Species of Agra of the Amazons Region. 361 
the number to 28. Since then, the researches of numerous 
collectors in Brazil and New Granada, Moritz in Venezuela, 
Sallé in Mexico, Bar in Cayenne, and myself in the region of the 
Amazons, have enabled the Baron de Chaudoir, our best autho- 
rity on the genus, to examine himself and describe 87 species, 
which added to others described by various authors, but unknown 
to M. de Chaudoir, make a total of 124 species described up 
to the present time. To this number I am enabled to add, in 
the following pages, 16, in addition to the large number pre- 
viously described by M. de Chaudoir from my collection, which 
advances the number of known species to 1406. The general 
rarity and retiring habits of the 4gre, together with the extreme 
narrowness of the range of most of the species, lead us to think 
that even this large number will eventually be greatly augmented, 
so that Agra will become one of the most numerous genera of 
the Geodephaga. The species form a most difficult study ; this, 
however, is saying little, for in the present state of the science all 
the large groups of the Geodephaga are in the same predicament ; 
the slight but sure differences in the general form, colour and 
sculpture, outline of head and thorax, requiring great labour and 
patience, and a well-trained eye, to discriminate one species from 
another. That the great majority of Agre so discriminated are 
perfectly sound species [ have no doubt whatever, this conclusion 
being grounded on the evident physiological and permanent 
separation of several of their closely-allied species which [ found 
inhabiting one and the same locality. Such species as 4. varians 
and 4. cytherea; A. chalcoplera and A. biseriata, are cases in 
point: the absolute structural characters which separate these are 
detected only by a practised eye, but they are constant, and the 
two groups of individuals in each case show, by the absence of con- 
necting forms, their physiological separation, or, in other words, the 
absence of intercrossing. In other cases the closely-allied forms 
inhabit distant localities ; the physiological test is not then appli- 
cable, and we have to decide on their probable distinctness from 
the analogy of the other cases. Some few species remain of which 
single specimens only were found ; these are described as distinct 
only when there is a certainty, or at least great probability, of 
their being so. When much doubt existed I have preferred de- 
scribing them as doubtful varieties, under the head of the next- 
allied species. 
In the 4gre@ well-marked secondary sexual characters, in most 
cases, offer excellent characters for distinguishing closely-allied 
species. The genus is remarkable amongst the Carabici on this 
DD2 
