Ng PHYTOLOGIA An international journal to expedite plant systematic, phytogeographical and ecological publication Vol. 72 March 1992 No. 3 CONTENTS _ NESOM, G.L., Revision of Erigeron sect. Linearifolii (Asteraceae: Astereae). _NESOM, G.L., A new species of Castilleja (Scrophulariaceae) from southcen- tral Texas with comments on other Texas taxa. .............-20005 209 | _-NESOM, G.L., New species and taxonomic evaluations of Mexican Castilleja (Pcrophulaacene; AAU hue ANA Aa 231 LIBRARY JUN 15 1992 NEW YORK BOTANICAL GARDEN Published’ by. Michael J. Warnock 185 Westridge Drive’ Hyntsville, Texas 77340 U.S.A. PHYTOLOGIA is printed on acid free paper. PHY TOLOGIA (ISSN 00319430) is published monthly with two volumes per year by Michael J. Warnock, 185 Westridge Drive, Huntsville, TX 77340. Second Class postage at Huntsville, TX. Copyright ©1991 by PHYTOLOGIA. Annual domestic individual subscription (12 issues): $36.00. Annual domestic institutional subscription (12 issues): $40.00. Foreign and/or airmail postage extra. Single copy sales: Current issue and back issues volume 67 to present, $3.50; Back issues (previous to volume 67), $3.00 (add $.50 per copy postage and handling US [$1.00 per copy foreign]). Back issue sales by volume: $17.00 per volume 42-66 (not all available as complete volumes); $21.00 per volume 67-present; add $2.00 per volume postage US ($4.00 per volume foreign). POSTMASTER: Send address changes to Phytologia, 185 Westridge Drive, Huntsville, TX 77340. Phytologia (March 1992) 72(3):157-208. REVISION OF ERIGERON SECT. LINEARIFOLII (ASTERACEAE: ASTEREAE) Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. U.S.A. ABSTRACT In a taxonomic study of Erigeron foliosus and its close relatives (sect. Linearifolit), fourteen species are recognized: LE. aequtfolius, E. angustatus, E. bioletti, E. blochmantae, E. breweri, E. elmeri, E. folto- sus, E. inornatus, E. mariposanus, E. miser, E. ozyphyllus, E. petrophilus, E. reductus comb. et stat. nov., and E. serpentinus sp. nov. Seven varieties are recognized in E. breweri: var. bisanctus var. nov., var. brewert, var. covillei comb. nov., var. ensenadensis var. nov., var. jacinteus, var. porphyreticus, and var. klamathensis var. nov. Five varieties are recognized in E. foltosus: var. confints, var. foliosus, var. franciscensis var. nov., var. hartwegti, and var. mendocinus comb. et stat. nov. Three varieties are recognized in E. tnornatus: var. ca- lidipetris, var. nov., var. keilii var. nov., and var. inornatus. Three varieties are recognized in E. petrophilus: var. petrophilus, var. viscidu- lus comb. nov., and var. sierrensis var. nov. Most of these have the greatest part of their range in California, from there extending east- ward into Nevada, southward into Baja California, and northward to Washington. Erigeron oryphyllus E. Greene is restricted to southern Arizona and adjacent Sonora. A detailed taxonomic treatment, in- cluding nomenclature and distribution maps, is provided for these taxa. Five other species peripherally related to the E. foltosus group are more briefly treated: Erigeron hyssopifolius of the northeastern United States and adjacent Canada; FE. chiangtt and E. scoparioides from northeast- ern México; and E. rhizomatus and E. lepidopodus from northwestern Mexico and the southwestern United States. KEY WORDS: Erigeron, Astereae, Asteraceae, California The group of taxa including Erigeron foltosus Nutt., E. brewert A. Gray, and E. inornatus (A. Gray) A. Gray has long been known to show complex 157 158 Pane O20 GTA volume 72(3):157-208 March 1992 patterns of variation. Compton (1934) published a short study of the E. foliosus complex in a narrow sense, but the group was treated in full by Cron- quist (1947) in his revision of the 133 North American species known at that time. A partial summary of Cronquist’s view of the taxonomy of this group is seen in his comment (1947, p. 281) that “The EF. foliosus-E. brewer: complex presents some of the most difficult problems in the genus Erigeron. There are a multitude of recognizable forms, often with relatively strong morphological characters, but these interlock and intergrade in such a way that taxonomic segregation becomes difficult, and a completely satisfactory treatment seems impossible.” With respect to the group at hand, Cronquist recognized six species, and his delimitations of the taxa have been essentially the only ones in use since then, because he contributed treatments of Erigeron consistent with the concepts of his revision to virtually all of the major floristic studies of the Pacific coast since 1947. The taxonomy presented here differs considerably from that of Cronquist (1947), and this is the first study to include detailed mapping of the geographic ranges of these taxa (Maps 1-9), which provides evidence significant in direct- ing the taxonomic decisions. All taxa recognized by Cronquist, except one, are recognized in the present study, and apart from eight previously unrecognized taxa, the other differences in treatment in large part reflect the recognition of new taxa and the elevation of varietal taxa to the rank of species. Such shifts have resulted from the observation that, in many cases, the taxa in- volved have discrete morphological and geographical ranges, they are broadly or completely sympatric with putatively conspecific taxa, and intergradation among these taxa is less than previously supposed. As regarded here, a large component of the confusing variation has resulted from previously unrecog- nized entities as well as parallel tendencies of variability among some of the taxa. Even with this seemingly less conservative approach, it is clear, as also seen by Cronquist, that several of the most widespread species are complex and comprise a number of varieties (e.g., as treated here, E. foliosus with five varieties and FE. brewer: with seven varieties). In previous studies (e.g., Nesom 1990), I have tried to be consistent in the application of names at the varietal rank, referring these to closely similar taxa that are geographically contiguous and intergrading at their zone of contact. In the Erigeron foliosus group, however, some of the species comprise sets of allopatric taxa separated by relatively small differences; some of these are intergrading, but the intergradation is more likely the reflection of parallel variability resulting from common ancestry rather than gene flow. A similar pattern of variability has been observed among some of taxa of the E. eatonit A. Gray group (Strother & Ferlatte 1989). As has been the trend in the present study, an alternative treatment might consider some of these allopatric population systems with small but distinctive differences as separate species, requiring higher rank for some of the taxa recognized here as varieties (as Nesom: Revision of Erigeron sect. Linearifolu 159 noted below). The present study treats in detail only the taxa of the Erigeron foltosus group, which are clearly closely related among themselves. Putatively related species (Nesom 1989) that are geographically and morphologically widely pe- ripheral are briefly included, but these need to be investigated in more detail to determine their degree of relationship. Although the major patterns of vari- ation within the E. foliosus group seem to be fairly clear, problems remain in two areas: (1) the definition of some of the taxa (particularly within E. brew- ert and EL. petrophilus E. Greene), for which field work in specific geographic areas will be critical in future studies, and (2) the pattern of phylogenetic re- lationships among the taxa. I have more confidence that the taxa recognized here are real entities in nature than I do that they are related as proposed. With further detailed examination, the taxonomy advanced here will almost certainly be modified. The concepts presented in this study are derived from examination of more than 2000 specimens, representing the whole geographic range of the group, although the taxa are primarily centered in California. Using the keys and distribution maps, identifications can be made with relative confidence. The nature of the characters that delimit the varietal taxa, however, will be easier to interpret with reference specimens for comparison. Chromosome counts have been made for eleven of the taxa included here (as documented below; all recorded here in “gametophytic” form, regardless of how they were reported). All of the taxa counted, except one of the peripheral species, are diploid. The discovery that polyploids were more common would not be unexpected, since they are known to be common in Erigeroneven within single species without taxonomic subdivision. Their presence, however, in whatever pattern, would probably not alter the taxonomic structure presented here. The taxa are treated alphabetically. Authorships for all names mentioned in the text can be found in the formal nomenclature. Designations of geo- graphic provinces in California follow Hickman (1989). Specimens are cited for new or rare taxa and elsewhere when necessary to clarify critical differences with previous taxonomic treatments. Taxa that are more common and more easily recognized are documented by detailed mapping; a single symbol may represent numerous collections for commonly collected taxa. Erigeron sect. Lineartfolii (G. Don) Nesom, Phytologia 67:79. 1989. Erigeron sect. Lineartfolii (G. Don) Nesom, Phytologia 67:79. 1989. BA- SIONYM: Aster sect. Linearifoliti G. Don in Loudon, Hort. Brit. 346. 1830. LECTOTYPE (Sundberg & Jones 1987): Aster gramintfolius Banks ez Pursh (= Erigeron hyssopifolitus Michx.). Sundberg & Jones (1987) lectotypified Aster sect. Lineartfolii with A. gramintfolius. Their 160 PHY TOLOCGTA volume 72(3):157-208 March 1992 choice of a type was apparently intended solely to preserve an established name within the genus Aster. Without an understanding of the taxon- omy of Erigeron, however, they merely transferred the problem from one genus to another (Aster to Erigeron). See comments below regarding the relationship of EF. hyssopifolius to the E. foltosus group). Erigeron sect. Pycnophyllum Crong., Brittonia 6:141. 1947. TYPE: Erigeron foltosus Nutt. (see additional comments following E. hys- sopifolius regarding Erigeron sect. Lineartfolit). Perennial herbs, glandular or eglandular, with or without eglandular vesti- ture, the stems prostrate or sprawling to decumbent ascending or erect, arising from fibrous rooted rhizomes or a woody taproot. Leaves all cauline, filiform to oblanceolate, entire, usually densely arranged and markedly longer than the internodes, not reduced upwards. Heads solitary or few in corymboid clus- ters; phyllaries strongly graduated in 3-5 series, usually with a strongly orange resinous midrib, rarely with 3 parallel veins; receptacles smooth to shallowly foveolate. Ray flowers absent or present in a single series, the corollas white or drying bluish, at maturity often slightly coiling at the tips. Disc corollas numerous, narrowly funnelform, the lower half sometimes slightly indurated; style branches with shallowly triangular collecting appendages ca. 0.1 mm long. Achenes compressed, with 2(-4) strongly orange resinous ribs, the sur- faces usually shiny; pappus mostly of 20-35(-50) bristles, usually with a few, outer setae. Chromosome numbers, n=9, 18. Plants of this section are recognized particularly by the long, ascending, decumbent, or prostrate basal portions of their stems, the narrow (usually), entire, and densely arranged cauline leaves, lack of persistent basal leaves, graduated phyllaries, minute collecting appendages of the disc style branches, and achenes with shiny surfaces and strongly orange resinous ribs. Further, the heads are solitary in some taxa, but they typically are more numerous and occur at the tips of branches arising near the apex of the stems, strongly reminiscent of the capitular arrangement in Heterotheca and Chrysopsis. Sect. Linearifolu has radiated almost exclusively within the boundaries of California; only two of the taxa lie completely outside of the state (one in Arizona and one in northwest Mexico); several extend to the south into Baja California, east into Nevada, and north into Oregon and Washington. KEY TO THE SPECIES OF THE ERIGERON FOLIOSUS GROUP OF SECT. LINEARIFOLII em tay OW OLE A DBO MEE ote tare. sce or seh sie craves crn tig ead eer (9) aay diowers: present. Get .'5 0 a hosee Pe ee ee ee (2) Nesom: Revision of Erigeron sect. Linearifoliu 161 2. Stem vestiture a mixture of eglandular and minutely but distinctly stipitate glandular hairs. ..........-+-seeeeeeeeees E. aequtfolius 2. Stem vestiture various, but without stipitate glandular hairs. .. (3) 3. Stems essentially glabrous to sparsely strigose with short, straight, ascending- appressed, hatte: qos cdranks sisi Fe eee ewiamed «i Pac Ouele wv alee weirs (5) 3. Stem vestiture pubescence short villous or hirsute to hirsutulous, often densely so, with spreading to spreading-deflexed hairs. ............ (4) 4. Stem vestiture puberulous to closely villosulous with crinkly hairs; uchenes: plalrotis.( ianated . makleits eetsle 2 asian /els's alors.» ore E. blochmaniae 4. Stem vestiture densely hirsutulous with straight, stiffly spreading, usually slightly deflexed hairs; achenes strigose. ....... E. breweri 5. Stems with axillary tufts of small leaves at most nodes; phyllaries with prominent, broad, scarious margins. ..........+-..+0+- E. mariposanus 5. Stems without axillary tufts of small leaves; phyllaries without prominent SCALIOUS, TBE RUNSS: te Feel is eee tae orice oa. tee eee clean pee a eine (6) 6. Stems prostrate to decumbent trailing, arising from slender, fibrous rooted jrhimomies? Lee ceri re cere + ce ne Sete a eens E. elmeri 6. Stems basally ascending but essentially erect, arising from a woody taproaty wae ee aa a eee ee ee eels oe Mee eg anes eee ate (7) 7. Middle and upper leaves much shorter than the internodes, eciliate. 5 leak spare Cate eee Aa Mae. Se ene eres aah ores ore E. oryphyllus 7. Middle and upper leaves much longer than the internodes, ciliate with ascending-Appressed Niairs. “Soe. nae Pers esos hee os vice au ve min aelniainn (8) 8. Leaves ciliate with thick based hairs; phyllaries usually with con- spicuous eglandular hairs, rarely glandular; ray flowers 15-50. .... w-Gisb nanan td Qi Rg a Ree ve a ee Ro ee ae RS Ne rer oe ee E. foltosus 8. Leaves ciliate with thin based hairs; phyllaries with barely percepti- ble eglandular hairs, densely granular glandular; ray flowers 9-13. a en aaa oahe eaten ore SURE Tele SO ep STON Paes Soe rede cate E. serpentinus 9. Leaves oblanceolate; stems with long ascending basal portions arising from a woody taproot! \ 215.0 a saree ee oon oa) Soe eens geet (11) 9. Leaves linear; stems erect from the base, arising directly from a multicipital rootcrown, or arising from slender, rhizomelike caudex branches. . (10) 162 Mab ie 13. 13. 15. 15. IPPs ite PHY POdnO'Gt A volume 72(3):157-208 March 1992 10. Stems 30-90 cm tall, basally erect, numerous, arising directly from e strongly woody TOOLCrOWD. ©. s..inn. 0s woe ee ee E. angustatus 10. Stems 8-20(-30) cm tall, basally erect or ascending, usually few and arising singly from thin, rhizomelike, caudex branches. E. reductus Pp lariesreriandiate cope cps. shais akan sini ede eiais e+ ur SE E. wnornatus Phivilarreay planiduler scr. teys pcp » visveis sialaimre steele (12) 12. Eglandular hairs on at least the lower half of the stems spreading. BN. 2h ah Bile Soe ni adv antes ote. ditty aud as scatter (14) 12. Eglandular hairs of the stems sparse and antrorsely appressed or completelymabsent. ccc. .% «ces eh sisrnecicier ose eine ce (13) Upper stems and leaves glandular. ...........:..0cceceeree. E. brolettiu Upper stems and leaves eglandular. ...................0 E. tnornatus tAotems and leaves eglandular. 22... 2.0.66. 0¢ eas sateen eee (15) 14. At least the upper stems and leaves distinctly glandular, sparsely to densely hirsute-villous with long, somewhat crinkly spreading IVES) bons aydye Sty ou aie yee tA choy 5-5 6k a 7m) adch aaro) sy aco, says." aka She Seerab ae ee Stems densely hirsutulous with short, slightly deflexed hairs of even length, or the hairs sometimes loosely antrorsely ascending. .......... SHEE 9S5 LOSRAR RES RECO ne ees E. inornatus var. keili Stems sparsely hispid-pilose with long, stiffly spreading hairs, or the hairs sometimes antrorsely ascending-appressed. E. petrophilus var. viscidulus 16. Stems 30-90 cm tall, erect; leaves 20-40 mm long; heads 12-15 mm UAC Oo pig bess Py pdt alsa corn an a bfuahensionhd ueverintent ik cena E. bioletti 16. Stems 5-20(-30) cm tall, basally ascending or decumbent; leaves (-25 mm long; heade (12mm wide. .....-..-. 2... 0+ «eee (17) Inner, phyllaries’(3.5-)4.0-5-0°mm long... vsi0< oi00+5 o0-s54mqe E. muser Inner phyllaries:ov0-7.0 mm long.) o..200-.2- crease ace cree E. petrophilus Nesom: Revision of Erigeron sect. Linearifolu 163 1. Erigeron aequtfolius H.M. Hall Erigeron aequifolius H.M. Hall, Univ. Calif. Publ. Bot. 6:174. 1915. TYPE: UNITED STATES. California: Tulare Co., ridges at Trout Meadows, Transition Zone, 6200 ft, 16 Jul 1908, H.M. & G.R. Hall 8386 (HOLO- TYPE: UC!; Isotypes: GH!,UC!,US!). Stems ascending-erect, 10-20 cm long, from slender, woody, rhizomelike caudex branches. Stems and leaves minutely granular glandular, sparsely short villous. Leaves narrowly elliptic to oblanceolate, 6-20 mm long, 1.5-3.0 mm wide. Heads 7-10 mm wide; phyllaries minutely glandular, sometimes with a few other nonglandular hairs at the base, inner phyllaries 3.5-4.0 mm long. Ray flowers 14-24(-30), the corollas 5-8 mm long, drying blue. Disc corol- las (2.8-)3.8-4.5 mm long. Achenes 2.2-2.5 mm long; pappus bristles 20-35. Chromosome number, n=9 (Semple 1985). California endemic, southern High Sierra Nevada (Map 1); ledges, crevices, 1560-2010 m; Jul-Aug. Cronquist’s prophecy regarding the potential validity of Erigeron aequifolius (1947) has held true. The species is now known by a number of collections of morphologically consistent plants, none of which show intermediacy with any other species. It probably is most closely related to E. brewert, with which it is completely sympatric, and to FE. elmeri (E. Greene) E. Greene, with which it is partially sympatric. 2. Erigeron angustatus E. Greene Erigeron angustatus E. Greene, Bull. Calif. Acad. Sci. 1:88. 