MUSTARD FAMILY 67 



ance and considerably unlike in character of pubescence, and in length and curvature of pods. 

 The material seems to represent, however, an involved complex, which w^e are, at present, unable 

 to resolve into definite phylogenetic lines. 



The entire series from Arabis secunda to A. pulchra admittedly represents very closely 

 related types, each of them possessing more or less of mystifying outliers. But our series as 

 worked out, is, we believe, in the interests of further investigation of the life history of these 

 plants along useful lines. We cite below some California stations for A. secimda. 



Locs. — Grouse Mt., Tracy 4838; Trinity Summit, Davy 5859; Klamath Eiver, Chandler; 

 Marble Mt., Siskiyou Co., Jepson 2834. 



Eefs. — Arabis secunda Howell, Erythea 3:33 (1895), type loc. Mt. Adams, Wash., Suhs- 

 dorf, Howell. A. arcuata Gray var. secmida Bob.; Gray, Syn. Fl. 1^:164 (1895). A. sparsiflora 

 Nutt. var. secunda Piper, Contrib. U. S. Nat. Herb. 11:294 (1906). A. Iiolboellii var. secunda 

 Jepson, Man. 430 (1925). 



13. A. retrofracta Graham. Northern Rock-cress. Biennial ; stem usually 

 solitary, erect, simple or branched, 7 to 24 inches high, pubescent with simple or 

 forked hairs; leaves pubescent with hairs more or less forked to stellate; blades of 

 the basal leaves oblanceolate, repand-dentate, % to I14 inches long, blades of the 

 cauline linear-oblong, acute, entire, I/2 to 1 inch long; flowers white or purple, 2 

 to 3 lines long; pods linear, straight or arcuate, II/2 to 2 inches long, more or less 

 reflexed or spreading, seldom strictly secund; pedicels hairy or rarely pubescent. 



Montane, 3500 to 6000 feet : Trinity Co. to Siskiyou Co.; Plumas Co. to Modoc 

 Co. North to Washington, east to Idaho, thence to British America. June. See 

 note under no. 12. 



Locs. — Dorleska, Trinity Co., Hall 8600; Quartz Valley, Siskiyou Co., Butler 610; Yreka, 

 Butler 592, 723, 728; Goosenest foothills, Siskiyou Co., Butler 1331; Goose Lake, C. C. Bruce 

 2247a; Forestdale, Modoc Co., M. S. Balcer ; Eagle Lake, J. Grinnell; Hot Springs Valley, Lassen 

 Peak, Jepson 12,287, 12,294. 



Eefs. — Arabis retrofracta Graham, Edinb. New. Phil. Jour. 7:345 (1829), type from the 

 region between Hudson Bay and Eocky Mts., Richardson. A. holboellii var. retrofracta Jepson, 

 Man. 429 (1925). A. holboellii B. & W. Bot. Cal. 1:33 (1876), in part. A. sparsiflora Nutt.; 

 T. & G. Fl. 1:81 (1838), type loc. near the sources of the Columbia Eiver, Nuttall. A. arcuata 

 Gray var. subvillosa Wats.; Gray, Syn. Fl. 1^:164 (1895). A. campyloloba Greene, Pitt. 4:192 

 (1900), type loc. Yreka, Greene. A. tenuis Greene, Pitt. 4:189 (1900), type loe. mts. of e. Wash., 

 SuTcsdorf. A. brucae Jones, Contrib. W. Bot. 14:37 (1912), type loc. hills near Davis Creek, C. C. 

 Bruce; apparently of the A. retrofracta aggregate; entire plant glabrous (ex char.), 



Arabis bolanderi Wats. Proc. Am. Acad. 22:467 (1887). Watson quotes three collections. 

 Of these we take as the type the first mentioned, "Yosemite Valley or Mono Pass (Bolander)". 

 The label on the Bolander specimen (Gray Herb.) reads in this manner: "with 6270 or 6273". 

 No. 6270 means Mariposa Grove, but no plant like this is known to us from that region. No. 6273 

 is recorded in the Bolander Field Book as "Arabis. Mono Pass". We take this to be the type 

 locality. The plant of the type collection bears only sterile pods of distinctive outline. These 

 plants are well matched by similar sterile specimens from the high central Sierra Nevada as 

 follows: Mt. Ealston, H. M. Evans (one individual of the collection showing, significantly, one 

 fertile pod) and Upper Echo Lake, Tahoe region, E. 0. Essig. This matter is more fully dis- 

 cussed by the writer in Madrono 1:254-255 (1929). It is our judgment that Watson's binomial 

 must be a synonym of A. retrofracta Graham. 



14. A. canescens Nutt. Gray Rock-cress. Stems slender, 1 or few from the 

 simple or branched crown of a taproot, 5 to 8 (or 17) inches high, finely puberu- 

 lent, scantily leafy but with small yet dense tufts of narrow leaves at base; blades 

 of leaves of the basal tufts narrow-oblanceolate, acute, 3 to 7 lines (or to 1% inches) 

 long, closely white-tomentulose; cauline leaves mostly few and scattered, their 

 blades linear, clasping; racemes short (1 to 2, or 4, inches long) ; flowers small, 1 

 to 11/2 lines long; petals white; pods pendulous, typically narrow (% to % line 

 wide) and straight-edged, strongly compressed, commonly glabrous, 1 to 1% 

 inches long, on pedicels 1% to 3 lines long; seeds winged, in 2 rows. 



Dry mountain slopes, 5000 to 8000 feet : Modoc Co. East to Wyoming, north 

 to eastern Oregon. May-June. 



Tax. note.— Little studied in the field and scantily represented in herbaria, this species is 

 probably limited in its typical form to the northern part of the Great Basin. It is perhaps not 



