115 
half of which had eaten canker-worms), Diptera and Coleoptera in 
nine and four respectively, and Hemiptera in seven. Earth-worms 
were found in five, myriapods (Geophilus) in but one, and Arachnida 
(mites ana spiders) in nine. Grass-like plants were taken by thirty- 
six, and fungi by twenty-nine. e 
Scanning the totals for each genus, a few results are noted which 
are worthy of ‘special remark. First, we observe that at least two 
very abundant genera, represented by specimens enough to give us 
a fair probability that their average food is correctly exhibited, can 
hardly be classed as carnivorous insects at all, namely, Harpalus, 
with its nineteen specimens and twelve per cent. of animal food, 
and Anisodactylus, with its thirty-one specimens and twenty-one 
per cent. of the same. Amara and Amphasia should probably be 
placed in the same category, six specimens of the first and five of 
the second having taken but twenty-three per cent. and seven per 
cent., respectively, of food of animal origin. ‘I'he excessively abundant 
Agonoderus ranks but little higher as a carnivorous insect, fifteen 
examples having derived only about one-third of their food from 
animal sources. On the other hand, twenty-three specimens of 
Chlenius and seventeen of Galerita had taken about nine-ienths of 
their food from insects, mites, myriapods and earth-worms. ‘Thir- 
teen specimens of Pterostichus had obtained three-fourths of theirs 
from similar sources, while Hivarthrus and Calathus, represented by 
seven and six specimens respectively, had averaged ninety-three per 
cent. and sixty-seven per cent. 
The fact has already been alluded to that the Carabide proper 
had eaten only animal food, and that nearly all this was of a fluid 
character. 
Second, we find the Carabide dividing into at least three tolerably 
distinct groups as respects their food: first, those which seem usually 
to seize their prey and suck its juices, and take vegetation rarely, 
if at all; second, those which take a much larger ratio of animal 
food than of vegetable, but masticate and swallow it, as a rule, in- 
cluding indigestible fragments; and third, those whose habit is 
essentially vegetarian, but which still take solid animal food in 
diminished ratios. A fourth group, consisting of Lebia and its allies, 
is perhaps obscurely indicated by the facts relating to the three 
specimens of~Loxopeza atriventris studied. This will probably be 
found to feed largely upon pollen and fungus spores, after the man- 
ner of the Coccinellide; and the fossorial Carabide will, perhaps, 
constitute a fifth. 
If we look now to the structures of these beetles for some expla- 
nation of their differences of habit, we shall find corresponding 
variations in the form and structure of the mandibles. Where the 
mandibies are long and curved, and are destitute of basal molar 
processes, but are provided at or near the middle of the cutting 
edge with processes relatively long and sharp, the beetle seems to 
feed substantially upon soft or liquid animal food. If they are of 
medium length, somewhat slender, broad at base and tapering dis- 
tally, with the tip acute, and provided with basal processes which 
are not especially prominent or sharp, the food is chiefly animal, 
but solid structures are masticated and swallowed, and some vegeta- 
