19 



legions, in rather alkaline deserts and the Reventi-arrecti which 

 aave considerable dilferentiation in the same regions but spread to 

 the north farther east. This last group amplifies into the Uliginosi 

 a cosmopolitan group of plains, prairies and open woods in temper- 

 ate regions. . This division on growing in cold meadows and dry 

 valleys passes into the Hypoglottides, and this again divides into the 

 Chaetodontes under the altered climatic conditions of the Columbia 

 drainage. But it should be noted that, all three groups grow in the 

 same region in the west, but each under its own peculiar ecological 

 conditions. Another offshoot of the Homalobi, close to the Flexuosi 

 is the Atrati of the interior juniper mesas, which branches into the 

 Strigulosi which belong mostly to the cool meadows and mesas and 

 pine forests of Mexico. 



The second main branch of the Homalobi is the Flexuosi which, 

 in the colder regions of the north in sweet soil branches into the 

 Argophylli, a group growing on dry benches and plains where there is 

 some winter snow. It shows pome differentiation into the Malaci 

 which also are replaced by Sarcocarpi of cold plains, and the Mollis- 

 simi which extend beyond the limits of snow on plains to central 

 Mexico, but in relatively cool regions. The Argophylli have some 

 species that have secured some alkali tolerance, such as the species 

 of the Navajo Basin. In the alkaline deserts and in clayey regions 

 where the soil is very poor the Flexuosi pass gradually into forms 

 with conspicuously inflated pods with rather thick walls, and the 

 Ocreati and Bisulcati without inflation but with much alteration of 

 walls by corrugations and sulcation at sutures and intrusion of the 

 dorsal suture. The Bisulcati branch into the Galegiformes, and these 

 f^gain early give off the Lonchocarpi. The Flexuosi pass into the 

 Hamosi which branch into the Leptocarpi, and these into the Micran- 

 thi, and the latter pass into the Didymocarpi the highest developed 

 species. This whole line of the Flexuosi is for the most part a line of 

 warm and variably desert regions. A very few species grow in mo- 

 derately cool regions such as bisulcatus, Drummondii, and some Ha- 

 racsi, but they are for the most part the result of long inhabitance of 

 hot regions of the south and subject to the greatest influence of 

 diverse ecological conditions. 



In the system which I have used it has been my object to get 

 related species together as far as a linear arrangement could do it. 

 find I have therefore inverted the order of relationship in parts of 

 groups to bring this about. 



In order to facilitate understanding the genetic relationship I 

 have appended a diagram of my view of the relationship of the groups. 



My diagram of and arrangement of species is based wholly 

 on relationship as indicated by the species of today. Whether this 

 is the actual line of descent is a different matter for no one knows 

 whether there has been retrogradation nor how much. A study of the 

 Asiatic species might throw some light on this matter but the geolo- 

 gical history is so localized that little benefit is liable to accrue from 

 this source. The differentiation of the genus except in rare cases 

 has manifestly been due to glacial and post-glacial agencies and these 

 factors are well worked out. 



GROUP DETAILS. 



1. Homalobi. It is a debatable matter as to which species is 

 the most primitive, but from its resemblance to Hosackia and Lathy- 

 rus I have assumed A. campestris to be the most primitive. But 

 for systematic reasons I have placed first those species with the 

 ventral suture the more arched and which belong in more humid re- 

 gions and mostly at the north. 



A. Wingatensis and tenellus fall together, with short or no stipe 

 and small flowers. It is here that the Debiles branch off. The sti- 



