14 ENTOMOLOGICAL NfiWS. [Jan.. *o5 



conclusively that the ants had nothing to do with the tube 

 building. I observed the caterpillar for two weeks. The tube 

 being added to only in the forenoon at the rate of one to two 

 inches per day. At no time was the caterpillar visible, and it 

 would build all around the leaves before feeding. One inter- 

 esting fact seemed to me worthy of note, that when no twig 

 was available to the nearest cluster of leaves, it would erect 

 the tube free in a straight line towards it, though the sense of 

 sight must be out of question. By some instinct the direction 

 of the nearest food is known though the caterpillar is encased. 



Ropronia, an anomalous Hymcnopteron. 



By J. Chester Bradley, Ithaca, N. Y. 

 In every natural scheme of classification in zoology one must 

 expect to find intermediate forms between the groups that tend 

 to link them together. Were all such forms that have existed 

 still in existence, classification would be impossible. It is only 

 by the loss of connecting links that we are able to define 

 groups at all. This loss may occur in two ways, either by 

 total extinction, or the link although in main preserved to us 

 may itself have specialized at least along certain lines during 

 the ages, so that the resulting form to-day may be very far 

 different from what the original link was. 



Let us consider a diagram in which A represents a type of 

 g ^ ^ animal in past ages. At D suppose a divergence 

 in descent which by multiplication along success- 

 ful lines of specialization has formed two large 

 families, B and C. Then D represents a form 

 which is a connecting link between these families. 

 This may in rare cases be preserved to us at E 

 without change, in which case the determination of 

 its true relations becomes a comparatively simple 

 matter. But suppose the link D has continued to 

 * specialize along unsuccessful Hues so that it has 



not flourished as B and C have. Many characters of B, and 

 many of C, may be retained, and others once characteristic of 

 A, may be retained, but lost in B and C. These latter may 



