54 



ratios July 28, had been 71 per cent, 83 per cent, 56 per cent, and 61 

 per cent, and it is possible that even in these plots the ant and aphis 

 population was, thru active dispersal under ground, rapidly be- 

 coming equalized with that of the other plots. The following year, 

 experiments which will be described in another section were made to 

 test the utility of heavy rolling immediately after each disking, to make 

 subterranean movement slower and more difificult and the search for 

 scattered eggs and root-lice less successful. 



In plowing the Bloomington field, unbroken strips three to five feet 

 wide had been left between the plots experimented with, but no suf- 

 ficient strip was left beside the second plot, and this was consequently 

 omitted from observation. The movements of the ants across the fur- 

 row separating the experimental plots from the unbroken strips were 

 observed continuously by Flint and Sanders in turn for the whole of 

 every night from July 16 to August 22 inclusive, a total of 38 days. 

 The ants, in crossing the furrows to escape from the plowed and har- 

 rowed plots followed fixed lines. Each of these was marked by the 

 observers as soon as established, and the number of ants seen in mo- 

 tion on it was noted and recorded for every night, equal times being 

 spent on the different lines. Migration lines were established as fol- 

 lows by the ants moving out of the several plots: Plot 1 (check) 24 

 lines; Plot 3, 53 lines; Plot 4, 58 lines; Plot 5, 60 lines,— a total of 195 

 lines observed, and an average of 57 lines for the experimental plots as 

 against 24 for the check. These lines were not all in use at any one 

 time. They seemed to be established by separate colonies, and al- 

 tho 116 lines were in operation on the first day 'of the observation, 

 others were added from time to time as follows : 1 1 on the second 

 night, 14 on the third, 8 on the fourth, 7 on the fifth, 1 on the sixth, 

 none on the seventh, 3 on the eighth, 1 on the ninth, 4 on the tenth, 

 13 on the eleventh, 3 on the twelfth, and none on the thirteenth, the 

 remaining lines coming two or three at a time up to the 16th of 

 August, after which no new lines were started. Some of these later 

 lines doubtless illustrate the normal movements of ant colonies in a 

 corn field, but the fact that over 59 per cent of them were started on 

 the first night shows ^that most of this movement was caused by the 

 stirring of the ground in the experimental plots. The number of mi- 

 gration lines running from the experimental plots averaged, it will be 

 noted, 2]A, times those from the check, and the difference in the mi- 

 gration movements of the ants becomes still more evident when we 

 compare the numbers of ants seen by Flint and Sanders on these lines. 

 They amount to a total of 2.860 for the check (Plot 1), 8.973 for Plot 



