5o8 Journal of Agricultural Research voi. x, no. io 



obtained when chips are soaked for 24 hours. (3) The smaller the 

 chips and the finer the powder, the greater is the quantity of extract 

 removed. And (4) the larger the volume of water used as a solvent, the 

 greater is the quantity of extract removed — that is, 10 gm. of chips 

 soaked for 24 hours in 3,000 c. c. of water yielded 32.1 per cent more 

 extract than 10 gm. soaked for the same period in 250 c. c. of water. 



It now remains to be determined whether or not experimental results 

 obtained by using quassia extracts on insects will support the preceding 

 data. Preliminary experiments soon proved that quassia extracts are 

 efficient on aphids only; therefore the results dealing with the effective- 

 ness of these extracts on other insects are briefly discussed under the 

 heading, "Pharmacological effects of quassiin." 



In the various experiments in which many experimental formulas con- 

 taining quassia extracts were used for the purpose of eliminating all those 

 formulas found to be inefficient in the laboratory, the following aphids 

 were employed: Tulip-tree aphids (Macrosiphum liriodendri Mon.), rose 

 aphids {Macrosiphum rosae L.), nasturtium aphids {Aphis rumicis L.), 

 cabbage aphids {Aphis brassicae L.) on kale, pea aphids {Macrosiphum 

 pisi h.), aphids {Aphis sp.) on bladder senna {Colutea arhorescens L.), 

 woolly beech aphids {Phyllaphis fagi h.), and those aphids {Chaitophorus 

 populicola Thos.) found on Carolina poplars. The following leaves, 

 branches, and entire plants, each bearing many aphids, were collected 

 between 7 and 8 a. m. and were placed in bottles of water on a long table 

 by windows : Leaves of tulip trees {Liriodendron tulipifera) , nasturtiums 

 {Tropaeolum spp.), and kale {Brassica oleracea viridis) ; branches of rose- 

 bushes {Rosea spp.), bladder senna, beech trees {Fagus americana), and 

 Carolina poplars {Populus deltoides) ; and entire sweet-pea plants {Lathyrus 

 odoratus) in small pots. The aphids were then sprayed with an atomizer, 

 and the bottles and pots, with their contents, were so arranged on the table 

 that each of them received an equal share of light. The insects were 

 counted before any of them died, and at regular intervals throughout the 

 day those remaining alive were recorded. The three following interfering 

 factors were usually present: (i) To a limited degree the aphids left the 

 leaves and branches and crawled toward the windows; (2) most of the 

 spray solutions were slightly repellent and certainly the less effective ones 

 caused a small percentage of the aphids to migrate from the leaves and 

 branches; (3) the tulip-tree leaves late in the afternoon showed evidence 

 of drying, which consequently caused the remaining live insects to leave 

 them sooner or later. The first factor is insignificant when comparing 

 various results obtained in the laboratory; but, when these results are 

 compared with those obtained outside the laboratory, it must be taken 

 into account. Inside and outside the laboratory the second factor 

 probably has the same weight; but, when the mortaHty of aphids sprayed 

 is compared with that of those not sprayed, a small probable error should 

 usually be allowed. To overcome most of the error caused by the third 