1885. TYPE: UNITED STATES. California: Napa Co., “dry hills on either side of Napa Valley, Jul-Oct” [Napa, 13 Aug 1874], E. Greene 389 (not located; probable duplicate: GH!). Although the type of Z£. angustatus is the same as one of the syntypes of E. inornatus A. Gray var. angustatus A. Gray (= E. reductus [Cronq]. Nesom var. angustatus [|A. Gray] Nesom in the present study), there is no indication that Greene intended his name as a new combination based on Gray’s earlier one. None of the sheets of this taxon at ND-G (cited below), all identified by Greene as E. angustatus, appear to be type material. I have not designated the GH specimen as lectotype, because | believe Greene’s original collection eventually will be located. Stems 3-9 dm tall, erect, arising directly from the crown of a thick taproot. Stems and leaves mostly glabrous, the margins widely and sparsely ciliate. Leaves linear, 1-6 cm long, 0.5-1.0(-2.0) mm wide. Heads 8-12 mm wide; phyllaries granular-glandular, without other vestiture, the inner 5.5-7.5 mm 164 PHYTOLOGIA volume 72(3):157-208 March 1992 ORs TE ENDER . aequifolius oO E. elmeri & E. blochmaniae @® E. angustatus A E. serpentinus Map 1. Erigeron aequifolius, E. elmeri, E. blochmaniae, E. angustatus, and E. serpentinus. Nesom: Revision of Erigeron sect. Linearifoltu 165 long, usually purple at the apex. Ray flowers absent. Disc corollas 4.2-6.0 mm long. Achenes 2.4-2.8 mm long; pappus bristles 26-38. California endemic, southern Outer North Coast Ranges (Map 1); serpen- tine areas, often in loose rocks, usually in brushy vegetation, ca. 90-150 m; May-Sep. Additional collections examined: UNITED STATES. California: Lake Co., no other locality data, 15 May 1893, Blankenship s.n. (UC); Whispering Pines Resort, 30 Jun 1940, Baker 9748 (CAS,UC); foot of “Rabbit Hill,” just SW of center of Middletown, 17 Aug 1971, Hamann & Dearing s.n. (JEPS). Napa Co.: St. Helena, 10 Jul 1891, E.L. Greene s.n. (ND-G 4 sheets); E side of Mt. St. Helena, 24 Aug 1941, Hoover 5572 (JEPS); Sage Canyon, Napa Range, 27 Aug 1933, Howell 11625 (CAS); Soda Creek Canyon, Napa Valley between Napa City and Yountville, 7 Aug 1938, Tracy 16089(DS,LL,UC). Sonoma Co.: Cazadero, side of stream, 24 Jun 1943, Cooke s.n. (UC); Dutch Bill Creek, 5 Sep 1947, Hoffman 1249 (UC); between Occidental and Graton, 26 Jul 1936, Howell 12669 (CAS); 2 mi E of Occidental, 5 Jul 1946, Mason & Grant 12898 (UC). This narrowly endemic species has not been recognized since its original description by Greene, but it is distinctive in morphology and habitat. It was included by Cronquist (1947) with Erigeron inornatus var. angustatus ( E. reductus var. angustatus in the current study), which produces stems that are stolonlike and nearly filiform at the base, with no well developed root crown. Erigeron angustatus also has been confused with LF. biolettu E. Greene, which is eradiate and somewhat similar in habit, though different in vestiture and habitat. The stems and leaves of E. angustatus are glabrous (vs. glandular), the leaves narrower (mostly linear and 0.5-1.0 mm vs. narrowly oblanceolate and 2-4 mm wide), and the stems arise directly and erect from the crown, in contrast to the long, ascending basal portions that are characteristic of E. brolettu. The two taxa are partially sympatric, and no intermediates have been seen in this study. Additional comments are given following E. reductus and E. serpentinus Nesom. 3. Erigeron biolettu E. Greene Erigeron biolettii E. Greene, Man. San Franc. Bay Region 181. 1894. LEC- TOTYPE (Cronquist 1947): UNITED STATES. California: Sonoma Co., base of Hood’s Peak, 8 Jul 1892, F.T. Bioletti s.n.(ND-G!). Erigeron tnornatus (A. Gray) A. Gray var. bioletti (E. Greene) Jeps., Fl. W. Mid. Calif. 569. 1901. Greene also cited a collection from Howell Mt. (Napa Co.) by Jepson. Stems 3-9 dm tall, erect or the basal portions usually long, ascending, and somewhat caudexlike, usually purplish at the base. Stems and leaves 166 PHY HOD OCT A volume 72(3):157-208 March 1992 densely glandular, often also sparsely hispid or hispid-pilose. Leaves narrowly oblanceolate, 2-4 cm long, 2-4(-5) mm wide. Heads 12-15 mm wide; phyllaries densely glandular, the inner 6.0-8.0 mm long, usually purple at the apex. Ray flowers absent. Disc corollas 4.5-6.5 mm long. Achenes 2.0-2.5 mm long; pappus bristles 22-38. California endemic, Outer North Coast Ranges (Map 2); dry slopes, rocks and ledges along rivers, 30-700(-1030) m; Jun-Sep. Representative collections examined: UNITED STATES. California: Hum- boldt Co.: Mattole River at Union Mattole, 4 Aug 1929, Kildale 8777 (DS-2 sheets); 5 mi N of Mendocino-Humboldt county line, S. Fork of Eel River, 9 Aug 1936, Munz 14881 (UC); Humboldt Co. State Park, Eel River, 9 Jul 1923, Petrson 3877 (UC); along S. Fork Eel River, 3-5 mi above its mouth, 23 Jul 1912, Tracy 8797 (UC); near Garberville, along S. Fork Eel River, 9 Sep 1933, Tracy 18076 (UC). Marin Co.: Pt. Reyes P.O., Jul 1903, Elmer 4649 (DS,MO). Mendocino Co.: Red Mt., 15 Jul 1969, Gankin et al. 2666 (CAS); along Eel River at Richardson (redwood) grove, 26 Jul 1924, Heller 18887 (DS,MO); Hale’s Camp, near Redwood Highway, 14 Jul 1931, Jones 29145 (UC); Navarro River near Philo, 11 Jul 1909, McMurphy 512 (DS); Red Mountain Creek Bluff, on S. Fork Eel River, 8 mi S of north county boundary, 7 Aug 1938, Tracy 16100 (DS,UC). Napa Co.: near St. Helena Sanitarium, 2 Jul 1916, Abrams 5750 (DS); Howell Mountain, Aug 1888, T. Brandegee s.n. (UC); Alta Loma Ranch, 7-14 Sep 1906, Chandler 7134 (UC); Howell Mt., Napa River Basin, 1893, Jepson s.n. (JEPS); ridge SE of Mt. Veeder, 21 Aug 1951, Raven 3867 (CAS). Solano Co.: Cordelia, 17 Aug 1920, Campbell s.n. (CAS); Stone Quarry Hill, Cordelia, 7 Aug 1901, Jepson 1743(JEPS). Sonoma Co.: near summit of Mt. Hood, 27 Jul 1948, Baker 12079 (CAS); Hope Val at summit above Coleman’s Ranch, near Kenwood, Aug 1950, Baker 12309 (UC); Russian River, below Guerneville, 22 Jul 1896, Davy 4147 (UC); Guernewood Park, 25 Jun 1909, Hall 8518 (DS-2 sheets,;MO,UC); Sonoma Creek Canyon near Kenwood Springs, 9 Oct 1932, Howell 10811 (CAS); dry hills near Agua Caliente, 1 Aug 1933, Jepson 16551 (JEPS); Duncan’s Mills, 17 Jul 1882, Jones 3620 (CAS,DS-2 sheets,MO,UC). Plants of the northern segment of the range of Erigeron bioletti: (Map 2) produce slightly broader leaves more densely arranged on the stems than those of the southern segment. The differentiation appears to be so slight, however, that there is no justification for applying more than a single name. Collections identified by Cronquist (1947) as E. biolettii from Tuolumne and Fresno counties are slightly glandular plants of E. inornatus. Of the varieties of Erigeron inornatus maintained by Cronquist (1947), E. broletta: might be held at that rank more justifiably than the others. The few collections, however, that might be interpreted as intermediate between BE. tnornatus and E. biolettw (e.g., Tracy 16100) occur where they apparently are closely sympatric in a small area of northern Mendocino Co., in contrast to Nesom: Revision of Erigeron sect. Linearifoli 167 . biolettii E. inornatus O var. inornatus ae fe: yy : oe | ey — es = 4 A var. calidipetris &A var. keilii Map 2. Erigeron inornatus and E. biolettii. The range of var. tornatus extends through Oregon into southcentral and southeastern Washington, and into immediately adjacent Nevada. 168 PH Y TODOGIA volume 72(3):157-208 March 1992 their otherwise relatively wide geographic ranges. The stems and leaves of Erigeron bioletti are always glandular, but eglan- dular hairs range from absent to present in variable densities. The vestiture of plants of a few collections even appears to approach that of FE. petrophilus, but the two species are distinct in habit. Erigeron bioletti is also similar to E. angustatus and partly sympatric with it, but the latter has linear leaves and is glabrous or nearly so. 4. Erigeron blochmaniae E. Greene Erigeron blochmaniae E. Greene, Pittonia 3:25. 1896. LECTOTYPE (Cron- quist 1947): UNITED STATES. California: Santa Barbara Co., Santa Maria, beach and sand a few miles inland, “summer,” /.M. Blochman s.n.(ND-G!). Erigeron foliosus Nutt. var. blochmaniae (E. Greene) H.M. Hall, Univ. Calif. Publ. Bot. 3:91. 1907. Greene cited no specimens in the original description, but the ND-G sheet selected fits the protologue and was marked “Type” by Greene. Stems 4-8 dm tall, erect or usually ascending and caudexlike at the base (some labels note a “rhizomatous” habit). Stems and leaves densely and closely puberulous to villosulous with minute, stiffly crisped hairs. Leaves lin- ear to narrowly oblanceolate, 1.5-3.0 cm long, 1-3 mm wide. Heads 9-14 mm wide; phyllaries densely glandular and densely canescent-hirsutulous, linear- lanceolate, with a broad orange middle area and thick, white lateral zones, the inner 4.5-6.0 mm long. Ray flowers 45-72, the corollas 8-11 mm long, dry- ing white to blue. Disc corollas 4.6-5.0 mm long. Achenes 2.2-2.8 mm long; pappus bristles 21-36. Chromosome number, n=9 (Keil et al. 1988). California endemic, Central Coast in San Luis Obispo and Santa Barbara cos. (Map 1); sand dunes and hills, 3-30 m; Jul-Oct. Cronquist (1947) observed that this taxon might justifiably be treated as a separate species but kept it within Erigeron foliosus because of putative in- tergradation with var. foliosus. The specimens that may have been regarded as intermediates were perhaps those treated here as E. foliosus var. franciscen- sis Nesom (discussed further following the latter), which among the varieties of E. folosus is most similar to E. blochmaniae, but these two taxa do not intergrade. Erigeron blochmaniae is restricted to the coastal sands of Santa Barbara and San Luis Obispo counties and is recognized particularly by its large heads, distinctive stem vestiture, and completely glabrous achenes. 5. Erigeron breweri A. Gray Erigeron brewert A. Gray, Proc. Amer. Acad. Arts 6:541. 1865. LECTO- TYPE (designated here): UNITED STATES. California: Mariposa Co., Nesom: Revision of Erigeron sect. Linearifolu 169 Yosemite Valley, 4000 ft, 1863, W.H. Brewer 1651 (GH; Isolectotypes: UC!,US!; Probable isolectotypes: UC-2 sheets). Gray also cited in the protologue a collection from Nevada (C.L. Anderson 94132-GH). Cron- quist (1947) cited only the Brewer collection but apparently saw only the US specimen. Stems ascending-erect, 20-75 cm long, arising from slender, rhizomelike or caudexlike branches, usually densely invested (sparsely in var. klamathensis) with short, even length, spreading-deflexed hairs, eglandular. Leaves linear to narrowly oblanceolate or oblanceolate, entire, 5-40 mm long. Heads 8-15 mm wide, in a terminal cluster; phyllaries glandular, sometimes also with nonglandular hairs. Ray flowers 12-45, the corollas 4-7 mm, drying blue, less commonly white to pink. Disc corollas 3.5-6.0 mm long. Achenes 2.2-2.8 mm long; pappus bristles 22-46. Key to the varieties of E. breweri 1. Phyllaries densely glandular, nonglandular hairs lacking or very sparse. . ae i re ote a ae (2) 1. Phyllaries slightly glandular to eglandular, with prominent nonglandular HIE: phat Sete ook, = ogee ae etme teen = ae SUNN Ratha UTS tehe wo 3a (4) 2. Stems prostrate to procumbent or decumbent, 7-15 cm; leaves 5-12 mim long; Sot tee oes es meilae waite wie var. jactnteus 2. Stems commonly basally ascending but the upper portions erect, usually longer than 20 cm; leaves mostly longer than 15 mm. . (3) 3. Hairs of stems and leaves 0.1-0.2 mm long; inner phyllaries with green, Aster-like apical areas; stems arising from slender, fibrous rooted, rhi- zomelike bases, without a strongly developed woody root. .var. breweri 3. Hairs of stems and leaves 0.5-1.0 mm long; inner phyllaries with broad, white thickened margins, lacking a distinctly demarcated, green apical area; stems arising from slender but woody basal offsets, these from a strongly developed, woody root. ..........+-..2++++> var. klamathensis 4. Stems 20-30 cm, distinctly wiry and brittle; phyllaries with long, thick based, stiffly spreading, translucent hairs, distinctly glandular BE WEN ee eee x not ee ts See eS Serer errs var. porphyreticus 4. Stems (30-)40-75 cm, thickening toward the base, not wiry or brittle; phyllaries with short, white, ascending to appressed hairs, eglandu- Dare ‘asses See ees casey ee cee te ede oc ra ati ea ero (5) 170 PRY POLroe ra volume 72(3):157-208 March 1992 5. Hairs of stems and phyllaries barely longer than broad; plants of Baja Wpltiormmm Norte ors ee. cess cece ce asco ane var. ensenadensis 5. Hairs of stems and phyllaries longer; plants of California. ............ (6) 6. Phyllaries with thick based, ascending hairs, the hairs dense on the outermost, sharply reduced in density on the middle and inner. .. SR, FC Ree eT ene en eee a var. coviller 6. Phyllaries with thin based, appressed-ascending hairs, the hairs grad- ually and slightly reduced in density from the outer toward the Th Cos erie. ee Ri ces RRR aS ane I A oe var. bisanctus a. Erigeron breweri A. Gray var. bisanctus Nesom, var. nov. TYPE: UNITED STATES. California: Los Angeles Co., San Antonio Canyon near Claremont, 2 Aug 1903, C.F. Baker 8451 (HOLOTYPE: LL); Iso- types: CAS!,UC!). These specimens were distributed as a new species proposed by E.L. Greene, using the epithet “antoninus,” which was ap- parently never validly published. An additional sheet of Baker 3451 (CAS) bears a plant of E. foliosus var. foliosus. Erigeronti brewert A. Gray var. coville: (E. Greene) Nesom sim- ilis sed phyllariis trichomatibus appressis ad basim tenuibus vestitis differt. California endemic, San Gabriel and San Bernardino Mts. (Map 3); open, dry slopes and washes, 370-1520 m; May-Sep. Additional collections examined: UNITED STATES. California: Los An- geles Co.: San Antonio Canyon, 2 Aug 1903, Baker 8658 (LL,UC); San Dimas and San Gabriel Divide, Cucamonga Quad., 20 Sep 1935, Horton 248 (UC); San Gabriel Canyon, 28 Jun 1892, Hutchinson s.n. (JEPS). San Bernardino Co.: 5 mi NE of Upland, 14 Jun 1927, Blake 9867 (LL); E fork of the Santa Ana River, 3 mi below CA Hwy 38, 9 Jul 1974, Helmkamp s.n. (UCR); Santa Ana River, 28 Aug 1922, Munz 6382 (UC-2 sheets); San Gabriel Mts., Big Dalton Canyon, chaparral, 10 Jun 1984, Swinney s.n. (UCR); 2 mi E of Clare- mont, 28 May 1933, Wheeler 1748 (LL); Santa Ana River drainage, Forest Service Road 1N64, between the 2 forks of Deer Creek, 18 Jul 1990, White s.n. (TEX,UCR). These plants have tall, thick, and relatively unbranched (until the capit- ulescence, near the apex) stems similar to var. covillei, but the phyllaries are densely invested with thin based, appressed hairs distributed somewhat differ- ently. The two apparently grow in close proximity (e.g., Baker 9658 [MO] from the San Antonio Canyon, the type locality of var. bisanctus, is var. covillet). Revision of Erigeron sect. Linearifolu 17 . breweri @ var. bisanctus ~ : es eels # O var. covillei praca: = & var. porphyreticus =: Map 3. Erigeron brewer: (var. porphyreticus, var. coville:, and var. btsanctus). 172 PHY. TOL O'G-LA volume 72(3):157-208 March 1992 Some plants of Erigeron foliosus in the the same region have phyllaries with vestiture nearly identical to var. bisanctus, and seems probable that gene flow occurs between these two taxa. b. Erigeron brewert A. Gray var. brewerr Chromosome number, n=9 (Solbrig et al. 1969; Semple et al. 1989). California, northern to southern High Sierra Nevada, Nevada in the Lake Tahoe region, apparently disjunct in the San Bernardino Mts. (Map 4); open, rocky sites in a variety of vegetation types, (1220-)1800-2930(-3050) m; Jun- Sep. As interpreted here, var. brewer: has a discrete geographic distribution in the eastcentral Sierra Nevada, with an outlying segment to the southwest in the San Bernardino Mountains. It remains one of the most complex taxa of the section, deserving further study and perhaps ultimately better treated as two varieties rather than one. Plants of the typical form produce thin, woody rhizomes with relatively short (12-40 cm), ascending stems and the leaves are distinctly oblanceolate. In Fresno and Tulare cos., many of the plants produce thicker stem bases, taller (30-60 cm) and more erect stems, and more linear leaves; only the tall form is present at the southern extremity of the range in Kern Co. The two forms are mapped with different symbols (Map 4), although many of the determinations are arbitrary in Fresno and Tulare cos., where the intergradation seems to be gradual. The taller plants were included by Cron- quist (1947) in var. covillez, but while they have its habit, perhaps reflecting the genetic influence of that variety, they produce the relatively shorter, merely glandular phyllaries distinctive of var. breweri. Further, their sierran habitats and range, together with their apparent complete intergradation with the typ- ical form, make it difficult to virtually impossible to recognize more than a single taxon. The plants identified as var. brewer: from the San Bernardino Mountains are most like the typical form, apparently long disjunct from the main range. They apparently occur in several localities, but a number of collections have been made from along the South Fork of the Santa Ana River and in the vicinity of Bear Valley. A collection of typical var. breweri (CAS, on the label: Mt. Eddy, Jul 1880, Lemmon s.n.) is almost certainly mislabeled, as no other collection of that taxon has been made anywhere near Siskyou County. This collection of var. breweri may have been made in the vicinity of Mt. Lassen (see additional comments following var. klamathensis). c. Erigeron breweri A. Gray var. covillei (E. Greene) Nesom, comb. nov. BASIONYM: Erigeron covillei E. Greene, Erythea 3:20. 1895. LEC- TOTYPE (designated here): UNITED STATES. California: Inyo Co., Revision of Erigeron sect. Lineartfolu 173 eaWe KE. breweri A var. jacinteus e var. breweri (typical) O var. breweri (tall form) Map 4. Erigeron breweri (var. breweri - the typical form and the “tall form” are mapped with different symbols - and var. jacinteus). 174 Pay TOL OrGel A volume 72(3):157-208 March 1992 near Crystal Spring, Coso Mountains, 12 Jun 1891, F.V. Coville & F. Funston 981 (ND-G!; Isolectotypes: UC-fragment [as cited by Cronquist 1947], US!). Erigeron foliosus Nutt. var. covillei (E. Greene) Compton, Bull. So. Calif. Acad. Sci. 33:51. 1934. California endemic, eastern margin of the Sierra Nevada, Mojave Desert, San Gabriel Mts., San Bernardino Mts., San Jacinto Mts. (Map 3); open, rocky sites in sagebrush, chaparral, and juniper, 980-1830 m; May-Sep. In the view of the present treatment, the dense, spreading-deflexed stem pubescence of these plants places them with Erigeron brewer: rather than E. foliosus. Further, they are broadly sympatric with E. foltosus var. foliosus and few if any unequivocal intermediates have been seen. Two collections from the area of Independence in Inyo County, otherwise similar to var. covzllez, are unusual in their foltosus-like, antrorsely appressed-ascending stem pubescence: Raven 9946 (JEPS) and Alezander & Kellogg 2974 (LL-3 sheets, MO,UC). Intergrades between var. coville: and E. brewerz var. breweri, however, were noted by Cronquist (1947) as well as in the present study. Var. covillez is strictly a taxon of the Mojave Desert, forming a southern segment of the range of E. breweri (Map 3). It is clearly separated from the southern end of the range of var. breweri, but intermediates appear to occur with the “tall form” in the Independence Creek area of Inyo County. The possible influence of var. coville: on var. brewerz is discussed following the latter. In addition, where the range of var. coville: meets that of var. porphyreticus in San Bernardino Co., there are apparent intermediates between them. d. Erigeron breweri A. Gray var. ensenadensis Nesom, var. nov. TYPE: MEXICO. Baja California Norte: Canon 10 mi N of San Vicente, near Mpio. Ensenada, 3 May 1969, D. Wilken & W. Werner 7432 (HOLO- TYPE: UC): Differt a Erigeronte breweri A. Gray var. covillei (E. Greene) Nesom caulibus ac phyllariis scabridiusculis. Northwestern Baja California Norte, México (not mapped); chaparral, desert scrub, ca. 150-300 m; May-Jul. Additional collections examined: MEXICO. Baja California Norte: Guada- lupe, 2 Jun 1893, Brandegee s.n. (UC); ca. 7 mi N of Santo Tomas along road to Ensenada, 29 Jun 1962, Wiggins & Thomas 415 (DS). These plants are similar to those of var. covillei, but the hairs of the stems and phyllaries are extremely short, barely longer than broad, and they are long disjunct (ca. 120 miles) from the southernmost populations of var. covillei in Riverside Co., California. Nesom: Revision of Erigeron sect. Linearifolu 175 e. Erigeron brewert A. Gray var. jacinteus (H.M. Hall) Cronq., Brittonia 6:284. 