Castilleja applegatei var. fragilis (Zeile) N.H.Holmgren

Authority

Holmgren, Noel H. 1971. A taxonomic revision of the Castilleja viscidula group. Mem. New York Bot. Gard. 21: 1-63.

Family

Scrophulariaceae

Scientific Name

Castilleja applegatei var. fragilis (Zeile) N.H.Holmgren

Description

Species Description - Stems relatively tall, (2-)3-6 dm high, often branched. Leaves quite variable in length, 1.8-4.5 (-7) cm long, the main leaves often all entire or sometimes with a pair of slender lateral lobes on the uppermost leaves. Infiorescence usually narrow and long, the bracts and flowers appressed. Bracts distally red, sometimes yellowish to salmon pink. Calyx (15-) 16-22 (-26) mm long, the abaxial cleft (4-) 6-10 (-12) mm deep, the adaxial cleft (5-) 7-10 (-12) mm deep, the segments (1-)2-5(-7) mm long, the tube usually villose at the base. Corolla (22-)25-33(-37) mm long; galea (9-) 11-16(-18) mm long, dorsally hght green to reddish, glandular-puberulent to sometimes scarcely viscid-puberulent, its lateral margins sometimes puberulent; lower lip greenish to pinkish or violet, the teeth sometimes brownish purple; tube (11-) 12-19(-21) mm long, loosely pilose to villose. Stamens exserted beyond the galea tip, the upper anthers 2-2.75(-3.25) mm long. Stigma 0.25-0.6 mm thick. Capsules variable in length, 7-15 (-17) mm long, ovoid to ellipsoid. Chromosome numbers: 2n=24 and 48.

Discussion

Castilleja pinetorum Fernald, Erythea 6: 50. 1898. Oregon, Klamath County, Swan Lake Valley, pine woods, 30 Jun 1896, Applegate 415 (GH, isotypes, DS, PH). The labels on the isotypes provide the additional information: "Brookside Ranch, head of Cabin Creek, near Picnic Grounds, in dry yellow pine woods."

Casiilleja brooksii Eastwood, Proc. Calif. Acad. Sci. Ill 2: 288. (3 Jun) 1902. Califomia, Wesno County, Bubba Creek of King's River, 1-13 Jul 1899, Eastwood s. a. (CAS 268875). This .specimen is rcjirescntative of the southern Sierra Nevada element, which differs from typical var fragilis in having a tendency toward divided leaves.

Casiilleja. bakeri Greene ex C. F. Baker, nom. nudum. West Am. PI. 1: 16. (Fall of) 1902. Nevada, Orm.shy County, King's Canyon, 1700-2000 m, 1 Jun 1902, Baker 889 (GH, NY, UC, US). West American Plants is merely a list of determinations of collections for distribution by Baker, who intended to publish "full notes on all the species representcfl in these series.

(p. 15), but failed to do so in this instance. I have included it in the list of synonymy because it has been recorded in the Gray Herbarium Card Index and appears in several publications.

Castilleja trisecia Greene, Leafl. Bot. Observ. Crit, 1: 78. 1904. Califomia, Tulare County, Hockett's Meadow, 18 Jul 1904, Baker 4431 (ND, isotypes, G H , CAS, NY, RSA). This type locality is not very far from that of Miss Eastwood's C. brooksii and the plants are very much alike.

Castilleja pinetorum var fragilis Zeile in Jepson, Man. Fl. PI. Calif. 938. 1925. Zeile typified this with the nomen nudum C fragilis Eastw. ex C. F. Baker, West Am. PI. 3: 4. 1904.

Castilleja dolichostylis Eastw. Leafl. W. Bot. 3: 88. 1941. California, Tehama County, near Govemment Flat, 9 Jul 1941, Eastwood & Howell 0837 (CAS 290387, isotypes CAS, GH, PH, US).

