408 Phylum Arthropoda 



a black surface and spread out thinly. Under these conditions it is prac- 

 tically impossible to find the early molted skins. 



Among the horsefly larvae which have been studied, the majority re- 

 quire a year or longer for development, and must, therefore, be properly 

 cared for during the winter months. They will live if kept in a warm 

 room and fed during the winter, but certain species, notably Tabanus 

 atratus, will pupate and emerge during mid winter if so treated, resulting 

 in a strictly abnormal life history. Overwintering in a nearly normal 

 manner may be accomplished by keeping the larvae about 4 feet under- 

 ground. The overwintering chamber which the writer has used success- 

 fully for many species and over a period of eight years consists of a 

 hole of which the floor is 4 feet beneath the soil surface and measures 

 4x6 feet. It has brick reinforced walls and a concrete cover 6 inches 

 thick. The cover is 1 foot below the soil surface and covered with a foot 

 of soil. The entrance is an opening at one end measuring 2x4 feet and 

 provided with two doors, one at the level of the concrete cover, and 

 one at the surface. The top door is made water tight. The hole„neces- 

 sarily must be located on high ground so that water will not seep in. 

 Larvae should be placed in such winter quarters by the average date for 

 the first killing frost in the locality, and left until after danger of frost 

 has passed in the spring. The writer has repeatedly overwintered hun- 

 dreds of larvae in such a chamber, leaving them there from October 15 

 to April 1. During that period the mortality has averaged about 2%. 

 There is practically no evaporation from the sand in the jelly glass cages 

 and the temperature (at Fayetteville, Arkansas) remains between 50 

 and 6o° F., even though temperatures as low as -io° occur above the 

 surface. Such an underground chamber incidentally is ideal for rearing 

 many species of subterranean insects, and for overwintering reptiles and 

 other animals which normally hibernate during the cold season. 



For most species no special care is required at the time of pupation. 

 Almost invariably the larvae crawl completely out of the sand to pupate, 

 the pupa, after the pupal molt, lying on top of the sand. Species reared 

 in soil usually pupate beneath the surface. Larvae of T. atratus ap- 

 parently pupate more readily if moved into soil-filled cages when nearly 

 full grown. 



Adult flies to be mounted should be removed from the rearing cages 

 as soon after emergence as they are dry and firm. If left an extra hour 

 they are likely to beat their wings off, and injure other parts, by buzzing 

 about in the sand. Those to be kept alive should be transferred to the 

 oviposition cages already described. 



Feeding caged flies on blood is difficult and only a very small per- 

 centage may be induced to feed. Cylindrical screen wire cages 5 inches 

 long and 2 inches in diameter, capped with muslin at both ends, have 



