32 
cordia, within a few days, amounted to upwards of 4,000 mules and 
horses, principally the former. 
Although frequently causing more or less trouble and loss, the gnats 
did not again appear, generally, and in such countless myriads until 
1882, although they caused serious injury in Tensas Parish in 1873 and 
1874, and doubtless in other localities also. 
But in 1882 they were more destructive to stock than ever before. 
The deer were driven from the woods, and frequently took refuge from 
their tormenters in the smokes, built by planters for the protection ot 
their cattle $ when in their agony they would allow people to rub the 
gnats from their bodies, and would even lay down in the glowing em- 
bers, or hot ashes, in their frantic endeavors to seek relief. 
In 1884 the gnats again appeared in great numbers, and were fully as 
destructive as in 1882. Throughout Franklin Parish, Louisiana, within a 
week from their first appearance, they had caused the death of 3,200 
head of stock. And for the first time in the history of the pest, they 
attacked horses and mules on the streets, and in the stables, in the city 
of Vicksburg, Miss. 
No general outbreak took place in 1885, yet they appeared in Tensas 
and Franklin Parishes in sufficient numbers to kill quite a number of 
mules. 
During the present season, although the gnats appeared pretty gen- 
erally throughout the country between the mouth of the Arkansas and 
that of the Bed Kiver, and westward to the Washita, and along the 
Yazoo River in Mississippi, no fatality to stock had been reported up 
to April 10, and there had been little or no suspension of work on 
plantations on account of gnats. 
Generally speaking, the Southern Buffalo-gnat may be said to infest 
the low, flat, wooded country adjacent to the Mississippi River and its 
tributaries, from the mouth of the Red River in Louisiana as far north 
at least as Southern Missouri. 
I have found nothing to indicate that these gnats originate in large 
streams, or even in small ones in hilly localities, although the latter 
may have both a swift current and a rocky bed. The fact of adult 
gnats occurring in such localities, even in destructive numbers, is not of 
itself sufficient proof of their having originated there, as they maybe 
carried long distances, and in immense numbers, by a strong wind. 
Furthermore, I have found no indication of their origin in other than 
perennial streams, although many intermittent bayous and small lakes 
were closely examined with this point in view. 
From the foregoing, we are forced to the conclusion that these gnats 
follow the tendency of others of the genus, and breed exclusively in 
the running water of small streams. But besides this, there is another 
equally essential element, viz, something to which the insect can at- 
tach itself during the adolescent stages. As no rocks are found in these 
bayous and small streams, we find the larvae utilizing wholly or partly 