1947. BASIONYM: Erigeron jacinteus H.M. Hall, Univ. Calif. Publ. Bot. 1:127. 1902. TYPE: UNITED STATES. California: River- side Co., San Jacinto Mountains, rocky ridges and peaks near Tauquitz, 8826 ft, Jun 1901, H.M. Hall 2322 (HOLOTYPE: UC!; Isotype: JEPS). California endemic, San Jacinto, San Bernardino, and San Gabriel, Mts. (Map 4); open, rocky crests and slopes, 2680-2900 m; Jun-Sep. Var. jacinteus is recognized by the dense vestiture of the stems and leaves, very small leaves, and short, nearly prostrate stems. It is very similar to var. brewerz, but in the San Bernardino Mountains, where they both occur, var. jacinteus occurs at higher elevations (var. brewerz is found at 1210-2580 meters). In a sense, var. jacinteus is related to the rest of E. brewerias E. miser A. Gray is to E. petrophilus; further study might show that var. jacinteus is distinct as a species and deserves coordinate taxonomic rank with EF. breweri. Five varieties of Erigeron brewerz occur in the San Bernardino Mountains. These taxa appear to be altitudinally segregated, from highest to lowest as follows: var. jacinteus, var. breweri, var. porphyreticus, and var. coviller and var. bisanctus the lowermost, which occur at roughly similar elevations. {. Erigeron breweri A. Gray var. klamathensis Nesom, var. nov. TYPE: UNITED STATES. California: Humboldt Co., Trinity Summit, 2 mi SE of Devil’s Hole, exposed rocky points in woods, westerly exposure, 26 Jul 1935, J.P. Tracy 15515(HOLOTYPE: UC!; Isotypes: JEPS!,MO!,TEX!). Differt a Erigeronte folioso Nutt. var. confini (Howell) Jeps. tri- chomatibus patentibus caulium ac foliorum, et foliis oblanceolatis brevioribusque. Differt a Erigeronte brewerz A. Gray var. breweri radice lignea, trichomatibus longioribus, et phyllariis sine apicibus viridis. California, primarily Klamath Ranges in Siskyou, Trinity, and Shasta cos., adjacent Oregon in Curry and Josephine cos. (Map 5); open, rocky slopes, rock crevices, 1580-2140 m; Jun-Sep. Additional collections examined: UNITED STATES. California: Del Norte- Siskyou Co. line, Little Grayback, 14 Aug 1934, Lee 1109(UC). Humboldt Co.: Trinity Summit, 25 Sep 1919, Tracy 5299 (UC); Trinity Summit at Wa- ter Dog Lakes, 26 Jul 1935, Tracy 14271 (UC); Trinity Summit, Devil’s Hole, 17 Jul 1932, Tracy 10696 (UC). Siskyou Co.: top of grade between Sawyer’s Bar and Etna, 8 Sep 1948, Alerander & Kellogg 5602 (UC); slopes SE of White Mt., 1 Jul 1961, Dempster 1832 (UC); Baldy Mt., 13 Jul 1950, Hoff- man 3552 (UC); Marble Mts., Elk Mt., 6 Aug 1939, Howell 15101 (CAS,LL); Devil’s Backbone, 17 Jul 1902, Jepson 2071 (JEPS), 22 Jul 1902, Jepson 2101 176 Poy fot CoG! A volume 72(3):157-208 March 1992 ian @ E. breweri var. klamathensis Map 5. Erigeron brewerz var. klamathensis. Nesom: - Revision of Erigeron sect. Lineartfoli 177 (JEPS); trail from English Lake to Hancock Lake toward Marble Mt., 19 Aug 1928, Kildale 6527(DS); at Del Norte Co. line, Little Grayback, 14 Aug 1934, Lee 1109 (DS); Mt. Eddy, near Shasta, Jul 1880, Lemmon s.n. (UC); Salmon Mts., high lake basins, vicinity of English Peak, 19 Aug 1968, Oetttnger 588 (UC). Trinity Co.: Trinity Summit, near Ranger Station, 24 Jul 1935, Tracy 14204 (DS,JEPS,MO,UC); Siskyou Mts., Rock Gulch off Jaynes Canyon, 6 Aug 1934, Wheeler 3055 (DS-2 sheets, UCR). Oregon. Curry Co.: banks of Rogue River at Brushy Bar, 27 Apr 1947, Baker 3819 (UC); Rocky Peak, 12 mi SE of Port Orford, 25 Jul 1919, Peck 8982 (MO). Josephine Co.: Mt. Grayback, Siskyou Mts., 2 Aug 1935, Tracy 12460 (MO); Baby Foot Lake area near Cave Junction, 23 Jul 1980, Vale s.n. (CAS). These plants form the northernmost segment of the range of Erigeron breweri. Cronquist (1947) included a Lemmon collection of var. klamathensis (Shasta Co.: Lassen’s Peak, 1870’s, Lemmon s.n. [JEPS,UC]) in var. brew- ert, but he did not recognize E. brewer: from any localities further north. It is likely that the label of these plants was interchanged with a collection by Lemmon (of typical var. breweri, said on the label to be from Mt. Eddy, but far out of range for that taxon). Var. klamathensts is similar to var. brewerz in its merely glandular phyllar- ies but more like var. porphyreticus in its strongly woody bases. The relatively sparse, stiffly hirsute vestiture of long hairs in var. klamathensis is not en- countered in any of the other varieties of the species. Many collections of var. klamathensis have been identified previously as E. foliosus var. confints, which usually is nearly glabrous to sparsely strigose and produces longer, lin- ear leaves. Some plants of the two taxa, however, are very similar in habit, and they may prove to be closely related (see additional comments following var. confinis). g. Erigeron brewert A. Gray var. porphyreticus (M.E. Jones) Crongq., Brit- tonia 6:283. 1947. BASIONYM: Erigeron porphyreticus M.E. Jones, Contr. West. Bot. 8:33. 1898. TYPE: UNITED STATES. Nevada: Mineral Co., Hawthorne, Big Indian Canyon, 27 May 1897, M.E. Jones s.n. (HOLOTYPE: POM; Isotypes: MO,US!). Erigeron foliosus Nutt. var. porphyreticus (M.E. Jones) Compton, Bull. So. Calif. Acad. Sci. 33:53. 1934. Erigeron petrocallis E. Greene, Erythea 3:21. 1895. LECTOTYPE (designated here): UNITED STATES. Nevada: Humboldt Co., West Humboldt Mountains, Jul 1894, E.L. Greene s.n. (ND-G!). A different collection by Greene has been annotated by previous workers as type material, but the specimen cited here is the only one that agrees with the protologue. 178 PRY MOt.0.G iA volume 72(3):157-208 March 1992 Chromosome number, n=9 (Montgomery & Yang 1960, the voucher appar- ently cultivated from “Rancho Santa Ana”). Western half of Nevada to the eastern edge of the Sierra Nevada in Cali- fornia, Mono Co. south into the Mojave Desert (Map 3); open, rocky sites in sagebrush to ponderosa pine, (1220-)1520-2530 m; May-Aug(-Sep). Duplicates of a collection of typical Erigeron breweri var. porphyreticus were distributed and identified by Munz with the varietal name “roosiz;” the specimens were marked as types, but the name apparently was never published. Var. porphyreticus is similar to var. brewerz in the production of slender, rhizomatous branches, spreading-deflexed stem vestiture, and blue ligules; it differs in its wirier stems, generally greater development of flowering branches, larger heads, and hirsute-hispid phyllaries with broad stramineous-indurated margins, lacking green, Aster-like tips. For the most part, the ranges of the two taxa are separate (Maps 3 and 4), but they are sympatric in Mono Co. and northern Inyo Co. and adjacent Nevada, as well as in the San Bernardino Mountains, and they appear to intergrade at least in Mono and Inyo cos. In the Kingston Range of San Bernardino Co., where numerous collections of Erigeron brewer: have been made, most of the plants can be referred to var. porphyreticus, as identified here, but some have smaller heads and have relatively few nonglandular hairs. Were these not in populations of other typical var. porphyreticus, they would be best identified as var. breweri. Some of the plants included here with var. porphyreticus (cited below) from San Bernardino Co. were identified by Cronquist (1947) as var. covillei. Var. porphyreticus in this southwestern portion of this county occurs at 1220-2440 meters elevation, higher for the most part than var. covillei. Specimens of var. porphyreticus examined from southwestern San Bernardino Co.: San Bernardino Mts.: Bear Valley, 5 Aug 1902, Abrams 2886 (UC), Jul 1909, Davidson 2169 (JEPS); Bear Lake, 8 Jul 1931, Clokey 5317 (UC); San Gorgonio Wilderness Area, Lower South Fork Meadow, 15 Aug 1976, David- son & Thorne 4716 (UC); Hesperia, 14 Jun 1895, Parish 3603 (JEPS,UC-2 sheets). 6. Erigeron elmeri (E. Greene) E. Greene Erigeron elmeri (E. Greene) E. Greene, Fl. Franc., part IV, p. 393. 1897. BASIONYM: Aster elmeri E. Greene, Pittonia 2:170. 1891. LECTO- TYPE (Cronquist 1947): UNITED STATES. California: Tuolumne Co., Grand Canyon of the Tuolumne River, summer [1891, as on the label], V. Chesnut & E. Drew s.n. (ND-G!). Ertgeron breweri A. Gray var. elmer: (E. Greene) Crongq., Brittonia 6:284. 1947. The protologue notes that the type collection was made in “summer 1890,” but I have seen no other specimen that might be the one referred to by Greene, nor is Nesom: Revision of Erigeron sect. Linearifoli 179 there any other such in ND-G (fide B. Hellenthal). Both the protologue and collection label note that the specimen was collected in the “Grand Canyon of the Tuolumne” by Chesnut & Drew, the only difference being that of one year in date. Although it is not annotated in any way, this is evidently the specimen that Cronquist (1947) referred to as the type, and his choice is accepted here. It seems most likely that either the label or the publication of Aster elmeri was inadvertently supplied with the wrong date. Stems prostrate-ascending to nearly erect, 6-20 cm long, from slender, woody rhizomes or rhizomelike caudex branches. Stems and leaves eglan- dular, sparsely to moderately strigose with thin-based, closely appressed to ascending-appressed hairs. Leaves narrowly oblanceolate or oblong, 5-20 mm long, 1-2 mm wide. Heads 7-10 mm wide; phyllaries minutely glandular, with- out other vestiture, inner phyllaries 3.5-5.0 mm long, often purple-tipped. Ray flowers 12-21, the corollas 6-9 mm long, drying bluish. Disc corollas 3.5-4.0 mm long. Achenes 1.6-2.5 mm long; pappus bristles 18-26. California endemic, central High Sierra Nevada (Map 1); ledges, crevices, cracks, talus, 1300-3300 m; Jun-Sep. Erigeron elmerz is distinguished by its low habit with wiry, trailing to as- cending stems, very small leaves, sparsely strigose vestiture, small heads, and pink to lavender ligules. It has been treated as a variety of E. brewer: because of putative intergradation between the two, but the occurrence of intermedi- ates has not been confirmed in the present study. In contrast, among the 30 separate collections of EF. elmeri examined, only a single one has stem vesti- ture that might be considered to approach that of FE. brewerz (Lake Tenaya, Yosemite, Eastwood 462-UC). On some branches of this collection, the stem hairs are ascending-appressed; on others, there is a mixture of spreading and ascending hairs. This atypical variation may reflect an influx of genes from E. brewerz, but the two taxa are sympatric and, with the possible exception of a few plants, they retain their integrity, providing convincing evidence of genetic isolation. A specimen cited by Cronquist as var. elmer: from Tulare Co. (Hall & Babcock 5697-UC), which has spreading stem pubescence, also is stipitate-glandular and clearly is a collection of FE. aequzfoltus. 7. Erigeron foltosus Nutt. Erigeron foliosus Nutt., Trans. Amer. Philos. Soc. 2, 7:309. 1840. TYPE: UNITED STATES. California: Santa Barbara Co., near Santa Barbara, [1836], 7. Nuttall s.n. (Possible isotype: PH fide Cronquist 1947). Stems 20-100 cm long, often arising from thin, rhizomelike or caudexlike branches, erect but commonly basally ascending, branched near the apex; 180 PHY TOPO GIA volume 72(3):157-208 March 1992 stems, leaves and phyllaries glabrate to sparsely invested with appressed hairs. Leaves linear to oblanceolate, entire, 10-65 mm long, 1-5(-10) mm wide. Heads 10-16 mm wide, in a loose terminal cluster, rarely solitary; phyllaries glabrous to moderately hairy, sometimes glandular. Ray flowers 15-60, the corollas 5-15 mm long, usually blue. Disc corollas 3.5-5.5 mm long. Achenes (1.8-)2.4-3.0 mm long; pappus bristles 20-34. In the present treatment, Erigeron foliosus is distinguished from the other taxonomically complex, radiate species of the section, E. brewerz, by its glabrate to sparsely strigose stems, as contrasted with the densely hispidulous stems of the latter. It is perhaps an oversimplified hypothesis, but the radiate E. elmeri and E. ozyphyllus are similar to E. foliosus in stem vestiture and may be most closely related to it. The varieties of Erigeron foliosus are separated primarily by relatively small differences in vestiture. For the most part, the taxa are represented by a series of essentially allopatric population systems (Maps 6 and 7). Var. mendocinus (E. Greene) Nesom and var. confinis as well as var. foltosus and var. francis- censis have slightly overlapping or at least parapatric distributions and form intermediates in the areas where they meet. Key to the varieties of E. foliosus 1. Phyllaries densely and prominently glandular, 0.8-1.0 mm wide, the mar- gins usually with a broad scarious rim. ..........-.s000+s+-=s00eee (3) 1. Phyllaries eglandular or sparsely and obscurely glandular, 0.5-0.8 mm wide, the margins usually thick or with only a narrow scarious rim. (2) 2. Leaves (3-)4-6 cm long usually with acute apices, often mostly ori- ented in one direction to one side of the stem; inner phyllaries (4.0-) SA DECGA I) ye 62 Wl (aS 112 Rete ee cece ae eR Ae var. hartwegu 2. Leaves 2-4(-5) cm long usually with rounded apices, without a promi- nent unidirectional orientation; inner phyllaries 3.2-4.5 mm long. ee kaiacc te Sens Be A Akde ee See var. foltosus 3. Phyllaries without eglandular hairs; leaves mostly 1-2 mm wide. ........ OES ce Tie TNC Tee ee ee roe ieee var. confinis 3. Phyllaries moderately to densely strigose-hirsute; leaves mostly 2-4 mm WAGGA ie: elas gale pst iene nies CPO Aor nies bye Loh a ee (4) 4. Leaf lamina sparsely to moderately strigose; phyllaries with a promi- nent, raised, orange resinous midvein; ray corollas 7-10 mm long. CR Fete REN cote Lh UNO aya uate enc SA Oe eee one var. franciscensis Nesom: Revision of Erigeron sect. Linearifolu 181 E. foliosus A var. confinis ® var. mendocinus Map 6. Erigeron foliosus (var. mendocinus and var. confints). 182 Pon YO LOsG 1 A volume 72(3):157-208 March 1992 EE EQ £. foliosus ) ! . hae ; © var. hartwegii e var. franciscensis 4 var. foliosus (narrow-leaved for var. foliosus (broad-leaved form Map 7. Erigeron foliosus (var. foliosus, var. hartwegti, and var. franciscensis). For var. foliosus, the broad leaved forms and narrow leaved forms are mapped separately to show their nearly complete sympatry; many of the symbols were assigned arbitrarily where intermediates in leaf width were involved. The range of var. foliosus extends into northwestern Baja California Norte. Nesom: Revision of Erigeron sect. Linearzfolu 183 4. Leaf lamina glabrous or glabrate, margins and midvein ascending cil- iate; phyllaries usually without a distinct, orange resinous midvein; CHV COFOM ARN ei SEUTL NOTE as Boss noo aise sia nfo oni dinutn var. mendocinus a. Erigeron foliosus Nutt. var. confinis (Howell) Jeps., Man. Fl. Pl. Calif. 1056. 1925. BASIONYM: Erigeron confinis Howell, Erythea 3:35. 1895. LECTOTYPE (designated here): UNITED STATES. Oregon: Josephine Co., Siskyou Mts., near Waldo, Jul 1887, T.J. Howell 1507 (WS; Isolectotypes: MO!,MI,NY,UC!,US!). The protologue says “On high rocky ridges of the S. Mts., July 1886” but specimens matching the locality description are “1887.” The sheets so cited (Cronquist marked the MO and UC sheets as “isotype”) were distributed by Howell with the unpublished epithet “siskyouensis.” Inner phyllaries 4.0-5.0 mm long. Ray corollas 9-12 mm long. Disc corollas 3.5-5.0 mm long. Klamath Ranges of California and adjacent southwestern Oregon (Map 6); rocky sites, chaparral to fir woods, over a variety of soils but often on serpentine, (rarely to 15-)(150-)750-2000(-2200) m; May-Aug. Plants of Erigeron foliosus var. confinis typically are low (1-2 dm tall), linear leaved, and have densely glandular phyllaries without other vestiture. There are apparent intermediates with var. mendocinus where their ranges meet, particularly in the region of the Smith River in Del Norte Co. and the Klamath River in Humboldt Co. Some plants identified as var. confinis are very similar in habit to E. breweri var. klamathensts (particularly in Siskyou and Trinity cos., e.g., Tracy 19196-UC), and the possibility of gene flow or very recent common ancestry between them should be investigated. The two taxa are sympatric over a considerable part of their ranges (Maps 5 and 6). b. Erigeron foliosus Nutt. var. foliosus Erigeron stenophyllus Nutt., J. Acad. Nat. Sci. Philad. 2, 1 (Pl. Gamb.): 176. 1847. TYPE: UNITED STATES. California: Monterey Co., near Monterey, [1842], W. Gambel s.n. (Probable isotype: GH!); not Hook. & Arn. 1836; not A. Gray 1857; not D.C. Eaton 1871. Erigeron foliosus Nutt. var. stenophyllus (Nutt.) A. Gray, Bot. Calif. 330. 1876. Erigeron nuttallii Heller, nom. nov. Bull. Torrey Bot. Club 25:628. 1898. Diplopappus occidentalis Hook. & Aru., Bot. Beechey’s Voy., Suppl. 350. 1839. TYPE: UNITED STATES. California: [{ca. 1830], D. Douglas s.n. (not seen). Erigeron douglast Torr. & Gray, nom. 184 PHYTOLOGIA volume 72(3):157-208 March 1992 | nov., Fl. N. Amer. 2:177. 1841. Not Erzgeron occidentalis Nutt., | 1840. Torrey & Gray were unaware that Erigeron foliosus Nutt., the | earliest available name for this species but which they treated asa | “Species unknown to us,” was the same as Diplopappus occidentalis. | They indicated that they had seen the original specimen collected | by Douglas. The identity of FE. occidentals Nutt. has not been | ascertained, but with “very numerous, red” rays and “subserrulate” leaves, it clearly is not part of the FE. foliosus group. Based on the description, albeit relatively generalized, and the information regarding the collecting localities of Douglas (Jepson 1933), it seems clear that the plants of his collection are E. foliosus var. foliosus, in the sense of the present study. Gray (Synopt. Fl. N. Amer. 1(2}:215. 1884) recognized the synonymy of FE. douglasw with EL. foliosus. Erigeron foltosus Nutt. var. tenuzsstmus A. Gray, Synopt. Fl. N. Amer. 1(2):215. 1884. LECTOTYPE (designated here): MEXICO. Baja California Norte. Hanson’s Ranch, 30 Jul 1883, C.R. Orcutt 1000 (GH!; Isolectotype: UC!). In the original description, Gray’s cita- tion was “San Diego Co., on the Mexican border and within lower California, Parry, Palmer, Orcutt.” Erigeron tenuissimus E. Greene, Pittonia 3:25. 1896. LECTOTYPE | (designated here): UNITED STATES. California: Ventura Co., | near Ventura, Jun 1893, Miss A. Symes s.n. (ND-G!; Isolectotype: — UC, as cited by Cronquist, but not located in the present study). Erigeron striatus E. Greene, Bull. So. Calif. Acad. Sci. 1:39. 1902. | LECTOTYPE (designated here): UNITED STATES. California: San Bernardino Co., Huston Flat, Transition Zone of the San Bernar- | dino Mts., Aug 1900, W.R. Shaw 36 (ND-G!; Isolectotype: DS!). Erigeron setchellit Jeps., Fl. W. Mid. Calif. (ed. 1) 568. 1901. LEC- — TOTYPE (designated here): UNITED STATES. California: [San Joaquin Co.], “arid plains of the Lower San Joaquin” [somewhere | between Mossdale (near present day Manteca) and Oakdale], 27 Jun 1896, W.A. Setchell & W.L. Jepson s.n. (JEPS!; Isolectotype: JEPS!). Locality information from Jepson’s field notebook was sup- plied by Mrs. A.Q. Howard. Erigeron fragilis E. Greene, Bull. So. Calif. Acad. Sci. 1:39. 