94 University of California Puhlications. [Entomology 



tion of the points where the connectives have arisen ; secondly, 

 for conditions within the wing favorable to the production of 

 this specialization; and thirdly, for evidence of its utility, by 

 reason of which natural selection will tend to maintain the 

 structure when once produced. In regard to the last item little 

 need be said, as the utility in strengthening the wing is very 

 evident; the fact that the two connectives whose positions 

 and structure are the most uniform are the ones where the 

 utility is the most evident, is distinctly confirmatory of this 

 view. 



There is one fact in the matter of the location of the con- 

 nectives that is at least very suggestive. When the wingpad 

 of one of these insects is pulled a little to one side, it will be 

 seen that the organ does not bend in a straight line, but along 



a curve corresponding to that 

 shown in the accompanying dia- 

 gram (Fig. 29). The reason for 

 this is evidently the difference in 

 the stiffness of the different parts 

 of the wing, dependent on its 

 T^T^ on T^• 1, • 1- f shape. Now with the wingpads 



FIG. 29. Diagram showing line oi ir =i 



n® m'h^ '" ^^"^ ^^^^ "* Odonatid placed as they are, a current in 



the water, or the movement of 

 the insect, would result in a slight bending of these organs 

 back and forth. Along the bending line the cells would have 

 a slightly different environment, and it is possible that the 

 cells of this region might become more and more sensitive 

 to this change until they responded by the production of fully 

 perfected connectives. 



How this might be brought about is explainable in two ways. 

 It may be that we have here simply a direct response to 

 mechanical stimulation, which is known in both plants and 

 animals to produce, under proper conditions, very decided 

 hypertrophies. Another, and perhaps sounder, explanation is 

 that the bend resulted in a slight displacement of the principal 

 veins where the bend occurs, so that cross veins occurring at 

 these points are accordingly strengthened. A support to this 

 explanation is the fact, already pointed out, that in the Prot- 

 odonata the primary and first posterior are somewhat approxi- 

 mated in Avhat will be the arculus region; further, the very 

 evident bend of the two veins in front of the primary at the 



