66 THEODORE H. PRISON. 



for subsequent work along this line. In 1917, I secured four 

 colonies out of fourteen experiments, or nearly twenty-nine per 

 cent. Besides this, in four of the other fourteen experiments, 

 the queens at one time or another gave evidence of some interest 

 in the nest. In 1919, out of nineteen experiments the queens 

 started colonies in thirteen of my artificial nests. This gave a 

 successful percentage of seventy-seven plus in 1919. Two of the 

 experiments of 1919 considered as failures should not really be 

 so counted, for in one box one queen was killed and the other 

 died of a nematode infection, and in the other instance both 

 queens escaped. Out of thirteen experiments in 1920, ten must 

 be considered as successful, or seventy-seven per cent. This 

 same year, in two of the other three experiments considered as 

 failures, the queens went so far as to make egg cells, and another 

 colony might have been started if at that time it had been possible 

 to continue these experiments longer in the spring. A survey of 

 Table I. will show also that in 1917, 1919 and 1920 there were ten 

 experiments in which but one queen was placed in. each artificial 

 nest. Colonies were started in seven of these ten experiments. 

 The queens in two of the other three experiments of this nature 

 never manifested any interest in the nest. One of the three, 

 however, at one time took an interest in the nest, but finally 

 abandoned it without laying eggs. All in all, I secured as good 

 results, in proportion to the number of experiments, when one 

 queen was used as when there were two. 



SUMMARY. 



Summarizing, I find that it is possible to induce the queens of 

 certain species of North American bumblebees to start colonies 

 in confinement when either one or two queens are used in the 

 experiments. Plath (1923) achieved similar results in 1922 by the 

 confinement of a single queen assisted by "one to three workers." 

 In 1919 and 1920, in starting colonies under controlled conditions 

 an average success of over seventy per cent, was attained. 

 These results are indicative that the style and make-up of the 

 artificial nest used in 1919 and 1920, as well as the procedure 

 followed, is the most likely to produce results of any method 

 thus far described. Furthermore, for the first time normal 



