4 University of California Publications. [BOTANY 



This is excellent in the case of Collomia and Polemonium, being 

 so reliable that the use of other characters of habit and foliage 

 are almost superfluous. 



Taking the order as a whole, Collomia may, in the first place, 

 be well marked off by the turbinate, accrescent calyx, the tube 

 possessing replicate sinuses, and not being distended by the 

 mature capsule. Polemonium is equally distinct with its entirely 

 herbaceous calyx, accrescent, and not distended by the capsule. 

 The other genera cannot be so easily separated by a single char- 

 acter as the foregoing. The pungent, unequally lobed calyx of 

 Navarretia is found to hold with all species agreeing in respect to 

 general habit and foliage. This is true also of Linanthus, but the 

 character of the foliage is necessary to separate some species of 

 this genus from those of Gilia. Still, the cylindrical to urn- shaped 

 calyx, with equal lobes, and conspicuous, broad, hyaline spaces 

 between the ribs, the tube equalling the capsule, but distended, 

 and usually ruptured by the mature fruit, is a primary feature of 

 typical forms. Phlox, to a still greater degree than the last two 

 genera, must be distinguished by the foliage and habit rather 

 than by the calyx, and is the only genus where the form of the 

 corolla is of value. The strictly salver- form corolla with narrow 

 orifice is a constant and, therefore, a reliable character. 



After separating the five genera which I have just considered, 

 there still remains a large number of forms, some exceedingly 

 variable and others less so. Looked at superficially, these forms 

 fall into groups which are easily recognized at a glance; but a 

 more careful study shows them to be so intimately related that 

 they must be of one genus. This sixth genus may well take the 

 inclusive name of Gilia, and cover the subgenera corresponding 

 to these unmistakable, but superficial, groups. If connecting 

 links had not been found, these groups would doubtless still be 

 kept as distinct genera, as they were formerly considered. 



With the question of genera disposed of, there remains the 

 more difficult task of classifying under each its several species. 

 Leaving considerable latitude for differences of opinion, there are 

 from one hundred and twenty-five to one hundred and seventy- 

 five species on the Pacific side of the Rocky Mountains, and new 

 species are continually being described. It is, therefore, with full 



