-38- 
is likely to yield a larger volume of effective kerosene extract 
than the seme weight of pyrethrum flowers. Pyrethruirt extracts are 
more effective in paralysing flies, derris extracts more effective 
in killing them. Kerosene extracts of the samples of Tephrosia 
at hand "'ere not so promising as were those of derris and cube. 
Both kerosene extracts and acetone extracts of ro ten one- bearing 
plants were tested to find out the relations between chemical com- 
position of the roots and effectiveness of the extracts against 
houseflies. Rotenone content "?es usually a good index of insecti- 
cidal value, but in some samples of derris their extracts -ere 
more effective than their rotenone content indicated. The value 
of these aberrant samples was indicated by the percentage of total 
acetone extractives or methoxyl content. In the samples of derris 
and cube studied, effectiveness was correlated on the rhole better 
with methoxyl content than with any other single chemical deter- 
mination. 
Jones, Campbell, and Sullivan ( 155 ) in 1935 reported tests 
on 32 semples of both kerosene and acetone extracts of Tephrosia , 
collected in different parts of the United States, against houseflies. 
The relative effectiveness of the kerosene extracts ^as similar to 
that of the acetone extracts, and the latter "es ^ell correlated 
with the degree of blue or blue- green color given by the Durham test, 
which the effectiveness of a sample of Tephrosia -can be roughly 
predicted. The insecticidal results were not ^-ell correlated with 
other chemical determinations. A comparison of kerosene extracts 
of a sample of derris root (rotenone, 1 percent) ,_ ith those of a 
sample of T. la ti dens (rotenone, 0,4 percent) from Florida showed 
that 1 part by weight of the derris was equivalent in effect against 
houseflies to more than 8 parts by weight of the Tephrosia . Tests 
with acetone extracts against houseflies showed that average commer- 
cial samples of derris and cube are about 10 times as toxic as a 
sample of T. Virginians root (rotenone, 0.3 percent). 
The same authors ( 154 ) in 1935 compared results of toxicity to 
houseflies of extracts of six samples of derris root, five samples 
of cube root, one sample of haiari stem, and one sample of Tephrosia 
virginjpna root with the values obtained on these samples by certain 
chemical determinations. The auantities of rotenone -present in the 
samples were too lo w to account for all the toxicity. In more than 
half the samples the figures by the Gross-Smith test, considered 
as representing the sum of rotenone and deguelin, agreed with the 
toxicity value, but in other samples thev ^ere lower. Total- 
extractive values ^ere higher than toxicity, and values based on the 
methoxyl content of the extract "'ere somewhat closer but '-ere also 
too high. When an approximate value for toxicarol was subtracted 
from the methoxyl figures, the results agreed more closely '"ith the 
toxicity figures in general than did the results of other determination; 
It was impossible, however, on the basis of the present results, to 
recommend unreservedly any one of these chemical determinations as a 
measure of the insecticidal effectiveness of rotenone-bearing plants. 
