75 



oiilv such forms of insects that were aerial oi- al least in that 

 StaiiO at the tiiii(> tliev were u^c(\ in the experiments. In these 

 tests only half-grown or larucr nymphs wei-e taken. Xo oth^r 

 food than land insects was i^iven rhem and these in most cases 

 were dropped in the feeding' jai's alive. Some of the nyni])lis 

 W(>re cai-ried tlironiiii in this way for one or two nionlts, or initil 

 thev finally emerged. In some cases this feeding period was 

 carried on for from two to five weeks. All the different spe- 

 cies of insects fed were readily eaten, except most of tlie ants, 

 especially tlios(> strong in formic acid, and some apliids. 



Beside adnlt inso^cts, caterpillars, maggots an,d grnbs of vari- 

 ous kinds, also other small crcatnres, snch as spiders, land Crus- 

 taceans (shrimps and sowbngs), millijiedes, earthworms, and 

 nematodes were fed to the nym])hs. All these were eagerly 

 snatched at and eaten n]». 



Often when no other food was available, snuill hits of fresh 

 heef seemed to be relished. 



That dragonfly nymphs do feed on land insects under nat- 

 ural conditions is borne ont by the findings in the contents of 

 the alimentary canal of the 253 nymphs dissected, as listed 

 above. Out of this number, 20 nymphs were found to contain 

 aerial forms, tlins representing 7.9 per cent, of the number dis- 

 sected ; or, if this nnmber is expressed in nnits, to conform with 

 the tabulated statement, then the land insects consnmed Ity these 

 nymphs constitute 7.9 per cent, of their food. 



Field Observation and Examination of the Contents of 

 THE Alimentary Canal of Adults. 



In this series of dissections 218 Pdiiffda and 24 Aiia.v (ex- 

 cluding those whose alimentary canal was found empty) were 

 taken. To these figures may be added a number of dragonflies 

 that were caught with the victim still more or less intact in 

 their mouth parts. As the mouth parts are really part of the 

 digestive tract, the insects thus found will be included in the lis* 

 of those found in the alimentary canal proper. 



As there seems to be a slight difference in the food habits 

 between the Pantala and Anax, as indicated by the observations 

 so far made, the findings of the two species will therefore be 

 given under se})arate tables, as follows: 



