5-seriatim imbricatis acutis atropurpureis dorso-puberulis; receptaculo plano; corollis purpureis ca. 6-7 mm. ob te 1h listinctis;d tibus limbi parvis patentibus; oOo longis oO 24 ROBINSON achaeniis nigris 5 mm. longis praecipue in angulis hispidulis; pappi setis ca. 25 sublaevibus corollam subaequantibus.—BraziL: Sao Paulo: grassy summit of the peak of Jaragua, near T aipas, Feb. 19, 1922, Prof. & Mrs. E. W. D. Holway; no. 1574 (tyPE, in Herb. Univ. Minn., phot. and fragm. Gr.). This species may well be placed near E. Vauthierianum DC., which it resembles in many of its traits, but from which it differs conspicuously in its alternate, shorter, and less clearly petiolate leaves, as well as in its somewhat more firm and less scarious phyllaries. E. (§ Cylindrocephala) lebrijense, spec. nov., fruticosum 1.2- 2.1m. altum; ramis juventate hexagonis brevissime sordideque in- curvo-puberulis maturitate subteretibus leviter costulatis glabratis; foliis oppcsitis petiolatis rhombeo-ovatis acuminatis grosse crenato- serratis basi acutis in petiolum decurrentibus utrinque glabris viridi- bus paullo supra basin 3-nervatis pellucido-stictis et -lineolatis firmiter membranaceis usque ad 12 cm. longis et 6 cm. latis; petiolo 2-3 em. longo; panicula (juvenili) late subpyramidato-ccnvexis 10 cm. alta et 16 cm. crassa; capitulis 15-20-floris 8 mm. longis ad apices ramulorum paniculae glomeratis subsessilibus; involucri squamis 4—5-seriatim imbricatis albidis vel fuscescentibrs oblengis ciliolatis apice rotun- datis 3-5-viridi-nervatis; corollis pallide violaceis gracilibus 3.5-4 mm. longis paullo sursum ampliatis; dentibus limbi_ brevissimis; achaeniis 2 mm. Icngis nigrescentibus (angulis pallidis) sublaevibus; pappi setis 30-35 albis laevibus lucidulis corollam aequantibus.— Cotoms1a: Dept. Santander: rocky hillside, valley of the upper Rio Lebrija, northwest of Bucaramanga, alt. 400-700 m., Dec. 29, 1926, Killip & Smith, no. 16,322 (ryPE, in Gray Herb.). In some respects this plant is clese to E. leucolithogenum, described below, but has subsessile instead of long-pedicelled heads. Further- more, its leaves are acute at base and in transmitted light are dis- tinctly though somewhat irregularly marked with a dot-and-dash punctation. Like E. leucolithogenum it presents a transition between § Cylindrocephala and § Subimbricata. E. (§ CyLinprocepnata) LEPTocEPHALUM DC, Prod. v. 148 (1836). This species varies much in the amount of its pubescence. The differences in this respect are sufficiently conspicuous to warrant the following recognition: _ Var. a. typicum, glabriusculum; foliis subtus glabris vel sparse breviterque pubentibus crebre conspicueque puncticulatis.—L. lepto- _ eephalum DC. 1. c.—Colombia to Peru. Var. 8. hypomalac var. nov., foliis subtus griseo-tomentellis, iculatione omnino obscurata.—E. leptocephalum Sch.-Bip. RECORDS OF THE EUPATORIEAE,—VII 25 Bull. Soc. Bot. Fr. xii. 81 (1865), & Linnaea, xxxiv. 535 (1865-66), not exactly DC.—Prru: Urubamba Valley at Cedrobamba, 2200 m.., Aug. 1927, Prof. F. L. Herrera, no. 1572 (type, in Gray Herb.). Bortvia: Prov. Larecaja: in woods, Queliguaya, vicinity of Sorata, in temperate region, alt. 2700 m., Sept. 1858, Mandon, no. 247 (Univ. Gen., Bot. Gard. Gen., Par., small fragm. Gr.). E. (§ Cylindrocephala) leucolithogenum, spec. nov., fruticosum 6-9 dm. altum; caule succulento tereti fusco-viridi glabro albo- medulloso; ramis recurvantibus usque ad 2 m. vel ultra longitudine; foliis oppesitis graciliter petiolatis ovatis acutis subremote mucronu- lato-denticulatis vel serrulatis vel leviter sinuato-dentatis subpin- natim 3—5-nervatis prope basin utrinque glabris viridibus membrana- ceis opacis vel subpellucide venulosis usque ad 15 em. longis et 10 em. latis; petiolo 2.5-3.3 em. longo; panicula terminali pyramidata 2 dm. alta et crassa; pedicellis oppositis usque ad 1.5 cm. longis paucibracteolatis incurvo-puberulis; capitulis racemose dispositis ca. 18-floris 1 em. longis; involucri squamis ca. 24 ca. 5-seriatim imbricatis apicem versus angustatis sed obtusis ciliolatis albis sed conspicue viridi-trinervatis; corollis viridi-albis cum limbo roseo graciliter tubulatis vix sursum ampliatis sine faucibus distinctis; limbi dentibus 5 anguste deltoideis ca. 0.3 mm. longis; styli ramis filiformibus; achaeniis immaturis 2.5 mm. longis glabris; pappi setis ca. 35 albis laevibus——Cotoms1a: Dept. Santander: in dense woods on the western slope of Mount Pena Blanca, near Charta, alt. 2500—- 2600 m., Feb. 11, 1927, Killip & Smith, no. 19,281 (ryPE, in Gray Herb.); also Rio Suraté Valley above Surata, alt. 2000-2300 m., Jan. 5-6, 1927, Killip & Smith, no. 16,657 (Gr.). -In some respects this species approaches E. chrysostictum Robinson and E. diaphanophlebium Robinson, but it differs from both in its much looser involucre, which, though multiseriate, exhibits in its open also in the lanceolate- angulatis; internodiis plerisque 1. cm. longis; pilis sordidis attenuatis articulatis; foliis oppositis breviter petiolatis oblongis i : tis vel subacutis viridibus sparse pilosis planiusculis glandulari-atomiferis subtus paullo pallidioribus in costa sordide setoso-villosis tenuiter prominulenterque reticulato- 26 ROBINSON venulosis 6-8.5 cm. longis 2-4 cm. latis; petiolo dense sordideque setoso-piloso 5-13 mm. longo; corymbis terminalibus densis multi- capitulatis convexis 7-10 cm. diametro plus minusve viscosis; pedicel- lis plerisque 2-4 mm. longis gracilibus; capitulis 5-floris ca. 8 mm. longis; involucri squamis ca. 10 valde inaequalibus, extimis lanceo- latis acutis, interioribus lineari-oblongis obtusis ciliatis medio striato- nervatis purpureo-brunnescentibus margine tenuioribus _ pallidis; corollis glabris; tubo proprio ca. 1.5 mm. longo; faucibus cylindratis ca. 3.5 mm. longis albis; dentibus limbi purpureis; styli ramis albis; achaeniis (juvenilibus) 3 mm. longis; pappi setis ca. 35 inaequalibus, longioribus corollam subaequantibus.