THE BIHDS OF AUSTRALIA. 
had no standing in nature. As an instance, the present association of 
species was quoted, as one worker had grouped together all the species 
generally referred to Accipiter and Astur under one generic name. This 
was combated on the score of convenience, “ as these were natural genera.” 
Such inconsistent argument does not please me. If genera be natural, 
convenience cannot enter into a discussion, and if convenience govern, 
then there can be no recourse to natural genera. 
As I view genera, they are associations of species showing natural 
affinity. When two species show discord I separate them and thereby indicate 
that there is a cause for investigation. 
In the present case we have a series of birds classed under Accipiter 
which do not show aU the same relationship to each other, but which 
have one common structm*al feature, viz., the lengthened middle toe. 
Another series classed under Astur show very similar structural features 
throughout, save that they lack the elongation of the middle toe. It 
must be conceded by any thinking person that on the score of convenience 
these two names should be upheld. It must be convenient to be able 
to place a bird in a smaller group by such a feature. However, it has 
been argued by thinking workers that convenience is gained by lumping 
these so that whether the long middle toe is present or not no one is 
the wiser without criticising double the number of species. I cannot under- 
stand that kind of convenience and consequently do not utilise it. 
I prefer to be able to identify a bird by means of features which 
can be recognised at sight. Thus the lengthened middle toe enables me 
to regard the bird as a representative of Accipiter sensu lato. Then I 
conclude, should further criticism necessitate it, I would subdivide that 
into different subgenera or sections. In this Order I am convinced that 
anatomical study is necessary to determine the limits of the genera, so 
in the present case am regarding the Australian representative of Accipiter 
as referable to Accipiter sensu lato. I would, however, regard the species 
cirrhocephalus VieiUot as very closely allied to the preceding genus Urospiza, 
and would not be surprised to find it, on anatomical examination, to be 
more nearly related to that genus than to Accipiter where it is now placed. 
It agrees in coloration with Urospiza and passes through the same plumage 
changes. In order to emphasize this peculiarity, I propose for Sparvius 
cirrhocephalus VieiUot the new subgeneric name 
Paraspizias. 
Since this determination I have noted that Gurney {Ihis 1875, p. 474) 
wrote: “There remains but one other section of the Sparrow-Hawks included 
by Mr. Sharpe in the genus Accipiter which appears to me to be distinctly 
74 
