METAMEKIC SEGMENTATION AND HOMOLOGY. 
235 
the series.^ It is difficult to see how some 300 segments could 
be intercalated in the trunk of a snake such as python 
without the formation of some zone of growth ; yet no such 
zone is present. But it may be answered that the appearance 
in some specimens of incompletely double vertebrae shows 
how new segments could be added one by one. Such 
occasional abnormalities, however, do not necessarily indicate 
the mode of formation of new segments ; they can be more 
easily explained as due to partial fusion and incomplete 
development, or as pathological phenomena induced by early 
injury. 
We have seen to what absurd conclusions the theory of 
intercalation would lead us if applied to the vertebrates as 
an explanation of the variation in position of the paired litnbs 
and girdles. The whole theory is based on too narrow and 
rigid a conception of homology. It assumes that a structure 
A is definitely and unalterably related to a given segment x, 
which can be traced from one animal to another ; that 
wherever a is found there also must be x. When a changes 
its ])Osition it is then necessary to suppose, for the sake of 
preserving its strict homology, that x has moved with it. 
This results in the paradox that to preserve the homology of 
the limb, we are obliged to sacrifice the homology of whole 
regions — in some cases of nearly the whole trunk. And even 
then the strict homology of the limb (say the pectoral of 
Torpedo) is not really saved. The sacrifice has been all in vain, 
for new segments have been added to it (seventeen in the case 
of Torpedo as compared with Scymnus ; see Diagram 2) . It is 
really quite futile to attempt to define the homology of an 
organ in a segmented body by its ordinal position in the series 
of segments. 
1 Segments appear to be always added at the posterior end ; but they 
may be retarded in their development at the anterior end, as in the case 
of the head somites of the Craniate. 
' From the point of view of a study of variation Bateson rightly 
maintains ‘’the impossibility of applying a scheme of homology 
between individual segments” p. 128 (1). • . 