—CoLomsB1a: Dept. Norte de Santander: edge of woods between Mutiscua and Pamplona, alt. 3400 m., Feb. 23, 1927, Killip & Smith, no. 19,739 (TyPE, in Gray Herb.). Also the same as a lower shrub (3-9 dm. high), Péramo de Romeral, alt. 3800-4200 m., Jan. 30, 1927, Killip & Smith, no. 18,655 (Gr.). A species which in many respects approaches E. baccharoides HBK. and E. arbutifolium Benth. but differs from both in the contour, toothing and membranaceous texture of the leaves, the copious indument of its branches, and in its slightly fewer-flowered heads. E. neruFoLIuM Robinson, Contrib. Gray Herb. Ixv. 49 (1922). The range of this species, hitherto known only from northern Vene- zuela, may now be extended on the basis of the following collection: CotomBia: Dept. Norte de Santander: on open hillside between Pamplona and La Isla, alt. 2000-2500 m., Killip & Smith, no. 19,799 (Gr.). The Colombian material corresponds well in all important details with the Venezuelan, but its leaves tend to be larger, becoming in some cases as much as 14.5 cm. long and 4.5 cm. wide. The flowers are reported as being greenish white. It should be mentioned that the occurrence of this broad-leaved form of E. neriifoliwm tends to reduce the differences by which the later E. leucomyelum Robinson of western Venezuela was separated from this species. However, E. leucomyelum differed not merely in the size of its leaves but in several other respects, since it has a somewhat flatter and denser inflorescence and essentially sessile leaves, the blade being decurrent _ (at least narrowly so) on the petioliform base almost or quite to the stem, which is not the case in E. neriifolium. E. (§ CytinprocerHata) Orpicnyanum Klatt, Abh. naturf. Gesell. z. Halle, xv. 324 (1882). This species has thus far been a problem. It was described from material collected by d’Orbigny in ___ Corrientes cong has been variously interpreted by Hieronymus, : Chodat, Hassler and others. Last summer the writer found in the - RECORDS OF THE EUPATORIEAE,—VII 27 herbarium at the Botanic Garden of Geneva an unnamed isotype of the species and was permitted most kindly to borrow it for more detailed study in Cambridge. When carefully examined and com- pared it was found to be precisely the plant later described from Paraguay by Morong as E. densiflorum. This identity is from geo- graphic considerations by no means surprising. Furthermore, Morong’s description mentions no character inconsistent with the earlier diagnosis of Klatt except in stating that the heads are about 20-flowered instead of 12-flowered as described by Klatt. However, dissection of a head from d’Orbigny’s Corrientes material showed it to have precisely 20 florets and indeed the number of florets in the head of a Eupatorium is in general subject to considerable fluctuatio even upon the same individual. Hassler in Fedde, Rep. Spec. Nov. xiv. 281 (1916) reduces E. densiflorum Morong and several other species to varieties of E. oxylepis DC., but this seems unwarranted by any conclusive intergradation between the Paraguayan plant with lanceolate to lance-linear acute, obscurely puberulent to sub- glabrous leaves and shorter heads with blunter phyllaries, on the one hand, and the Brazilian plant with ovate-elliptic very obtuse, shorter, and commonly velvety leaves, as well as longer heads and sharper phyllaries, on the other. Dr. Hassler, however, calls atten- tion to a smooth or nearly smooth form of the Paraguayan plant. This smoother plant under the new interpretation of specific identities may be called: E. Orbignyanum forma glabratum (Hassl.), comb. nov. (E. oaylepis var. densiflorum forma glabratum Hassl. |. ¢.). E.(§C linium) ornatilobum, spec. nov., herbaceum perenne humile 8-23 cm. altum; caulibus 1-6 vel suberectis vel conspicue decumbentibus gracilibus subteretibus glandulari-puberulis vel rarius villosulis infra foliosissimis supra nudiusculis; foliis oppositis (supre- mis paucis saepe alternis) in axillis saepe proliferis vel lanceolatis et crenatis vel rhombeo-ovatis et profunde crenato-lobatis utroque acutis supra strigillosis vel sparse breviterque hirtellis subtus glabrius- culis vel inconspicue hirtellis vel praecipue in nervis pilosis 9-22 mm. longis 3-14 mm. latis; crenaturis vel lobis saepissime utroque latere 2(-4); petiolo plerumque 2-7 mm. longo; corymbis terminalibus valde convexis vel etiam subglobosis 2-3(-5) cm. diametro; pedicellis mm. longis 6 mm. diametro; involucri subturbinato-campanulati tubulatis paullo sursum ampliatis 3 mm. longis; faucibus plerumque vix distinctis; tubo albo vel viridi-albo glabro; lobis 5 deltoideis lavandulaceis vel roseis dorso ornatissime pappilloso-puberulis ca. 28 ROBINSON 0.4 mm. longis; styli ramis clavatis corollae limbo similiter coloratis; achaeniis nigris glabris deorsum decrescentibus 1.8-2 mm. longis; pappi setis ca. 25 albis corollam aequantibus.—CoLomBia: Dept. Santander: edge of Paéramo de las Vegas, alt. 3300-3700 m., Dec. 20-21, 1926, Killip & Smith, nos. 15,687 (rypE, in Gray Herb.), 15,744 (Gr.); Péramo de Mogotocoro, near Vetas, alt. 3700-3800 m., Jan. 18, 1927, Killip & Smith, no. 17,606 (Gr.); western slope of Paéramo Rico, alt. 3300-3600 m., Jan. 15-19, 1927, Killip & Smith, nos. 17,720 (Gr.), 17,740 (Gr.). Colombia without locality: Purdie (K.). This species, being one of several rather similar paramo types. of § Conoclinium, seems most nearly related to EF. humile (Benth.) Hieron., which, however, has oval or elliptic, merely crenate and not lobed leaves, obtuse or rounded at each end, and much more coarsely hairy. There are also other differences. For instance /. ernatilobum tends to be more leafy-stemmed and to be somewhat nigrescent in drying. E. § Cylindrocephala) perglabrum, spec. nov., fruticosum vel arborescens glabrum I-5 m. altum; ramis primo angvlatis tardius subteretibus folicsis; internodiis plerisque 1-3.5 cm. | ngis; foliis oppositis breviter petiolatis lanceolatis utroque attenuatis vel acumi- natis utroque latere crenato-serrulatis apice basique integris utrinque viridibus glabris impunctatis supra basin 3-nervatis minute pellucide- que reticulato-venulosis 5-10 em. longis 12-28 mm. latis; petiolo 7-14 mm. longo; corymbo convexo composito 7-14 em. diametro basin versus foliato-bracteato multicapitulato; ramis ramulisque angulatis et saepe compressis; capitulis 5-floris ca. 12 mm. longis 3-4 mm. crassis in glomerulos 3—5-capitulatos arcte sessilibus; in- volucri squamis 14 ca. 4-seriatim imbricatis valde inaequalibus pallide viridibus saepe purpureo-marginatis dorso glabris plerumque 