1902. LECTOTYPE (designated here): UNITED STATES. California: Orange Co., Trabuco Canyon, 16 Jun 1901, L. Abrams 1801 (ND- G!; Isolectotype: DS!). Erigeron foltosus Nutt. forma grinnellu Cronq., Brittonia 6:279. 1947. TYPE: UNITED STATES. California: Los Angeles Co., Mt. Wil- son, 4 Jul 1904, F. Grinnell s.n. (HOLOTYPE: CAS!). Nesom: Revision of Erigeron sect. Lineartfolit 185 Inner phyllaries 3.2-4.5 mm long. Ray flowers 15-49, 6-10 mm long. Disc corollas 3.5-4.0(-5.0) mm long. Chromosome number, n=9 (Semple 1985; Strother 1972; Solbrig et al. 1964; Raven et al. 1960). California, widespread in the southwestern part of the state, adjacent Baja California Norte, México (Map 7); dry, open, rocky or grassy slopes, chaparral, oak to pine or pine-fir woodlands, (40-)300-2250(-2900) m; May-Aug(-Sep). Plants of Erigeron foliosus with linear leaves (var. stenophyllus) have tra- ditionally been separated from those with much broader, oblanceolate leaves (var. foliosus), but when confronted with a large number of specimens, the correspondingly large number of intermediates and the nearly exact congru- ence of the geographic ranges of the two putative varieties makes it impossible to separate them in an unarbitrary taxonomy. In the view here, var. foliosus comprises a series of plants with leaves varying from filiform to oblanceolate and broadly oblanceolate (the latter in extreme form in forma grinnellit). The habit of Erigeron foliosus var. foliosus is somewhat variable, but it is most commonly similar to that of E. breweri var. coville1, with stems basally ascending and gradually thickened towards the base. Complete root systems have been collected only rarely. Some plants identified here as var. folzosus, particularly in Tulare and Kern counties (e.g., S. Fork King’s River, Jepson 778A-JEPS), have stems densely invested with minute, slightly spreading- ascending hairs and phyllaries densely glandular. These may have originated as hybrids between var. foliosus and Erigeron brewert. See other comments following var. franciscensts regarding variability in var. folosus. c. Erigeron foliosus Nutt. var. franciscensis Nesom, var. nov. TYPE: UNITED STATES. California: San Mateo Co., Lake Merced, Jun 1908, K. Brandegee s.n. (HOLOTYPE: UC!). A Erigeronte folioso Nutt. var. folioso phyllariis longioribus glandulosis sed trichomatibus non glandulosis densioribus. Inner phyllaries (4.0-)5.0-6.0 mm long. Ray flowers 28-48, 7-10 mm long. Disc corollas 4.5-5.0 mm long. California endemic, primarily the San Francisco Bay Area (Map 7); grassy dunes, chaparral, oak woodlands, 2-800 m; May-Oct. Representative collections examined: UNITED STATES. California: Ala- meda Co.: Oakland, 9 Jul 1881, Jones 2374 (CAS,DS); 25 mi N of Cedar Mts, 13 Oct 1929, Mason 5511 (UC). Contra Costa Co.: Mount Diablo, Inner Black Hills E of the meridian, 5 Sep 1932, Bowerman 1550 (UC); hills between Antioch and Marsh Creek, 16 Jan 1907, K. Brandegee s.n. (UC); Walnut Creek, 20 Jun 1892, Greene s.n. (UC). Marin Co.: San Francisco, Jul 1908, K. Brandegee s.n. (UC-2 sheets); near Chinese Camp, 25 Jun 1944, 186 PHY Taree volume 72(3):157-208 March 1992 Howell 19663 (CAS,UC); Tiburon Peninsula, Paradise Drive, 2.4 mi NW of road to Naval Net Depot, 29 Jun 1961, Penalosa 1871 (CAS,UC). Monterey Co., Pajaro Hills, Jun-Jul 1899, Chandler 487 (UC); Pt. Lobos, 6 Aug 1912, Eastwood 1568 (CAS); mt. above Carmel Valley, 7 Jul 1962, Howitt 1374 (CAS). San Joaquin Co., Lower San Joaquin River, 27 Jun 1896, Jepson s.n. | (JEPS). San Mateo Co.: N end of Sawyer Ridge, San Francisco Watershed | Reserve, 23 Jul 1983, Buck 448 (UC); Belmont, 17 Jun 1893, Davy 804 (UC); near Searsville Dam, 14 Jun 1895, Dudley s.n. (DS); E-W canyon draining into Lake Merced, just N of the San Francisco Golf and Country Club golf course, 17 Jul 1950, Raven 2608 (CAS,UC). Santa Clara Co.: Saratoga, Sep 1893, Davy 329 (UC); Black Mountain, Jul 1903, Elmer 4589 (CAS,DS,MO,UC). Santa Cruz Co.: Bear Creek Canyon, ca. 4 mi from Boulder Creek, 6 Aug 1953, Hesse 1178 (JEPS); Empire Grade, Ben Lomond Mt., 15 Jul 1953, Thomas 3478 (DS,JEPS). Solano Co., S end of basalt hill at Thomasson, 16 Jun 1937, Jepson 18415 (JEPS). Stanislaus Co., Oakdale, 30 Jun 1896, Jepson s.n. (JEPS). Var. franciscensis is closely similar to var. foliosus, but for the most part the two taxa are geographically well separated. A few collections that techni- — cally would be identified as var. franciscensts are well within the range of var. foltosus (e.g., Los Angeles Co., Leona, Michener & Bioletti s.n.-UC; San Luis Obispo Co., Paso Robles country, Blochman s.n.-UC); other scattered collec- — tions identified and mapped as var. foliosus have densely glandular phyllaries, but the phyllaries are shorter (3.2-4.5 mm long vs. [4.0-]5.0-6.0 mm long in var. franciscensts), generally more sparsely invested with non glandular hairs, without prominently scarious margins, and the plants appear to be closely | linked to the variability in others of more typical morphology. The two vari- eties appear to be sympatric in northern Monterey County, and intermediates occur there. In both var. foltosus and var. franciscensis, the midvein of each phyllary is prominently raised and orange-resinous, but the midvein is low and barely discernible as orange in var. mendocinus; in var. hartwegii and var. confinis it is more variable, sometimes prominently orange, but often not. d. Erigeron fohosus Nutt. var. hartwegi (E. Greene) Jeps., Man. Fl. Pl. Calf. 1056. 1925. BASIONYM: Erigeron hartwegii E. Greene, Erythea 3:21. 1895. LECTOTYPE (Cronquist 1947): UNITED STATES. Cal- | ifornia: Locality and date unspecified, K.T. Hartweg 850 (NY!; Isolec- totype: GH!). In the original description, Greene noted that this taxon had been “first collected by Hartweg, afterwards by Fremont, Bigelow, — and by the present writer.” Bentham (Pl. Hartweg. 316. 1849.) earlier — had identified this same collection as E. decumbens Nutt. The nota- tion “Hartweg 1774” denotes the number of sequential listing in Plantae Nesom: Revision of Erigeron sect. Linearifoliu 187 Hartwegianae rather than the collection number. Erigeron blasdalei E. Greene, Erythea 3:124. 1895. LECTOTYPE (designated here): UNITED STATES. California: Calaveras Co., dry rocky soil, Stanislaus River near McCormic’s Bridge, 10 Jun 1895, W.C. Blasdale s.n. (UC!; Probable isolectotype: ND-G!). The latter specimen is marked only as “Erigeron blasdalei,” but the plants match those on the UC sheet. Inner phyllaries (4-)5-6 mm long. Ray flowers 26-50, 8-13 mm long. Disc corollas 4.0-5.5 mm long. California endemic, north and central Sierra Nevada Foothills (Map 7); rocky sites, commonly along rivers, hills in oak or pine-oak woodlands, 100- 600 m; Apr-Jul. Var. hartwegii has a geographically discrete distribution and the plants are very constant in morphology. Further comments are given following Erigeron mariposanus Congdon. e. Erigeron foliosus Nutt. var. mendocinus (E. Greene) Nesom, comb. et stat. nov. BASIONYM: Erigeron mendocinus E. Greene, Leafl. Bot. Obs. Crit. 2:9. 1909. LECTOTYPE (designated here): UNITED STATES. California: Mendocino Co., Big River, Jul 1903, J. McMurphy 358 (CAS!). Only a single collection was noted in the protologue, but its place of deposition was not specified. Inner phyllaries 5.0-6.5 mm long. Ray flowers 25-48, 10-15 mm long. Disc corollas 3.8-5.0 mm long. California endemic, North Coast Ranges (Map 6); bars, banks, and ledges along rivers, dry slopes, 30-750 m; May-Aug. Representative collections examined: UNITED STATES. California: Del Norte Co.: Smith River near Myrtle Creek, 24 Jun 1938, Parks & Tracy 11458 (UC); Smith River near Van Deventer Ranch, 29 Jun 1938, Van Deventer 127 (JEPS). Humboldt Co.: Van Duzen River, mouth of Grizzly Creek, 11 Jul 1916, Abrams 6055 (DS,LL,UC); Mad River, 10 Jul 1888, Chesnut & Drew s.n. (UC); South Fork Eel River, Upper Look Prairie, 29 May 1934, Constance 720 (JEPS); near Alder Point, 22 Jun 1937, Eastwood & Howell 4731 (CAS); Junction, E of Arcata, 16 Jul 1931, Jones 29149 (MO,UC); near Ettersburg, 25 Jun 1927, Kildale 3585 (DS); Bridgeville, 23 Jul 1928, Kildale 5694 (DS); near Hydesville, bluffs along Van Duzen River, 2 Aug 1901, Tracy 1241 (UC); Grasshopper Ridge, head of Canoe Creek, 4 Sep 1916, Tracy 4756 (UC); Mad River gravel bar at Essex, 11 Jun 1924, Tracy 6695 (UC); Bull Creek to Mattole Road, near ridge summit, 4 Jul 1930, Tracy 8829 (DS,JEPS,LL,UC); Trinity River near Willow Creek, 17 May 1931, Tracy 9433 (UC); Garberville, 18 Jul 188 PHYTOLOGIA volume 72(3):157-208 March 1992 _ 1942, Tracy 17284 (UC); Blue Slide on Van Duzen River, 24 Jul 1951, Tracy | II 19625 (UC). Mendocino Co.: near Mendocino, May 1898, Brown 786 (MO); Fort Bragg, 1914, Mathews s.n. (JEPS); Fort Bragg, 8 Jul 1933, Stahelin s.n. (UC). These plants have been previously identified, for the most part, as Erigeron foltosus var. hartwegii, but the two are geographically and morphologically distinct. 8. Erigeron inornatus (A. Gray) A. Gray Erigeron tnornatus (A. Gray) A. Gray, Proc. Amer. Acad. Arts 16:88. 1881. BASIONYM: Erigeron foltosus Nutt. var. nornatus A. Gray, Bot. Calif. 1:330. 1876. LECTOTYPE (Cronquist 1947): UNITED STATES. Cali- fornia: Mendocino Co., Eel River, hot, sunny slopes, 1868-69, A. Kellogg & Harford 847 (GH!; Isolectotypes: MO!,NY,US!). Stems 2-7(-9) dm long, branched near the apex, basally decumbent but quickly erect, arising from a woody taproot, at least the upper portions glabrous. Leaves narrowly oblong-oblanceolate, (1-)2-6 cm long, 2-7 mm wide, the lam- ina glabrous or sparsely strigose, the lower usually with short, upcurved hairs, margins sparsely and widely ciliate with short, upturned hairs. Heads 7-12 mm wide, 5-many in a terminal, flat topped cluster; phyllaries glabrous, rarely minutely glandular, with broad, stramineous lateral areas and a thin, orange midvein, the inner 4.5-5.5 mm long. Ray flowers absent. Disc corollas 3.8-4.5 mm long. Achenes 2.0-3.0 mm long; pappus of 28-45(-60) bristles. Key to the varieties of E. inornatus 1. Upper stems densely hirsutulous with slightly deflexed hairs, sometimes densely invested with strigose ascending hairs. .............. var. keri 1. Upper stems usually sparsely strigose to glabrous. ..............006- (2) 2. Stems 10-20 cm long, decumbent-ascending, with spreading-deflexed hairs below the middle, appressed above the middle; middle and lower leaves with hispid lamina and margins with stiff, spreading cilia; inner phyllaries 5.0-6.5 mm long. ........... var. calidipetris 2. Stems 30-70(-90) cm long, basally ascending, the upper portions erect, with appressed hairs below the middle; middle and lower leaves with short strigose lamina and eciliate margins; inner phyl- laries) 4,525.5 mim long. .:00s Joc agate Perea eee var. tnornatus Nesom: Revision of Erigeron sect. Linearifoliu 189 a. Erigeron inornatus (A. Gray) A. Gray var. inornatus Erigeron douglasit Torr. & Gray var. eradiatus A. Gray, Pac. R.R. Rep. 12(2):52. 1860. TYPE: UNITED STATES. Washington. [Yakima Co.,] sandy pine forest on the tableland east of Mount Adams, J.G. Cooper s.n. (not seen). Hrigeron eradiatus (A. Gray) Piper, Contr. U.S. Natl. Herb. 11:568. 1906. Erigeron inornatus (A. Gray) A. Gray forma subradiatus S.F. Blake, J. Washington Acad. Sci. 19:270. 1921. TYPE: UNITED STATES. California: Placer Co., Tahoe Tavern, 1900 m, 15 Aug 1927, S.F. Blake 10802(HOLOTYPE: US!; Isotypes: GH-2 sheets!,LL-2 sheets!). Erigeron inornatus (A. Gray) A. Gray forma pseudoradiatus S.F. Blake, J. Washington Acad. Sci. 28:487. 1938. TYPE: UNITED STATES. Nevada: Washoe Co., public campground, Mt. Rose, 7 mi W of Reno Hot Springs, 1700 m, 12 Jul 1937, W.A. Archer 5580 (HOLO- TYPE: US!; Isotype: LL!). Chromosome number, n=9 (Semple 1985; Strother 1983; Anderson et al. 1974; Raven et al. 1960). Widespread in northern California (Map 2), to Oregon, southern Washing- ton, and westcentral Nevada in Douglas, Ormsby, and Washoe cos.; chaparral to oak, pine, or pine-fir woods, (150-)750-2300 m but mostly 1050-1900 m; Jun-Sep(-Oct). Erigeron inornatus has traditionally included varietal taxa recognized as E. angustatus, E. biolettiu, E. reductus, and EL. petrophilus in the present treat- ment. Erigeron petrophilus is broadly sympatric with E. tnornatus (Maps 2 and 8), and L£. angustatus is geographically and morphologically distinct. Some intermediates apparently are formed between LZ. inornatus and both E. biolettz: and E. reductus (see the latter two for further comments). Erigeron tnornatus has the geographic widest range of all the taxa included here in sect. Linearzfolii, and typical plants (with glabrous stems, leaves, and phyllaries) occur over its whole range. Plants with minutely glandular phyl- laries, however, occur sporadically through the range and are within normal variability for the species. Some of the atypical variation in the species oc- curs in the sierra of eastcentral California and adjacent Nevada. As noted by Cronquist (1947), some plants there produce phyllaries and uppermost cauline surfaces that are slightly glandular as well as phyllaries and stems that are sparsely hirsutulous or minutely strigose (e.g., Tuolumne Co., Congdon s.n.; Placer Co., Yates 5996, Washoe Co., Archer 5595). The origin of this variation is not clear, but such plants are best placed in FE. nornatus. Further comments on these plants are given following E. tnornatus var. kezltt. From the vicinity of Mt. Rose, near Reno in Washoe Co., Nevada, a col- lection of plants with very small ligules produced from hermaphroditic flowers 190 PHY TOL OGIA volume 72(3):157-208 March 1992 with abortive stamens has been named Erigeron inornatus forma pseudora- diatus. Cronquist (1947) regarded forma pseudoradzatus as a hybrid between typically radiate E. brewerz and typically rayless E. tnornatus, both of which also are known from the vicinity. Two other collections of more or less the same morphological form also were made at about the same time (Archer 6401-with longer, apparently normal rays, and 6402) from the same area (Galena Creek), although each of these collections apparently represents a different biotype. It would be useful and interesting to reassemble (primarily from DS,RENO, and UC) a set of collections from this area (mostly near Galena Creek) and to evaluate them in more detail, in conjunction with field studies. Among these collections are numerous atypical plants that can be identified most closely as E. breweri var. porphyreticus and FE. tnornatus var. tnornatus: Archer 5580, 5592, 5595, 6400, 6401, 6402, and Lehenbauer 890. Erigeron inornatus forma subradiatus, a similar variant, was collected in adjacent California (Placer Co.). Erigeron inornatus occurs in the area of Donner Lake, the type locality of E. miser, and a plant collected from there by E.L. Greene (Aug 1883-MO) has features that suggest it may be a hybrid between E. inornatus and E. miser. b. Erigeron inornatus (A. Gray) A. Gray var. calidipetris Nesom, var. nov. TYPE: UNITED STATES. California: Siskyou Co., Porcupine Lake (a small vernal depression in an indurate lava flow), T41N, R38E, Section 10, 4280 ft, moist margin of pond at high water mark, set in Abies concolor-Pinus ponderosa forest, 6 Aug 1987, D.W. Taylor 9241 (HOLOTYPE: UC!). Differt a Erigeronte inornato (A. Gray) A. Gray typico caulibus brevioribus decumbentibusque trichomatibus patentibus, foliis in- feris laminis hispidis ac marginibus ciliatis, et phyllariis longioribus. California endemic, Modoc Plateau and eastern margin of the Cascade Ranges, in Modoc, Siskyou, and Shasta cos., apparently disjunct to northern Butte Co. (Map 2); sandy habitats, commonly in pine or pine-oak woodlands, ca. 1100-1400 m; Jun-Aug. Additional collections examined: UNITED STATES. California: Butte Co.: Jonesville, 6 Aug 1930, Copeland 490 (CAS,DS,MO). Modoc Co., sandy flat in lava beds, no other data, 1893, M.S. Baker s.n. (JEPS). Siskyou Co.: McCloud, 21 Jun 1923, Bethel s.n. (CAS); Fowler Camp, McCloud River, 28 Jun 1938, Dearing 2179 (CAS); toward Black Fox, 15 Aug 1899, Dudley s.n. (DS); Bigelow’s, McCloud River, 25 Jul 1921, Eastwood 10809 (CAS). Shasta Co.: 3-4 mi E of McCloud and Bartle on Hwy 89, knob cone and yellow pine association, loose sandy soil, 10 Aug 1941, Ferris & Lorraine 10498a (CAS,DS). Nesom: Revision of Erigeron sect. Linearifolu 191 Var. calidipetris is different from typical Erigeron inornatus in vestiture and habit and has a more limited geographical range compared to the typical variety. At least in the area of McCloud, the two taxa appear to be sympatric, and the collection from Butte County is well within the range of var. tnornatus. | Ferris & Lorraine 104938 may have been collected in Siskyou County rather than Shasta, but the label says Shasta Co. With more detailed study, var. calidipetris may prove to deserve rank as a species. Further comments are given following £. petrophilus var. viscidulus (A. Gray) Nesom, to which it may be closely related. c. Erigeron inornatus (A. Gray) A. Gray var. keilii Nesom, var. nov. TYPE: UNITED STATES. California: Tulare Co., ca. 3.5 mi from Balch Park, Sequoia National Forest, dry slope with incense cedar and Arctostaphylos, 31 Jul 1950, R.S. Ferris & L. Lorraine 12849 (HOLO- TYPE: DS!; Isotypes: CAS!,UC!). A Erigeronte brewerz1 A. Gray var. brewer: floribus radii caren- tibus praecipue differt; a E. tnornato (A. Gray) A. Gray caulibus dense minute hispidulis et phyllariis dense brevistrigosis vel glan- dulosis differt. California endemic, southern Sierra Nevada Foothills of Fresno and Tulare cos. (Map 2); dry, grassy slopes and meadows, areas of conifer woodlands, 1200-2200 m; Jun-Sep. Additional collections examined: UNITED STATES. California: Fresno Co.: Sequoia Mills, 19 Jul 1892, Brandegee s.n. (UC); W edge of Lake Hume, 19 Jul 1961, Breedlove 798 (DS); Bubb’s Creek, S. Fork of King’s River, 1-13 Jul 1899, Eastwood s.n. (CAS); grade to King’s River Canyon from General Grant Natl. Park, 26 Jul 1942, Ferris & Lorraine 10882 (DS); Tehipite Valley, 6-10 Jul 1900, Hall & Chandler 486 (DS, MO); King’s River Canyon, jct of Middle and S. Fork, 29 Jun 1940, Rose 40676 (DS). Tulare Co.: NE of Springville along Bear Creek Road, Mountain Home State Forest, Methuselah Group Campground to vicinity of Dogwood Meadow, 2 Aug 1986, Keil 197153- B (TEX); NE of Springville along Bear Creek Road, Mountain Home State Forest, Methuselah Group Campground; montane mixed conifer forest around meadow, 25 Jul 1991, Keil 19662 (OBI,TEX). The collections of Erigeron inornatus var. keilii have been made from a relatively small area of Fresno and Tulare counties, near the southernmost extension of the geographic range of E. inornatus. Var. keili is taxonomi- cally positioned with the latter primarily because of its rayless heads, but in its dense, typically spreading stem vestiture, it is more like the plants of E. brewerz with which it is sympatric. The rayless plants might justifiably be treated as a variety of E. breweri on the basis of their vestiture, but among 192 PREY On O-G TA volume 72(3):157-208 March 1992 other taxa within sect. Linearifolii, the absence of ray flowers is indicative of differentiation between species. The plants of var. kezlz were first called to my attention by the recent collections of Dr. David Keil of California Polytechnic State University. Ac- cording to Dr. Keil’s collection data, the plants of 19713-B (rayless) were collected from a mixed population with both rayless and purple rayed forms (Keil 19718-A, radiate; identified here as the “tall form” of E. brewer: var. brewerz), although most populations in the area comprised only rayless forms. Variation in the presence or absence of ray flowers was not noted in the data of the other collections of var. kezlzt. The stem vestiture of var. kezlzzis dense and composed of minute, spreading hairs that vary from slightly deflexed to ascending. Among the specimens of var. kewlw, the stem hairs are ascending on Hall & Chandler 486, Ferris & Lorraine 10882, and Rose 40676; the others have spreading to slightly deflexed hairs. The phyllaries are densely short strigose and eglandular on Ferris & Lorraine 10882, Rose 40676, and the type, while the phyllaries are glandular without other hairs on Hall & Chandler 486, Breedlove 798, Eastwood s.n., and Brandegee s.n. Plants with strigose, eglandular phyllaries are designated here as typical var. keiliz, with the hypothesis that those with merely glandular, EHrigeron brewerv-like ones have acquired these features through gene flow with E. brew- ert. The EF. brewerz-like stem vestiture also supports such an hypothesis, as does the mixed population of radiate/eradiate plants noted by Keil. Addition- ally, plants in a collection of radiate E. breweri sympatric with var. ketlzi in the vicinity of Hume Lake in Fresno Co. (17 Jul 1948, Rose 48214-CAS) produce densely strigose-hispidulous phyllaries similar to those of var. kezliz, suggest- ing that the gene flow may be reciprocal. Further north, however, where var. tnornatus is more common and sympatric with var. breweri, similar hybrids have not been collected (see notes, however, following E. inornatus on pos- sible hybrids from westcentral Nevada referred to by Blake as E. inornatus var. pseudoradtatus). Occasional variants of typical E. inornatus var. inorna- tus from further north may produce a lightly strigose vestiture on stems and phyllaries but none with hairs so dense as var. keilzi. 