44 BULLETIN 1147, U. S. DEPARTMENT OF AGRICULTURE. 
with acid lead arsenate (sample 39) . After having been dried by an 
electric fan, the leaves were fed to 50 large hungry silkworms. When 
the silkworms had ceased eating, they were removed to clean cages 
where the feces, contaminated as little as possible, were collected and 
subsequently analyzed. The next morning a sample of 34 dead and 
dying silkworms was thoroughly washed for five minutes in running 
tap water, then, one worm at a time, in six different washes, the first 
four consisting of hydrochloric acid (2 per cent) and distilled water 
and the last two of distilled water alone. A pencil brush was used for 
scrubbing them. Analysis of the sixth wash showed the presence of no 
arsenic. These experiments were repeated five times. To determine 
how much of the arsenic had passed through the intestinal walls, the 
alimentary canals of three sets were removed by careful dissections. 
The results of the analyses of these samples were as follows: 84 
entire silkworms yielded 2.66 milligrams of arsenic oxid, being 54 
per cent water-soluble; 72 silkworms with alimentary canals removed 
yielded 0.89 milligram of arsenic oxid, being 36.7 per cent water- 
soluble; the alimentary canals of these 72 silkworms yielded 1.03 
milligrams of arsenic oxid, being 55.9 per cent water-soluble; and the 
2.18 grams of dried feces from these 72 silkworms yielded 0.45 milli- 
gram of arsenic oxid. According to the figures obtained from these 
72 silkworms, 37.6 per cent of the total arsenic eaten had passed 
through the walls of the alimentary canals, 43.4 per cent of it was 
retained inside these canals, and 19 per cent of it was voided with the 
feces. Reaction (pH) of water extract from the larvae was neutral 
(7); from the alimentary canals, slightly alkaline (7.1); from the 
larvae with the alimentary canals removed, slightly acid (6.2) ; and 
from the feces, acid. 
The foregoing experiments were repeated on a larger scale by feed- 
ing 13 arsenicals sprayed on leaves to caterpillars of the catalna- 
sphinx moth (Ceratomia catalpae Bdv.). The results obtained in- 
dicate the following : (a) As a general rule, the higher the percentages 
of water-soluble arsenic in the larvae and feces, the higher the rates of 
toxicity of those arsenicals; (b) the percentage of water-soluble 
arsenic in the arsenical ingested usually has little to do with the rate 
of toxicitj^; (c) the amount of arsenic found per caterpillar is fairly 
constant for all the arsenicals used; (d) the higher the ratio of total 
arsenic (per 100 grams of larval material or feces) found in the larvae 
to that found in the feces, the higher the rate of toxicity; (e) the 
reaction (pH) of water extracts from the larvae fed various arsenicals 
seems to bear no relation to the" rate of toxicity. 
In 1920 the preceding experiments were repeated on a much larger 
scale, using the following insects: Honeybees, 2 sets (each of 100); 
silkworms, 3 sets (each of about 25) ; Ceratomia, 2 sets (each of about 
25) . The procedure followed was the same as that in the preliminary 
tests, but, in order to determine the percentage of arsenic actually 
made soluble by the juices of the insects, the percentage under 
" control results" in Table 21 was subtracted from the percentage of 
arsenic found soluble in the bodies of insects. Since the solubility of 
a minute quantity of arsenic in 500 cubic centimeters of water 
proved to be greater than that of a larger quantity, an amount of 
arsenic approximating the average amount found in a sample of the 
insects analyzed was employed as a control. 
