662 



THE AMERICAN NATURALIST [Vol. XLV 



to laciniation is all that could be observed. Perhaps the 

 result of the cultural experiments, when a large number 

 of rosettes can come under observation, will shed more 

 light upon this point. But there was no danger of con- 

 fusing the adult plants of the species and of the laciniate 

 variety, for the latter presented a most striking appear- 

 ance (Fig. 3). At first glance it seemed as if caterpillars 

 or grasshoppers had destroyed the greater part of each 

 leaf ; the midrib and larger veins and here and there a 

 small portion of the blade being left. A closer examina- 

 tion revealed the fact that the rosette leaves differed least 

 from those of the normal plant. The leaves a and b in 

 Fig. 6 show this sufficiently. The latter represents a 



rosette leaf of the var. laciniata, the former one of the 

 species. The next leaves possess a marked laciniation 

 (Fig. 6 c, d, etc.). As one proceeds higher up the stem, 

 the laciniation becomes more pronounced, till finally but 

 a figment is left of the leaves in the region of the inflores- 



