HOW AGKlCULTrHF, MODIFIES INSECT CONTROL. 31 



cent; the parasitism in lian<i;ing squares on tlie first was 5.79 per 

 cent, on the second 11.53 j)er cent. Here the only variable condi- 

 tions were soil, possibl}" weeds, and time of plant destruction. The 

 parasitism in the two classes of forms was dianu'trically reversed. 



At Terrell, Tex., were two fields on the sandy prairie, both planted 

 in March, but having; different weeds present. The weevil infesta- 

 tion August 20 on (Hie was 05.2 per cent, on the other 97.5 per cent, 

 while the parasitism in hanging squares on the lust was 29.5 per cent 

 and on the second 25.6 per cent. The variables were field surround- 

 ings and weevil abundance. 



The unknown influence which entered most of these examples was 

 very probably tlio relative abundance of the different species of para- 

 sites. This ma}' best be illustrated by the hanging squares from the 

 timbers and bottoms at Arlington, which are quoted above. In the 

 timbers the determinable parasites proved to be 16 Eurytoma tyloder- 

 matis, 10 Microhracon melUtor, 6 Ceramhycohius cyaniceps, 5 Micro- 

 dontomerus anthonorni, and 3 Catolaccus spp. In the bottoms there 

 were 17 Ceramhycohius cyaniceps, 13 Microdontomerus anthonorni, 10 

 Eurytoma tylodermatis, 8 Catolaccus spp., and 7 Microhracon mellitor. 

 The rank of the species was almost entirely reversed. 



Probably the most important point in the entire set of examples 

 is that the earliest crop had the most parasites. To show this in 

 another way we ma}^ refer to the conditions on the experimental farm 

 at Dallas. The first part of the field to put on squares was the first 

 part to show parasites. On July 8 infested squares were to be found 

 in six plats, but only on this earliest plat was there any parasitism — 

 5.7 per cent. On July 19 it and the adjacent plat were still consid- 

 erably in the lead. 



That the earliest field should show the highest parasitism was 

 expected by the writers in view of the early spring observations. 

 The parasites in hibernation, whether on the boll weevil or on winter 

 cohosts, all reached maturity in the latter half of March at Dallas. 

 It was reasoned that cotton, squaring and attacked by April 15, 

 would get the hibernated parasites in any part of the State; that 

 cotton squaring and attacked by May 15 would get the first genera- 

 tion of parasites from the cohosts, and so on. It is reasonable to 

 expect that cotton with squares infested in season to attract hiber- 

 nated parasites or a new brood from cohosts will fare better than 

 cotton that commences squaring when all the parasites are concen- 

 trated upon neighboring cohosts. This cotton must wait until the 

 period of the favored cohosts begins to wane before the parasites will 

 begin to seek new scenes of activity. Although it was so reasoned, 

 it was hardly expected that there would be sufficient proof to warrant 

 voicing the proposition. 



