REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 93 
method II. Four sets of flies were employed, each set consist- 
ing of six males and six females, the sexes being tested separately 
as usual. Every test consisted of three trials which were aver- 
aged to obtain the index for that test. Each of the four sets of 
flies was tested once for each of the three colors, the successive 
tests coming at intervals of 2 hours. For each set, however, the 
arrangement of the colors in the series was varied. ‘Thus for set 
A it was red, green, violet; for set B, red, violet, green; for set C, 
violet, green; red; and for set D, green, red, violet. The results 
of these tests when averaged together for the four groups were as 
follows: males—violet, 64, red, 29.8, green, 24.5; females—violet, 
81.8, red, 64.6, green, 57.2. In every one of the four sets violet 
was first in each complete test for males and females. As be- 
tween green and red, red won in three out of the four sets for 
both males and females. Thus it would appear from this ex- 
periment that the colors are effective in the order—violet, red, 
green. 
A further and better test was later made according to the fol- 
lowing method. Two groups of insects each consisting of ten 
males and ten females, were picked out and designated as Group 
A and Group B. Each group was now tested six times at two 
hour intervals, with three trials to a test. In this case, however, 
the three trials constituting a test were not all of one color. 
Instead, there was a single trial for each of the three colors in 
every test. Furthermore, in each of the six tests the arrange- 
ment of the colors was altered according to a set plan. Group 
B was treated in exactly the same manner, except that the se- 
quence of the color arrangements for each test in the series was 
reversed.. Thus for test one the Group A arrangement was V, 
Gal fortest2ev,k, G: for test's R, GY Vetor test 4 KR, V~G; 
for test 5 G, V, R; and for test 6 G, R, V. For Group B, the 
series began G, R, V and ended with V, G, R. At the end of 
the tests the indices for all the trials of a given color were aver- 
aged together in Group A and Group B. Finally the averages 
thus obtained for Group A and for Group B were averaged. 
* The tube in this experiment was only divided into four sections instead of 
the usual five.. 
