232 



THE AMERICAN NATURALIST [Vol. XL VII 



The characteristic whose inheritance it was finally decided to 

 study was the ratio of the third antennal joint to the fourth 

 antennal joint. This ratio offered the advantage of a fairly large 

 range, i. e., from 1.25 to 1.75; and the mode of a fraternity was 

 soon observed to vary from about 1.35 to 1.55. 



The measurements were made upon the mother at the same 

 time with the offspring. The offspring were measured when it 

 was obvious that they were mature and soon to breed. The 

 insects were first etherized and the measurements were then 

 made with a micrometer eyepiece, with a magnification such that 

 the third antennal joint averaged about 32 units. 



Results 



The following table gives the numerical data derived from the 

 measurements made : 



The ratios that are derived from measurements of the off- 

 spring are grouped into classes, and the frequency of the classes 

 shown graphically in Figs. 1-7. 



First, it appears that all aphids fall into two classes, winged 

 and wingless. While the winged mothers had a smaller antennal 

 ratio than the wingless mothers, the antennal ratio shows practi- 

 cally no difference in the winged and wingless offspring. Thus 

 10-winged offspring of a wingless mother give an average 

 antennal index of 1.54, and 5 wingless offspring of the same 

 mother give an average antennal index of 1.48. From a wingless 

 mother with antennal index of 1.46 were derived 13 wingless 

 offspring with an average index of 1.49, while from another 

 wingless mother with antennal index of 1.50 were derived 22 

 winged offspring with an antennal index of 1.49. We may com- 

 pare the antennal indices of the two lots of offspring whether 

 they happen to be winged or not. 



The nature of the food plant may be, on the other hand, of 

 importance for the antennal index. Thus of two mothers with 

 practically the same antennal index, one was fed on opium and 

 the other on nasturtium. The progeny of the first finds its 

 mode at 1.50 to 1.59 ; of the second at 1.30-1.39. It was not 

 possible to determine from comparative studies whether there 

 is uniformly a reduction of the index in the offspring of nastur- 

 tium-fed mothers, or whether this result was due to the fact 

 that the nasturtium-fed mothers belonged to a special strain 

 with a low index. In our ignorance it is clearly permissible to 



