52 



lice had been diminished by 27 per cent and the number of root-lice 

 themselves by 9 per cent. A much more marked difference was that 

 between the check plot and plots 4 and 5. If these two plots are taken 

 together as one double plot, the numbers from plot 1 also being doubled, 

 of course, for comparison, the treatment of three diskings of plots 

 4 and 5 with a 20-inch disk and harrowing once or twice appears to 

 have reduced the number of infested hills, as shown six days after 

 planting, from 102 to 29 (71 per cent), and the number of ants from 

 4,400 to 740 (83 per cent) ; the number of hills infested by root-lice 

 from 58 to 26 (86 ])er cent ), and the number of root-lice from 1,550 to 

 593 (61 per cent). There is no very clear reason why Plot 4 should 

 have been less heavily infested than Plot 2, since 4 differed from 2 

 only by one additional harrowing. Both received the same treatment 

 otherwise; but there was an interval of three days — July 10-13 — be- 

 tween the plowing of Plot 2 and the first disking, while Plot 4 was 

 first disked the next day, July 13, after the plowing was done. It 

 seems possible that the shower of rain which fell at 11 a. m. July 12, 

 may have so compacted the earth and disturbed and practically para- 

 lyzed the ants that these insects were less able to reach and reinhabit 

 their homes ; or it may be that the three days' interval between the 

 first tearing up of their quarters and the first disking enabled them to 

 put their aft'airs substantially to rights. 



A more perplexing exception is presented by Plot 3, which, it will 

 be remembered, dift'ers from all the others by the fact that it was 

 disked with a 16-inch disk which stirred the ground to an average 

 depth of two inches less than the 20-inch disk used on the other plots. 

 The toothed harrow worked also to an estimated depth of an inch less 

 than in the other plots. The effect of this difference would be to leave 

 the deeper two or three inches of the ground turned up by the plow 

 undisturbed by the subsequent operations. Possibly for this reason 

 but little effect seems to have been produced by the disk on Plot 3, 

 which differed from the check July 28 by a reduction of only 23>4 

 per cent in the number of hills infested by ants and of 6 per cent in the 

 number of ants in 90 such hills; while in respect to the aphis infesta- 

 tion the figures for this plot were greater than those for the check. 

 At the next digging, three days later (July 31), it is true, these latter 

 figures fell much below those of the check — a 35 per cent reduction for 

 the number of hills infested by root-lice and a 59 per cent reduction 

 for the number of root-lice themselves — and it is possible that the first 

 discrepancy is simply one of those "chance" occurrences common in 



