154 BOTANY OF THE DEATH VALLEY EXPEDITION. 



Type specimen in the United States National Herbarium, No. 1049, Death Valley 

 Expedition; collected June 25, 1891, on the west slope of the divide between 

 Kcrnville and Havilah, Kern County, California, by Frederick V. Covillo. 



The species most nearly resembles in its vegetative characters N. viscidula, JV". 

 atractyloides, and N. heierodoxa, but it is readily distinguished from all these by its 

 leaves, the terminal lobe alone in K. seiiloha having from 10 to 20 teetli on each side. 

 The leaves of the others are not differentiated into a terminal lobe and have rarely 

 more than I teeth on each side throughout their entire length. 



Navarretia schottii Torr. Bot. Mex. Bound. 145 (1859). Type locality, "in the 

 Colorado Desert, Sonora," Mexico. 



In Panamint Valley (No. G87), and near the Vegas Wash (No. 1900). See notes on 

 Navarretia setosissima punctata. 



Navarretia setosissima Torr. & Gr. Rot. Ives Exped. 22 (1801). Type locality, 

 "Rocky hill sides, Camp 67." This camp was situated on the banks of the Colorado 

 River at the mouth of Diamond River, in northwestern Arizona, directly north of 

 Peach Springs, a station on the Atlantic and Pacific Railway. 



Vegas Valley, Nevada (No. 1887). One, plant among these specimens is an inter- 

 grade between the type form and the; variety described below. See notes on Navar- 

 retia setosissima punctata, 



Navarretia setosissima punctata Coville, Proc. Biol. Soc. "Wash. vi. 72 (1892), 

 under Gilia — Gray in herb. Type locality as given below. Platk XIV. 



"Flowers and fruit larger than in the type form; corolla with tube about 10 mm. 

 long, its lobes 7 to 10 mm. long, white, with purple dots sometimes arranged in lon- 

 gitudinal lines, and a pair of golden spots at about the middle; capsule 6 to 9 mm. 

 long, often with 10 seeds in each of the 3 cells. 



"The plant differs from the typo form in the characters above mentioned. In 

 G. setosissima the corolla tube has about the same length, but the lobes are much 

 smaller (3 to 5 mm. long) and cream-colored, with neither purple nor yellow mark- 

 ings, and the capsule is commonly about 5 mm. long with about 5 seeds in a cell. 

 This variety holds the same relation to the type form that G. maitheivsii does to 

 G. schottii, except that in the case of the latter two species the differentiation appears 

 to be complete, while in the former integrades in size and coloration occur. The 

 flowers of G. setosissima and its variety are regular, erect, and with straight stamens, 

 while those of the other two species are irregular, inserted at an angle or even hori- 

 zontally, and have ascending stamens. Iu herbarium specimens this irregularity is 

 often obscured, and G. schottii is frequently confounded with G. setosissima. Both 

 G. schottii and G. matthewsii are, however, readily distinguishable from G. setosissima 

 and its variety by a vegetative character, which was originally pointed out by 

 Watson, but which was afterward lost sight of. In the former the lateral bristles 

 of the leaf arise singly, in the latter in twos (rarely singly or in threes), from each 

 hair tubercle. Tliis character is constant. 



"These four plants are very interesting from the standpoint of their genealogical 

 interrelation. The parent form probably was, or was very similar to, (/. setosissima; 

 from this G. schottii developed; and then, from both these, plants with larger, strik- 

 ingly colored corollas differentiated, G. setosissima punctata and G. matthewsii, re- 

 spectively. The name adopted for the variety is one used on herbarium specimens 

 by Dr. Gray, but never published. 



"Type specimen in the United States National Herbarium, No. 710, Death Valley 

 Expedition; collected April 21, 1891, in Surprise Canon, Panamint Mountains, Cali- 

 fornia, by Frederick V. Coville." 



Collected also iu Panauiiut Valley (Xo. 6G8), and near Keelor (No. 8G4). 



