516 REPORT OF THE COMMISSIONER OF AGRICULTURE. 



insects seems to take place, however, only during an inundation, and, 

 granting the connection between the two phenomena, the causes for it 

 are yet obscure. It was by the elucidation of this problem that we 

 hoped to discover some means of preventing the injiiry of the flies by 

 preventing the multiplication of the larvfe. 



Inundations in the lower Mississippi Valley are not occasioned by 

 local rains, but by the immense volume of water brought doAvn by 

 the river and its more northern tributaries, and such overflows first 

 take place in the northern regions infested by the Buffalo Gnats, and 

 not in the southern. The earlier appearance of these insects in the 

 South would seem to invalidate the prevailing belief that an overflow 

 brings them. Similar conditions prevail in Hungary, where a closely 

 allied insect does so much injury to all kinds of live stock. There 

 the gnats appear every spring in varying numbers, forming local 

 swarms which move about with the wind: but no general invasion 

 takes place until the river Danube inundates the region infested. 



Is it not probable that swarms of these gnats are forced by the con- 

 ditions consequent upon an inundation to extend their flight beyond 

 their usual haunts to the more elevated and drier regions, and that 

 in this fact we have at least one of the causes of the connection? 

 Small swarms, otherwise local and unobserved, would thus, during a 

 period of high water, be forced to band together in such immense 

 armies. There must be other reasons, not yet clearly demonstrated, 

 why these insects appear in such vast swarms with an overflow, and 

 this problem can only be solved by a critical study of many breeding- 

 places during several seasons over the whole region involved. 



Some peculiarities of the swarms of Buffalo Gnats have been ob- 

 served, and these may, by closer study in future, throve some light 

 upon the problem. It is to be noted that all the specimens compos- 

 ing these swarms are females, and that not one male has been found 

 among them either here or in Europe. There is every reason to 

 believe that none of the females composing the blood-thirsty swarms 

 return to the loca-lities where they were born and developed. Expe- 

 rience indicates that once gorged with blood they die. The swarms 

 dwindle in proportion as they are carried away or move from their 

 breeding-places. 



Close investigation with the microscope has failed to reveal any 

 eggs in the ovaries of the females composing these swarms, and if 

 they deposit eggs at all it is before congregating to attack animals. 



These singular facts invite speculation and theory, but it were un- 

 wise to indulge in these before we have learned more about the eggs, 

 when and where deposited, and whether the females depositing them 

 are in any way different from those comprising the svv^arras. Dr. 

 Fritz Miiller has published in the Archivos do Museu Nacioiial do 

 Rio de Janeiro, Vol. IV, p. 47, pi. IV-VII, * some very interesting 

 observations on another fly {Paltosoma tot^rentiiim), the larva of 

 which is only found in the torrents and cascades of certain streams 

 descending the mountains of Brazil. There the pupae fastened by 

 the flat venter to the rocks under water, and change into the i^er- 

 fect flies. He found by opening the mature pupae that there are al- 

 ways two forms of females associated with one form of male. Tlio 

 one form of female possesses a rudimentary mouth, only fit to sip 



* Eeviews of his paper appeared in Kosmos, Vol. VIII, pp. 37-42; Nature, July 7, 

 1881, p. 214; EntomologifiV s Monthly Magazine, February, 1881, p, 20G and pp. 130- 

 132, and March, 1881, pp. 225, 226. 