3-nervatis; corollis viridescente vel rubescente albis a basi ad limbum paullo gradatimque ampliatis 5 mm. longis sine faucibus distinctis; dentibus limbi 5 deltoideis 0.5 mm. longis; antheris apice distincte appendiculatis; achaeniis gracilibus 4.5 mm. longis in angulis paullo scabrato-hirtellis; pappi setis ca. 30 ca. 4 mm. longis albis sublaevibus. —Cotompta: Dept. Santander: edge of woods, Rio Suraté valley, above Suraté, alt. 2000-2300 m., Jan. 5-6, 1927, Killip & Smith, rr 16,550 (Type, in Gray Herb.); open hillside, vicinity of California, t. 3000 m., Jan. 11-27, 1927, Killip & Smith, no. 16,929 (Gr.); : thickets, vicinity of Charta, alt. 2000-2600 m., Killip & Smith, nos. Cae (Gr.); 18,918 (Gr.) and 19,237 (Gr.). A neat and very smooth shrubby or arborescent species with a copious acre green foliage and ample corymbs of slender sessile RECORDS OF THE EUPATORIEAE,—VII 29 heads, apparently frequent in Santander, but not yet seen from else- where. In many respects E. perglabrum is similar to E. leptocephalum DC. but differs in its completely glabrous, much more strongly costate-angulate branchlets and pedicels, and also in its more finely and continuously crenate-serrulate (instead of rather sharply and remotely toothed) impunctate leaves. (§ SuBrmpricaTa) Streetz Robinson, Proc. Am. Acad. lv. 36 (1919). This species has been known thus far only from northern Venezuela about Caracas and Colonia Tovar. However, in the rich. Colombian collection recently made by Messrs. Killip & Smith there is a plant which appears at once conspecific and varietally distin- guishable from the Venezuelan form, the differential traits being as follows: Var. x. typicum, foliis basi rotundatis; involucri squamis ciliatis sed margine excepta vix pubescentibus.—Northern Venezuela as stated. Var. 8. columbiense, var. nov., foliis basi acutiusculis vel acutis vel etiam breviter acuminatis; involucri squamis non solum copiose ciliatis sed etiam saltim apicem versus dorso_villoso-hirtellis.— CoLomBia: Dept. Santander: vicinity of Las Vegas, alt. 2600-3000 m., Dec. 21-23, 1926, Killip & Smith, no. 15,933 (Type, in Gray Herb.). E. (§§ ConociiniuM) sTorEcHADIFOLIUM L. f., Suppl. 355 (1781); J. E, Sm. Ie. iii. t. 69 (1791); Robinson, Proc. Am. Acad. liv. 324 (1918). The material of this species collected by Messrs. Killip & Smith falls into two well marked varieties, as follows: Var. x. typicum, foliis oblongis plerumque obtusis crebre crenu- latis conspicue discoloribus supra viridibus plus minusve griseo- lanulatis subtus albo-lanatis; involucro basin versus albo-lanato.— Type collected by Mutis (herb. Linn. Soc., phot. Gr.). This variety was collected by Messrs Killip & Smith at various points in Dept. Santander, nos. 16,883, 17,243, 18,184, and 18,717 (all in Gr.). Var. 8. pamplonense, var. nov., foliis oblongo-lanceolatis plerum- que acutis integerrimis vel paullo undulatis margine subrevolutis multo minus bicoloribus quam apud var. typicum supra subglabratis viridibus subtus obscure viridibus maturitate vix griseo-lanulatis; involucro viridi basin versus arachnoideo-hirto—CotomBia: Dept. Norte de Santander: open hillside between Pamplona and La Isla, alt. 2000-2500 m., Feb. 27, 1927, Killip & Smith, no. 19,795 (TyPE, in er Herb.). ; E. (§ Cylindrocephala) suratense, spec. nov., fruticosum 1 8 2.4 m. altum; caule basin versus usque ad 1.2 em. diametro apicem 30 ROBINSON versus (post Daticatanieda glanduloso-hirtello viridi flexuoso; ramis ap eae ndentibus; foliis oppositis breviter petiolatis lanceolato-ovatis acuminatis breviter remoteque serrulatis vel in- tegriusculis basi acutis a basi 3-nervatis utrinque laete viridibus supra in nervis hirtellis subtus in nervis venisque tenuiter patenterque pilosis et in superficie fusco-punctatis usque ad 10 em. longis et 3 em. latis membranaceis; petiolo gracili patenter glandulari-hirto ca. 1 em. longo; corymbis terminalibus amplis compositis usque ad 3 dm, -diametro foliaceo-bracteatis; ramulis inflorescentiae ultimis plerisque 1-5 em. longis nudis gracilibus glandulari-hirtellis apice 3(-5)- capitulatis; capitulis ca. 12-floris 11.5 mm. longis 5 mm. diametro plerisque sessilibus vel subsessilibus ternis vel spicatim quinis dis- positis; involucri squamis ca. 20 ca. 4—5-seriatim imbricatis albidis (1-)3-viridi-nervatis apicem rotundatum versus viridibus_hirtellis margine eroso-ciliolatis; corollis albis; tubo paullo sursum ampliato glabro 5 mm. longo; dentibus limbi 5 deltoideis 0.5 mm. longis; achaeniis 4 mm. longis gracilibus in angulis minute scabratis; pappl setis ca. 40 albis sublaevibus—Cotomp1a: Dept. Santander: valley of the Rio Suraté above Suratd, alt. 2000-2300 m., Jan. 5-6, 1927, Killip & Smith, no. 16,533 (Tyee, in Gray Herb.). A highly char- acteristic species of neat appearance, in habit somewhat resembling E. perglabrum described above, but with pubescent foliage and larger more numerously flowered heads, which are sometimes spicately disposed at the ends of the branches in the much more open corymbi- form inflorescence. There may be associated with E. suratense somewhat doubtfully a shade variety, the distinctions between it and the typical variety being as follows: Var. a. typicum, foliis lanceolato-ovatis flavido-viridibus subtus fusco-punctatis; ramulis pedicellisque patenter pilosulis; capitulis plerisque ad apices ramulorum inflorescentiae ternis vel quinis; corollae limbo pallide roseo.—Hab. and excicc. as above cited. Var. 8. aphanostictum, var. nov., foliis ovatis viridibus subtus rarius binis vel trinis; corollae limbo albo.