400 Wn-:(JAND: Some Species of Bjdexs 



I 



The value of the different characters from a taxonomic stand- 

 pohit has, hi general, been found to agree with that placed upon 

 them by previous authors. The most important of all is, of 

 course, the achenc, and by this one' character alone most of the 

 species may be recognized. Very much dependence cannot be 

 placed upon the leaves, because of the unusually great variation 

 produced by environment, through which influence both the size 

 and the incisions arc modified. Two other structures, the in- 

 volucre and the corolla, have, however, been used much more 

 than by previous writers. 



The involucre varies principally in the size and number of 



I 



bracts, and is important for specific distinction rather than to show 

 lines or strains of variation. In this respect it differs from the 

 corolla, in which the pale yellow. 4-toothed character is found 

 practically constant among a number of related species, while the 

 orange, 5 -toothed type represents another group, both evidently 

 distinct lines of development. But the corolla does not show all 

 the lines of variation or even the primary ones. To determine 

 these one must look to the achcne. The number of awns is found 

 to be mucli more constant than was before supposed. Only one 



I 



species, B. lacvis, really shows a variation in number without a 

 corresponding change in other characters. Some allowance must, 

 of course, be made for the fact that the outer achcncs usually 

 show a shortening of the awns which process is often carried so 

 far that some are entirely obliterated. 



In determining specific differences the direction of the barbs 



n I 



upon the margins of the achene seems to be of more importance 

 than upon the awns. So far as known, no specimens of species 

 ha\'ing truly downwardly barbed achcnes have been found with 

 upwardly barbed awns ; moreov^er, it seems possible that when erect 

 barbs are found in B, vidauocarpa and /)'. coiinata they are to be con- 

 sidered as a continuation of the hairs of the achenc out upon the 

 awns, thus excluding the real barbs, and not as a reversal of the 



I 



barbs themselves. Whether this is really the case has not as yet been 



'I 



definitely settled. In some cases the erect barbs are nearly as slender 

 as the hairs on the achene, but more often they are stouter. The 

 hairs and warts upon the achencs, although more or less charac- 

 teristic in some species, arc not of special diagnostic value. 



