200 Journal of Agricultural Research voi. x. No. 4 



the same day. The caterpillars varied considerably in size, particularly 

 in the tests with the larger forms. Every reasonable effort was made 

 to eliminate errors which might result from such variations in size or in 

 the vitality. To accomplish this, the numerous tents and accompanying 

 foliage were in some cases left piled together overnight. Most of the 

 caterpillars would then collect on some prominent branch, and thus a 

 very uniform mixing would be obtained; otherwise, care was taken to 

 secure this uniformity when introducing the caterpillars to the sprayed 

 foliage. 



In all cases the solutions were applied to the foHage with an ordinary 

 I -quart glass- jar hand sprayer, care being taken to keep the material 

 agitated and to get the solution thoroughly applied and evenly distrib- 

 uted. The solution was allowed to dry on the foliage before introducing 

 the caterpillars. The specimens of sprayed foliage were transferred to 

 vessels containing water, so that the foliage would remain fresh and 

 attractive. Approximately 1,000 caterpillars were used in each test. 

 From time to time a few would drop from the foliage and crawl away. 

 These were always discarded. 



Every day, Sundays excepted, following the introduction of the cater- 

 pillars, the dead larvae (drop) were collected from the oilcloth spread 

 under the vessels containing the foliage. The twig was first jarred 

 sharply to dislodge any dead caterpillars that might be caught in the 

 web. The caterpillars were picked up one at a time, carefully brushed 

 with a camel's-hair brush, and placed in a paper sack, which was labeled 

 with the test number, date, and total drop for the day. The excrement 

 was then collected and run through a fine soil sieve to remove any bits of 

 foreign material. Pieces of leaves, molted skins, etc., were carefully 

 removed. The excrement was then placed in glass vials, properly labeled, 

 and filed for analysis, after which the oilcloth was washed and dried. 



The arsenates used in the experiments were prepared by one of us, in 

 order that as pure a compound as possible could be assured. Lead 

 hydrogen arsenate, basic lead arsenate, and calcium arsenate were 

 employed. The lead arsenates were prepared as directed by Robinson 

 and Tartar,* while the calcium arsenate, which is a salt of variable 

 composition, was prepared by a method recently devised at this Station. 

 It gave upon analysis the following composition : 



Per cent. 



Calcium oxid (CaO), total 28. 14 



Arsenic pentoxid (AS2O5), total 57- 91 



Water (HjO), by difference i3- 95 



This approximates closely the theoretical composition of CaH. AsO^. 



•Robinson, R. H., and Tartar, H. V. ths ARSE»fATBs op lead. Ore. Agr. Exp. Sta. Bui. 128, 32 

 p., 3 fig. 1915. 



