LIFE HISTORIES OF THE ODOXATA. 47 



also destroy miasmata, the fact that Odonats feed upon 

 them would seem to remove them from the category of 

 useful scavengers. As a matter of fact we have been un- 

 able to get them to eat anything in the least decayed. 

 Their preference is for fresh and living food. The Italian 

 entomologists recommend pieces of fresh fish for feeding 

 Odonata in captivity, but urge that they be supplied with 

 fresh sweet food. Biro'^ states that nearly fifty thousand 

 young fish were destroyed by a species of Libellulina in a 

 pond in Hungary. Riley and Howard '-^^ mention a case, 

 on the authority of Mr. W. L. Jones, of Atlanta, Georgia, 

 where the larva) of Anax Junius were found feeding on 

 young carp. In our several aquaria we have studied the 

 apjietites of our larvae under varying circumstances, and 

 feel prepared to state that, even with thousands of Culex 

 larvse provided for them, the young Odonats will exhaust 

 every other source of food first and then attack each other. 

 This latter habit is a serious one when artificial propaga- 

 tion is to be considered. Poulton^-'' thinks that cannil)al- 

 ism among larvse frequently arises from scarcity of food, 

 but our experience leads us to believe that they are canni- 

 bals from choice. 



Species of Agabus geoffria and allied genera, and Xo- 

 tonecta undulata, furnish the young dragon fly with dainty 

 food, and while there remain any of these the Culex 

 larvffi are unmolested. The slender, soft bodied " wrig- 

 glers " usually succeed in eluding the ratlier ponderous 

 apparatus evidently intended for larger, hard bodied game. 



As the Odonat grows towards maturity (Plate V., 

 Fig. 2) its form becomes more graceful and its movements 

 less clumsy, but it is still sluggisli and })refers to remain 

 at the muddy l)ottom. Butler ^^ thus describes this habit : 



