267 



as well to consider the autumn flowering of some species as in 

 the regular course of events. This is especially true of the vio- 

 lets. There is a place near North Attleboro, Massachusetts, 

 where Viola laiiceolata 2.nA Viola canina, van MiiJilenbergii, grow 

 in great profusion. Not a year passes that I do not gather hun- 

 dreds of blossoms from this locality towards the last weeks of 

 September and the first of October. I expect to find them in 

 blossom then, and should be surprised if I did not. They are 

 not in such great profusion as in spring, of course, but are still 

 quite common. In the same way, along the middle of October, 

 Viola pedata is found in considerable abundance along some local- 

 ities near Assawompsett Pond near Middleborough, Massachu- 

 setts. Other species have been invariably found but never in 

 great numbers : Viola primulcefolia and Viola matllata, which 

 are found during the last weeks of September at North Attlebo- 

 ro, Massachusetts. Viola striata I have seen in blossom at Gran- 

 ville, Ohio, during the autumn of 1885, but never had occasion 

 to look for the same again. Two species of violets, which from 

 their frequency in spring I should have found in their respective 

 haunts in autumn if they were in flower, Viola sagittata and 

 Viola piihescens, I have never discovered in flower at that time 

 of the year. It is not an uncommon feature of certain of 

 e fall violets not to develop all their petals typically. This is 

 especially true of Viola lanceolata, and in a certain measure of 

 V. cticullata. Although most of the blossoms appear to be per- 

 fectly normal, those developed in colder weather show a tendency 

 of the lateral and upper petals to become abortive, so that the 

 lower petal in one flower may be well developed, the lateral 

 petals may be reduced to one-third their normal size, and the up- 

 per petals may be visible only after the surrounding sepals have 

 been dissected away. It will be noticed that in the process of 

 this abortion those petals which are most necessary to guide the 

 insects to the honey are aborted last. The semi-aborted petals 

 have also a tendency to project straight forward in line with the 

 sepals, and not to expand as in ordinary flowers, /// effect increas- 

 ing the size of the lower petal with which they are thus made 

 contiguous. 



A few plants were found in blossom this year which Jo not 



th 