9. Erigeron mariposanus Erigeron mariposanus Congdon, Erythea 7:185. 1900. LECTOTYPE (des- ignated here): UNITED STATES. California: Mariposa Co., Mariposa, 16 Jul 1899, J. W. Congdon s.n. (UC!; Isolectotype: DS!). Congdon cited only one collection from the single date but did not specify its place of deposition. Stems 15-28 cm long, decumbent-ascending, sometimes branched near the Nesom: Revision of Erigeron sect. Linearifolu 193 _ apex, sometimes arising from short (1-4 cm), thin, rhizomelike or caudexlike branches; stems, leaves, and phyllaries sparsely to moderately strigose with loosely appressed hairs, eglandular. Leaves oblanceolate, entire, (25-)30-45 mm long, (2-)5-8 mm wide, reduced in size near the heads, axillary tufts of small leaves prominent at most nodes. Heads 8-12 mm wide, solitary or 2-3 on pedicels 15-20 mm long; inner phyllaries 4.0-5.0 mm long, 0.8-1.0 mm wide, with broad, scarious, nearly winglike margins, the outer phyllaries sparsely strigose with thin based hairs, with a prominent orange resinous midvein. Ray flowers 18-22, the corollas 7-9 mm long, bluish. Disc corollas 4.0-4.5 mm long. Achenes ca. 2.0-2.4 mm long; pappus bristles 28-32. California endemic, central Sierra Nevada Foothills in Mariposa County (Map 9); habitat unknown, ca. 600-800 m; Jun-Aug. Additional collections examined: UNITED STATES. California: Mariposa Co., Mariposa: Congdon s.n. 30 Jun 1892 (UC, on sheet with 16 Jul 1900); Congdon s.n. 9 Aug 1892 (DS); Congdon s.n. 19 Jun 1900 (UC on sheet with lectotype); Congdon s.n. 16 Jul 1900 (UC, on sheet with 30 Jun 1892). These are very distinctive plants represented by several collections other than the type (over eight years, all made by Congdon). All are from “Mari- posa,” with no other collection data, but the present city of Mariposa is located at the southernmost margin of the range of Erigeron foliosus var. hartwegit. Erigeron mariposanus previously has been treated as a synonym of var. hartwegii, but the two are extremely different in morphology, the former immediately and most easily distinguished from var. hartwegw as well as all other varieties of FE. foliosus by its short stems, short, broadly oblanceolate leaves, tufts of axillary leaves produced at most nodes, and inner phyllaries with broad, scarious margins. Details of contrast between var. hartwegii and E. mariposanus are provided in the following couplet. 1. Stems 30-85 cm long, without axillary tufts of small leaves; leaves narrowly oblanceolate to linear lanceolate, 35-80 mm long, 1-2(-4) mm wide, 17-40 times longer than wide; phyllaries without prominent scarious margins; PAV HOMETE Ose ne aes sie ens ete a ican Saas E. foliosus var. hartwegii 1. Stems 15-20 cm long, axillary tufts of small leaves produced at most nodes; leaves oblanceolate, (25-)30-45 mm long, (2-)5-8 mm wide (at widest point), the lower and middle 5-9(-12) times longer than wide; inner phyllaries with prominent, broad, scarious margins; ray flowers RSet eee pmo mea ae cc car + eis eR ee eee. © ae E. mariposanus It may be suspected that Erigeron mariposanus occurs in some specialized habitat and that it may now even be extinct. In any case, the formal taxo- nomic recognition of these plants should stimulate an attempt to relocate the population and determine its present status. 194 PHY TOLOG fA volume 72(3):157-208 March 1992 9. Erigeron miser A. Gray Erigeron miser A. Gray, Proc. Amer. Acad. Arts 13:372. 1878. LECTO- TYPE (designated here): UNITED STATES. California: [Placer Co.], crevices of rocks at Donner Lake, 14 Oct 1874, E.L. Greene 466 (GH!). Gray also cited a collection from the summit of Mt. Stanford made in Sep 1877 by Lemmon, Gray, and Hooker; this specimen is mounted on the same sheet as Greene 466. Stems decumbent-ascending to ascending-erect, 5-25 cm long, slender, woody, and caudexlike basally, arising from a woody root, the stems and leaves vil- lous with stiff, white hairs, also minutely glandular. Leaves narrowly obovate, 7-16 mm long, 1.0-3.5 mm wide. Heads 7-12 mm wide; phyllaries densely granular-glandular, without other vestiture, the inner (3.5-)4.0-5.0 mm long. Ray flowers absent. Disc corollas 3.2-4.5 mm long. Achenes 2.0-2.5 mm long; pappus bristles 18-25(-28). , California endemic, northern High Sierra Nevada (Map 8); talus, rock | crevices, 1950-2290 m; Jul-Oct. See comments following Erigeron petrophilus and E. inornatus. 10. Erigeron oryphyllus E. Greene Erigeron ozyphyllus E. Greene, Erythea 3:20. 1895. LECTOTYPE (des- | ignated here): UNITED STATES. Arizona, Mohave Co., Yucca, May | 1884, M.E. Jones s.n. (US!; Isolectotype: PH). Stems shiny glabrous, 5-25 dm tall, sharply ascending basally from short, caudexlike branches, numerous, arising from a thick woody taproot. Leaves sparsely and minutely strigose, eglandular, filiform to linear oblanceolate, 0.5- 1.0(-2.0) mm wide, the lower 2-5(-10) cm long, quickly reduced in length up- wards, widely separated and not at all overlapping. Heads 7-12 mm wide, mostly solitary on long, bracteate peduncles, less commonly the peduncles shorter and the heads in loose clusters of 2-3; phyllaries granular-glandular, without other vestiture or the outer sometimes sparsely and minutely strigose, the inner 4-5(-6) mm long, with (1-)3 orange veins. Ray flowers 12-27(-40), the corollas 6-9 mm long, drying bluish. Disc corollas 3.0-4.0 mm long. Achenes ca. 2.0 mm long, with 2-4(-5) orange-resinous nerves; pappus bristles 17-25. Chromosome number, n=9 (Pinkava & Keil 1977). Southern Arizona in Maricopa, Mohave, Pinal, and Yuma cos., adjacent northwestern Sonora, México (not mapped); rocky hillsides around seeps or by streams, often with Acacza- Yucca-Coleogyne, 700-1100 m; (Feb-)May-Jun, Oct-Nov. This species was included by Nesom (1989) in Erigeronsect. Spinost (Alexan- der) Nesom & Sundberg on the basis of its putatively close relationship with Nesom: Revision of Erigeron sect. Linearzfoli ee les E. petrophilus var. petrophilus e Ae eee @ Ayay ee i SNe Ye © var. sierrensis . .e A. @ i © var. viscidulus =) J} ( & EF. miser peels ia a Een EE ~ ty ae ait Map 8. Erigeron petrophilus and E. miser. 196 BAY POLO GIA volume 72(3):157-208 March 1992. E. ortegae S.F. Blake (= Aster spinosus Benth.). The latter species has now. been segregated as the monotypic genus Chloracantha Nesom et al., appar-| ently more closely related to the genus Boltonia L’Herit. and even Heterotheca Cass. than to either Erigeron L. or Aster L. (Nesom et al. 1991). The posi- tion of EF. bye: Sundberg & Nesom, which was also included in this section, remains equivocal (Sundberg & Nesom 1990) but apparently it is not close to E. ozyphyllus. Cronquist (1947) did not recognize the close relationship of Erigeron ozy- phyllus to the E. foliosus group but suggested instead a relationship with E£. arenartoides (A. Gray) Rydb. The latter, however, produces persistent basal leaves and caudex branches with densely clustered, persistent petiole bases and its ray corollas do not coil at maturity; it has been placed with one other species in Erigeron sect. Arenartozdes (Rydb.) Nesom (Nesom 1989). Among the other taxa of sect. Linearifolu, Erigeron ozyphyllus is most similar in habit and vestiture to E. serpentinus, but each of these probably has been independently derived from the ancestral stock that produced the | widespread and variable F. foliosus. Erigeron oryphyllus differs from all other | species of the section in its cauline leaves that are sharply reduced in length | upwards, and it is geographically set apart as well from the rest of the E. foliosus group. 11. Erigeron petrophilus E. Greene Erigeron petrophilus E. Greene, Pittonia 1:218. 1888. LECTOTYPE (des- ignated here): UNITED STATES. California: Contra Costa Co., high rocks above Wild Cat Creek, near Berkeley, Aug 1887, E.L. Greene s.n. (ND-G!; Isolectotype: US!). Among several collections of this species at ND-G, all identified by Greene, this sheet is the earliest one collected. It is likely that the date of ”1881” cited in the protologue resulted from a typographical error for “1887.” Cronquist (1947) suggested that this collection might be the type but he cited only the US sheet. Stems decumbent-ascending, sometimes described as “sprawling,” mostly 1-2 dm long, (up to 3 dm and nearly erect in Monterey Co.), from rhizome- like bases, at least the upper stems and leaves glandular (eglandular or very sparsely glandular in var. visczdulus, see comments below), sparsely to densely villous with loose, thin, white hairs. Leaves narrowly oblong to oblanceolate, 10-25 mm long, 1-5 mm wide, usually with long, spreading cilia along the prox- imal margins. Heads 8-12 mm wide, (1-)2-5(-10) in a definite corymb produced at the branch apex; phyllaries glandular, uncommonly with a few, spreading nonglandular hairs, the inner 5.5-7.0(-8.0) mm long, the apices sometimes pur- ple, sometimes loose. Ray flowers absent. Disc corollas 4-6 mm long. Achenes (2.0-)2.5-3.0 mm long; pappus of 22-30(-35) bristles. Nesom: Revision of Erigeron sect. Linearifoli 197 Erigeron petrophilus is distinguished by its ascending stems arising from distinct, fibrous rooted, rhizomelike bases, its few headed but generally com- _ pact, corymboid capitulescences, absence of ray flowers, and by its distinctive vestiture. Vars. petrophilus and sierrensis Nesom produce a densely glandular vestiture, and at least the stems are invested with loose, white, spreading, _nonglandular hairs variable in density. The observations by Cronquist (1947) regarding the distinction between Erigeron petrophilus and E. miser remain generally justifiable. Hrigeron miser is a narrow endemic at high elevations, in contrast to the relatively wide geographic distribution at lower elevations of E. petrophilus. The former is distinguished morphologically primarily by its smaller heads, but there are also overlapping differences in leaf size, disc corolla length, shape of the style appendages, and the number of pappus bristles. Erigeron miser might also be treated as taxonomically coordinate with the three varieties of E. petrophilus, particularly as it terminates the eastern “arm” of the distribution of E. petrophilus, where it is situated as a fourth, closely related, allopatric population system (Map 8). Key to the varieties of E. petrophilus 1. Stems and leaves eglandular or very obscurely glandular, the eglandular Hatrsymostly stilt, straight. OF CUTVED.). « ori0.0.6.00 sie e s.sine var. visctdulus 1. Stems and leaves densely glandular, the eglandular hairs usually loose, CANEEE, SSE “ayetetevendc cnfieyctepecic anc Peal) a er aoe eA (2) 2. Phyllaries with a distinctly expanded, purplish area at the apex. .. ee re ere oi Moe cotes ee sari Nec swell es var. petrophilus 2. Phyllary apices not purplish and differently colored from the lower FRCS ACRE ers 5 Se Ryo aia al ayo wae hae wid Sahl ores Wd ios var. sterrensts a. Erigeron petrophilus E. Greene var. petrophilus Chromosome number, n=9 (Strother 1983, reported as E. miser). California endemic, South to North Coast Ranges, San Francisco Bay area (Map 8); rocky slopes, crevices, and talus, (530-)690-2100 m; May-Sep. Some of the plants from the southernmost area of the range of var. petrophi- lus, particularly in Monterey and Santa Cruz counties, produce atypically tall and erect stems and are superficially similar to EF. bioletti, but most others in the same area are shorter. A few collections are densely glandular but completely lack eglandular hairs (e.g., Marin Co., Elmer 4649-CAS; San | \ 198 PHY TOLOGIA volume 72(3):157-208 March 1992) Benito Co., Eastwood 6748-CAS); these are superficially like E. bzolett, but they produce relatively short stems, and smal] heads and leaves, and clearly | belong with EF. petrophilus. | | | b. Erigeron petrophilus E. Greene var. sierrensis Nesom, var. nov. TYPE: UNITED STATES. California: Yuba Co., Peterson Ridge Road, ca. 4 mi SE of Challenge, yellow pine forest, dry, red, rocky soil, 1000 m, 10 Aug 1983, L. Ahart 4272 (HOLOTYPE: TEX!; Isotype: MO!). Differt a Erigeronte petrophilo E. Greene var. petrophilo prae- cipue phyllariis absque apicibus purpuratis et habitationibus in sierris. California endemic, northern Sierra Nevada Foothills (Map 8); ledges, other rocky sites in areas of pine and pine oak woods, most commonly on serpentine, | 360-1910 m; Jul-Aug. Additional collections examined: UNITED STATES. California: Butte — Co.: Forbestown Res., ca. 5 mi NE of Forbestown, 24 Jul 1983, Ahart 4206 (MO,TEX); near Pulga, 26 Jun 1951, Howell 27525 (CAS); 2 mi W of Pulga, | 19 Aug 1951, Howell 28161(CAS). El Dorado Co.: Junction Dam, W of Union | Valley Reservoir, 20 Jul 1968, Stebbins 6805(CAS); foot of Jaybird Falls, above — Jaybird Power House on canyon of Silver Creek, 20 Jul 1968, Stebbins 6820 | (CAS). Plumas Co.: Jamison Creek, 27 Jun 1951, Howell 27625 (CAS); ca. 20 mi SW of Quincy, 1/4 mi E of Frenchman Hill, 27 Sep 1980, Taylor et al. 3346A (MO); Red Hill Lookout, ca. 3 air mi NE of Belden, 25 Aug 1981, Taylor 4238 (MO); Rocky Point, head of No Ear Bar Trail, Middle Fork Feather River, among rocks in red clay, yellow pine forest with Quercus kelloggii, 21 Jul 1955, Weatherby 1650 (UC). c. Erigeron petrophilus E. Greene var. viscidulus (A. Gray) Nesom, comb. nov. BASIONYM: Erzgeron inornatus (A. Gray) A. Gray var. viscidulus A. Gray, Syn. Fl. 1(2):215. 1884. LECTOTYPE (designated here): UNITED STATES. California: [Siskyou Co.?], mountains about the headwaters of the Sacramento River, 1 Sep 1882, C.G. Pringle 250 [or 14579| (GH!; Isolectotypes: MO!,US). Erigeron viscidulus (A. Gray) E. Greene, Pittonia 1:174. 1888. Cronquist (1947) selected the Pringle collection over another syntype (Kellogg & Harford s.n.), but he did not choose between the GH and US specimens. Erigeron decumbens Eastw., Bot. Gaz. (Crawfordsville) 41:290. 1906; not Nutt. 1840. TYPE: UNITED STATES. California: Siskyou Co., Mt. Eddy, 1400 m, 17 Aug 1903, E.B. Copeland s.n. (HOLO- TYPE:-CAS!): Nesom: Revision of Erigeron sect. Linearifoltu 199 Klamath Ranges of California and immediately adjacent Oregon (Map 8); - open, rocky slopes, ledges, talus, sometimes on serpentine, 1490-2650 m; Jul- Sep. Representative collections examined: UNITED STATES. California: Trin- ity Co.: Lower Canyon Lake, 5 Aug 1948, Alezander 5400 (UC); summit of Scott Mtn., 2 Sep 1958, Bactgalup: 6857 (JEPS); near top of Granite Peak, 8 Aug 1926, Baker 267a (UC); Trinity Alps, above Ward Lake, 19 Aug 1966, Ferlatte 450 (JEPS); Siligo Peak-Summit Lake area, Dear Creek Pass, 16 Aug 1974, Ferlatte 1660 (JEPS); summits near Dorleska, Jul 1909, Hall 8588 (UC); Potato Mt., W slope, 2 Aug 1935, Tracy 14487 (UC); Devil’s Canyon Mts., head of White’s Creek, 6 Aug 1935, Tracy 14579 (JEPS,UC). Siskyou Co.: near Toad, S Siskyou County, 12 Aug 1911, Alezander & Kellogg 307 (UC); Mt. Eddy, 30 Aug 1912, Eastwood 2034 (UC); E side of Mt. Eddy, 28 Aug 1914, Heller 11787(UC); Scott Mts., Blue Point near Scott Mt. Public Campground, 8-9 Aug 1953, Kellogg 123 (UC); side of Mt. Shasta, Jul [no year], Lemmon s.n. (UC); Caribou Lake, Salmon-Trinity Alps Primitive Area, 27 Jul 1955, Wiggins 18560 (UC). Oregon: Jackson Co.: Mt. Ashland, S of Ashland, 29 Jun 1986, Ertter 6457 (MO,UC); serpentine slopes of Red Mt., Siskyou Mts., 29 Jul 1935, Thompson 12374 (CAS,MO,UC). A collection of typical var. viscidulus by J.G. Lemmon (s.n. - JEPS,UC) from “Sierra Valley” is almost certainly mislabeled if the locality is meant to be from Sierra County. The cauline vestiture of var. viscidulus is variable, comprising stiff hairs that primarily are spreading to slightly deflexed hairs but that vary to antrorsely appressed. Plants of this taxon with spreading vestiture were included by Cronquist (1947) within Erigeron petrophilus, and some (especially in Trin- ity Co.) even approach typical FE. petrophilus in the length, looseness, and density of the nonglanduler hairs, although none produce the corresponding dense glandularity. Plants with ascending stem hairs, including the lectotypic Pringle collection, were maintained by Cronquist as £. inornatus var. visctdu- lus. Part of the origin of the appressed stem hairs is perhaps found in the area of Scott Mountain (Trinity Co.), where apparent intermediates between var. viscidulus and E. reductus var. reductus occur. These plants seem to grade into typical var. viscidulus (with spreading vestiture). However, among nu- merous collections examined from Mt. Eddy (Siskyou Co.), well away from known populations of E. reductus, are plants with spreading stem hairs as well as others with appressed hairs. The morphology of such plants of var. visctdu- lus approaches that of E. tnornatus and probably has provided the basis for considering the two taxa conspecific. The two, however, are sympatric and occur in generally the same kinds of habitats. While they are best regarded as separate species, it is possible that gene flow is occurring between them. Further, the variability accepted here within var. viscidulus leaves open the possibility that it may be polyphyletic in origin. 200 PHY MOU OuGun A volume 72(3):157-208 March 1992 12. Erigeron reductus (Crong.) Nesom Erigeron reductus (Cronq.) Nesom, comb. et stat. nov. BASIONYM: Erigeron inornatus (A. Gray) A. Gray var. reductus Cronq., Brittonia 6:288. 