TRANSFORMATIONS OF BUFFALO GNATS. 21 
front at molting, and the remainder of the skin toward the caudal 
extremity, the tracheae being withdrawn through the spiracular open- 
ings. The number of molts has not been determined. 
MIGRATORY HABITS AND LOCOMOTION OF THE LARVAE. 
On being suddenly disturbed or on the diminishing of the current, 
the larvae let themselves be carried down stream steadied by the silken 
thread previously described, which they have attached to some sta- 
tionary object, and seek a more favorable situation. In many in- 
stances the larvae have been observed almost to regain their former 
position, by winding up the thread with the thoracic proleg and 
mouth parts, but as a rule they are unable to work against the current. 
The method of locomotion in still water is similar to that of geo- 
metrid larvae. By looping the body, they bring the anal extremity 
forward beside the proleg, which is then released on the caudal sucker 
again obtaining a firm hold. 
A heavy rain causing a sudden swiftening of the current will 
often entirely change the distribution of the larvae. This fact is of 
considerable economic importance, as it may account for the sudden 
appearance of Simulium in localities usually exempt. This was well 
illustrated in the following instance, observed by the writer at Spar- 
tanburg, S. C, A fair-sized stream had been free from larvae for 
two months or more, when one of the tributaries near its source be- 
came heavily infested with young and half -grown larvae. A heavy 
rain and consequent washout carried practically every larva from 
the tributary down the large stream for nearly a mile, where they be- 
came successfully established, causing a subsequent heavy infestation. 
Young larvae seem to migrate more readily than full-grown larvae. 
Several experiments were tried in regard to this. A vessel containing 
larvae of "all sizes was placed under a running faucet, causing the 
vessel to overflow. The full-grown larvae as a rule remained in the 
comparatively still water within the vessel, often spinning their 
cocoons and successfully pupating. The young and half-grown 
larvae seemed invariably to detach themselves and would be found 
within two hours gathered on the surface over which the water was 
flowing. 
At one time the attention of fish culturists was drawn to Simulium, 
as the larvae were stated to cause the death of young trout by en- 
tangling them in a silken web. This was disproved in the American 
Entomologist and Botanist, in 1870, by Mrs. Sarah J. McBride, 1 who 
clearly demonstrated that the "death web of young trout" was a 
myth, and that the threads were not strong enough, or interwoven, 
so that fish could be entangled in them. 
1 See Bibliography, p. 35. 
