insecticides. In 1932 appeared a second paper on Tephrosia, this one 
by Tattersfield and Gimingham ( 391 ) • In 1940 Tattersfield and co- 
workers ( 395 ) described the insecticidal properties of Annona and 
Mundulea and of some fish-poison plants (394) • 
In 1931 Roark and Keenan ( 347 ) published a long list of plants 
found in India that were reputed to have insecticidal value. Since 
1927 several papers on the insecticidal properties of Indian plants 
have appeared, and since 1933 several others on plants found in China, 
Worsley ( 431-434 ) between 1934 and 1939 wrote about the insecti- 
cidal properties of some East African plants, 
METHODS OF RESEARCH 
In the search for insecticidal plants no dependable guide has been 
found. The only procedure is the trial-and-error method. Botanical 
classification is not a trustworthy guide, because hundreds of plant 
species must be examined to find one that is sufficiently pr< rising 
to warrant intensive chemical and toxico logical studies. Evidence of 
this is the fact that the most important commercial insecticides of 
plant origin are distributed in five different families as follows j 
Insect powder or pyrethrum, Asteraceae; anabasine, Chenopodiaceae; 
derris, cube, and timbo, Fabaoeaej hellebore, Liliaceae; and nicotine, 
Solanaceae, 
Much time has been spent in insecticide laboratories trying to 
find an ideal test insect, which would correspond to the white rat 
commonly used in other fields of research, A method used by Mclndoo 
and coworkers ( 260 ) in a study of Derris in 1919 was to extract the 
plant materials with several solvents and to test the extracts against 
several species of insects, A later method, used by Tattersfield and 
coworkers ( 392 ) in a study of Tephrosia in 1925, was to use only two 
solvents (alcohol and water) and only one species of insect, the bean 
aphid, A still more recent method is to use only one solvent (alcohol 
or acetone) and only one species of insect, such as the bean aphid, 
housefly, mosquito larva, codling moth larva, or Japanese beetle, 
Roark ( 345 ) has pointed out that many of the tests on insecti- 
cidal plants are inadequate because the material was not tested 
against the proper insect, all parts of the plant were not tested, or 
the proper solvent was not used for extraction. He considers that the 
examination of plants for insectioidal value should be prosecuted as 
follows i 
(a) The plant should be identified by a botanist who 
is a specialist in the order or family to which it belongs. 