1947. TYPE: UNITED STATES. California: Placer Co., on rocks, Yuba River near Cisco, 5600 ft, Transition Zone, 3 Jul 1910, H.M. Hall 8741 (HOLOTYPE: UC!; Isotype: US!). Stems 8-20(-30) cm tall, basally erect or ascending, few, arising from long, slender, rhizomelike caudex branches, glabrous. Leaves linear, 8-25(-35) mm long, 0.5-1.0(-1.5) mm wide, glabrous or sparsely strigose on the lamina, the margins prominently ascending ciliate with thin based hairs. Heads 8-10 mm wide, solitary or 2-4 on short peduncles; phyllaries minutely but prominently granular-glandular, without other hairs, the inner 4.0-6.0 mm long. Ray flow- ers absent. Disc corollas 3.5-4.5 mm long. Achenes 2.2-2.8 mm long; pappus bristles 20-61. Erigeron reductus is composed of two allopatric but very similar entities united by their distinctive habit, highly reduced vestiture, and rayless heads. Both taxa have previously been treated as varieties of FE. inornatus, but there is evidence that FE. wnornatus and var. reductus intergrade in Trinity Co. (see comments below). Cronquist (1947) also recognized var. reductus and var. angustatus as most closely related to each other, although his concept of the latter was broader than that here (see comment below). As treated here, E. reductus has a geographic distribution similar to that of E. petrophilus, forming a northern “hook” around the Sacramento Valley (Map 9). Key to the varieties of E. reductus 1, Phyllary apices purplish (variably so in the Tamalpais region); pappus bsistless(382)46-6i1), 024) Ae aE AP, SR ee var. angustatus 1. Phyllary apices greenish; pappus bristles 20-30. ............ var. reductus a. Erigeron reductus (Cronq.) Nesom var. angustatus (A. Gray) Nesom Erigeron reductus (Cronq.) Nesom var. angustatus (A. Gray) Nesom, comb. nov. BASIONYM: Erigeron inornatus (A. Gray) A. Gray var. angustatus A. Gray, Syn. Fl. 1(2):215. 1884. LECTOTYPE (Cronquist 1947): UNITED STATES. California: Mendocino Co., Red Mountain, 2 Jul 1869, A. Kellogg & Harford 899 (GH!; Isolectotype: US!). Cronquist noted that the CAS sheet of Kellogg & Harford 399 bears a specimen of E. broletti. In the original description, Gray also cited “Greene, Napa Co.” (Greene’s no. 339 at GH), but this is EZ. angustatus, the species. Nesom: Revision of Erigeron sect. Linearifolu A E. mariposanus E. reductus @ var. angustatus O var. reductus Map 9. Erigeron reductus and E. mariposanus. 201 202 PAE YT OTE OsG: IGA volume 72(3):157-208 March 1992 California endemic, San Francisco Bay area, Outer and High North Coast Ranges (Map 9); rocky sites, commonly on serpentine, often in pine or pine-oak woods, 600-1360 m; Jun-Aug. Specimens examined: UNITED STATES. California: Marin Co.: Carson Ridge summit at head of Big Carson Canyon, 16 Jul 1943, Bacigalup: 2772 (DS); Westpoint Road, Tamalpais, Jul 1907, K. Brandegee s.n. (UC); Mt. Tamalpais, 2 Aug 1896, Eastwood s.n. (JEPS,UC), Sep 1898, Eastwood s.n. (UC), 29 Jul 1912, Eastwood 1518 (CAS,MO,UC); burned area S of Barth’s Retreat, Mt. Tamalpais, 7 Jul 1946, Howell 22197 (CAS,DS,UC); N end of Tiburon Peninsula, above Vista de la Bahia, 18 Jun 1972, Howell 48875 (CAS); El Campo, Jul 1891, Michener & Bioletti 11388a(ND-G); S side of Mt. Tamal- pais, 14 Jul 1913, Suksdorf 538 (MO). Mendocino Co.: Little Red Mt., 6 Aug 1932, Tracy 10806 (JEPS,UC), 7 Aug 1932, Tracy 10816 (UC); Little Red Mt., 6 Aug 1932, Jepson 16504 (JEPS) and 16502 (JEPS); Red Mt. near Ukiah, 24 Jul 1909, McMurphy s.n. (DS); Red Mt. N, 1 Jul 1982, Smith & Wheeler 7563 (CAS) and 7576 (CAS). San Mateo Co.: Crystal Springs Lake, 9 Nov 1930, Howell s.n. (CAS). Sonoma Co.: 3 mi N of Occidental on Bohemian Hwy, 5 Jun 1981, Best s.n. (CAS); Monte Rio, Aug 1920, Campbell s.n. (CAS); Pepperwood Ranch, Franz Valley Road, 6 Jun 1981, DeNevers 1451 (CAS); Pepperwood Ranch, Dentist’s Lake, 20 Jun 1981, DeNevers 1684 (CAS; 2 mi E of Occidental, 5 Jul 1946, Mason & Grant 12898 (DS); unnamed ridge on W side of Dry Creek Valley, N of Crane Creek, S of Bradford Mt., 30 Jun 1985, Raiche 50565 (CAS). Tehama Co.: serpentine slope on NW side of Tedoc Mt., 17 Jun 1972, Heckard 2962 (JEPS). In Cronquist’s view (1947), this taxon included Erigeron angustatus and was subsumed as a variety of E. tnornatus. Erigeron angustatus (the species) is different in habit from E. reductus and has a narrower geographic range. b. Erigeron reductus (Cronq.) Nesom var. reductus California endemic, southern Klamath Ranges, northern High Sierra Nevada (Map 9); crevices and open, rocky sites, commonly on serpentine, 760-1940 (-2360) m; Jun-Aug. Specimens examined: UNITED STATES. California: Nevada Co.: 5.6 mi E of Emigrant Gap on US Hwy 40, 30 Jun 1940, Beach 800(CAS,JEPS,LL,UC); above Meadow Lake, 7 Aug 1892, Sonne 5 (UC, on sheet with FE. miser). Plumas Co.: slopes of Mt. Elwell above Long Lake, 10 Jul 1927, Bacigalupi 1688 (DS); Feather River region, Summit Peak above Round Lake, 15 Jul 1920, Head s.n. (CAS); Feather River region, Lake Center Camp, 15 Jul 1921, Head s.n. (CAS); Jamison Lake, 5 Sep 1932, Rose 82621 (CAS). Shasta Co.: Lamoine, 16 Jul 1910, Blankinship s.n. (JEPS). Siskyou Co.: Trinity Mts., road to Mumbo Basin, just below summit, 29 Jul 1967, McClintock s.n. (CAS). Nesom: Revision of Erigeron sect. Linearifoli 203 1955, Barbe 936 (UC); N of Coffee Creek, 16 Aug 1988, Dean 175 (UC); SW of Peanut, 17 Jul 1955, Howell 30701 (CAS); 1 mi S of Scott Mt. Summit, 20 Jul 1949, Keck 6377 (DS-2 sheets); head of Rush Creek, 20 Jul 1914, Yates 541 (JEPS). In Trinity County, plants of some collections of Erigeron reductus appear to vary toward E. inornatus (and vice versa), but identifications can generally be made on the basis of the habit. The stems of EZ. reductus usually arise singly from slender, rhizomelike caudex branches, while the stems of E. tnornatus are more numerous and arise from a woodier base, which is rarely collected. The plants in the northern segment of var. reductus range somewhat taller than typical, perhaps reflecting genetic input from EL. inornatus. Trinity Co.: 2 mi S of Scott Mtn Forest Camp, on Scott Mtn. Road, 3 Aug 13. Erigeron serpentinus Nesom Erigeron serpentinus Nesom, sp. nov. TYPE: UNITED STATES. Cali- fornia: Sonoma Co., Layton Mine, Austin Creek, serpentine soil, 30 May 1947, F.W. Hoffman 558 (HOLOTYPE: UC!). Differt a Erigeronte angustato E. Greene floribus radii caren- tibus et capitulis minoribus. Differt a Erigeronte folioso Nutt. var. franciscensi Nesom ac var. mendocino (E. Greene) Nesom caulibus numerosis ad apicem radicis statim orientibus, caulibus ac foliis glabris, et floribus radii paucioribus. Stems 4-5 dm tall, erect, numerous, arising directly from the crown atop a woody taproot. Stems and leaves mostly glabrous, leaves sparsely ascending ciliate with thin based hairs. Leaves linear, 2-4 cm long, 0.6-0.8 mm wide. Heads 9-12 mm wide; phyllaries minutely granular-glandular, with very few, minute, appressed non-glandular hairs, the inner 4.5-5.0 mm long. Ray flowers 9-13, the corollas 7-8 mm long, 1.5-2.5 mm wide, drying blue. Disc corollas 3.2-4.0 mm long. Mature achenes not seen; pappus bristles 26-32. California endemic, North Coast Ranges in Sonoma Oo., vicinity of Layton Mine along Austin Creek (Map 1); shrubby vegetation over serpentine, ca. 400- 600 m; May-Aug. Additional collection examined: UNITED STATES. California: Sonoma Co., near seep above Austin Creek, (below Layton Mine buildings), the Cedars, occasional in chaparral on serpentine, usually close to moisture, 6 Aug 1983, Ratche 30587 (JEPS). Erigeron serpentinus apparently is a narrow endemic restricted to serpen- tine outcrops in central Sonoma County. It is similar in habit and vestiture to E. angustatus, and probably most closely related to it, but plants of the latter 204 PRY BOG OG IA volume 72(3):157-208 March 1992 are rayless and they have larger heads and longer disc corollas. Erigeron an- gustatus also is endemic to serpentine exposures of the same region, although it is somewhat more widespread (Map 1). The collection information for E. serpentinus (30587) notes that the plants “apparently spread by rhizomes,” but the specimens do not show this. Instead, the numerous stems on the specimens collected arise densely and immediately erect from the crown. Among the radiate taxa with glabrous to sparse and appressed stem vesti- ture, the geographical range of Erigeron serpentinus occupies a region between that of E. foliosus var. franciscensis and var. mendocinus. It is strikingly differ- ent from both of the latter in vestiture as well as head morphology; scattered individuals of var. foliosus approach EF. serpentinus in habit, or in vestiture, or in the reduced number of ray flowers, but I have seen no specimen of var. foltosus that combines the distinctive features of E. serpentinus. SPECIES PERIPHERAL TO SECT. LINEARIFOLII In contrast to the group of taxa closely centered around Erigeron folzosus, the other species that most closely resemble them differ in significant features. These latter were included with the FE. foliosus group in Erigeron sect. Linear- ifolii (Nesom 1989), but after detailed study of the typical part of the section, it is more difficult to retain the peripheral elements together with the typical. The former are included and briefly treated here, as they may prove to be closely related, but it would not be surprising if they were found to repre- sent, even among themselves, a diverse group of taxa most closely related to disparate phyletic lines within the genus. 1. Erigeron hyssopifolius Michx. Erigeron hyssopifolius Michx., Fl. Bor. Amer. 2:123. 1803. TYPE: CANADA. Quebec: Lake Mistassini, 4. Michauz s.n. (GH photo!, UC photo, US photo!). Aster graminifolius Banks ez Pursh, Fl. Amer. Septent. 2:545. 1814. TYPE: [Canada]. at Hudson’s Bay, Aug, Sep, [year not specified], Dr. Richardson s.n. (“herb. Banks,” not seen ). Galatella gramini- foha (Banks ez Pursh) W.J. Hook., Fl. Bor. Amer. 2:15. 1834. Southeastern Canada and adjacent northeastern United States. Chro- mosome number, n=9 (Montgomery & Yang 1960; Gervais 1979; Love & Love 1980; Morton 1981; Semple 1980; Love & Love 1982; Chinnappa & Chmielewski 1987). Two varieties have been recognized within Erigeron hyssoptfolius, but | have not attempted to evaluate their status. Cronquist (1947) included the Nesom: Revision of Erigeron sect. Linearifolii 205 species within sect. Pycnophyllum (the E. foliosus group) on account of its nar- row leaves densely arranged on the stems, occasionally resembling FE. foliosus var. confinis, but as he noted (p. 277), “it certainly is not approached by any- thing else in the section.” It differs strongly from the species centered around E. folosus in several features: cauline leaves abruptly and sharply reduced in size, the heads solitary on bracteate peduncles 5-10 cm above the level of the leaves; 3 nerved phyllaries all nearly equal in length; and ligules of the ray corollas not coiling. Should E. hyssopifolius prove to be only distantly related to the E. foltosus group, Erigeron sect. Linearifoli probably would comprise only E. hyssopifolius, and sect. Pycnophyllum would be the correct name for the numerous taxa of the E. foliosus group. 2. Erigeron Ee dapedus (B. Rob. & Fern.) Nesom Erigeron lepidopodus (B. Rob. & Fern.) Nesom, Sida 9:31. 1981. BASIONYM: Aster lepidopodus B. Rob. & Fern., Proc. Amer. Acad. Arts 30:117. 1894. TYPE: MEXICO. Chihuahua: Pine forests about Chuchuichupa, 14 Jun 1891, C.V. Hartman 697 (HOLOTYPE: GH!; Isotypes: US-2 sheets!). Central Chihuahua, México. Chromosome number, n=9 (R.W. Spellenberg unpubl.). Erigeron lepidopodus and E. rhizomatus Cronq. are closely similar in geo- graphic range and morphology, and they are almost certainly related as sister species. Plants of both produce stems from rhizomelike caudex branches, with- out clustered basal leaves, but both produce long, noncoiling ray corollas and 'phyllaries that are nearly equal in length, unlike any taxa of the E. foliosus group. Cronquist (1947) placed E. rhizomatus as a member of Erigeron sect. Wyomuingia (A. Nelson) Crongq. 3. Erigeron rhizomatus Cronq. Erigeron rhizomatus Crongq., Brittonia 6:274. 1947. TYPE: UNITED STATES. New Mexico: McKinley Co., canyon S of Fort Wingate, 16 May 1943, Ripley & Barneby 5272 (HOLOTYPE: NY!). Southwestern New Mexico, United States. Chromosome number, n=9 (Ward & Spellenberg 1986). 4. Erigeron chiangti Nesom Erigeron chiangii Nesom, Sida 8:65. 1979. TYPE: MEXICO. Coahuila: Canon de la Madera, Sierra de la Madera, 29 Mar 1975, T. Wendt & E. Lott 842 (HOLOTYPE: LL!). 206 PHY POwmseCisA volume 72(3):157-208 March 1992 Central to northern Coahuila, México. Chromosome number, n=18 (Ne- som 1979). The relationships of Erigeron chiangu and E. scoparioides Nesom were not clearly understood at the times of their separate publications, but it now appears that they may be sister species. Both have a habit generally charac- teristic of the E. foliosus group, although the basal leaves of E. chiang are sometimes persistent. The latter was originally compared to the E. foliosus group, but both it and E. scoparioides produce features anomalous among those western species. Their potential relationship, however, to any more ge- ographically proximal species is equally as obscure. 5. Erigeron scoparioides Nesom Erigeron scoparioides Nesom, Phytologia 66:488. 1989. TYPE: MEXICO. Tamaulipas: Mpio. Bustamante, ca. 1 mi NW of Hwy 101 on road to Bus- tamante, pine-juniper woodlands, 1680 m, 2 Aug 1983, G. Nesom 4785 (HOLOTYPE: TEX!; Isotypes: ANSM!,CAS!,GH!,MEXU!,MICH!,NY!, US!). Westcentral Tamaulipas, Mexico. ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. J.L. Strother for their comments and review of the manuscript and the staffs of CAS,DS,JEPS,ND-G,NY, and UC for loans of specimens. Some specimens at GH,NY, and US were studied on visits there, and all of the Erigeron foliosus group at MO have been so studied and annotated. Barbara Hellenthal’s comments and help in locating specimens in the E.L. Greene Herbarium (ND-G) have been invaluable, and Dr. David Boufford confirmed the status of some of the type specimens at GH. Dr. David Keil’s review of the treatment of Erigeron for the “Jepson Flora” has also been helpful in making the present study more internally consistent in some respects. Dr. Arthur Cronquist read an early version of this manuscript (April 1991) and discussed some of its aspects with me, for which I am grateful. This paper is dedicated to his memory. LITERATURE CITED Anderson, L.C., D.W. Kyhos, T. Mosquin, A.M. Powell, & P.H. Raven. 1974. Chromosome numbers in Compositae. X. Haplopappus and other Aster- eae. Amer. J. Bot. 61:665-671. Nesom: Revision of Erigeron sect. Linearifolu 207 Chinnappa, C.C. & J.G. Chmielewski. 1987. Documented plant chromosome numbers 1987. 1. Miscellaneous counts from western North America. Sida 12:409-417. Compton, G. 1934. A revisional study of the species Erigeron foliosus Nutt. Bull. So. Calif. Acad. Sci. 33:50-54. Cronquist, A. 1947. Revision of the North American species of Erigeron, north of Mexico. Brittonia 6:121-302. Gervais, C. 1979. Liste annotée de nombres chromosomiques de la flora vasculaire du nord-est de |’Amerique. Naturaliste Canad. 106:451-461. Greene, E.L. 1894. Manual of Plants of the San Francisco Bay Region. Cubery & Company, San Francisco, California. Hickman, J.C. (ed.). 1989. Introduction to the Jepson Manual. Jepson Herbarium and Library, University of California at Berkeley, California. Jepson, W.L. 1933. David Douglas in California. Madrono 2:97-100. Keil, D.J., M.A. Luckow, & D.J. Pinkava. 1988. Chromosome studies in Asteraceae from the United States, Mexico, the West Indies, and South America. Amer. J. Bot. 75:652-668. Love, A. & D. Love. 1980. In IOPB chromosome number reports LXIX. Taxon 29:707-709. . 1982. In IOPB chromosome number reports LXXV. Taxon 31:344-360. Montgomery, F.H. & S-J. Yang. 1960. Cytological studies in the genus Erigeron. Canad. J. Bot. 38:381-386. Morton, J.K. 1981. Chromosome numbers in Compositae from Canada and the U.S.A. Bot. J. Linn. Soc. 82:357-368. Nesom, G.L. 1979. A new species of Erigeron (Compositae) from Coahuila. Sida 8:65-70. . 1989. Infrageneric taxonomy of New World Erigeron. Phytologia 67:67-93. . 1990. Taxonomy of Erigeron bellidiastrum (Asteraceae: Astereae), with a new variety. Phytologia 69:163-168. 208 PEPYS ROL OcGAA volume 72(3):157-208 March 1992 —______, Y. Suh, D.R. Morgan, S.D. Sundberg, & B.B. Simpson. 1991. Chlo- racantha, a new genus of American Astereae (Asteraceae). Phytologia 70:371-380. Pinkava, D.J. & D.J. Keil. 1977. Chromosome counts of Compositae from the United States and Mexico. Amer. J. Bot. 64:680-686. Raven, P.H., O.T. Solbrig, D.W. Kyhos, & R. Snow. 1960. Chromosome numbers in Compositae. I. Astereae. Amer. J. Bot. 47:124-132. Semple, J.C. 1980. In IOPB Chromosome number reports LXVII. Taxon 29:357-358. 1985. Chromosome number determinations in fam. Compositae tribe Astereae. Rhodora 87:517-527. —_______, J.G. Chmiewleski, & M.A. Lane. 1989. Chromosome number de- terminations in fam. Compositae, tribe Astereae. III. Additional counts and comments on generic limits and ancestral base numbers. Rhodora 91:296-314. Strother, J.L. 1972. Chromosome studies in western North American Com- positae. Amer. J. Bot. 59:242-247. 1983. More chromosome studies in Compositae. Amer. J. Bot. 70:1217-1224. & W.J. Ferlatte. 1989. Review of Erigeron eatoni and allied taxa (Compositae: Astereae). Madrono 35:77-91. Solbrig, O.T., L-C. Anderson, D.W. Kyhos, & P.H. Raven. 1969. Chromo- some numbers in Compositae. VIII. Astereae III. Amer. J. Bot. 56:348- 353. Solbrig, O.T., L.C. Anderson, D.W. Kyhos, P.H. Raven, & L. Rudenberg. 1964. Chromosome numbers in Compositae. V. Astereae IJ. Amer. J. Bot. 51:513-519. Sundberg, S.D. & A.G. Jones. 1987. Loudon’s Hortus Brittanicus (1830): An early source of sectional names, necessitating nomenclatural changes in many genera—Aster: A case in point. Taxon 36:97-98. Sundberg, S.D. & G.L. Nesom. 1990. A new species of Erigeron (Asteraceae: Astereae) from Chihuahua, México. Phytologia 69:278-281. Ward, D.E. & R.W. Spellenberg. 