19 
have been used in the key to genera. The nature of these differences can 
best be appreciated by a study of the figures given of the genitalia of each 
genotype in connexion with the comments on the genitalia under each 
generic heading. 
SEQUENCE OF GENERA 
In the present sequence of genera the writer has followed to a con- 
siderable extent the old order, based on the spined or unspined nature of 
the fore tibiae. In a linear system of classification the relationships of 
genera can never be other than inadequately indicated and there is little 
to be gained, therefore, by introducing an entirely new sequence. The 
writer has not attempted to work out any key indicating the phylogeny of 
the various genera; such keys, at the present time, are largely guess work 
and of little lasting value, and though the writer certainly cannot subscribe 
to the one given by Hampson (Cat. Lep. Phal. IV, 6), he does not feel that 
his views on the subject have sufficiently crystallized to enable him to offer 
anything better. 
The writer has, therefore, placed Euxoa at the commencement of his 
system of classification for no other reason than that it seems to be the 
present culmination of one line of development and of comparatively 
recent origin, containing as it does a large number of species with very 
similar male genitalia and in most cases a very characteristic frontal 
tubercle, as w 7 ell as complete and heavy fore-tibial spining and a fourth 
tarsal row of spines. Following Euxoa are the allied genera Chorizagrotis , 
Protexarnis, Loxagrotis, and Pseudorthosia, which all agree as far as spining 
is concerned. 
The next genus, Pseudoseptis, has been erected for one of the puzzling 
southwestern species and seems to have no particular affinities in North 
America; it owes its present position to its complete fore-tibial spining 
and the presence of a fourth row of tarsal spines. The same may be said of 
the genus Richia which is next on the list. 
The typical Agrotid group then follows, representing the end-point 
of another line of development and with no particularly close relationship 
to the preceding groups. This comprises the genera Trichosilia , Onycha - 
grotis, Eucoptocnemis , Proragrotis, Agrotis, Feltia, and Actebia , all with 
complete fore-tibial spining and fourth tarsal row. Protogygia is only 
slightly modified, but lacks the fourth tarsal row of spines. Spaelotis and 
Choephora are aberrant as far as genitalia are concerned and the former 
possibly shows affinities to the Euxoa group; in both, however, the tarsal 
spining is reduced. Eurois, Ochropleura, Anicla, and Euagrotis have the 
typical Agrotid form of harpe but show a reduction of the fore-tibial 
spining, the outer row being incomplete and the spines limited to the apical 
portion. 
At this point a break in the logical sequence follows and the aberrant 
southern genus Hemieuxoa, with no close relationships, as well as the two 
closely allied genera Metalepsis and Cerastis are interpolated. 
Taking up the Euxoa series again we have Hemipachnobia, a highly 
specialized genus with fourth row of tarsal spines and harpe of the Graphi- 
phora type, followed by Paradiarsia which would have been placed closer 
