F. W. Dry 85 



as I made all the observations myself there was not more than one 

 personal equation as a source of error. And I took it as a guiding rule 

 in estimating percentages of attacked plants, when the maggots had left 

 the plants, in doubtful cases to count as attacked only those plants of 

 which one or more leaves showed a scar as well as the supposed crumpled 

 appearance. This would mean that a few slightly attacked plants were 

 not counted as attacked, but it was as good a rule as I could find. 

 I am confident therefore that the error due to failure to recognize 

 attacked plants was a very small one. Thus by simply counting the 

 number of attacked plants in a row the percentage in that row could be 

 calculated. 



I have found by a laboratory experiment that one female fly can 

 lay enough eggs to bring about the characteristic crumpled leaf appear- 

 ance. And after examining a number of attacked plants with eggs or 

 maggots on them, I think that the average number of eggs laid on a 

 plant may be put at about fifteen, that estimate being, I believe, low 

 rather than high. I have estimated the number of eggs in a few midges 

 cut into microtome sections to be about a hundred. This suggests that 

 one midge may infect from one to, say, ten plants. Several times 

 I have found either two or three attacked plants adjacent to each other 

 in a row of swedes, separated from the next attacked plant in that row 

 be several hundred plants. It is probable, therefore, that these two 

 or three plants were selected by one and the same female for oviposition. 

 The number of swede plants in an acre after singling is about 20,000. 

 From these data, therefore, if it served any useful purpose, we could 

 calculate the number of female flies which would be needed to infect 

 any particular percentage of plants in that area. 



I should point out that in the operation of singling, which is per- 

 formed, as a rule, from four to six weeks after sowing, a large proportion 

 of the plants are knocked up with the hoe, and die. In early sown fields 

 singling may be finished before any or many of the flies have made 

 their appearance. In later sown fields, on the other hand, some plants 

 will be attacked before and some after singling, when the number of 

 plants is greatly reduced. And sometimes in the latest sown fields 

 singling may not even commence until all the flies of the first brood are 

 dead. This must be one reason why, as will shortly be shown to be 

 the case, the figures from late sown fields are lower than those from 

 early sown fields, for all the figures were obtained after singling had 

 taken place. 



My usual method was to find the number of attacked plants in each 