—Cotomsta: Dept. Sant- ander: thickets along river, vicinity of Suraté, alt. about 1700 m., iyo 4-10, 1927, Killip & Smith, no. 16,468 (Gr.). E. (§§ Cylindrocephala) Thurnii, spec. nov., fruticosum ut videtur _ ereetum scabrido-puberulum subrigidum; sabi subtereti fusco- = 5 mm. diametro medulloso dense breviterque puberulo et . atomifero; internodiis usque ad 5 em. longis; foliis oppositis breviter RECORDS OF THE EUPATORIEAE,—VII 31 petiolatis vel saepe subsessilibus lanceolatis integris vel undulato- crenatis acutis basi cuneatis a basi 3-nervatis supra lucidulis aliquid bullatis et rugosis in nervis venisque brunnescenti-hirtellis aliter glabris subtus vix pallidioribus prominulenter reticulatis breviter brunnescenti-pubescentibus plerisque 4-6 cm. longis 1-2 cm. latis textura firmis crassis rigidiusculis; panicula subcylindrica usque ad 4 dm. longa et 13 cm. crassa; ramis paniculae oppositis foliaceo- bracteatis curvato-adscendentibus ad apicem corymbiferis; corymbis valde convexis vel planiusculis ca. 4-6 cm. diametro plerisque 6-20- capitulatis densiusculis; pedicellis plerisque 3-8 mm. longis; capitulis ca. 25-floris maturitate 15 mm. longis et 7 mm. diametro; involucri squamis ca. 30 multiseriatim imbricatis pallidis arcte appressis textura firmis apice rotundis dorso flavescenti-brunnescentibus sub- lucidis 3-nervatis, extimis late ovatis 2.5 mm. longis et 1.5 mm. latis, intermediis elliptico-ovatis vel oblongis 3-5 mm. longis usque ad 2.5 mm. latis, intimis paucis angustissimis 9 mm. longis vix 0.7 mm. latis acutiusculis 1-nervatis; corollis subglabris sparse atomiferis caerulescenti-albis, tubo proprio et faucibus gradatim ampliatis vix distinguendis ca. 6 mm. longis; dentibus limbi anguste deltoideis 0.7 mm. longis; achaeniis gracilibus (angulis paullo scabratis exceptis) glabris (immaturis) ca. 4 mm. longis; pappi setis ca. 45 tenuiter capillaribus sublaevibus albidis ca. 6 mm. longis.—E. conyzoides var. Oliver, Trans. Linn. Soc. ser. 2, ii. 277 (1887), not Vahl.—Britis Gutana (?): Roraima district alt. about 1650 m., 1884, im Thurn, no. 95 (K.); Roraima on the lower savannah, Jan. 1910, E. Ule, no. 8790 (K.). The latter specimen being the more complete of the two clearly identical ones here cited has been the one chiefly used in drawing up the character and should therefore be regarded as the type, not- withstanding the fact that the species is named for Everard F. im Thurn, the original discoverer. a Though this plant was regarded by Oliver as a possible variety of E. conyzoides Vahl (i. e. E. odoratum L.) it clearly differs in many respects from that widely distributed and variable species. Its heads are somewhat larger, its phyllaries of firmer texture, Its leaves are subsessile or much more shortly and stoutly petioled, more rugose and of a somewhat rigid texture, its stem instead of being light in color and sparingly to copiously pubescent is dark brown and densely ; from the conclusion that it constitutes a distinct and endemic species of this highly peculiar region. As the Roraima district is a rather vaguely delimited area oo ROBINSON Venezuela it is by no means certain which of these countries included the station or stations at which im Thurn and Ule collected this plant. E. § Supmericata) Varcasianum DC. Prod. v. 155 (1836) ; Robinson, Proc. Am. Acad. ly. 39 (1919). This species, though reported as occurring in Colombia by Klatt in Engl. Bot. Jahrb. viii. 36 (1887) and by Heer, Mém. Soc. neuchat, Sci. Nat. v. 420 (1913), has heretofore been known to the writer only from Venezuela, where collected originally by Vargas (Conserv. Gen.) near Caracas and subsequently at Sanchorquiz by Eggers, no. 13,413 (Gr.) and at San Pablo de Mendoza, State of Trujillo, by Pittier, no. 10,745 (Gr.). There seemed grounds for doubting the reports from Colombia and it has lately been possible to examine in the herbarium of the British Museum a specimen of Lehmann’s no. 938 from the Central Cordillera of Popayan, the number upon which Klatt based his record. The specimen at the British Museum proved to be unquestionably FE. Stuebelii Hieron. Concerning the identity of the plant determined by Heer as FE. Vargasianum no information can be given, since it has not been found in any of the several large herbaria recently visited by the writer. However, it is now possible to record the Santander: in thicket between Chindcota and La Esmeralda, alt. 1000-1300 m., Mar. 19, 1927, Killip & Smith, no. 20,883 (Gr.). This specimen appears to be an exact match for the one collected at Sanchorquiz by Eggers. It is furthermore from a region where other Venezuelan elements have been found in the flora, which would be quite improbable about Popayan where the species was wrongly reported by Klatt. Mikanta ancuiaris H. & B. Pl. Aequin. ii. 87, t. 106 (1809). This species was based on a plant collected near Gonzanama, Prov. Loja, Ecuador, by Humboldt & Bonpland, no. 3434 (Par., phot. Gr.). Thus far it is known to the writer only from this original material from southern Ecuador and from similar and as it appears conspecific material collected near the Monzon River, Dept. Huanuco, Peru, by Weberbauer, no. 3640 (Brl., fragm. Gr.). It is a plant which is nearly glabrous, as originally described, and has triangular-hastate leaves and white pappus. The species was long ago reported by Bentham, Bot. Sulph. 113 (1844), as having been collected by Hinds at Tumaco, Dept. Narifia, Colombia ; and Baker in Mart. FI. Bras. vi. pt. 2, 245 (1876) cites it as collected in Dutch Guiana by Wull- schlaegel and by Hostmann, no. 164. On the usually excellent author- oe ity of Bentham, though without opportunity to study the plant RECORDS OF THE EUPATORIEAE,—VII 33 of Hinds, the species was included by the writer in his revision of the Mikanias of Colombia, Contrib. Gray Herb. Ixiv. 33 (1922). It was similarly included, on the authority of Baker, by Pulle in his very useful Enumeration of the Vascular Plants of Surinam, 458 1906). While at the Kew Herbarium in the summer of 1927, the writer had a chance to examine the Hinds specimen from Colombia, labelled M. angularis by Bentham and found it by no means this species but a characteristic specimen of M. punctata Klatt. On the sheet Mr. J. G. Baker had written in pencil “ = Wullschlaegel 1494 - Paramaribo,” thus indicating its identity with the plant he was in the Flora Brasiliensis recording as M. angularis from Dutch Guiana. As a result of the here corrected identification of the Hinds plant, M. angularis H. & B. may for the present be deleted from the flora of Colombia. Similarly, although the writer has not yet been able to check either the specimen of Wullschlaegel or Hostmann’s no. 164, it appears probable from Baker’s annotation that they were identical with the plant of Hinds and were therefore M. punctata Klatt. At all events the Guiana representation of M. angularis must be critically re-examined if records based upon it are to be trusted, especially as this species of the Ecuadorian and Peruvian Cordilleras would appear a rather improbable element in the Guiana flora, while, on the other hand, there would be’ considerable likelihood that M. punctata Klatt, a plant long known from Mexico to Venezuela, may extend to similar habitats furnished by the Guianas. M. antiogurensis Hieron. in Engl. Bot. Jahrb. xxviii. 