198 THYSANOPTERA AND ORTHOPTERA. 



found on them in the summer must have come to them the same season. 

 Search was made in the spikes of this plant, and cjuite a few thrips were 

 found, some of them on plants that were separated by about 100 meters 

 from all other vegetation of any kind. 



At first thought this might seem proof positive that one or more 

 thrips had flown at least this distance of 100 meters. But it is con- 

 ceivable that the insects were blown, or fell, into the water nearer shore, 

 and then drifted to the pickerel weed. If this is to occur, the thrips 

 must be able to float on the water and to remain there some time with- 

 out serious injury. To determine this, specimens of Thrips physopus 

 and Anthothrips verhasci were placed in water. Most of the former 

 species adhered to the surface film. At the end of 50 hours they were 

 removed, and in a few minutes were crawling rather actively about. 

 Some of this species, however, parted from the surface film and sank to 

 the bottom; when removed at the end of 50 hours they did not revive. 

 The specimens of Anthothrips verhasci all floated on the surface. I tried 

 to force them down into the water, but adhering air bubbles brought 

 them back every time to the surface. After 72 hours, they were re- 

 moved; one began to crawl in less than three minutes, and the rest 

 shortly afterwards. Other drowning experiments, performed for 

 another purpose, are described under the head of "Enemies and 

 Checks." 



It seems probable, therefore, that thrips could drop into the water 

 and float to plants at a considerable distance from the shore. If this 

 was the case at Rush Lake, and the thrips fell into the water near the 

 shore after only a short flight, which is the- point at issue, the plants 

 near the windward side of the lake would stand a better chance of being 

 infested than those on the leeward. Since thrips do not readily mi- 

 grate until forced to do so by drying tissue, as I have shown above, and 

 since the spikes of Pontederia last for several weeks, the plants near 

 "the windward side of the lake should, on the supposition of short flights 

 and drifting on the water, bear more thrips at any time during their 

 active season than plants on the leeward side. In Table IV, I give some 

 observations on this point. Examinations were all made in two days 

 time, so that seasonal variations in the abundance of the insects would 

 not vitiate results. Prevailing winds were, as usual in this region, 

 westerly. 



