220 Journal of Agricultural Research voi. xxn. no. 4 



take a new host (as actually happened in the experiments with Cyllene 

 ia oak). A few individuals may survive and continue the strain so 

 that it becomes, after a time, at least physiologically different and may 

 also develop correlated differences of color or structure. It can hardly 

 be said that such forms are much less distinct than in the case of the two 

 species Callidium antennatum in pine and C. janthinum in juniper; for 

 even though these have a slight color distinction and each is absolutely 

 restricted to its own host, they interbreed. On the other hand, in the 

 different forms of Hylotrupes ligneus, of which the eastern form in juniper 

 is constant in marking, the western form in redwood is quite variable, 

 as is also the Rocky Mountain form in Douglas fir. The juniper and 

 redwood forms interbreed, but all attempts to mate either of these with 

 the Douglas fir form have failed. All these forms can be furnished with 

 substitute hosts, but in the experiments in which this has been done the 

 original color pattern has resulted thus far. 



The grape and hickory strains of Cyllene pictus, although showing no 

 color differences, do not readily mate. Two species of Cyllene, C. pictiis 

 and C. rohiniae, are separable only as adults, by a slight difference in the 

 color pattern, yet in seasonal and biological habits they are strikingly dif- 

 ferent. It is conceivable that one of the two species originated through 

 the adoption of a new plant and continuous breeding in that plant. 



It may be asked. If one or two years' feeding in a new host results in 

 individuals which prefer that host, thus giving rise at least to new physio- 

 logical varieties, why does not this occur more frequently in nature? 

 That it does occur must be granted, as we have species living in many 

 host plants as well as those restricted to a species or genus, but that it 

 is not of more common occun-ence is believed to be due to the high mor- 

 tality in first-stage larvae in a new host rather than to absence of oviposi- 

 tion in the new host. Although the adults show a decided predilection 

 for a favored host in ovipositing and even, in certain species, a preference 

 for the plants in which the larvae have fed for one or two generations, 

 the instinct to oviposit seems to overbalance that of host selection, con- 

 sequently new hosts are frequently selected — possibly more frequently 

 in nature than is generally realized. As an example of this, take 

 Cyllene pichis requiring hickory cut during the winter. This con- 

 dition would be fully met in tops left during logging operations. When 

 the timber cutting ceased, a concentration of adults would be left with 

 none of the favored host plant available in the right condition. The 

 grape, osage orange, and hackberry strains collected at Hummelstown, 

 Pa., were in reality taken in a woods v/hich had been logged for hickory 

 and in which operations had ceased three years prior to the finding of 

 these strains. At Falls Church, Va., in June, 1920, adults of Neoclytus 

 erythrocephalus were observed ovipositing on pine logs. Much infested 

 ash, from the previous year, was lying about from which they had 

 emerged in great numbers. 



