PLANTS INSECTICIDAL PROPERTIES. 5 
and dusted or sprayed food thereafter was renewed daily. T 
by starting each set of experiments on the same day of the v 
the days i Sunday.- ) on which no records were taken always fell on 
the fifth, twelfth, and nineteenth days of the tests. Very little dis 
noticed, and parasitism was not bad among these larvse. 
Tent cat rpiUars. — Tent-caterpillar tents were collected in the 
fields from wild-cherry trees, and were later handled as already 
described for the webs of webworms. Dusted or sprayed wild-cherry- 
tree foliage was placed in the jars daily and counts were made daily 
as usual. Owing to the prevalence of the "wilt" or polyhedral 
disease, it was necessary to test these larvae while they were in the 
earliest instars, because during the last instar they were badly 
diseased. 
Catalpa caterpillars. — Catalpa caterpillars were collected on 
catalpa trees, and when transferred to the laboratory were trea 
as described for the webworms and tent caterpillars. 
Roaches. — These insects were reared in a specially constructed 
roach box which was very suitable for the purpose. On various dates 
20 or 25 were put in each of several small screen-wire cages in which 
they were dusted or fed poisoned foods. 
Grasshoppers. — Grasshoppers in the fourth, fifth, and sixth 
(adult) instars, caught in the fields, were either fed poisoned bran 
mash in the cheesecloth cages, already described, or were dusted in 
the small screen-wire cages. 
F >. — The insects to be tested were put in the large 
battery jars. S indies (20.32 centimeters) in diameter by 12 inches 
(30.48 centimeters) tall, each with a capacity of about 9.8 liters: then 
a 1-gram cone of powder was placed in each jar; next the powder 
was ignited: and finally a glass cover, almost airtight, was placed 
over each jar. 
RESULTS OBTAINED BY THE WRITERS. 
The writers have tested 232 preparations from 54 species of plants 
(not including tests of which the results have already been pub- 
lished) against a total number of 28 species of insects. Some of the 
more important results obtained are recorded in Tables 1 to 8. The 
plants first discussed are arranged alphabetically by genera, but 
under the subheading ,; Comparative results discussed " this arrange- 
ment is not maintained, although it is for the remainder of the dis- 
cussion under the heading " Discussion of the less important results 
obtained.'" 
For a more complete account of the species of plants discussed, 
the reader is referred to the " Catalogue of plants tested for or 
reported to possess insecticidal properties r ' and to the ''Index of 
botanical and common names of plants catalogued." 
DISCUSSION OF THE MORE IMPORTANT RESULTS OBTAINED. 
AMIANTHIUM OR CROW POISON. 
The writers obtained the following results by using amianthium 
or crow poison (Ohrosperma muscattoxicum). The puwdered bulbs 
