29 [ 449 1 



familiar example, eat almost every i^lant T\-liicli comes in their way, and 

 have been known to subsist, for a time, ni)on the dead remains of other 

 insects. The great diversity of foliage eaten by the different varieties 

 of the Batana ministra cannot, therefore, be admitted as proof of diffe- 

 rence of species. 



A much closer test of specific identity is the tolerance of a change of 

 food-plant. If two insects very similar, but yet with such differences 

 as to render their sjiecific identity doubtful, be found feeding upon diffe- 

 rent kinds of plants, and if, upon transferring each of them to the food- 

 plant of the other, they continue to feed and thrive, it is generally re- 

 garded as affording the strongest presumptive evidence that they are 

 only varieties of one and the same species ; and inversely, if they each 

 refuse to eat the food of the other, that they are specifically distinct. 

 This is no doubt, in the great majority of cases, a correct and sufficient 

 test. Can we go farther and hold it to be an unexceptionable rule f If 

 we assume it to be such, it will follow that some of the different forms 

 of the Data na ministra are distinct species f for we have repeatedly 

 tried to change the black walnut varietv to the apple tree, and the 

 walnut and sumach varieties each to the food of the other, and in every 

 instance they have persistently refused to eat. But this test does not 

 appear to us to be of such a nature as to make it infallable. If two 

 broods of some indiscriminate feeder, such as the larvae of the Spihsoma 

 virginica above referred to, should be so situated that they would be 

 compelled to feed, each upon some one species of plant, and if this re- 

 striction should be continued through many generations, it would seem 

 very probable that their tastes might become so confirmed that each 

 would refuse to eat the food of the other, especially if the two plants 

 were verj" unlike each other. 



It will have been observed that the three trees to wliich we have 

 above referred as the food-plants of the larvce of Datana — the apple, 

 the sumach, and the black walnut — belong to as many distinct families 

 of plants, far removed from each other in their natural characters. But 

 where theii' food-plants are botanic^ally allied to each other, the larvae 

 can be transferred from one to the other without difficulty. I received, 

 last summer, a number of half-grown larvi© which had, been found upon 

 the quince, and these were reared to maturity upon the leaves of the 

 apple. 



But let us proceed to describe some of the principal variations which 

 occur in these insects, both in the larv* and imago states, and see how 

 far these variations seem to depend upon diversity of food-plants. 



Without going into any extensive detail, we will take three of the 

 most characteristic varieties, those found ui)on the aj^pie, the sumach, 

 and the black walnut, and point out the most striking differences. The 

 larvjie described below are to be understood to have arrived at maturity. 