1986. Chromosome counts of angiosperms of western North America. Phytologia 61:119-125. Phytologia (March 1992) 72(3):209-230. A NEW SPECIES OF CASTILLEJA (SCROPHULARIACEAE) FROM SOUTHCENTRAL TEXAS WITH COMMENTS ON OTHER TEXAS TAXA Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT Castilleja genevievana sp. nov. is described from Crockett, Pecos, Upton, and Val Verde cos., Texas, and one closely adjacent locality in Coahuila, México. It is most closely related to C. integra and C. purpurea var. citrina but is allopatric with both and morphologically distinctive in its entire leaves, unbranched trichomes, and yellow flo- ral bracts and calyces. Castilleja wootonti (the type from southeastern New Mexico) is the correct name for the species in Jeff Davis Co., Texas, previously known as C. ciliata and assumed to be a narrow en- demic. Castilleja elongata, which has been recognized primarily from Brewster Co., Texas, and a proposed but yet unpublished species en- demic to Jeff Davis Co. are interpreted as merely intergrading elements of the widespread and variable C. integra. Castilleja latebracteata (the type from Texas) is a synonym of Castilleja rigida (the type from Chi- huahua), a species primarily distributed in northern México. Castilleja tortifolia (the type from Texas) is conspecific with the earlier named C. mezicana, also most abundant in northern Mexico. KEY WORDS: Castilleja, Scrophulariaceae, Texas, México I. A new species from southcentral Texas In his taxonomic treatment of Texas Castilleja, Holmgren (1970) noted that a population of C. integra A. Gray in Crockett Co. represented a geographi- cally outlying segment different from the rest of the species in its “yellowish inflorescences and shorter floral measurements” and might prove to be dis- tinct. These yellow flowered plants occur on the western edge of the Edward’s Plateau of southcentral Texas, where they are now known from Pecos, Upton, and Val Verde cos., as well as Crockett Co. and from one closely adjacent 209 210 Perey LOL OVG TA volume 72(3):209-230 March 1992 locality in Coahuila, México (Map 1). They are completely geographically separated from C. integra, which in Texas is restricted to the trans-Pecos area and which produces red bracts and calyces as well as a different vestiture. The Crockett Co. plants are more similar in vestiture, bracteal morphology, and coloration to C. purpurea (Nutt.) G. Don var. citrina (Penn.) Shinners, which occurs primarily from central Texas northward and which has a different leaf and floral morphology. Specimens of these geographically intermediate plants have most often been identified with caveats as C. integra or as C. purpurea var. citrina, and indeed they appear to be most closely related to both of these latter two taxa. As discussed below, however, they are distinct from both and may be justifiably regarded as a separate species. Castilleja genevievana Nesom, sp. nov. TYPE: UNITED STATES. Texas: Crockett Co., Hwy 290, 1.5 mi WSW of jct with Hwy I-10, ca. 24 mi W of Ozona (jct of I-10 and Hwy 163); area of steep limestone hills with domi- nant junipers, 2500 ft elevation, ca. 80 plants in an area of ca. 250 square yards, road bank and among scattered shrubs to fenceline; 5 April 1992, Guy & Genevieve Nesom 03 with Julia Nesom (HOLOTYPE: TEX!; Iso- types: ANSM!,BRIT!,COLO!,GH!,MO!,NMC!,NY!,PH!,SRSC!,SWT!, TAES!,UC!,UNM!,UTEP!). Castilleyae wntegrae A. Gray ac C. purpureae (Nutt.) G. Don var. citrinae (Penn.) Shinners similis sed ab uterque proprius habi- tatione et distributione geographica, vestimento trichomatum non ramosis, foliis plerumque integris, bracteis ac calycibus flavis, bracte- orum lobis anguste lanceolatis, et corollarum lobis infernis com- parate brevioribus. Plants perennial from strongly developed woody roots. Stems branched at base, the branches erect or ascending, 15-45 cm tall, loosely lanate with a moderate to dense investment of multicellular, unbranched, flattened, and vitreous hairs 0.5-2.0 mm long, these most commonly becoming somewhat tangled, often deflexed or retrorsely oriented, glandular hairs absent. Leaves linear-lanceolate, strongly 3 veined, 3-7 cm long, 4-8 mm wide, entire or the uppermost (rarely from midstem upwards) with a pair of linear or filiform lobes originating on the distal third of the blade, the adaxial surfaces moderately to densely invested with short, stiffly erect hairs, the abaxial surfaces glabrous or very sparsely hairy, margins densely long ciliate or at least with hairs distinctly longer than on the surfaces. Mature inflorescence 5-10 cm long, elongating at fruit maturity to 10-20 cm. Floral bracts 2-3 cm long, yellow-green on the basal 2/3, distally yellow varying uncommonly to orangish yellow or rarely to red, 8-12 mm wide, ovate to broadly oblanceolate obovate, entire or with a pair Nesom: Castilleja in southcentral Texas 211 GEEZ Castilleja genevievana \_ Castilleja purpurea @ var. citrina © var. lindheimeri xX var. purpurea Map 1. Distribution of Castilleja purpurea, C. genevievana, and C. integra in Texas. 212 PHY TOLOGTIA volume 72(3):209-230 March 1992 BIZeqyul efajiyseVD © BueAaIAIUasB efel[4seD & Map 2. Detailed distribution of Castilleja genevievana and C. tntegrain Texas. Two collections of C. integra from Deaf Smith County are not shown. Nesom: Castilleja in southcentral Texas 213 of narrowly lanceolate lobes originating slightly above the middle, the upper bracts densely stipitate glandular, without other hairs. Calyces 23-27 mm long, yellow to orangish yellow at least distally, the primary lobes 10-15 mm long, the secondary lobes 4-7 mm long, the primary lobes 1.7-2.5 times longer, densely stipitate glandular distally, without other hairs. Corollas 24-3] mm long, the 3 teeth of the lower lip 2-4 mm long, usually distinctly thickened, distally yellow, the tube light yellow, the galea 7-12 mm long, 1/3-2/5 the corolla length, with yellow to white or pink margins, the dorsal surface green to yellowish, densely short glandular; stigma and style exserted 1-4 mm, the stigma weakly to strongly recurved at maturity, with ovate branches. Mature fruits ovate, 10-15 mm long, 6-8 mm wide. Color photograph in Warnock (1970, p. 125, no. 4), as “C. citrina.” Rocky, limestone hills of southcentral Texas and adjacent Coahuila, México, vegetation dominated by juniper, 1700-2800 ft; midMarch through midJune. Additional specimens examined: MEXICO. Coahuila: Mpio. Villa Acuia, open country between Rancho Santo Domingo and Hacienda Piedra Blanca, 4 Jul 1936, Wynd & Mueller 487 (GH). UNITED STATES. Texas: Crockett Co.: 23 mi W of Ozona, 30 Apr 1970, Correll & Correll 38524 (LL); Salviastrum Mesa, 32 airline mi NW of Ozona, frequent on talus slope, 23 Apr 1947, Cory 58424 (SMU); 25 mi W of Ozona, 7 May 1947, McVaugh 8208 (SMU,TEX-2 sheets); Read Ranch, 14 mi SE of Ozona, 3 Jul 1961, Read A-18 (SRSC); 7 mi S of Ozona on Hwy 163, 1 Jun 1963, Read 141 (SRSC); 15 mi E of Pecos River on Hwy 290, top of Batchlor Hill, 15 Apr 1964, Read 593, 594, and 595 (SRSC); 20 mi W of Ozona, 14 Apr 1949, Tharp & Havard 49358 (TEX); 14.8 mi W of Ozona, 2400 ft, 14 Mar 1949, Turner & Warnock 274 (LL,SMU,SRSC); 14 mi N of Juno in limestone soil, 2000 ft, 1 May 1949, Warnock & Turner 765 (SMU,SRSC); 5 mi E of Pecos River on Hwy 290, limestone bluff above Live Oak Creek, 11 _ May 1947, Whitehouse 18689 (SMU). Pecos Co.: Sheffield, 20 Apr 1931, Jones 18287 (MO); 14 mi N of Sheffield, 1 Jun 1957, Warnock 14907 (LL,SRSC). Upton Co., no other locality data, 7 May 1938, Cory s.n. (TEX). Val Verde Co.: limestone bluffs above Pecos River, 26 mi N of Langtry, 30 Mar 1947, Warnock 47-279 (SRSC-2 sheets); 2 mi N of Devil’s River crossing along Hwy 163, 1 May 1949, Warnock & Turner 759 (SRSC); 22 mi N of Comstock on Hwy 163, 1 May 1949, Warnock & Turner 765 (SRSC). The new species is named for my yellow haired daughter, an avid outdoors- man and happy participant in the collection of the type. Castilleja genevievana is restricted to the southwestern corner of the Edward’s Plateau (Maps 1 and 2). At the western margin of its range, there is a corresponding, abrupt vege- tational discontinuity at the edge of the juniper dominated landscape moving westward into areas of Larrea. The geographic position of C. genevievana intermediate between its two close relatives is matched by its habitats at an intermediate range of elevation (1700-2800 feet). Castilleja integra occurs at 214 PHY TOLOGTA volume 72(3):209-230 March 1992 € dvW eyeuyy efeT[yse @ Map 3. Distribution of Castilleja lanata in Texas. Nesom: Castilleja in southcentral Texas 215 elevations of (2800-)3200-8600 feet in Texas; C. purpurea ranges over a number of different vegetational zones in Texas but occurs at 300-1750 feet in eleva- tion, with rare outlyers (var. citrina in Ector and Schleicher Cos.) up to 2500 feet. Castilleja genevievana, C. purpurea, and C. integra are similar in their perennial duration, lightly lanate stem vestiture, 3 veined leaves with distinctly reduced adaxial vestiture, and the dimensions and morphology of their bracts and flowers. Further, the stem and leaf hairs of all three have a strong tendency to be retrorsely oriented. Castilleja genevievana differs from C. integra in the yellow (vs. red) pigmentation of its floral bracts and calyces and in its vestiture of unbranched hairs (see discussion below, following C. elongata). In contrast, C. genevievana is similar to C. purpurea var. citrina in its pigmentation and unbranched hairs, as well as the narrowly lanceolate bracteal lobes. It differs from C. purpurea in its entire leaves, although rare plants of C. genevievana may produce lobed leaves on the upper half of the stems. The floral bracts of most plants of Castilleja genevievana are distally yellow; they vary to orangish yellow or rarely slightly orangish (label data of Mc Vaugh 8208) or rarely vary within populations from “yellow, orange, to scarlet” (label data of Warnock 47-279) or from yellow (“common,” Read 598) to yellowish orange to reddish (“rare,” Read 594) or orangish (“rare,” Read 595). Of the ca. 80 plants of the population from which the type collection was made, all except two produced bracts and calyces with a bright, lemon-yellow color. The two exceptions were distinctly orangish. Castilleja genevievana is probably most closely related to C. purpurea var. citrina, in view of their predominately yellow pigmentation (almost certainly a specialization within this group of related taxa). Castilleja genevievana and var. citrina produce only unbranched hairs (an unspecialized condition, see discussion below) and never bear the distinctive branched hairs characteris- tic of C. integra and C. lanata A. Gray. The new species might have been reasonably included as a fourth variety of C. purpurea, closely related to var. citrina, but the other three recognized varieties differ from C. genevievana in their consistent production of lobed leaves (1-2|-3] pairs of lobes) from the bottom to top of the stems (vs. leaves entire in C. genevievana) and corollas with the lower lip of three thin and at least distally petaloid, ovate lobes 4-7 mm long (vs. thickened, lanceolate, and 2-4 mm in C. genevievana). Further observations on variation within C. purpurea are given below. According to Holmgren (1971, p. 26), “Most of the closely related species in Castilleja probably could not stand the test of sympatric maintenance.” The same may be true for C. integra, C. purpurea, and C. genevievana, because they are geographically separated. Each, however, is immediately distinguishable by morphological features, and all three can justifiably be recognized at an equivalent taxonomic rank. 216 Pe yeteOriiOcG A volume 72(3):209-230 March 1992 II. Variation in Castilleja purpurea in Texas The geographic range of Castilleja purpurea extends from central Texas (Map 1) through Oklahoma to southern Kansas. Yellow bracted plants (var. citrina) occur from Texas to the northern extremity of the range of the species; purple bracted ones (var. purpurea) are found from southern Oklahoma to northcentral Texas; red to orange bracted ones (var. lindhetmeri [A. Gray] Shinners) are endemic to southcentral Texas. Plants within single populations of var. purpurea are strikingly variable in coloration: the bracts and calyces are predominately pinkish purple but vary from red, to reddish orange, burnt orange, peach, light yellow, creamy (very light orange-yellow), and rarely white. Similar variation has been noted on the labels of collections from Eastland Co. (McVaugh 8338-TEX), Hamilton Co. (Correll & Johnston 21098-LL), Hill Co. (Lundell 11309, 11310, 11311, 11812, 113818, 11314-all LL), McLennan Co. (Nesom 7276-TEX), and other localities. In addition, a great deal of infrapopulation variation occurs in bract and flower size: the bracts 20-40 mm long, the corollas 25-40 mm long. Similar variation in size also occurs in the other varieties of Castilleja purpurea. Plants of var. indhetmerz produce predominately reddish orange to orange bracts and calyces, but a red color is commonly produced similar to that of typical Castilleja integra. A citrinalike yellow occurs rarely: in one population (Comal Co., Nesom 7284-TEX), one plant with three branches of orange flow- ers also produced a single branch with yellow flowers; in another (Travis Co., Nesom 7275-TEX), two yellow-bracted plants occurred in a predominately red bracted population, and several individuals were intermediate. I have tenta- tively treated these as populational variants of var. lindheimeri. The typical coloration of var. lindhetmer: might be regarded as only a small subset of that occurring in var. purpurea (except for the citrina yellow, which does not appear as normal variation within var. purpurea), but there appears to be a relatively abrupt geographic transition between the two (Map 1), and they can be justifiably maintained as weakly delimited varieties. Populations of var. citrina are considerably less variable in coloration than in the other two varieties of Castilleja purpurea. The bracts and calyces are tipped with bright lemon-yellow, uncommonly with yellowish orange. In the southern portion of the range of the species, var. citrina and var. lindheimeri apparently are essentially allopatric or nearly so (Map 1); the ranges of var. citrina and var. purpurea, however, more closely approach each other. In McCulloch and Menard cos., near the juncture of the two taxa, populations of var. purpurea are even more variable than normal and include plants with yellow bracts and calyces that do not appear elsewhere in its range. One such population (3 mi N of Brady, McCulloch Co., Nesom 7282-TEX) produced an astounding array of individuals with red, pink, purple, orange, yellow, and white bracts and calyces. Along Highway 190 for about 25 miles, from Nesom: Castilleja in southcentral Texas Pallitg near Brady (McCulloch Co.) towards Menard (Menard Co.), populations of normally variable var. purpurea alternate with presumably hybrid populations producing yellow variants. Westward from about six miles east of Menard, all plants are yellow bracted var. citrina, without any evidence of introgression from var. purpurea. Genes of var. citrina apparently are present in populations of var. purpurea in their region of contact, but not vice versa, as if var. cztrzna were serving only as the pollen parent of the putative hybrids. Chromosome counts and field observations at additional localities will be necessary in a more critical assessment of the relationship between all three varieties of Castilleja purpurea. Castilleja purpurea (var. purpurea and var. lindheimeri) commonly grows with C. indivisa Engelm. along the eastern margin of the Edward’s Plateau. The latter is native to eastern Texas, southeastern Oklahoma, and western Louisiana but is now also widely seeded by the Texas Highway Department for its beauty along roadsides. Castilleja indivisa is closely related to the widespread Mexican species C. scorzoneraefolia Kunth and relatives (see Ne- som 1992), including the Texas species C. rigida Eastwood (see below), but where C. indivisa grows with C. purpurea, hybrids between the two species are common. At one such roadside locality in Coryell County (Nesom 7279/7280- TEX), intermediates and apparent backcrosses toward both parents produced a wide range of combinations of morphological features, including bract and calyx color, corolla, calyx, and leaf morphology, duration (judging from the root development), and habit. At several localities in Hays County (e.g., Ne- som 7288-TEX), where C. indivisa was not artificially seeded, hybrids between it and C. purpurea were relatively common, and no individuals of the former without genetic influence of the latter could be found. III. Two putatively rare and endemic taxa of trans-Pecos Texas The status of Castilleja elongata Pennell In his study of the Scrophulariaceae of trans-Pecos Texas, Pennell (1940) described Castilleja elongata, which he considered to be a narrowly endemic species. It has generally been acknowledged to occur on Emory Peak of the Chisos Mountains in Brewster Co., although specimens from the Davis Moun- tains of Jeff Davis Co. have also been identified as C. elongata. The original description of Castilleja elongata was based only on the type specimen, and Pennell’s sole comment regarding it (1940, p. 307) was that it “would seem to be the representative of the group of Castilleja mintata Dougl. [ex Hook.] in the mountains of the Big Bend district of southern Trans-Pecos Texas.” In contrast, with many collections now available for study from trans- Pecos Texas, I can find no feature to separate C. elongata from the widespread 218 Pony tO OSG aes volume 72(3):209-230 March 1992 and variable C. integra. Holmgren (1970) distinguished C. elongata from C. integra by its stems “finely villose” (vs. “whitish-tomentose or lanate”) and its bracts “not glandular-pubescent” (vs. “glandular-puberulent on the colored portions”). All C. «ntegra-like plants of Brewster Co., however, produce bracts that are prominently glandular on the distal portion with short, capitate hairs, these usually mixed with nonglandular hairs, and the stems of such plants are highly variable in the amount of vestiture produced. In the present study, I have examined 58 collections (not including duplicates) of C. integra from Brewster Co.; fifteen of these are from the Chisos Mountains. Castilleja integra in Brewster Co. occurs at 3600-6200 feet, outside of the Chisos Mountains, where it occurs at 5500-7500 feet; elsewhere in Texas it has been collected at elevations of 5000-7500(-8600) feet in Jeff Davis Co., 4150- 7500(-8000) feet in Culberson Co., and 5900-7000 feet in El Paso Co. At the higher elevations, plants tend to be more variable in height, ranging somewhat taller (to 1 meter or more in the Chisos Mountains), with a somewhat reduced amount of vestiture, and they produce calyces and corollas that range to near the low end of the range of length for the species. Within these atypical populations, however, the intergradation in stature, vestiture, and floral size with plants of lower elevations appears to be complete. The same phenomena were observed in Castilleja integra in New Mexico by Standley (1909, p. 86), who noted that in the Organ Mountains, “it grows at an altitude of about 6000 feet, on shaded slopes, where it is a tall plant and but little branched. Farther north, in the region about Santa Fe it is very low and much branched, growing on the open mesas and foothills, but at about the same altitude. The size of the flowers and bracts varies noticeably in different plants ... Most of these variations, perhaps, are due to climatic conditions, especially temperature and the water content of the soil.” Castilleja integra occurs from Arizona to Colorado, New Mexico, and Texas and in the Mexican states of Durango, Sonora, Chihuahua, Coahuila, and Nuevo Leon. Over almost its entire geographic range, it produces red to scarlet floral bracts, calyces, and corolla margins. There are three named varieties of the species, all red bracted, at the northern and northwestern edges of its range: from Arizona, var. gloriosa (Britt.) Cockerell (The Southwest 2:134. 1900.); from Colorado, var. gracilis Cockerell (Bull. Torr. Bot. Club 17:35. 1890.); and from New Mexico, var. intermedia Cockerell (The Southwest 2:134. 