STUDIES OF VARIATION IN INSECTS 327 



winged creatures, are not of sufficient advantage or disadvan- 

 tage to the individuals to lead to a weeding out (by death) or 

 saving of such varying individuals by immediate selective action. 

 Whatever the rigor and danger of the out-doors bee life, these 

 variations seem to be insufficient to cut any figure in the per- 

 sistence or non-persistence of any individual in the face of this 

 rigor. 



Still other cases in point are those revealed by our study of 

 the variation in the pattern of various insects with incomplete 

 metamorphosis, as the leaf-hopper Tettigonia sp. (p, 287), the 

 Capsid flower-bug, (p. 291), the water-boatman Corisa sp. (p. 

 293) and the variation in structure of other insects with incom- 

 plete matamorphosis, as the variation in number of tibial spines 

 of the red-legged locust Melano^his fei-mtir-rubrtim (p. 301), in 

 the periodical cicada. Cicada seftendecim (p. 306), etc. In all 

 these cases variation of much range and variety is found in 

 series of the adult individuals, during which a more or less pro- 

 tracted post-embryonic development have been exposed to the 

 struggle for existence, with their patterns and superficial struc- 

 tural characteristics in practically the same conditions as found 

 in the adult stage. This variation has existed for the most 

 part, all through the exposed life of the individual^ and has had 

 its chance to influence for weal or woe the fate of the individ- 

 ual. How much influence have these variations exerted? 



Our case which most nearly seems to illustrate determinate 

 variation is that of the variation of the flower-beetle, Diabrotica 

 soror (p. 274 et seq.). Among a thousand individuals col- 

 lected on the university campus in 1895, a certain condition of 

 variation in the elytral pattern exists as represented graphic- 

 ally by figure 53. In 1901 and 1902, other thousands col- 

 lected from the same place and examined to determine the con- 

 dition of the variation in this pattern show a distinctly different 

 status, as illustrated in figures 51 and 52. (To be sure that a 

 series of 1,000 individuals really reveals the conditions of this 

 pattern variation, repeated series of 1,000 individuals each were 

 examined and found practically identical.) The difference in 

 the variation status between the 1895 lot and the 1901-1902 

 lots consists in the dominance in 1901-1902 of one of the two 



