376 PROCEEDINGS OF THE ACADEMY OF [1892. 



Although the flowers of Campanula rotundifoUa are classed as 

 proterandrous, the pollen is not ejected from the anther cells till 

 after the corolla has opened and the hairy pistil has been developed 

 considerably beyond the line of the anthers, though most of the 

 genus seem to discharge their pollen while still in the bud. The 

 stamens wither soon after the anthers have discharged their pollen. 

 In the flowers in this large branch the stigmas do not exj^and till 

 the fourth day after flowering. In the quiet atmosphere of the room 

 and in the absence of all insects that are usually considered aids in 

 fertilization, there seems to be no pollen on the stigmatic surfaces, 

 but capsules are all enlarging, and the young seeds swelling as if 

 the fertilization of the flowers had been perfected. It is difficult to 

 believe that in some manner fertilization has not taken place. Only 

 the full ripening of the seed could positively prove this point. 

 Unfortunately I had not the opportunity to test it. 



The variation of the color in the flowers of this branch may be 

 noted. During the week that I had them under observation, 

 there may have been between two and three hundred flowers. 

 Some were nearly white, others of a rosy purple, the majority blue. 

 The observations were concluded on the 15th of July. 



CORNUS CANADENSIS. 



Some of the exotic species of Contxs are dioecious, but there is no 

 record of dioecism in any of the American species. The plant is very 

 common on Mt. Desert Island. Near Northeast Harbor I spent 

 several hours, July 27th, examining these plants particularly, 

 amongst other things, and would frequently find large patches that 

 were evidently the production of several years by underground 

 stolons, entirely barren. Other patches would have a single berry 

 in the central portion of the cyme and all the others barren. 

 Other patches were abundantly prolific. It is a safe inference that 

 some plants are monoecious and others wholly dioecious. 



As it is well understood, the leaves are really in opposite pairs, the 

 verticil being formed by the suppression of the internodes and axil- 

 lary bud.«. One specimen was found in wliich two axillary buds 

 had produced branches, and these two again produced each two 

 more from their apices. These four branches were terminated by 

 four heads of flowers, each with its four milk-white bracts which in 

 the midst of so many companion plants with numerous red berries 

 had a unique effect. 