Castilleja wherryana Pennell, Proc. Acad. Nat. Sci. Philadelphia 99: 180. 1947. Oregon, Baker County, Dooley Mountain, S of Salisbury, "open pine woods on sandstone," 5000 ft, 4 Jul 1931, Pennell 15454 (PH 819655, isotypes, US, MO, NY).

Castilleja latifoliata Pennell ex Edwin, Leafl. W . Bot. 9: 46. 1959. Ne\ada, Washoe County, S of Mount Rose, Route 27, granitic ridge, 9000-9300 ft alt, 24 Jul 1940, Pennell 26267 (PH 90214, isotypes CAS, UT, UTC). I could not find the holotype at the Academy in Philadelphia.

Type. Cahfornia, Sisson, 14 Aug 1903, Copeland 3883 (CAS, DS, GH, JEPS, MO, NY, RSA). The specimens of the type appear to be strongly modified by grazing and subsequent second growth.

The range of Castilleja applegatei var fragilis overlaps the ranges of more than twenty other species of CastiUeja. A more detailed observation reveals that in actuality only ten of these grow within altitudinal range for possible cross pollination. Of these ten species I find evidence of hybridization between var fragilis and the following six species: C. chromosa A. Nels., C. disticha Eastw., C. linariaefolia Benth., C. miniata Dougl. ex Hook., C. pruinosa Fernald, and C. subinclusa Greene. Perhaps further field work will uncover natural hybrids involving the other four species, which are C. angustifolia (Nutt.) G. Don, C. covilleana Henderson, C. pilosa (S. Wats.) Rydb., and C. rustica Piper. The following is a discussion of the six hybrid combinations. The sympatric range of var fragilis and C. chromosa forms a narrow crescent along the eastern front of the Sierra Nevada and across northern Nevada and adjacent Oregon, where, for some unknown reason, they both display predominantly yellowish inflorescences. C. chromosa rarely comes into pollinating range with var fragilis. They are usually kept apart by a spatial difference in altitude and a temporal difference in flowering, var fragilis being at higher elevations and flowering later. In the Pine Forest Range I observed var fragilis (Holmgren & Reveal 1210) and C. chromosa (1218, NY, UTC, BRY) growing together with hybrids (1219, NY, UTC, BRY).

The distribution of C. disticha, a closely related species, overlaps with that of var fragilis, in the Mineral King area of Tulare County, California. Here the distinctions between the two species break down in a manner suggesting introgression. To m y knowledge this is the only occurrence of sympatry among the taxa of the C. viscidida group.

I have found no evidence of present-day hybridization between var fragilis and C. linariaefolia. However, morphological and cytological evidence coupled with distributional relations between these two species and C. dissitiflora gives me reason to believe that C. dissitiflora m a y have descended from a hybrid population involving the above mentioned species (see discussion on phylogeny).

Castilleja miniata, the most widespread species of Castilleja, is likely to be found wherever var fragilis is, but in a more moist and shaded habitat. The type of C. excelsa Eastw. is much like C. miniata, differing mostly in having a sparse vesture of gland-tipped hairs. The only source of such hairs would be through introgression with var fragilis. A n example of introgression more nearly approaching typical var fragilis is evidenced in a large, entire-leaved, nearly glabrous collection (Heller 9850) from near Donner Pass in the Sierra Nevada.

In the region of the Klamath Vlountains, the northern Coast Range of California, and the northern Sierra Nevada, the geographical distributions of var fragilis and C. pruinosa overlap. The two species also share similar habitat preferences and flowering periodicit3^ Typical C. pruinosa has a dense pubescence on the herbage consisting of dendritic (branched) hairs and with no gland-tipped hairs below the inflorescence. Frequently, however, C. pruinosa is found with fewer branched hairs that are interspersed with gland-tipped hairs. In this region of overlapping distribution var fragilis is rarely found without at least a few forked hairs along the leaf margins. Just about every level of introgression can be found between the two species. The types of C. angustifolia var adenophora Fernald, C. muscipula Eastw., and its var armeniaca Eastw. are essentially C. pruinosa with various degrees of glandulosity very likely introduced through introgression from var fragilis. Castilleja muscipula var angustifolia Eastw. can be regarded as typical of C. pruinosa. It appears from Heckard's (1968) counts that C. pruinosa is a tetraploid and var fragilis in this region is diploid. This m a y account for the fact that these species have not swamped each other out.