580 (1901). When revising the Mikanias of Colombia some years ago, the author, Contrib. Gray Herb. Ixiv. 35 (1922), reduced this species to the nearly related Brazilian M. lanuginosa DC., interpreting that species widely both as to range and variability. Material subsequently available and a re-examination of the type of M. lanuginosa make it seem probable that the Cordilleran plant of Venezuela, Colombia, Peru and Bolivia, though in many respects similar, may be separated with fair definiteness from the Brazilian plant, which appears to have thinner more flexible and distinctly toothed leaves with more deeply cordate base and more rounded sinus. Its leaves are more strongly bicolorous and the pubescence is less closely matted than in the Cordilleran plant. The latter has thicker firmer subentire leaves less deeply cordate, with less developed basal lobes and closer pubescence. Unless further collecting should disclose tendencies toward intergradation not evident at — these plants would seem best treated as separate species. MM. may be regarded as covering also the plant passing mae the herbarium 34 ROBINSON name M. Karsteniana Klotzsch, a nomen subnudum, mentioned but without adequate diagnosis by Hieronymus, |. c. 581. The slight distinctions of petiole, pubescence and pappus, upon which Hierony- mus relied in attempting a specific distinction between his M. antioquiensis and the undescribed M. Karsteniana, do not hold satisfactorily. To M. antioquiensis may be referred also a presumably Brazilian specimen, namely Ule’s no. 8797, collected on his Amazonas Expedition, in the Roraima region along river banks. M. (§ Thyrsigerae) arthroclada, spec. nov., fruticosa volubilis breviter denseque fulvo-tomentella; caulibus teretibus; internodiis 6 cm. vel ultra longis; ramis flexuosis gracilibus conspicue nodosis; internodiis ramorum plerisque 6-18 mm. longis; foliis oppositis parvis breviter petiolatis ovato-oblongis obtusis crenato-serratis basi subcuneatis integris paullo supra basin pinnatim nervatis supra cum glandulis globosis sessilibus ornatis subtus in nervis brevissime puberulis obscure puncticulatis 10-17 mm. longis 5-9 mm. latis; petiolo 2-5 mm. longo; paniculis ramos terminantibus pyramidatis 3-5 cm. altis et crassis fulvo-tomentellis; capitulis (submaturis) 3.5-4 mm. longis breviter pedicellatis; involucri squamis oblongis apice rotundatis lacerato-ciliatis dorso plus minusve fulvo-hirtellis ca. 3 mm. longis; corollis albis limbo hirtis; achaeniis (valde im- maturis) nigris ca. 0.7 mm. longis; pappi setis ca. 25 albis vix scab- ratis.—Prru: Dept. Cuzco: above Cosnipata between the Tambos Trecruces and Tambomayo, alt. 2500-2600 m., May 7, 1912, Prof. A. Weberbauer, no. 6972 (typE, Field Mus., phot. Gr.). Readily distinguished from other Peruvian species of this genus by its very small leaves and its very short crisp, somewhat stiff or setulose rufous indument. The leaves are early deciduous from the lower parts of the curved branches, which then show conspicuous thick nodes. M. Baprert DC. Prod. v. 194 (1836). This species, thus far recorded only from the Lesser Antilles and Trinidad, is much too closely approached by M. Hookeriana DC. (I. ec. 195) of British Guiana. After an examination of the types and several subsequent collections of both of these species, it has been impossible to discover a single constant difference. It is true, the arrangement of the sessile heads at or near the tips of the panicle-branches is some- what more spicate in its tendency in the plant of British Guiana and terminal and glomerate in the plant of the Antilles, but this trait shows such variability in both plants and even in some cases _ on the same individual that it forms a poor differential character. It seems probable that M. Hookeriana should be reduced to M. RECORDS OF THE EUPATORIEAE,—VII 35 Badieri, which was simultaneously published and has a slight priority of position. It is, in this connection, of interest that a specimen specifically indistinguishable from the plants here discussed has been found in a loan from the Missouri Botanical Garden bearing the following data: Tehuantepec, Oaxaca, Mexico, C. R. Orcutt, no. 3434. Unless some confusion of labeling has here occurred this forms a notable extension of range and adds an interesting and characteristic species to the recorded flora of Mexico. M. Gonzatezit Robinson & Greenman, Proc. Bost. Soc. Nat. Hist. xxix. 107 (1899). This little known plant, hitherto recorded only from Vera Cruz, Mexico, may now be extended in range as follows: Cotomsia: Dept. Santander: vicinity of Tona, alt. 1900-2100 m., in woods, Feb. 17, 1927, Killip & Smith, no. 19,469 (Gr.). The species has rather closely the habit of M. cordifolia (L. f.) Willd. but somewhat larger heads, flattish (uncrisped) corolla-teeth, broader and blunter phyllaries, more conspicuous foliaceous bracts, less prominently angled branches and larger smoother leaves. M. § Racemosae) Hovustonrana (L.) Robinson, Proc. Am. Acad. xlii. 47 (1906). This species was founded on material collected by Dr. Houston in Vera Cruz, Mexico. Happily the original material is still preserved in the British Museum and there is an excellent photograph of it in the Gray Herbarium. It shows a plant with rather broadly ovate leaves, apparently of rather soft probably carne- ous texture, which become thin in drying and show on both surfaces only a rather obscure and loose reticulation of mostly immersed veins and veinlets. Essentially the same plant is now known from Oaxaca, Chiapas, Guatemala, British Honduras, and Costa Rica. In Costa Rica the reticulation of the leaves tends to become a little more prominent. In Colombia near Las Nubes, H. H. Smith collected a plant which it has seemed best to refer to M. Houstoniana, though it has somewhat more narrowly ovate leaves of slightly firmer texture. Nearly related plants of a more puzzling nature have