NOVEMBEE 29, 1912] 



SCIENCE 



753 



literature of evolution, where it has figured so 

 largely. 



23.1 50.0 111 33Z 

 3aS +/.2 50.0 56.0 



/)l«a7i 



4fc.7 3^.1 33.3 385- 

 «33 klS 7/.4 iS-O 



Fig. 6. The survival rate in the extreme cases 

 in the Bumpus sparrow ease to test periodic selec- 

 tion. The rates plotted are the rates made up by 

 averaging those for the corresponding extreme 

 classes of all the dimensions, each expressed in 

 percentages of length. First line of numbers, the 

 extreme classes; second line, survival of females; 

 third line, survival of males. 



The weight of the sparrows shows a secular 

 selection in a negative direction, but in per- 

 centage of length, the selection seems indiffer- 



233 

 71 



ISJ 

 41.S 



2 ha 

 45 



271 

 45 



Fig. 7. Survival polygon for weight in the 

 Bumpus sparrow case plotted in four classes. 

 First line of numbers, weight in grams; second 

 line, survival rates. 



ent. The selection in absolute weight, there- 

 fore, is due to the correlation with length 

 which is selected. For this reason and be- 

 cause it is not a dimension, I have not plotted 

 it with the other characteristics (Figs. 7 

 and 8). 



The application of this method to Cramp- 

 ton's observations on the length of antennae 



in female pupae of the moth Philosamia cyn- 

 thia, is shown in Figs. 9 and 10. This shows 

 clearly the amount of selection and brings out 

 the fact, which is otherwise overlooked, that 



Fig. 8. Survival polygon for length in the 

 Bumpus sparrow ease plotted in four classes. 

 First line of numbers, length; second line, survival 

 rate. 



Fig. 9. The survival polygon of the length of 

 left antennas in female pupte in Crampton's ex- 

 periment plotted in four classes. First line of 

 numbers, length of antenna in millimeters; second 

 line, survival rate. 



Fig. 10. The survival polygon of the length of 

 left antenna in female pupEe in Crampton's ex- 

 periment plotted in eight classes. First line of 

 numbers, length of antenna in millimeters; second 

 line, survival rate. 



