236 TRELEASE— SPECIES IN AGAVE. [April 22, 



specific, even if either of them might properly bear the name 

 attached to it. Jacobi clearly saw the value of spine characters in 

 this genus, in his study of its representatives in the garden col- 

 lections of his day, and he applied them as consistently as he could 

 in his descriptions ; but unfortunately the material on which these 

 were based was often immature and the descriptions are too fre- 

 quently generalized. Of the many illustrations of Agave, some- 

 times exquisitely colored, almost none approximates the truth in 

 these details any more closely than a studio-made volcano ap- 

 proaches the true declivities of its cinder-cone and foot-slope. Not- 

 withstanding all of its defects — some of them very real and serious 

 — photography now ensures the truthful picturing of minutiae that 

 the eye of the describer may mistake and the pencil of the delineator 

 is quite likely to misrepresent. Even by its aid, however, part 

 truths may appear as truths, and a real fact may enter as an unreal 

 specific character. It is for this reason that my own conception 

 of specific identities and differences in this genus oscillates as my 

 study proceeds : the leaf characters of a first specimen being most 

 commonly ignored until more and difTerent material forces their 

 recognition ; but with increasing evidence that, chosen from mature 

 leaves of adult plants, and used with judgment, they are dependable. 

 Obviously, no species in any group can be considered as fully 

 defined until all of its characters are known ; and no species is 

 satisfactorily described until its characters have been tersely brought 

 together in contrast with those of its allies, as is painfully evident 

 whenever a new form is confronted with published descriptions, 

 which may be equivalent to saying that no species can be satis- 

 factorily described until all of its closely related congeners are 

 known. Obviously, too, species based on incomplete material are 

 more likely to prove capable of ultimate subdivision than those of 

 which all characters are represented when the first description is 

 drawn ; and in placing reliance on minutiae, more than the usual 

 need exists for using these with the judgment derivable only from 

 experience, and for selecting with the greatest care their adult state. 

 With such care, and subject to these restrictions, the species of this 

 difficult genus, even in their vegetating form, appear to be capable 