recently been collected in the Departments of Santander and Norte de Santander, having still narrower ovate-oblong rather than ovate leaves of even firmer texture and provided both above and beneath with an exserted reticulation of somewhat double character (i. e. a finer within the - areolae of a coarser). In all other respects, these plants agree rather closely with M. Houstoniana. It is possible that further collecting may show these plants distinct and capable of maintaining the perceptible differences which now separate them. It appears rather more probable, however, that they will be found to merge when the now scarcely known territory between their stations has been more 36 ROBINSON perfectly explored, for their distinctions are chiefly vegetative and of a kind easily bridged. For this reason it seems best, for the present, to give them status as probable varieties of the same species, as follows: Var. «. typica, foliis late ovatis vel ovato-ellipticis subcarneis; venulis immersis vel leviter prominulentibus; capitulis racemosis graciliter pedicellatis——M. Houstoniana (L.) Robinson, 1. e. Var. 6. santandrensis, var. nov., foliis angustius ovatis vel ovato-oblongis firmiusculis subcoriaceis utrinque duplice prominulen- terque reticulatis; capitulis aut racemosis et graciliter pedicellatis aut subspicatis et subsessilibus—Cotompr1a: Dept. Santander: an herbaceous vine, with pale green bracts and yellow-green pappus, on wooded banks of the Rio de la Baja, La Baja, alt. 2200-2300 m Jan. 26, 1927, Killip & Smith, no. 18,290 (typ, Gr.); vicinity of La Baja, alt. 3000 m., Jan. 14-31, 1927, Killip & Smith, 18,129 (Gr.); vicinity of Las Vegas, alt. 2600-3000 m., Killip & Smith, no. 16,127 (Gr.). Dept. Norte de Santander: in dense woods between Pamplona and Toledo, alt. 2800-3000 m., Feb. 27-28, Killip & Smith, no. 19,961 (Gr Of this plant two differing forms are noticeable, as follows: Forma a. normalis, capitulis racemosis graciliter pedicellatis. —Distrib. and exsice. as above stated. Forma b. apodocephala, forma nov., capitulis subspicatis aut sessilibus aut brevissime pedicellatis——CoLtompr1a: Dept. Santander: woody vine in dense forest, Rio Suraté Valley, alt. 2000-2300 m., Jan. 5-6, 1927, Killip & Smith, no. 16,727 (Gr.). This form with subspicate heads naturally approaches § Spiei- formes and particularly resembles M. leiostachya var. irrasa Robinson. However, it is clearly a mere form of the variety just described and like all forms of M. Houstoniana, whether herbaceous or more or less lignescent, has fistulose stems, while M. leiostachya and its variety have stems with a solid pith and a somewhat firmer character, being much less inclined to crush and flatten in drying. M. (§ Spictrormes) Kixurem Robinson, Soatiie Gray Herb. Ixxvii. 53 (1926). Hieronymus in Engl. Bot. Jahrb. xix. 48 (1894) recorded M. psilostachya DC. near see in the Dept. El Cauca, on the basis of Lehmann’s no. 7495. When revising the Mikanias of Colombia, the writer, 1. c. Ixiv. 30 (1922), alluded to this same number of which he had seen a single sheet in the herbarium of the New York Botanical Garden and explained that it seemed an unusually smoot form of the species and that the New York sheet bore quite different data, indicating that the plant had come not from El Cauca but RECORDS OF THE EUPATORIEAE,—VII 37 from Dept. Cundinamarca. Subsequently he has been able to examine three excellent sheets of Lehmann’s no. 7495, two of them belonging to the Kew Herbarium and one recently received in ex- change at the Gray Herbarium. «The labels of these all agree with the data cited by Hieronymus, but the plant is clearly quite distinct from M. psilostachya DC. and is in fact M. Killipii Robinson, the type of which came also from Dept. El Cauca. M. LinpLeyana DC. Prod. v. 195 (1836). This species was based on a specimen from Guiana, belonging in the Lindley herbarium. In recent years the writer has examined four specimens which appear to be referable to this species and which considerably extend its recorded range. They are as follows: (1) Forest of Plantation, Spieringshoek, Surinam, Samuels, no. 38 (Gr.), bringing the plant into the flora of Dutch Guiana, though not listed by Pulle. (2) A portion of the material representing Fendler’s no. 2348 in the Kew Herbarium (where mixed with M. gracilipes Robinson), labeled as from near Colonia Tovar, Venezuela. (3) Climbing high on trees in wet woods, Frijoli Station of the Panama Railroad, Feb. 6, 1862, Sutton Hayes, no. 490 (K., BM.). (4) A whitish-flowered plant, at Jurué Miry, Rio Jurué, Amazonas, Brazil, July, 1901, E. Ule, no. 5679 (K.). It seems rather remarkable that a species of fairly characteristic habit, ranging from the Guianas to Panama and north- western Brazil, should not have been reported more frequently. Being a climber of low wet woodlands, it is probably rather difficult to reach and to collect. M. (§ Corymbosae) lutescens, spec. nov., fruticosa volubilis; caule gracili subtereti glabro post exsiccationem flavido-brunneo; dentatis vel -lobulatis vel trifidis basi sinu patenti cordatis a basi 5-7-nervatis tenuibus supra in nervis puberulis subtus parce setoso- villosis utrinque viridibus concoloribus 3-6 cm. longis 1.5-3 cm. latis; dentibus mucronulatis; petiolis 1-2 cm. longis, eis ejusdem jugi basi per lineam crassiusculam plus minusve setoso-ciliatam trans- junctis; corymbis compositis ‘lias busca, ramis inflorescentiae ultimis plerumque 3-7-capitulatis; pedicellis 1.5-3 mm. longis; bracteolis herbaceis ovatis integris ad vel versus summum pedicelli affixis; capitulis ca. 11 mm. longis; involucri squamis lineari- lanceolatis acutis ca. 8 mm. longis glabris flavescenti-viridibus basi angustatis incrassatisque; corollis lutescentibus glabris; tubo proprio 2.7 mm. longo gracili; faucibus subcylindrato-campanulatis 2.5 mm. longis; dentibus limbi lanceolatis acutis 1.8 mm. longis; achaenlis 38 ROBINSON glabris 3 mm. longis; pappi setis ca. 60 albis laevibus.