REPORT OF THE ENTOMOLOGIST. 515 



phide. Strong tar water killed them; diluted it proved harmless. 

 KeroseiiG emulsion diluted to contain 5 per cent, kerosene was effect- 

 ive; three ounces of carbon- bisulphide in 7 quarts of water proved 

 fatal within ten minutes; the other insecticides were ineffective. It 

 would be very costly to put enough of these materials in the water 

 to produce the desired effect. 



Other experiments in smaller creeks, in which numerous larvae of 

 the Turkey Gnat were observed, were carried out in a different way. 

 The materials tried were freshly burned lime, emulsion of kerosene, 

 powdered pyrethrum, carbon-bisulphide, powdered cocculus indicus, 

 and tobacco soap. With the exception of the lime, which was thrown 

 into the water in pieces of the size of an Irish potato, all the others 

 were in a watery solution or suspension. Repeated trials with all the 

 chemicals produced the same effect. As soon as the larvee came in 

 contact with any of the insecticides they would immediately loosen 

 their hold upon the leaf and drop down-stream. When the insecti- 

 cides became so much diluted as not to incommode the larvae any 

 longer, these would again fasten to leaves. By using a larger amount 

 of the various substances many larvae were killed, as well as m.ost of 

 the small fish and aquatic insects. 



But if the breeding-places in the creeks have to be searched out to 

 apply the insecticides, it would be much more simple to remove all 

 the logs, sticks, and leaves. All the fences across the branches should 

 be removed, or rather should be replaced by wire fences, which would 

 neither impede the current nor catch as many sticks and leaves. Logs 

 and larger twigs, if not embedded too deep in the mud of the creek 

 bed or banks, will always be removed by any high water; a very com- 

 mon occurrence in the Buffalo Gnat region. Old leaves, made heavy 

 by the adhering mud, would also be carried away by any high water 

 if the obstructions in these creeks were removed, and with these sticks 

 and leaves many if not most of the larvae would be carried away 

 either into the main rivers or the lower level of the creeks or lakes, 

 where there is no current and where they would soon perish. 



If the general opinion that broken levees are to blame for the de- 

 structive swarms of Buffalo Gnats prove to be the correct one, the 

 restoration of such levees would, within a few years at most, restore 

 the former immunity from these insects. This time would be materi- 

 ally hastened by the removal of obstructions in all such parts of the 

 bayous where they would come in contact with the swiftest current. 



OVERFLOWS AND BUFFALO GNATS. 



It is very generally claimed by the inhabitants of the infested re- 

 gion that as long as the States bordering upon the Mississippi River 

 had a perfect levee system, which prevented the water from escaping 

 into the inland bayous, no damage was occasioned by Buffalo Gnats, 

 not even in districts now badly infested. It is further claimed that 

 the Buffalo Gnats appear with every overflow, and only with an over- 

 flow if such overflow occur at the proper season and with the proper 

 temperature, viz, during the first continuous warm days of March, 

 April, or May. 



The chronological data already given seem to prove such assertions 

 correct. Too much weight should not, however, be attached to these 

 data. The region is as yet rather thinly settled, and no systematic 

 records of the appearance of Buffalo Gnats in injurious numbers have 

 ever been kept. A general and widespread 'appearance of these 



