THE ANCESTRY OF THE CHORDATA. 545 
number of these is liable to great variations, not even being 
constant in the species. For example, certain deer and also 
certain sheep have specifically more horns than two; and in 
the case of Iceland sheep the horns may be three, four, or five 
(Youatt, ‘The Sheep’). By the nature of the case none of 
these repetitions can be atavistic ; and it is interesting to notice 
how, just as it was shown that irregular repetitions of parts 
about the axes of symmetry of the body often take up regular 
secondary relations to them, recurring either in segmental 
pairs or in radial symmetry, so these minor repetitions take up 
regular relations (secondary in some cases, probably primitive 
in others) to the axes of the limb or part of the body in which 
they occur. Thus the ossifications in the Crinoid stem or 
the Starfish arm are so regularly related to the axis of the 
part that in the latter case they have suggested to Haeckel his 
extraordinary view of the phylogeny of the group, appearing to 
him precisely similar to the segmentation of a Chatopod. The 
case of the scales of fishes and the hairs and markings of cater- 
pillars should perhaps have been more properly quoted in the 
former connection, as being an instance of irregular repetitions 
which have become definitely related to the symmetry, as in 
the case of the Sturgeon, and among caterpillars the Tussocks 
and the Spherigide. One very curious instance may be 
quoted of a series of repetitions which, though essentially 
arranged with reference to the axis of a limb, have yet a defi- 
nite relation to the long axis of the body. This instance is 
that of the Vertebrate tail, which has often been adduced by 
opponents of the Annelid theory of Vertebrate descent. Now, 
the structures which repeat themselves in the Vertebrate tail 
with great variability of number, namely, the vertebre with 
their neural and hemal arches, the segmental vessels and 
nerves, &c., are precisely those structures upon whose repeti- 
tion in the trunk the view of the primitive character of the 
segmentation of the Vertebrata mainly depends. 
In the foregoing pages the attempt has been made to show 
that greater or less repetition of various structures is one of 
the chief factors in the composition of animal forms, that these 