Castilleja subinclusa, of the western foothills of the Sierra Nevada from Butte County to Tulare County, sometimes forms hybrids with var fragilis. An example of this is H o w e U d Barneby 29510 (CAS) from 3 miles north of Bagby, Mariposa County, which, in addition to the gland-tipped hairs of var fragilis, bears the characteristic scabrous-hispidulous pubescence, the long narrow leaves and deeper abaxial calyx cleft of C. subinclusa.

Although var fragilis exhibits an extremely wide range of variation in any region of its distribution and even in its local populations, there are some discernible regional patterns. However, these variation patterns are too slight to warrant infraspecific recognition. The following characteristics under discussion are to be interpreted only as being statistically more prevalent and not jiresent to the exclusion of those characteristics found to predominate elsewhere.

The type specimen for the synonym C. pinetorum from southern Oregon exemphfies an entire-leaved race which predominates in an area extending south from Klamath and Lake counties of Oregon to the Yolla Bolly Mountains of the North Coast Ranges of California and to the Lake Tahoe region in the northern Sierra Nevada. Besides being predominantly entire-leaved, this race is usually tall-stemmed and has a longer and narrower inflorescence than the other races. The few chromosome counts from this region have turned out to be diploid (2n = 24, Heckard, 1968). The rest of var fragilis appears to be tetraploid (2n = 4i8, Heckard, 1968). If diploidy correlates with the morphological differences outlined above, it would give enough added weight to consider it as a taxonomic variety. A more thorough cytological study is necessary before this should be done, however.

The region of the Warner Mountains of northeastern California and adjacent Oregon, extending eastward to the Sonoma, Jackson, Pine Forest, and Santa Rosa ranges of northern Nevada and the Steens Mountains of southern Oregon, is inhabited by a race of var fragilis exhibiting a predominance of yellow and orangish yellow bracts. As I mentioned earlier, this curious color predominance also exists in C. chromosa in the same region, where the distributions of these two species overlap. Early collections from the Steens Mountains were regarded by Pennell (1941) as his C. glandulifera, very likely because of their similarities in coloration. Plants from the Steens Mountains become short in stature along an altitudinal cline, and those at high elevations sometimes resemble C. glandulifera in aspect. On the basis of all other characters, this local population definitely is var fragilis.

In central Idaho there is a high mountain ecotype that resembles var viscida in aspect. The floral dimensions, however, show it to be well within the bounds of var fragilis. An ecocline can be traced in Washington and Boise counties between the typical var fragiUs in the foothills and the alpine ecotype of the high mountains. The latter completely replaces the former in the Seven Devils Mountains and the Sawtooth Range. The Sawtooth Range population represents the easternmost extension of the distribution of var fragilis, where it most closely approaches the distribution of var viscida. Three facts given here, i.e., the adaptation to high mountain habitats, the resemblance in aspect, and the geographical proximity, can be interpreted as suggesting that this race of var fragilis is not far removed from the ancestral stock that gave rise to var viscida.

Distribution

Habitat and distribution. Collected in flower from (April) early July to mid August (September) at elevations from 1000 feet in the Klamath Mountains to over 11,000 feet in the Sierra Nevad-a; in dry sagebrush and open coniferous forests of foothills to near timberline. Central and southwestern Idaho, eastern and southwestern Oregon, northern inner Coast Range and the Sierra Nevada of Cahfornia, and western and northern Nevada (Fig. 7).

United States of America North America|