A494 PASSERIFORMES CHAP. 
D. Oscines. 
Of this group, or the true singing birds, the anatomy is even 
less worked out than that of the Suwboscines, and consequently 
the relationships are in many cases extremely doubtful. The 
Hirundinidae stand distinctly apart, with no other Familhes 
near them, though some similarity of habits, and possibly of 
structure, may be recognised in certain JMZuscicapidae (especially 
if Artamus be included among them), and to a less extent in 
a few Sturnidae; yet almost as much resemblance may be 
perceived in some Zyrannidae, which (not being Oscinine) can 
have no affinity to the Mirwndinidae. The likeness is therefore 
probably one of analogy only, and, if so, of no taxonomic value. 
Here again it must be stated that the Oscines hardly attain more 
than the rank of a Family, and that in the Passeriformes the 
“Families” have not the same value as in the foregoing Orders. 
The <Alaudidae are generally recognisable by the casing of 
their metatarsi (p. 496), but in other respects they seem to shew an 
affinity to the JM€otacillidae (Anthinae) on the one hand, and to 
the Fringillidae on the other. The wing-characters, on which 
reliance was formerly placed, prove to be wholly untrustworthy. 
As to the rest of the “ Families” of Oscines, it is impossible 
to indicate their probable relations by placing them in sequence, 
or even at present to group more than a few of them with any 
degree of assurance. Some have not yet been defined at all, and 
few of them sufficiently for anyone to be confident as to their 
limits. Thus in the MJuscicapidae hardly any two systematic 
ornithologists will agree as to which genera should be included 
and which excluded. If restricted to Old World forms, there is 
still no defined boundary between them and the Campephagidae, 
while there is perhaps even less distinction between them and 
the Sylviidae (or Sylviinae, if these last be included among the 
Turdidae). Moreover, by some taxonomers the Mniotiltidae have 
been referred to the Juscicapidae, and though there is an obvious 
distinction between them in the number of the primaries, the 
instance of the <Alaudidae (already mentioned) shews that this 
may be of very slight importance. 
Some other “Families” may perhaps be grouped with less 
risk of error. The Tvroglodytidae, Certhiidae, Sittidae and 
Paridae are possibly such; but the limits of the last-named are 
