348 
COLEOPTERA. 
India where they must occur. The structure of the larvae is held to 
support the view that this family is more nearly related to the Pythidce 
or Pyrochroidce than to any other Heteromera. Besides the metallic 
blue Autocrates cenea, Westw., of the Himalayas, Trictenotoma 
Grayi, 8m., occurs in South India, T. childreni, Gray, in the Khasis, 
and T. mniszechi, Deyr., in the Himalayas (Ann. Soc. Ent., France, 1875, 
p. LIX). Westwood (Cab. Or. Entom., 1847) figures T. childreni, Gray, 
T. temfletonii, Westw., and T. cenea, Parry, and discusses the characters 
on which he separates these as a distinct family. 
PHYTOPHAGA. 
The tendency in classification at present is to a complexity of 
families, especially in Coleoptera, and while this is possibly justified 
from structural characters, it is certain that there is not as yet sufficient 
material available to define so many families ; to all but the student of 
systematic entomology, the old broad families embracing insects 
allied in structure and habits are still the most natural and the simplest 
in actual working. We have accordingly adhered to the three fami¬ 
lies composing this series, the plant-feeding beetles ; the Bruchids are 
seed-eating, the Chrysomelids live on green plants, the Cerambycids 
in the woody tissues of plants. This makes but three families and to 
place an insect in one of these three, places it as far as these habits go. 
Modern classification makes two or more of Bruchids, 13 or more of 
Chrysomelids, and two of Cerambycids, with a great tendency to make 
more. 
The series is distinguished by the apparently four-jointed tarsi 
usually with at least one joint expanded and pubescent beneath, and 
the absence of a prolongation of the head as a rostrum. It is, in prac ¬ 
tice in the field, a peculiarly homogeneous series, the three families 
sharply distinct in all our common species. It is easy to put a Scelo- 
donta down as a weevil however, though it has no distinct beak, because 
it resembles the leaf-eating smaller weevils in which the rostrum is 
not much developed and actually the limits of these two series do, as 
they should, shade into each other. 