—Perrv: Dept. Ayacucho: Prov. Huanta: twining shrub with pale yellow flowers, in evergreen bush-wood, Choimacota Valley, alt. 2900 m., Feb. 28- Mar. 10, 1926, Prof. A. Weberbauer, no. 7574 (TPE, Field Mus., ISOTYPE, Gr.). Probably related to M. brachyphylla Hieron., which, however, has much smaller, thicker, more densely puberulent, less toothed leaves, shorter, broader, more obtuse and more pubescent phyllaries, and many other differing features. M. (§ Spiciformes) ochroleuca, spec. nov., volubilis gracilis ut videtur herbacea; caule tereti primo crispe minuteque puberulo tardius laeviusculo fistuloso; internodiis usque ad 16 cm. longis; foliis oppositis graciliter petiolatis deltoideo-ovatis profunde hastatis — utrinque viridibus glabris tenerrimis 7-8 cm. longis 5-6.5 em. latis ima a basi 3(—5)-nervatis margine undulatis; lobis sadieditien acutis vel subacutis; petiolis flexuosis 3-3.5 cm. longis, eis ejusdem jugi a linea puberula vix incrassata transversa conjunctis; paniculis subdensis ovoideis terminalibus et lateralibus 7-13 cm. altis 6-10 cm. diametro puberulis, ramis paniculae subdivaricatis basi iterum ramosis; ramulis ultimis 1-2 cm. longis spiciformibus; capitulis sessilibus vel breviter pedicellatis 7 mm. longis; bracteolis parvis lanceolato-linearibus 1-1.5 mm. longis; involucri squamis viridibus oblanceolato-oblongis acutis basi paullo incrassatis sub- glabris 4.5-5 mm. longis; corollis albo-luteis; tubo proprio gracili glabro ca. 3 mm. longo; faucibus vix ullis; limbi dentibus oblongo- ovatis obtusis patentibus ca. 1 mm. longis; achaeniis nigris 2.2 mm. longis; pappi setis 30-40 carneis corollam aequantibus.—BraZIL: a climbing plant with whitish-yellow flowers, Seringal S. Francisco, Rio Acre (coll. on Amazonas Exped.), July 1911, E. Ule, no. 9886 (K., phot. and small fragm. Gr.). In many respects similar to M. hastifolia Bak. (known to the writer only from the original descrip- tion), but having the spiciform branches much shorter and tending to branch toward the base, the heads larger, and the achenes more an twice as long. M. (§ Thyrsigerae) pachydictya, spec. nov., fruticosa volubilis inflorescentiis puberulis exceptis glaberrima; ramis —* flexuosis -acuminatis integris basi rotundatis firmiter coriaceis pallide viridibus concoloribus glabris prope basin pinnatim 5-nervatis supra grosse RECORDS OF THE EUPATORIEAE,—VII 39 panicula terminali laxa composita foliaceo-bracteata crispe sordideque hirtella; ramis ramulisque divaricatis; ramulis ultimis plerisque 3-(spicatim)5-capitulatis; capitulis 1 cm. longis sessilibus; bracteolis m. longis subulatis; involucri squamis anguste oblongis 5-5.5 mm. longis apice rotundatis ciliatis dorso brunnescentibus vix nervatis apicem versus subhirtellis aliter glabris; corollae tubo proprio gracili 1.5-2 mm. longo glandulari-atomifero; faucibus campanulato- subcylindratis ca. 2 mm. longis; dentibus limbi deltoideis 0.5 mm. longis; achaeniis gracilibus 4.5 mm. longis atrobrunneis in faciebus paullo granulato-puberulis; pappi setis ca. 40 brunnescentibus corol- lam aequantibus.—CotomB1a: Dept. Norte.de Santander: edge of woods, eastern slope of Péramo del Hatico between Toledo and Pamp- lona, alt. 2200 m., Mar. 12, 1927, Killip & Smith, no. 20, 564 (TYPE, in Gray Herb.). ; This species in its inflorescence and some other features approaches M. clematiflora Rusby, but differs in having much thicker, firmer and very strongly reticulate leaves, shorter petioles, puberulent inflores- cence, and somewhat smaller heads. M. (§ Racemosae) roraimensis, spec. nov., fruticosa volubilis ubique glaberrima; caule tereti fistuloso post exsiccationem pur- purascenti-brunneo costulato; internodiis 3-5 cm. longis; foliis oppositis breviter petiolatis obovatis vel oblanceolato-spatulatis breviter subacuminatis (apice vero obtuso) integerrimis basin versus attenuatis paullo supra basin trinervatis concoloribus crassiusculis subcoriaceis obsolete puncticulatis 5-6 cm. longis 1.8-2.4 cm. latis; petiolo ca. 5 mm. longo; racemis firmiusculis terminalibus et opposite in axillis superioribus dispositis pedunculatis 4—6 cm. longis; pedicellis 3-6 mm. longis patentibus vel saepissime paullo recurvatis; bracteolis linearibus pedicellos basi subtendentibus et subaequantibus; capitulis ca. 1 cm. longis nutantibus; involucri squamis anguste oblongis apice -rotundatis; corollis albidis; tubo proprio gracili 1.8 mm. longo; faucibus campanulatis ca. 1 mm. altis; dentibus limbi deltoideis ca. 0.5 mm. longis; achaeniis gracilibus glabris 4 mm. longis; pappi setis 40-50 brunnescenti-albidis corollam aequantibus.—Braziu: in Ror- aima, Jan. 1910, E. Ule, no. 8798 (K., phot. and small fragm. Gr.). This species agrees in many points with the description of M. Sprucei Bak. in Mart. Fl. Bras. vi. pt. 2, 264 (1876), but Baker states that his species had the heads all sessile, with bractlets at the base of the involucre, the corolla-limb equal to the tube, the achenes only 1 line (= 2 mm.) long, the petioles 4-6 lines (= 8-12 mm.) long, and the leaves more than half again as large as those shown in the plant here described. With these very definitely stated differences it 40 ROBINSON seems highly improbable that the two plants can prove conspecific though they must have some habital similarity. M. Sprucei came from San Carlos on the Rio Negro, a locality in southern Venezuela, probably over 600 km. from the Roraima region where Ule collected the plant here in question. M. (§ Racemosar) scapra DC. Prod. v. 190 (1836). The range of this species, thus far recorded from Brazil, Guiana, and Trinidad, as stated by Urban, Symb. Ant. v. 218 (1907), may now be extended upon the basis of the following specimens: (1) VENEZUELA: very high climber, near stream, Ciudad Bolivar, L. H. & E. Z. Bailey, no. 1356 (Gr.). (2) Panama: shrubby, climbing over shrubs and trees near the City of Panama, Aug. 1862, Sutton Hayes, no. 677 (K.). M. (§ Thyrsigerae) Smithii, spec. nov., fruticosa volubilis; caule tereti costulato brunneo plus minusve villoso vel puberulo vel glabrius- culo albo-medulloso; internodiis 5-15 em. longis; foliis oppositis petiolatis ovato-oblongis vel ovatis acuminatis integris vel obsolete sparsissimeque cuspidato-denticulatis basi rotundatis pinnatim 5- nervatis utrinque viridibus supra lucidulis glaberrimis vel paullo setulosis subtus obscuris leviter duplicato-reticulatis praecipue in nervis hirtellis 5-9 em. longis 3-4.5 cm. latis; petiolis gracilibus 1-2 cm. longis plus minusve setuloso-villosis; paniculis terminalibus ovoideis compositis foliaceo-bracteatis pubescenti-hirtellis; capitulis ca. 8 mm. longis pedicellatis; pedicellis filiformibus plerisque 3-5 mm. longis; bracteola plerumque prope apicem pedicelli gesta; involucri squamis laete viridibus anguste oblongis tenuibus subglabris apice rotundatis basi gibbosis; corollis albis vel flavidi-albis ; tubo proprio gracili glabro 1.5 mm. longo; faucibus turbinati-campanulatis 1.5 mm. longis; dentibus limbi acutis ca. 1 mm. longis; achaeniis gracilibus 3 mm. longis obscure hirtellis; pappi setis ca. 30 albis vel albidis corollam subaequantibus.