14 Rhodora [January 



one or more entire leaves, and still other plants with all the leaves 

 entire. It is extremely difficult to make out how these last differ from 

 V. sororia. The situation is quite the same when we consider the 

 character of pubescence. We find specimens of V. sororia with 

 dense, almost matted, pubescence on the petioles, others with sparse 

 spreading pubescence, others with mere traces of pubescence just 

 below the blade, and this on only one or two leaves, and finally 

 other plants quite glabrous. Somewhere in this shifting series we 

 pass from V. sororia to V. papilionacea. In several other species 

 of Viola we find pubescence just as variable a character. V. blanda 

 is sometimes quite pubescent, but often nearly or wholly glabrous ; 

 so is V. rotundifolia ; so is V. rcnifolia. Under these circumstances 

 it would seem that in the genus Viola, whatever may be the case else- 

 where, pubescence, unless correlated with other characters, is not a 

 satisfactory basis on which to found a species. Possibly these cor- 

 related characters may yet appear. Meanwhile, as I am loth to 

 make new names or new combinations, I provisionally speak of 

 these five forms under division "C " as species. 



3. V. latiuscula, Greene (Pittonia, v. 93, Nov. 1902), was 

 founded upon specimens from Twin Mts., West Rutland, Vermont. 

 The plant proves to be of frequent occurrence in dry sterile soil in 

 western New England and adjacent New York. It was collected 

 by the writer in May, 1899, on the dry open sand-plains near Fort 

 Ethan Allen, Essex, Vermont. On revisiting the station for mature 

 leaves and fruit in June, 1903, after over fifty days of rainless 

 weather, I found that the plant had completely disappeared. But on 

 a third visit, September 11, after a rainy August, large plants in abun- 

 dance bearing copious fruit were to be had. The plant had become 

 as completely dormant during the summer drought as during the 

 winter cold. Further facts tend to show that certain species of 

 stemless violet are capable of leading a double existence, passing, 

 with intervals of rest between, through a vernal and an autumnal 

 stage that differ from each other in foliage, flower and fruit, — thus 

 posing as a sort of Jekyll and Hyde among plants.^ 



' Viola rotundifolia, Michx. well illustrates this ; there is a marked contrast 

 between the April plant with its small leaves and yellow flowers, and the August 

 plant with its large leaves and chocolate cleistogamous flowers on branching 

 peduncles (really, disguised stolons). Pursh described the latter form as a distinct 

 species, V. clandestina; though he opined its connection with Michaux's species. 

 But for this opinion he was taken to task by Schweinitz, who stoutly asserts that 

 V. rotundifolia "can have no affinity 7uhateTer^^ with V. clandestina ! (Am. Jour. 

 Sci. V. 63.) 



