72G PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxvi. 



out of veinlets that are longitudinally pla(^ed; or, as illustrated in the 

 following diagram (tig. 20). 



A few .strong cross veins for binding the longitudinal veins together 

 would be developed first. In the wider spaces between the longitudinal 

 veins the cells would take on hexagonal form as soon as necessitv for 

 economy in the disposition of vein-forming substance arose, and hexa- 

 gons would, of necessity, fall into regular series or rows. This con- 

 dition would early be attained — has been attained or surpassed by 

 almost all living Odonata. Meanwhile the reduction among cross 

 veins goes on apace, and the fiuidamental thing in this reduction con- 

 sists in the better marshaling of the hypodermal cells at the proper 

 moment of transformation. 



This brings us to the practical point, that cross veins are constant in 

 number and position only when they have attained individual responsi- 

 bility — when each has its own share in the stress of the wing stroke. 

 Utility determines survival among cross veins as among species. No 

 better illustration could be found than the two cubito-anal cross veins 

 before the triangle in the wings of the Libellulida>. The lirst of these 

 is formed, as we have seen, about the descending portion of the anal 



Fig. 20. — Diagram illustrating the emergence of strong cross veins. 



trachea, and that is a sufficient reason why in the I'cduction of cross 

 veins it never disappears. The second in the fore wing l)ecomes braced 

 against the front angle of the triangle and acquires new responsibility, 

 which insures its preservation; one would as soon expect to miss those 

 othei- cross veins which l)ound the triangle. But in the hind wing 

 this same cross Acin acquires no such importance — rather loses through 

 the recession of the triangle any importance it may have had to begin 

 with, so it early disappears. In the more generalized Li])ellulida^ it 

 is always present; in the more specialized it is uniformly a])sent, Imt 

 in a few, in this respect transitional genera, it is present or absent 

 inditl'erently. Similarly in the more specialized Libellulidte ])ut one 

 cross vein persists in the space which the bridge incloses, and there is 

 a very good reason why that one never disappears — it is formed about 

 the descending trachea which precedes the bridge. 



Very similar are the antenodal and postnodal cross veins, of which 

 so nuich use has ])een made in systematic work. In the fore wing of 

 Neut'de.schna the antenodals are more than forty, the number being 

 inconstant; in Paehydiplax they are always six. In the hind wing of 

 Nmraeacluui th<> antenodals ai-e about twenty-tive; in that of Tet/ra- 



