38 BULLETIN 1147, U. S. DEPARTMENT OF AGRICULTURE. 
and comparing the toxicity of the three Paris greens with that 
of the lead arsenate, it is 11.5 per cent more. London purple 
(sample 90) has an average percentage of toxicity of 63.9, bemg 
practically the same as that of zinc arsenite. While this sample 
killed all of the webworms tested within 20 days, only about 90 per 
cent of those fed zinc arsenite died during the same period of time. 
RELATIVE TOXICITY OF NEW ARSENATES. 
In making a comparison of the relative toxicity of new arsenates, 
three commercial products and three pure laboratory products were 
used. The commercial acid lead arsenate (sample 39) was taken as 
a standard by which to judge the relative toxicity of the other 
products. Trie two other commercial products (sample 70, acid 
lead arsenate made by a new process, and sample 62, magnesium 
arsenate) and the laboratory sample of barium arsenate (sample 71) 
are practically new, while the laboratory samples of arsenates of 
aluminum (sample 73) and of copper and barium (sample 74) are 
totally new, as far as known. 
In 1919 the following insects were tested: Silkworms, 1 set of 50; 
webworms (H. cunea), 1 set (variation 124-195, average 152); web- 
worms (H. textor), 1 set (189-514, average 314); potato-beetle larvae, 
2 sets (150-355, average 260); and grasshoppers, 2 sets (181-305, 
average 265). After deducting the mortalities of the controls, the 
following figures were obtained. When silkworms, webworms (H. 
cunea), and potato-beetle larvae were tested, the average percentages 
of -toxicity were : Sample 39 (acid lead), 58.2; sample 70 (acid lead, 
new process), 57.3; and sample 62 (magnesium), 59.8. When silk- 
worms, webworms (both species), potato-beetle larvae, and grasshop- 
pers were tested, the percentages were: Sample 39, 58.8; and sample 
62, 54.2. When webworms (both species), potato-beetle larvae, and 
grasshoppers were tested, the percentages were : Sample 39,55; sam- 
ple 71 (barium), 43.6; and sample 74 (copper and barium), 48.9. 
When webworms (both species) and potato-beetle larvae were tested, 
the percentages were: Sample 39, 57.2; and sample 73 (alumi- 
num), 34.6. 
In 1920 these experiments were repeated, with the results shown 
in Table 18, as well as the following: Silkworms, 2 sets (each of 50); 
webworms (H. cunea), 2 sets (variation 647-897, average 776); 
webworms (H. textor), 1 set (189-514, average 314) ; honeybees, 2 sets 
(each of 50); and tent caterpillars, 3 sets (240-556, average 337). 
