Crotch.] * t) [Feb. ;, 



Dacne also this structure goes gradually out. It may be noticed also 

 that the tarsal characters are constant in Rhynchophora, Tetramera, and 

 Heteromera, all the variations by a diminished number of joints belong- 

 ing to the sub-series Clavicomes. 



Pentamera. 



A. Ventral segments 6, 1-3 connate, 1st visible only 



at the sides Adephaga. 



B. 1st ventral segment visible in the middle. 



a. Antennae clavate, anterior coxal cavities variable Clavicomes. 



b. Antennae lamellate, cavities closed Lamellicomes. 



c. Antennae serrate or filiform, cavities open Serricomes. 



The natural arrangement would be Lamellicomes, Clavicomes, Ade- 

 phaga, Serricomes. 



These groups are fairly exact ; the difficulty lies between Clavicomes 

 and Serricomes, and no clearly dividing line has yet been expressed. I 

 think, however, that the antennae afford a moderately good character ; 

 thus the clavate members of Anobiidce and Gleridm cannot be referred to 

 Clavicomes as some of their closest allies are obviously pectinate, but 

 Cis and Bitoma [Lyctus] seem to me undoubted Clavicomes. 



Adephaga. 



A. Eyes 2, antennae filiform. 



B. Posterior coxae separate, anterior coxal cavities closed (except in 



Trachypachys and Carabini). 



C. Antennae inserted on the front Gicindelidce. 



Ca. Antennae inserted under the front Carabidce. 



Ba. Posterior coxae contiguous. 

 C. Anterior coxal cavities closed, posterior coxae 



not reaching the margin PseudomorpMda>. 



Ca. Cavities open, coxae reaching the margin. 



D. Posterior coxae large transverse Amphizoidas. 



Da. Posterior coxae very large rounded. 



E. Posterior coxae forming a plate covering 



the abdomen Haliplidce. 



Ea. Posterior coxae connate with the small 



metasternum Bytiscidm. 



Aa. Eyes two, feebly granulated, sometimes want- 

 ing ; antennae moniliform. 

 Posterior coxae small, sub-triangular, promi- 

 nent, distant Rhyssodidce. 



Ab. Eyes four, antennae short, stout. 



Anterior coxae globular. Gyriniche. 



Whether the first six families would not properly be considered sub- 

 families of Adephaga, is to me doubtful ; the dividing lines break down 

 at many points, e. g., Elaphrus is practically a Cicindelid, and Pelobius is 



