376 BOTANICAL GAZETTE [MAY 
rounded incurved apex: pod oblong-linear, pubescent like the plant, 
shortly acute, 10-18 mm. long, straight or slightly curved, usually 
recurved but sometimes ascending, one-celled, the dorsal suture 
being but slightly intruded: seeds few (about 5). 
Perhaps this finds its nearest relatives in A. atratus Wats. and A. obscurus 
ats. No. 1887, collected July 2, 1912, above the “Hot Hole” of the East 
Bruneau, Owyhee Co., Idaho, is the type. It was quite inconspicuous because 
of its color and its fasta of creeping among the short grasses of the sagebrush 
plains. 
GERANIUM CAESPITOSUM gracile, n. comb.—G. gracile Engelm. 
in Gray Pl. Fendl. 27. 1849; G. atropurpureum Heller, Bull. Torr. 
Bot. Club 25:195. 1898; G. furcatum Hanks, N. A. Fl. 25:16. 
1907. 
Specimens of G. ‘caespitosum when they show a tendency to become glan- 
dular pubescent on the pedicels have been considered as G. gracile (G. atropur- 
pureum), and if this glandulosity extends to the stem they are G. furcatum. A 
series of specimens can be so arranged as to show all degrees of glandulosity to 
the complete lack of it. The proposed variety, therefore, rests primarily upon 
the slenderer and more erect stems, the narrower petals, and usually a trace of 
glandulosity. Where the glandulosity becomes marked throughout, it merges 
into G. Parryi (Engelm.) Heller. 
It is somewhat singular that there should be any misunderstanding in 
regard to Geranium caespitosum James. Admitting that the original printing 
of the name did not publish the species, Dr. Gray’s diagnosis in Pl. Fendl. 25 
and Dr. TRELEASE’s in Bost. Soc. Nat. Hist. 4:72 fixed the plant to which this 
name must apply. 
Specimens answering to this description are not rare in the herbaria and 
are always non-glandular and with the pubescence of the stem (whether sparse 
or abundant) more or less retrose. The plant is always cespitose, growing in 
the form of a turf or mat from which short assurgent stems arise. The new 
characterization in the N. A. Fl. 25:15 would seem to be without warrant, and 
at best that description presents merely one of the variants of G. Fremontii Torr. 
Specimens wholly typical of G. caespitosum are BAKER 448, Colorado; 
Baker, EARLE, and Tracy 407, Colorado; 1155 from Colorado, distrib. by 
the New York Bot. Gard. as one of the type nos. of G. Cowenii Rydb.; BuFFUM, 
Wyoming; NELSON 8591, Wyoming; METCALF 194, New Mexico; MACDOUGAL 
118, Arizona. Besides G. Cowenii, G. marginale Rydb. must undoubtedly be 
referred to G. caespitosum as here understood. 
GENTIANA AFFINIS major, n. var.—Leaves uniformly narrower: 
calyx lobes oblong-lanceolate, nearly equal and all approximately 
