43 
[Conn. 
of variation within a single order. As a starting point a Campode- 
oid form is taken. This is the most widely distributed and has 
frequently been pointed out as the closest representative of the an- 
cestral insect living at the present day. Starting with the Campo- 
deoid type the different families of beetles have been studied as 
far as is possible with our present knowledge of them. The follow- 
ing are the most important points presented by the study of this 
group. 
1. With the exception of the Campodeoid type of larvae, which is 
found in a number of families, all beetle larvae are secondary mod- 
ifications which have been introduced during the larval life of the 
beetles, and have never been represented by any adult features. 
They are therefore of no value in teaching the history of beetles 
except in their larval stages. They do not represent ancestral 
stages. They ma}q however, and frequently do, teach relationship, 
since the presence of a similar larva may indicate a recent com- 
mon ancestor. 
2. It is possible, amid the immense variety of larvae, to recog- 
nize four somewhat distinct types : the Campodeoid tj 7 pe, a type 
slightly and variously modified from the Campodeoid type, a Scar- 
abid t}q)e and a maggot-like tj 7 pe like that of the weevils. In many 
cases it is possible to determine definitely the sort of conditions 
that have produced the present type. 
3. The division of larvae into t} 7 pes seems to have no relation to 
the classification of adult insects into sub-orders. None of the 
classifications of adult beetles into sub-orders runs in any way par- 
allel to the natural division of larvae into groups. The classifica- 
tion of the families of larvae does, however, run parallel to the 
classification of the families of adults, so that it is usually possi- 
ble to tell from the structure of a larva to what family it belongs. 
To this rule there are many exceptions, some of which are easily 
explained by differences in habit. The exceptions are most com- 
mon in the low degraded types of larvae. The classification of 
families into sub-families and genera seems also as a rule to run 
parallel with the classification of adults, though there are many ex- 
ceptions to this rule. The exceptions are such as to indicate that 
in some cases the adult classifications are at fault, and in other 
cases that there is really no parallel between the two stages. From 
this we can draw the conclusion that the present larval types of 
beetles are about as old as families but not much older. 