—Cotompra: Dept. Santander: Mesa de los Santos, alt. 1500 m., Dee. 11-15, 1926, Killip & Smith, nos. 15,200 (ryPE, in Gray Herb.), 15,118 (Gr.), 15,429 (Gr.). The last mentioned number shows in many places especially along the leaf- nerves, on the pedicels and phyllaries, etc., felt-like growths of the sort produced by eriophyid mites. Traced by the author’s key to the Colombian Mikanias (Contrib. Gray Herb. Ixiv. 32) this species falls near M. Ruiziana Poepp. and M. lanuginosa DC., from both of which it is to be readily distinguished by its smaller, thinner, greener leaves and lack of the copious velvety or silky indument characteristic of the species mentioned. . M..( Spiciformes) Steinbachii, spec. nov., semivolubilis 3 m. _ alta; caule tereti fistuloso post exsiccationem costulato; internodiis RECORDS OF THE EUPATORIEAE,—VII 41 ca. 1 dm. longis; foliis oppositis breviter petiolatis ovatis acuminatis integris cordatis supra atroviridibus glabris subtus paullo pallidioribus obscuris pinnatim 5-nervatis in venulis reticulatis minute tomentel- lis 10 em. longis 7 cm. latis; paniculis ramos breves paucibracteatos terminantibus ovoideis puberulis ca. 2 dm. longis 10-13 cm. diametro; spicis densis patentibus breviter pedunculatis plerisque 1-3 cm. longis 1.5 cm. crassis; capitulis omnibus sessilibus vel infimis brevissime pedicellatis patentibus 7-8 mm. longis; bracteolis ovatis vel ovato- lanceolatis acutis ca. 1.5 mm. longis; involucri squamis oblongis obtusis stramineo-viridibus vel aetate atrobrunneis glabris lucidulis striatulis 4.5 mm. longis; corollae tubo proprio gracili 1.5 mm. longo; faucibus aequilongis turbinato-campanulatis; dentibus limbi parvis recurvatis ca. 0.5 mm. longis; achaeniis gracilibus nigrescentibus 3 mm. longis; pappi setis ca. 25 primo argenteis aetate carneis corollam aequantibus.—Boutvia: Dept. Santa Cruz: Prov. Cercado: Bafiados del Piray, alt. 450 m., 18 July, 1924, Jose Steinbach, no. 6257 (K., phot. Gr.). M. (§ CorymposakE) TaraALLANA HBK. Nov. Gen. et Spec. iv. 137 (1820). Thus far this species seems to have been recorded only on the basis of the original material, said to have been obtained from the banks of the Daule River near Guayaquil, Ecuador, Tafalla (Par., phot. and fragm. Gr.). It is now possible to report indications of a second station, as follows: Peru: Pillao, Lima, from the herbarium of Ruiz & Pavon (K.). This supposed Peruvian material in the Kew Herbarium consists merely of a single large detached leaf and a fruiting corymb from which the corollas have fallen. While this material agrees closely with the type-material at Paris, even to the precise shade of the unusually reddish-brown pappus, it. must be borne in mind that there was much laxity in the geographic labelling of these early collections. In fact, both the Ecuadorian and the Peruvian stations will have to be subject to some doubt until they can be substantiated by rediscovery and more complete specimens. M. renax Sch.-Bip. Bonplandia, iv. 54 (1856), as a mere nomen nudum. When revising the Mikanias of Peru, the writer, Contrib. Gray Herb. Ixiv. 72-95 (1922), had seen no representation of this unpublished species. Subsequently he has been able to examine and to borrow for more detailed study the isotypes (Lechler’s no. 2355) from the Royal Botanic Gardens at Kew and from the Conservatory of Botany at Geneva. The species appears, in fact, to be a good one, pretty readily distinguishable from any previously described. It may be characterized and given proper standing, as follows: M. (§ Thyrsigerae) tenax Sch.-Bip. (hucusque ined.), volubilis 42 ROBINSON perennis vel verisimiliter fruticosa; caule subtereti post exsiccationem costulato-angulato brunneo juventate puberulo-scabrato maturitate sublaevi; internodiis plerisque 5-10 em. longis; foliis oppositis breviter petiolatis oblongis acuminatis integris basi subacutis utrinque viridi- bus glabris pinnatim 7-nervatis firmiter chartaceis 6-11 cm. longis 3-4 em. latis supra lucidulis subtus vix pallidioribus obscure punctatis; petiolo gracili puberulo ca. 1 em. longo; paniculis terminalibus et lateralibus sordide incurvo-puberulis basi foliaceo-bracteatis; capitulis 6-7 mm. longis ad apices paniculae ramulorum glomeratis sessilibus vel brevissime pedicellatis; involucri squamis oblongis obtusis ciliatis brunneis glabriusculis vix striatis 4 mm. longis 1.5 mm. latis; corollae tubo proprio gracili ca. 1.5 mm. longo; faucibus turbinatis ca. 2 mm. altis; dentibus limbi deltoideis ca. 0.8 mm. longis; achaeniis (immatu- ris) laevibus 2 mm. longis; pappi setis ca. 35 albidis vix scabratis.— PERU: in mountains near San Govan, July 1854, W. Lechler, no. 2355 (K., Gen., phot. and fragm. Gr.). Among the Pucca Mikanias hitherto described M. tenax may well be placed near M. lasiopoda Robinson (which differs in its setose indument, and in its distinctly punctate leaves which are cordate at the narrowed base) and M. caustolepis Robinson (which has more ovate leaves, relatively longer petioles, and peculiar fuscous- puberulent phyllaries). UHNIA ADENOLEPIS Robinson, Proc. Am. Acad. xlvii. 201 (1911). This species, thus far known only from the original material, collected by Pringle in the Chapala Mountains near Guadalajara, has been rediscovered by Mrs. Mexia at San Sebastian, Jalisco. Mrs. Mexia’s specimens, while manifestly conspecific with the type, show some variation, having leaves somewhat shorter, broader, and often sharply few-toothed. Some of the middle and lower leaves are subopposite or occasionally ternate. The pappus (still undiscolored) is bright white. The species is one which forms a transition to Brickellia, in which, with almost equal propriety, it might have been placed.