almost straight but the general characters of its group are not given 

 His Vesicarii contains no North American species. The chief character 

 is inflated calyx. Our A. oocalycis could be placed there. His Sesamei 

 contains no North American species. His Onobrychoidei contains 

 such wholly diverse species as A. caryocarpus (crassicarpus), Miss- 

 ouriensis, Labradoricus, adsurgens (nitidus), which belong to at least 

 three distinct groups. His Dissitiflori contains only A. gracilis. His 

 Hypoglottidei contains A. bidentatus and Hypoglottis (agrestis) which 

 species are not at all related. 



From this it is evident that no conception of North American 

 relationship in the genus was in the mind of DeCandolle. 



Phaca was kept up by him to contain such diverse species as 

 A. alpinus, triflorus, villosus, caespitosus, mollis, the Inflati, andinus, 

 etc. 



The only other serious attempt at classification worthy of mention 

 is that of Torrey and Gray in their Flora of North America, but they 

 only copied DeCandolle. Hooker in his Flora did the same. 



GRAY'S REVISION. 



Dr. Asa Gray in Proc. Am. Acad. 6 1864 made the first real 

 attempt at proper classification of the genus as to North America. 

 Tn 1871 Watson, profiting by a year's field work in the West, ampli- 

 fied it somewhat and corrected some details but for the most part 

 followed Gray. This was also the first attempt to group the species 

 n-enetically starting according to Gray's universal rule from the most 

 developed and going to the least altered species. 



Gray's system was bised on the fundamental idea of one and 

 two-celled pods as separating the two ser'es Astragalus proper anrl 

 Phaca. Astrrr^lns was the more developed because of being 2-celled 

 and was based on the greater intrusion of the dorsal suture if the 

 nods were not fully 2-celled. Phaca was unfortunately based on the 

 s itures being not at all intruded or the ventral the more produced, 

 ■'^ad he made three series, one to co-^ + '^in Astragalus proper, the other 

 Phaca as defined by Linnaeus in his -^recies of the Sp. PI. and rep- 

 r-^sented by A. alpinus (Phaca alpina) to contain those with the 

 ventral sut'ire the most intruded, and the third series which we may 

 call Homalobus with nerve-like sutures not at all intruded he wouM 

 have had a set of series which could not be improved upon, but I do 

 not think anything is gained by attempting to separate the genus 

 into three series. 



Under Astragalus proper we find him making 16 sections. His 

 supposedly most developed species were the Sarcocarpi because of the 

 fleshy and fully 2-celled pods. His emphases of the 2-celled feature 

 prevented him from seeing that this group is manifestly close to the 

 Argophylli which he puts last in his series and some of whose species 

 he put in Phaca (A. inflexus. Purshii and Utahensis'). In addition 

 this group is not as highly developed as the Didymocarpi and Micranthi 

 which he places below it. 



TTis next groun is the Diphysi containing A. lentiginosus which is 

 far removed from the rest of the Inflati to which it is related but which 

 nevertheless connects the Inflati with the Sarcocarpi. 



His third group is the Chaetodontes containing Spaldingii and 

 Lyallii but he seems to have failed to have recognized its close re- 

 lationship to the Didymocarpi, Micranthi and Reflexi which are the 

 most developed groups in the genus. To this we now add A. 

 Brauntoni. 



His fourth section is the Mollissimi which is wholly out of place, 

 being next to 'be Argophylli. 



His fifth section is the Uliginosi, the Ciceroidei of DeCandolle. 

 a well defined group containing A. Canadensis and Mortoni, and should 



