L. Harrison 
7 
The systematic worker is at present faced with two alternatives, 
when he comes to write a description of a new species. Either he must 
write a very long and detailed description, giving equal value to an 
infinite number of characters, some of which he may find later to be 
very variable; or he may try, in a brief account, to found his species 
upon a few characters of absolute diagnostic value. In the former 
case the work is laborious, which would not be of any consequence did 
the result justify it; but the longest description very frequently fails 
to characterise one of two closely related species, if it has been discussed 
without reference to the other. In the latter case there is, at present, 
no general agreement among workers, and the opinions of one man as 
to what is of absolute diagnostic value will differ widely from those of 
another. 
I should like to suggest a few methods and characters by means of 
which we may hope to have more satisfactory specific descriptions in 
the future, especially as I may in this way induce some discussion on 
the point. First I would wish to insist on the discussion of a new 
species with reference to its nearest relatives, if these can be accurately 
determined. When Giebel stated that his Nirmus angulicollis, from a 
petrel, found its nearest ally in N. fenestratus of the cuckoo, he wrote 
mischievous rubbish. The comparison of species by dumping them 
into artificially contrived groups may be almost as dangerous. Kellogg 
(1909, p. 119) writes of Lipeurus absitus from the hoatzin: “This new 
Lipeurus of the group clypeati sutura distincta (all the other members 
of which have been taken from strictly maritime and coast birds) is 
thoroughly distinct from the other species of the group.” This state¬ 
ment is, of course, absolutely correct; but the inference is that Lipeurus 
absitus is more closely related to these marine Lipeuri than to any 
others, which is certainly not the case. Obviously the group clypeati 
sutura distincta is not a natural one, or, at all events, its definition should 
be limited; and no good can come of comparing two unrelated species 
because they happen to fall within its artificial limits. I do not for a 
moment accuse Kellogg of bowing before the fetish of an inelastic group 
definition; but Neumann certainly has done so, on two separate occa¬ 
sions. He has (1914) attempted to transfer Goniodes lipogonus and 
G. complanatus to the genus Degeeriella, because they do not come 
within the verbal limits of the former generic definition. He gibes at 
Taschenberg, who previously encountered the same difficulty, for 
erecting a new genus Strongylocotes to receive these forms. Taschen- 
berg’s treatment had, at any rate, the merit of leaving the species 
among the forms to which they were most closely related. Neumann 
