CONTRIBUTIONS TO WESTERN BOTANY NO, 15 r:«l 



has the pods of the latter group and the leaves of the Tomier. Tl^he 'for- 

 mer group has flat and wide leaflets very obtuse to retuse. From the 

 leaf structure I would at once place cornuta with longipes where *I haue 

 always put it Then we have Miss Eastwood's C. nana, got at the type 

 locality of C. Palmerana which belongs here with the latter, and whicli 

 has the scale like upper bracts just as in C, Palmerana. Miss East- 

 wood's specimens of C nana are evidently young with pods not yet ixiYiy 

 developed, and to my mind are certainly C. Palmerana, which in tura 

 may deserve separation from C. longipes. C. Montrosae Payson, from tin: 

 leaves would seem to belong with C. oocarpa. To this species also Ije- 

 longs C, plocasperma and Coville's proposed stenospcrma which Payeon ig- 

 nored. Then we have C. obtusifolia to which belongs C. taurocranos Nelson, 

 and the van pubescens Nelson which is a form of the type. What we '^^hall 

 do with brevipes and gracilis I will leave for more field work. 'There 

 is no doubt that tlie splitters are seemingly justified in making so miii\' 

 of Cleomella, for the characters given seem abundantly distinct at fir<t 

 sight, but the more material one has the less weight he gives to di- - 

 tinctions recognized by Watson and Gray. It is manifest lliat the de- 

 velopemnt of filaments isi a matter of little weight. The leaf characters 

 vvould seem to be more permanent. But when we come to study them v>e 

 find them varying toward the linear group, but after all r^n it be nn - 

 sible that longipes can include comuta and Palmerana? If so we: fehall 

 have to reconstruct our ideas of Cleomella. The distribution of t e 

 genus is sporadic. On page 38 Payson speaks of C* "refexa". What 

 he means by that remains to be seen. There is no such species so far 

 as I can find. I am convinced that too much emphasis has l^een plncvl 

 on morphological characters, and that ultimately we j^hall have to V^'H 

 more than half of the recognized species. But this can be done only 

 by plant breeding. There is nothing to show that C. brevipes' is anv 

 thing but depauperate form of parvi flora. C. gracilis Brandegee Is 

 probably C. parviflora. C. Mojavensis Payson is manifestly a form of 

 I)locasperma. C. Hillmani is of course C. longipes. 



Recently I have had an opportunity to study certain species" in *fhc 

 field to determine the relative value of specific characters. The speciis 

 nbsen'ed are obtusifolia and plocasperma (oocarpa). Certain vegetafive 

 characters are constant. C. obtusifolia alwavs has the short obovate and 

 folded leaflets, and on the young parts are the papery and hair-like 

 stipules at least, though mostly the plants are conspsicuously pubescent 

 throughout and the stipules very prominent. The stipes are alwa^-s re- 

 flexed and the styles long. The pods vary from obliquely diamond- 

 shaped to almost transversely linear because of the development of the 

 horns or valves. The most marked character is in the habit whicTi is 



prostrate, forming mats often 4 feet wide on the ground because of tlic 

 multitude of branches spreading out wand-like m^er the ground. The 

 flower clusters are almost sessible and innumerable. Once' in a wliile 

 the central stem rises erect several feet but always surroufided" by th-; 

 spreading branches. No other Cleomella has this habit. Cleomella plo- 

 casperma on the other hand had the Tiablt oT Cleome'lutea,"that'-is,^ erect 



