OF WASHINGTON, VOLUME XI, 19Q9 107 



varies considerably. \Ye counted more than 60 in one leaf. They 

 are at first not easily visible unless the leaf is held against the 

 light, but gradually the color of the leaf tissue which surrounds 

 the eggs becomes more or less brownish, when the eggs can be 

 seen without difficulty. A fortnight later the first larvae were 

 seen. No observations were made on the number of molts or 

 of the development of the larvae, because, as stated above, the 

 oviposition extended more than two months and there were 

 always small and large larvae to be found on the same branch 

 of the willows. In the beginning of August the full-grown 

 larvae (pi. vm, fig. 5) were seen to drop to the ground and to 

 crawl off in search of a place for pupation, which was never on 

 the low ground where the willow trees grow, but on the slope 

 of the higher ground well above the height of a moderate 

 freshet. However, in this year comparatively few of the larvae 

 reached full maturity. There were many of them starved to 

 death in places where a complete defoliation had taken place. 

 Countless others were killed by what we consider as a fungoid 

 insect disease, so that their dead bodies lying on the ground or 

 hanging on the trees made a very unpleasant stench. 



Some observations were made on two species of parasites. 

 From numerous leaves beset with eggs which were collected in 

 the first week of June we obtained in our breeding jars a large 

 number of an egg parasite which seens to be undescribed, and 

 cannot at present be referred to by name. According to Mr. 

 Crawford it belongs to the Tetrastichidae. We never observed 

 in the field the issuing of these parasites from the eggs, nor did' 

 we ever see in the field a single specimen, but from the fact that 

 \\ e bred the same egg parasite from eggs collected in the latter 

 part of June and the early part of July, it must be concluded 

 that the parasite is able to undergo two generations in the eggs 

 of a single generation of the host. 



A second parasite, which has been named by Mr. Crawford 

 as Testrastichus hylotomce, was first observed by us issuing in 

 great numbers from the cocoons found in the ground. The 

 imago parasite was seen frequently by us in July, flying singly 

 around behind the saw-fly larvae. The latter manifestly were 1 

 aware of the proximity of an enemy, and were moving their 

 abdomens in the air in the fashion of many other saw-flies, try- 

 ing to frighten off the parasite. In this case it is evident that 

 this particular species of parasites hibernates in the body of the 

 -uw-flies and issues only the next spring. Of other enemies 

 only a single species of a Pentatomid was observed to suck the 

 half-grown saw-fly larvae. 



The saw-fly under consideration (pi. VITI. fig. 4) agrees per- 

 fectly well with specimens named in the U. S. National Museum 



