METAMEUIO SEGMENTATION AND HOMOLOGY. 
231 
Or to put the matter in a more general way — suppose we 
call the consecutive segments a B c n . . . In one case 
the limb may belong to segments a b , in another to segments 
B C, and in a third to segments c and so on. Again in one 
animal a limb may belong to segments E F, in another to 
segments c D E f g, and in a third to segments a B c n E fg H. 
How are we to reconcile these facts with our conception of 
homology ? 
Some anatomists seem to consider the paired limbs and 
limb-girdles as special organs developed outside the truly 
segmented region of the body, as it were independent of it, 
and therefore capable of moving from one region to another. 
But this view is quite inadmissible; as already pointed out 
above, it is directly contradicted by the well-established facts of 
anatomy and embryology. The limbs are as truly segmented 
as the myotomes and vertebral column, and are formed from 
the same materials. The segmentation may, of course, be 
obscured in the adult, just as it may be in the trunk or head, 
but it is obvious in development. Moreover, the limbs never 
really abandon the segments from which they originated (12). 
It certainly is not by a theory of the migration of ready- 
formed limb-material in the embryo that we can explain the 
shifting of limbs, or preserve our ideal of homology. 
Migration during ontogeny may indeed occur, but to a very 
limited extent. The direction and amount of the migration 
is faithfully recorded in the adult by the course of the nerves 
supplying the limb. For instance in the frog the straining 
backwards of the pelvic plexus shows that the base of the 
hind limb was moved back with the elongation of the pelvic 
girdle. In such fish as the whiting, where the pelvic fins have 
attained a jugular position actually in front of the pectorals, 
their real place of origin is betrayed by the nerves which pass 
backwards to the spinal cord, crossing those supplying the 
pectoral fins (see Diagram 4). 
Certain variations in the attachment of the pelvic girdle to 
the sacral vertebrae may be explained in the same way. In 
the salamander, as shown by von Jhering (17), the pelvic 
