194 AMERTCAN COLEOPTERA. 



scribed, but, so far as concerns our fauna, it gives a division into 

 what appear to be natural groups, and it is hoped may serve until 

 the tribe Bembidini shall have been studied in its entirety. 



Before passing to the definition of our species, it may be well to 

 briefly review the characters used in grouping them. 



In nearly all, the head, although not narrowed to a neck, is free 

 from the thorax for some distance behind the eyes. A single excep- 

 tion to this is seen in cenescens, in which it is inserted in the thorax 

 to the eves. This is shown on Plate VI, fig. 9. 



The mentum varies in the different species, being more or less 

 emarginate in front, with a tooth of varying prominence at the 

 bottom of the emargination. In rather more than half our species 

 there will be observed, behind the tooth, two large, forameniform 

 punctures (Plate VI, fig. 13), which are entirely absent in the 

 others (Plate VI, fig. 12). This character seems an important one. 

 It is not difficult to detect, even with a moderately powerful hand- 

 lens, and, as soon as recognized, its presence or absence can readily 

 be determined even in our smallest species. 



Several differences are observable in the palpi, especially in the 

 comparative length of the small, subulate terminal joint, but I have 

 been unable to make use of these. 



The antennae are subject to some variation. The second and third 

 joints, as already stated, are subequal in the majority of our species. 

 The second joint is, however, slightly shorter than the third in 

 nanus, falli and frontalis, while in scitulus and several other species 

 of the ninth group the reverse is the case. In rufotestaceus and 

 limit (Plate VI, fig. 14) the second is very distinctly longer than 

 the third, and in the last named species the seven following joints 

 are nearly equal and submoniliform, the eleventh being longer, and 

 the antennae resemble closely those of Anillus. 



In nearly all the eyes are huge or moderately large and promi- 

 nent, In ferrugineus, however, they are comparatively small and 

 somewhat flattened. This is also the case, to a less degree, in mordax. 



The thorax affords few characters of use in the division of the 

 genus into groups. It is usually transverse, subquadrate, with the 

 base as wide or wider than the apex, or more rarely slightly nar- 

 rower. The base is usually squarely truncate, sometimes more or 

 less obliquely so each side. In the fourth group, which includes 

 tripunctatvs, ineurvus and most of the more convex species, two 

 series are observable, based upon the presence or absence of three 



