Semi-Centennial Volume. 67 



smudges provided for cattle and thus became an easy mark for a shot. 

 These statements, also giving methods of protection and treatment, agree 

 substantially with accounts already mentioned or previously published. 

 However, he has known of no occurrence of the flies, at least in suffi- 

 cient numbers to have caused any complaint, during the past ten years 

 or more, excepting late in the spring of 1912, when the appearance of a 

 few examples seemed to have attracted just enough attention to furnish 

 evidence that the insect still existed in the region. 



In regard to the investigations of agents as first mentioned, both 

 were made on Bayou Pierre, which is considered to be an old channel of 

 the Red river, originating a few miles below Shreveport. South of 

 Wallace lake it forms the boundary line on which the parishes of De Soto 

 on the west side and of Caddo and Red River on the east side adjoin, and 

 it further divides the latter from Natchitoches parish west of the point 

 where the bayou empties into the Red river. The bayou is the outlet of 

 several lakes situated within the eastern border of De Soto parish. 



The agent who first visited the stream was Mr. Harper Dean, jr. 

 His notes are dated March 19-21, 1906, and the species of which he 

 found evidence only in the form of pupal cases is believed to have been 

 identical with the one actually collected in the adult stage by Mr. A. H. 

 Rosenfeld on the second investigation in the same vicinity. The deter- 

 mination of this species as Simidium pecuarum Riley, which is commonly 

 called the buffalo gnat, has lately been verified through examination of 

 the specimens by Dr. W. D. Hunter, Bureau of Entomology, U. S. De- 

 partment of Agriculture. This name will likely apply to the flies men- 

 tioned by Mr. Garrett as having occurred at Logansport, as well as in 

 the cases reported by Mr. Moreland, Doctor Dalrymple and Major Lee. 

 The results of Mr. Dean's investigation are given as he reported them: 



"On Monday, March 19, I secured the services of a guide and pro- 

 ceeded in a boat up the Bayou Pierre for about two miles from the house 

 which I used as headquarters. At this point I visited the 'Government 

 raft' that is thought to be the main breeding place for the gnats that 

 infest the Bayou Pierre section. This raft is about two miles long and 

 about 200 feet wide. It is at present nothing more than a mass of logs 

 jammed into all positions and all in different stages of decay, being 

 overgrown with tangled masses of vegetation. The water under this raft 

 is from 8 to 10 feet deep and during high water the raft is completely 

 submerged. A careful inspection of this raft failed to reveal a single 

 empty pupa case or any indication of the gnats whatever. There were 

 plenty of upright logs in it, yet not a sign of a previous brood of gnats 

 could be found. 



"After finishing the inspection of this raft, I proceeded slowly down 

 stream towards the starting point of the morning, and examined carefully 

 every log jam, projecting stump, brush pile, etc., in the stream, as well 

 as the Cottonwood, willow and cypress brakes along the banks, all of 

 which are subject to overflow. 



"The log jams showed no evidence whatever of a former brood. The 

 biush piles in some instances contained some few empty pupa cases of the 

 last brood, but practically all of the pupa cases were found on projecting 

 stumps in mid stream. All pupa cases found were on the current side of 

 the stream. These were exposed to the full sweep of the current and all 

 were attached to the timber at a height of two feet above the water level 

 at the time, indicating that the last brood, said to have appeared two 

 weeks previous to this date, occurred during a corresponding rise in the 

 stream. 



