608 THE AMERICAN NATURALIST [Vol. XLV 



It is pretty generally conceded, however, that the 

 corolla of flowers and many other conspicuous floral 

 structures possess a very important function in serving 

 to attract various insects. Many odors likewise serve the 

 same purpose, even to a very considerable extent, as 

 shown by the interesting experiments of Plateau and 

 others. 



In connection with experimental cotton breeding inves- 

 tigations in northern Georgia, the writer has had occa- 

 sion to give considerable attention to the visits of bees 

 and other insects among cotton blossoms in this region. 

 The large number of certain species of bees regularly 

 visiting cotton blossoms and the ease with which observa- 

 tions could be made in the field led the writer, during the 

 summer of 1910, to make a series of experiments in order 

 to learn more definitely, if possible, just how cotton 

 blossoms attract bees, whether mainly through the visual 



These experiments were made at the beginning of the 

 blooming season when blossoms were very scarce and 

 bees very numerous. The first experiment was begun in 

 the forenoon of July 26, and others followed throughout 

 July and early August. Each day the period of observa- 

 tions began at eight or nine o'clock and terminated about 

 twelve or one o'clock, at which time the blossoms were 

 beginning to close and bee visitors were much less fre- 

 quent. For each test a period of half an hour or an hour 

 was usually allowed, thus making the number of insect 

 visits sufficiently numerous for valid conclusions. 



By far the majority of bee visits were made by Metis- 

 sodes bimaculata, although other species of Melissodes 

 were no doubt casual visitors. The big wasp (Elis plum- 

 ipes Drury) and a Bomlms or an Entechnia occasionally 



