119 
imens, which stand at thirty-six per cent. The pollen eaten by each 
group was thirteen per cent.—the same in both. If we combine the 
two collections, and treat the thirty-nine specimens of both as a 
whole, we find that insect food is about a third of the entire amount, 
and that the other animal elements are only trivial. The function 
of the beetles of this family of limiting the multiplication of plant- 
lice is expressed by the fact that these insects compose a fourth of 
the food of this entire collection. The pollen of grasses and Com- 
posite make fourteen per cent., the spores of lichens four per cent., 
and those of fungi nearly half the whole (forty-five per cent.)- 
SuFFICIENCY oF Data. 
The food of the Coccinellide seems to be, on the whole, remark- 
ably simple and uniform, consisting wholly of spores of ‘the lower 
eryptogams, pollen grains, and plant lice, and varying but little from 
one genus to another. This similarity is likewise reflected in the 
mouth parts, which agree as closely in form and structure as do the 
ratios of the food. I have consequently little doubt that the data 
derived from the thirty-nine specimens here discussed, will be found 
sufficient for a correct general idea of the food of the family under 
ordinary circumstances. 
With respect to the Carabide, we have other proof. In a brief 
paper published by me in 1880, in Bulletin No. 3, Lllinois State 
Laboratory of Natural History, based on an examination of only 
twenty-eight specimens belonging to seventeen species, the conclu- 
. sion was announced that about one-half of the food of this family 
consisted of vegetation, and one-third of insects; and the vegetation 
was thought to be about equally divided between cryptogams, grasses 
and exogens. If these figures or those of the present paper were 
far wrong, the probabilities would by very slight indeed that the two 
estimates would agree, especially as no comparison whatever was 
made of the two sets of data, until the tables were completed in 
their present form. When, therefore, we find that the one hundred 
and seventy-five specimens of the present paper, belonging to thirty- 
eight species, were estimated to have taken fifty-seven per cent. of 
animal food, and thirty-six of insects, and that the ratios of crypto- 
gams, graminaceous plants and exogens are respectively five, eleven, 
and five, we must conclude that the above figures are a fair average of 
the ordinary food of the family. 
Recurring now and finally to the questions propounded at the 
commencement of this paper,* we have to note the replies which the 
facts collected enable us to make. 
As far as the Carabide are concerned, the answer must vary 
according to the genus and species—some being so far vegetarian 
in. habit that their function as checks upon insect life is only 
trivial in importance. Respecting those which are to be properly 
classed as insectivorous, it is plain from the foregoing data that a 
very sensible effect must be produced upon already existing oscilla- 
tions. So many species were found eating a great excess of cater- 
pillars in the orchard where canker-worms abounded, that we cannot 
doubt that they had been tempted from their usual regimen by the 
eeatbrenLOSs 
