156 THE BRITISH MUSEUM CATALOGUE OF BIRDS, VOL. Y. 
Sundevall’s system, equally with this latter, involves. He 
remarks (Vol. IV., p. 6) :— 
“In attempting to draw a hard and fast line between groups 
which nature has connected by intermediate forms, the difficulty 
of dealing with the connecting links has to be faced. I have 
preferred to accept a line which is capable of definition even in 
cases where forms, apparently nearly allied, must be separated, 
being convinced that any line, wherever drawn, must be subject 
to the same objection.” 
This is equivalent to the assumption that the line in question 
is as good as any other that can be drawn; but when I find 
that line, to give two instances, uniting a ground feeding Thrush 
like Cochoa, with Hemipus and Tephrodornis, and these again 
with Platylophus and Hypocolius, and on the other hand 
dividing off into different families Ceétia und Prinia, I confess 
that I, personally, think but-poorly of that line, and look 
forward to the drawing of a very different line, open to much 
less serious objections. 
Mr. Sharpe’s work is admirable, but, to my idea, the frame 
in which he has placed it is most artificial and wnnatural. 
But to return to the present volume. In a most interesting 
preface Mr. Seebohm gives some idea of the principles by 
which he has been guided in his work. He explains that he 
has rejected the old-fashioned axiom that genera must be 
founded upon structural characters, because he is convinced 
that these so-called structural characters have no generic value 
at all, and has fallen back for his generic characters upon colours, 
or pattern of colour, as a character which in fact dates further 
back than the shape of the wings, tail, or bill. 
Without stopping to enquire for the proof on which this last 
assertion is based, I may say that to my mind the results of 
this novel system appear in the highest degree unsatisfactory ; 
and that, for instance, each of Mr. Seebohm’s genera Geocichla, 
Turdus and Merula constitute a melancholy jumble of disere- 
pant forms. If we are to have citrina, Lath., monticola, Vig., and 
wardi., Jerd, in one genus, then at least let us return to the 
simplicity of Linué, and keeping only Turdus, discard, along 
with all the other genera, Geocichla and Merula. There is 
no logical standpoint between Turdus, solus and some fifteen or 
seventeen genera. 
But the idea of structual characters for genera is not the 
only old-fashioned notion that Mr. Seebohm has felt impelled to 
reject. He feels equally bound to reject those rules of the British 
Association Code of scientific nomenclature, which do not meet 
his approval. Of course he gives very weighty reasons for his 
violations of the law—reasons that, if considered only with 
reference to the particular instance, might well be allowed to 