1900.). The taxonomic status of these is not evaluated here, but only the first has generally been recognized in regional floristic treatments. I have, however, examined specimens from over the entire range of C. zntegra. Warnock (1977, p. 201) noted that “color variation is unusual” in Castilleja integra and provided a photograph of a rare variant with yellow bracts from a primarily red bracted population about 10 miles south of Alpine in Brewster Co. I have observed only three collections of such a yellow variant of C. integra: (1) Scott 2 (SRSC) from near Panther Junction in Brewster Co. - Scott 1 Nesom: Castilleja in southcentral Texas 219 (SRSC) from the same area is typical red bracted C. integra, (2) Steiger 1212 (NY) from Boot Canyon in the Chisos Mts., and (3) Stewart 1739 (GH) from the Sierra de la Encantada in northwestern Coahuila, México. The occurrence of these yellow bracted but otherwise typical plants of C. integra along the very eastern margin of the species range, facing the geographic ranges of two, yellow bracted, closely related taxa (C. genevievana and C. purpurea var. citrina), seems likely to be indicative of an earlier genetic connection between all three. Among all of these populations of Castilleja integra, one feature in partic- ular appears to indicate their genetic coherence. While most of the hairs of the stems and leaves are unbranched, at least some of them are branched. The basal cell varies in length from ca. 0.02-0.10(-0.40) mm, then branches into two thinner, divergent cells that are also variably elongated and contribute to the lanate appearance of the vestiture. I have not been able to discover a geographic pattern that would predict the relative abundance (branched/un- branched) of the branched hairs, which varies from about 0.1-10.0 percent (or sometimes greater on the leaves). All plants over the entire range of the species produce branched hairs at least at a low frequency. In trans-Pecos Texas, Castilleja lanatais the only other species with branched hairs similar to those of C. integra. The former is easily identifiable by the dense, close, white woolly pubescence of its stems and both surfaces of its leaves. All of the hairs are composed of a short stipe with 2-4, elongated lat- eral branches (up to 2 mm long) produced at the apex of the stipe and that form the mass of the conspicuous vestiture. Few other American species of Castilleja are known to produce branched hairs, among them C. dendridion Nesom and C. galehintontae Nesom from México (see Nesom 1992) and C. pruznosa Fern. and individuals of C. apple- gate: Fern. var. fragilis (Zeile) N. Holmg. of the northwestern United States. Except for C. dendridion, the hairs of these taxa are similar to those of C. in- tegra and C. lanata, but they tend to produce secondary branches. Holmgren (1971) considered the similarity in trichome morphology significant enough to support an hypothesis of introgression between the two northwestern U.S. taxa, which otherwise are relatively distantly related between themselves. It also seems plausible that the branched hairs of C. integra indicate a close re- lationship between it and C. lanata. Complicating this hypothesis somewhat is the observation that the rare endemic, C. organorum Standley of the Or- gan Mountains in southeastern New Mexico, also bears a low percentage of branched hairs similar in morphology to those of C. integra and C. lanata, both of which also occur in the Organ Mountains. The calyx of C. organo- rum, however, is asymmetrically divided, and its evolutionary relationships are probably more closely linked to C. wootonw Standley and C. linariifolia Benth. (see comments below). Castilleja lanata and C. integra are sympatric over the greatest parts of their geographic ranges. The former has a slightly broader range than the lat- 220 PRY OLOsG tA volume 72(3):209-230 March 1992 ter, occurring at lower elevations and further to the south in México. Castilleja integra occurs at higher elevations and slightly further to the north (northern New Mexico and southern Colorado), and even in the area where C. lanata does not occur, the vestiture of C. integra includes branched hairs at least at a low frequency. This may indicate that branched hairs were present in the ancestral populations of C. integra, rather than acquired through later hybridization and introgression with C. lanata, although the latter processes might account for the variation in abundance of branched hairs in C. integra. Castilleja integra, however, is more similar to C. genevievana and presumably shares with it a more recent common ancestor. Castilleja lanata is partially sympatric with C. genevzevana (Maps 1 and 3) with no evidence of gene flow. Chromosome numbers have been reported from numerous populations of Castilleja integra (Heckard 1958, 1968; Heckard & Chuang 1977; and Ward 1983, 1984). In each of Arizona, Colorado, and New Mexico, both diploids (n=12) and tetraploids (n=24) are known, and Heckard & Chuang (1977, p. 163) noted that “No cytogeographic pattern is apparent in the distribution of the two chromosome levels in this fairly wide-ranging species.” Lockwood & Forstner (1991) reported a hexaploid of C. integra from near Alpine and a diploid of C. elongata from the Chisos Mountains (both from Brewster Co., Texas). Unpublished chromosome counts (as specimen annotations in 1991 by Lockwood at SRSC) indicate that tetraploid populations of C. integra also occur in trans-Pecos Texas. The single reported chromosome count of C. lanata (Ward 1983) has been diploid. My interest in the systematics of Texas Castilleja was begun without knowl- edge of another very recent study that included some of the same plants re- ported on here. I studied the SRSC collections annotated by Mark Lockwood but did not examine his thesis (1991) until after the essential completion of my own manuscript. The strong contrasts in our approaches and conclusions, discussed below, are the result of totally independent studies. A strong point of disagreement between the present study and Lockwood’s (1991) concerns the nature of variability within what can be accepted as Castilleja integra. According to Lockwood (p. 86), “The seed-coat similari- ties between C. elongata and [plants from Mt. Livermore in the Davis Moun- tains identified by the yet unpublished epithet “livermorensis”| also suggest that these two species are very closely related, at least phenetically. This al- lows for the possibility that they may also represent relictual populations of a once widespread species that has undergone speciation in the isolated montane habitats of the Davis and Chisos Mountains. Castilleja integra is also closely | related to these two species and possibly shares a common ancestor; In contrast, I suggest that C. integra itself is the “once [and still] widespreac) " species” postulated by Lockwood and that some of its relatively high elevatio1 ! populations are slightly divergent from the morphological form more typica ! for the species at lower elevations. Nesom: Castilleja in southcentral Texas 221 Lockwood’s key to Castilleja integra, C. elongata, and “livermorensis” uses only aspects of vestiture and floral bract morphology that I find to be quan- titative and intergrading, as discussed above. His data show the calyces and corollas of C. integra to be longer than in the other two taxa, where only slight differences in size are shown. In contrast, I did not observe any discontinuous differences in size within what is here recognized as C. integra in Texas. With a more detailed knowledge of the cytogeographical variation, the high elevation forms of Castilleja integra in southwestern Texas might indeed prove to be relictual diploids, with intervening areas of the species range primarily filled by polyploids. Diploids of C. integra, however, are known to be scat- tered through a large part of the range outside of Texas. Further, regional differentiation apparently occurs quickly in the genus, as evidenced by nu- merous narrow endemics and many variable species. Given the wide range of elevations, habitats, and substrates known for C. integra, the chromosomal variation, the intergrading morphological variability, some of it possibly re- flecting genetic influence of other species, and the proposed but unevaluated varietal taxa already existing within C. integra, I can find no satisfactory or appropriate way to recognize segregate taxa, even at the varietal rank, within it from Texas. My view of the taxonomy of this species as it occurs in Texas is summarized below. Castilleja integra A. Gray in Emory, Rep. U.S. & Mez. Boundary Surv. 2(1):119. 1858. SYNTYPES: UNITED STATES. Gray cited collec- tions by Wright, Bigelow, Kreuzfeldt, Fendler, and E.K. Smith. The details of the necessary lectotypification will be discussed in a separate paper (Boufford & Nesom in prep.). | Castilleja elongata Pennell, Proc. Acad. Nat. Sci. Philad. 92:306. 1940. TYPE: UNITED STATES. Texas: Brewster Co., wooded slopes of northern Chisos Mountains, ca. 6000 ft, Aug 1934, T.L. Stezger | 1689 (HOLOTYPE: NY!). Plants perennial. Stems 15-45(-60)(-90 in Brewster Co.) cm tall, lanose to villous, with a mixture of simple and branched hairs, the stipes 0.02-0.08(-0.4) mm long, bearing 2(-4) branches, the simple hairs and branches 0.1-0.5(-0.8) mm long. Leaves 2-6(-8) cm long, linear to narrowly lanceolate, (1-)2-7(-9) cm long, (1-)2-6(-8) mm wide, the lower surfaces densely invested but not obscured by stiff, short, somewhat retrorsely oriented hairs, the longer of these also commonly apically branched. Bracts entire or with 2 lobes originating from middle to near the apex, the apices with capitate glandular hairs. Calyces (18-)21-35(-38) mm long, the primary lobes (7-)10-16(-18) mm long. Corollas (23-)28-42(-47) mm long, the galea (8-)10-17 mm long, the tube 17-27(-30) mm long. ' ; 222 PO YT OD OsG aie volume 72(3):209-230 March 1992 Limestone and igneous substrates; (2800-)3200-8000(-8600) ft; flowering late Mar-Sep(-Oct). The status of Castilleja ciliata Pennell A population of plants thought to be endemic to the Davis Mountains of Jeff Davis Co., Texas, has been recognized and maintained as Castilleja ciliata (Holmgren 1970). Comparative study of these, however, shows them to be con- specific with plants of the Sacramento and White Mountains of southeastern New Mexico, where they are known by the earlier name C. wootonw. These are tall perennials with glabrous or very sparsely hairy stems, except in the inflorescence and lowermost portions, which commonly are villous. The linear leaves also range from very sparsely pubescent to completely glabrous or with only short ciliate margins. Those from New Mexico tend to be slightly hairier than the Texas plants, but all are unquestionably the same species. No collec- tions of C. wootoni are known from the Guadalupe Mountains, which are in a nearly intermediate geographic position between the Davis and Sacramento Mountains. The only apparent influence from other species on any of the plants of Castilleja wootont examined in this study is observed in Worthington’s collec- tion 121384 (UTEP) from Otero Co., New Mexico. One of the three mounted stems is typical C. wootonit; the other two, while nearly identical to the first in overall morphology, have stems sparsely invested with unbranched, retrorsely appressed hairs. The source of this variation is equivocal, but the plants may be hybrids with C. integra, which is common in the same area. Pennell (1940) compared Castilleja ciliata to C. miniata, but the latter dif- fers in its nearly equally divided calyx (vs. unequally divided in C. wootonii). The key provided by Wooton & Standley (1915) contrasted C. wootonz with C. linaritfolia, the latter a widespread species of the western United States and according to Holmgren (1976) a member of subg. Castilleja, although it sometimes hybridizes with the relatively more dissimilar C. minzata (see notes by Holmgren [1984]). Plants of both C. linaritfolia and C. wootonti pro- duce glabrous or nearly glabrous stems, thick, linear-lanceolate leaves, and un- equally divided calyces. Although the former tends to produce more strongly unequal calyx lobes, it is clearly closely related, if not conspecific, with C. wootonu. The couplet in Wooton & Standley’s treatment compared these two taxa in features of their leaf morphology, but with many specimens of each taxon now at hand, the putative foliar differences appear to be completely overlapping. The following couplet provides an updated comparison of the two taxa. 1. Calyx clefts 10-20(-22) mm deep abaxially and 2-6(-12) mm adaxially, the abaxial clefts 3-5 times longer than the adaxial; bract margins short hispid; lower portion of the stems glabrous to villous. ... C. linariifolia Nesom: Castilleja in southcentral Texas 223 1. Calyx clefts 11-14 mm deep abaxially and 8-9 mm adaxially, the abaxial clefts 1.2-1.8 times longer than the adaxial; bract margins densely long- ciliate; lower portion of the stems villous. ................. C. wootoniu Castilleja wootonii Standley, Muhlenb. 5:84. 1909. TYPE: UNITED STATES. New Mexico: [Lincoln Co.,] White Mountains, 7400 ft, 25 Aug 1907, E.O. Wooton & P.C. Standley 3411(HOLOTYPE: NY! ex NMC, NMC- photo!). In the protologue is additional and conflicting collection data (“Gilmore’s ranch on Eagle Creek,” at “about 3400 feet”), but it seems reasonable that the NY specimen represents authentic type material. Castilleja ciliata Pennell, Proc. Acad. Nat. Sci. Philad. 92:306. 1940. TYPE: UNITED STATES. Texas: Jeff Davis Co., Davis Moun- tains, rocky slopes of Mt. Livermore, 2200 m, 6 Jun 1928, E.J. Palmer 84388 (HOLOTYPE: PH!; Isotype: NY!). The PH speci- men was annotated in an undated, unsigned note as “=C. wootoni Standl. fide Pennell.” Additional specimens examined: UNITED STATES. New Mexico: Lincoln Co.: White Mts.: Eagle Creek Canyon, ca. 7 mi NW of Ruidoso, locally common on shady aspects of S-facing exposed granite faces, pine-oak-Douglas fir, 11 Jul 1982, Soreng, Spellenberg, & Ward 2024, voucher for chromosome count of n=12, reported by Ward 1983 (NMC); N of Sierra Blanca on trail to MonJeau lookout, 8 Sep 1970, Spellenberg 2392 (NY); White Mt. Peak, 1897, Wooton s.n. (NMC) and 1901, Wooton s.n. (NMC); ca. 2.0 mi W of Sierra Blanca Ski Lodge, lower part of Apache Bow! Ski Area, open meadows, 11,000 ft, 28 Jun 1980, Worthington 6148 and 6149 (UTEP); ca. 1.0 air mi W of the Sierra Blanca Ski Resort, bottom of ravine through conifer forest near stream, 10,800 ft, 5 Jul 1981, Worthington 7227 (UTEP); entrance to Sierra Blanca Ski Area, ca. 9800 ft, 7 Aug 1984, Worthington 12807 (UTEP). Otero Co.: White Mts.: [no other data], Jun 1936, Hinckley 1024 (NY,TEX); divide above the Mescalero agency, 1895, Wooton s.n. (NMC); Sacramento Mts.: Cloudcroft pine-spruce-aspen, partly open, wooded slopes, 19 Aug 1968, Correll & Correll | 36130 (LL); Mescalero Apache Indian Reservation, Silver Lake, 2450 m, 3 Jul 1979, Freeman s.n. (UTEP); S of Cloudcroft, 21.1 mi along State Rd. 24, Cathey Canyon Vista, open grassy area in dry pine forest, 6 Jul 1983, | Hardison 11 (TEX); Cloudcroft, Fir Campground, rocky hillside with pine and fir, 17 Jul 1965, Iwen 199 (TEX); Sacramento’s Silver Campground, 8 Sep 1979, Patrick s.n. (UTEP); High Rolls, 19 Oct 1969, Smartt 43 (UTEP); Cloudcroft, 19 Oct 1969, Smartt 98 (UTEP); vicinity of Toboggan (an old settlement) ca. 3.5 mi W of Cloudcroft in bottom of canyon, pine-oak-Douglas fir, 10 Aug 1969, Spellenberg 2081 (NMC); Cloudcroft, 1899, Wooton s.n. (NMC); Mescalero reservation, 1905, Wooton s.n. (NMC); canyon between entrance to Sunspot and road S down Sacramento Canyon, N slope, limestone 224 PHY TOLOGTS volume 72(3):209-230 March 1992 soil, edge of Douglas Fir, 9100 ft, 28 Jul 1979, Worthington 4793 (UTEP); Bailey Canyon ca. 2 mi NW of Cloudcroft, mixed pine-fir forest, 8400 ft 4 : : : : ' ' : ¢ : : o] 4 4 : ' ‘ - ‘ ¥ *,. : , 4 . . ‘ r ; ' it ‘ 41> ) | \ : As é < y 7 ’ y ' t : AR, \ ; - ‘ , ' sy 7 \ 5 : ’ i 1 ‘ - ’ 7 ‘ if J : i , . ) A\ : ¥ \ ineeh - ' \) - ‘ . : i ( if i ; ; i 7 \y 7 me - - wt * \ hive | ‘eh ) At TARA ‘4 A ’ ' . t «\ Hy I ' i ; } \ 4 . ‘ ; ] a Le ) t Ng iy 4 - ‘ sf ’ : Ve . ‘ : \ ‘ Ve : Ny SANE ty k te ) ¥* 7 i ry 3 { i ’ \ * 7 yp ‘s . > 4¢ Ly . v /\ ; ’ Wr" ‘eeu ‘ ay eye . AN eA 4 ann y Ay yi | 7. ¥ ay ’ : ; i 5 ' HAAN he Yeni PANN. + en Ye ARG Aad j " oy ; ; - rae ae ‘ » 4 \ ) aus AY mar Aa : ay 1A ae a | A BN. POLARS th ANE eA aan Ng VU uly ACIP TC ULE aris a 6 en) ) sa. . % a, . : LAT AEA UASAN Vi CP Lay St , 4 ¥ a} 1 > ne , , AG 7 7 7 c ‘ i \ » > - - + nY, Varah : Ab ‘ ; ye FS wee * ‘ A 5 hy Sve" Information for Authors Articles from botanical systematics and ecology, including. biographical sketches, critical reviews, and summaries of literature will be considered for publication in PHYTOLOGIA. Manuscripts may be submitted either on computer diskette, or as typescript. Diskettes will be — returned to authors after action has been taken on the manuscript. Diskettes may be 5.25 inches or 3.5 inches and may be written in any IBM or MacIntosh compatible format. Typescript manuscripts should be single spaced and will be read into the computer using a page scanner. The scanner will read standard typewriter fonts but will not read dot matrix print. Manuscripts submitted in dot matrix print cannot be ac- cepted. Use underscore (not italics) for scientific names. Corrections made on typescript manuscripts must be complete and neat as the scan- ner will not read them otherwise. Language of manuscripts may be either English or Spanish. Figures will be reduced to fit within limits of text pages and therefore, should be submitted with an internal scale and have dimensions proportional to those for text pages. Legends for figures should be included in figures whenever possible. Each manuscript should have an abstract and key word list. Specimen citations should be — consistent throughout the manuscript. Serial titles should be cited with abbreviations used in Botanico Periodicum Huntianum. References cited only as part of nomenclatural summaries should not appear in — Literature Cited. Nomenclatural work should include one paragraph per basionym and must provide proper (as defined by the current /nterna- — tional Code of Botanical Nomenclature) citation of sources of epithets — and combinations. Authors should arrange for two workers in the appropriate field to review the manuscript before submission. Copies of reviews should be forwarded to the editor with the manuscript. Manuscripts will not be published without review. Cost of publication is currently $13.00 US per page for publication without reprints. Publication with 100 reprints is provided for $18.00 US per page, 200 reprints for $21.50 US per page. Page charges are due with manuscript and no paper will be published before payment is received in full. Reprints must be ordered and paid for in advance. Page charges will be determined on the basis of a typescript page (single spaced, 10 points, blank line between paragraphs) with all type inside a rectangle 143 mm (horizontal) by 219 mm (vertical), not including running head and page number. Tithe page should include title, author(s) name(s), and address(es). Two blank lines should appear above and below section headings (Abstract, Discussion, Literature Cited, etc.) in the manuscript. | No extra charge is made for line drawings provided they conform to limitations of size and proportion for normal text. Halftones require an extra charge of $10.00 US per page.