88 W. N. F. WOODLAND 
being the least differentiated. The muscles in the stump play 
no part in the process.’ It is also possible that the continuation 
of the lining epithelium of the canalis centralis of the spinal 
cord is produced by these histogenetic cells. On the other 
hand, it appears that the nerve trunks of the regenerated tail 
are produced by the growth into the regenerating tail of the 
torn ends of the trunks in the original tail, the posterior root 
ganglia of which ‘are increased in size or number owing to 
increase in size of the nerve bundles’. The preceding account, 
which I can confirm in full as regards the origin of the skin, 
muscles, fat layer, fat bundles, and cartilagmous tube, is thus 
in distinct opposition to the views of Fraisse, who believed that 
the skin, connective tissue, cartilaginous tube, and muscles of 
the regenerated tail are all derived ultimately from the corre- 
sponding tissues of the origmal tail—that new tissues can only 
be reproduced from tissues ike themselves. This belief is, in 
the main, not only disproved by actual observation, but is also 
contradicted by some of the results obtained from caudal 
regeneration under abnormal conditions now to be described. 
CAUDAL REGENERATION UNDER ABNORMAL CONDITIONS. 
Intervertebral Regeneration. Though Fraisse 
rightly came to the conclusion that the remnants of the old 
notochord (even if these be exposed by the injury) take no 
part in the formation of new skeletal tissue, yet since a more 
recent writer like Gadow (p. 494) is of opinion that ‘repro- 
duction of centra [in the regenerated tail] is precluded by the 
previous normal reduction of the chorda, around which alone 
proper bony centra could be formed’ (though Fraisse has 
shown that in the regenerated tail of Urodeles new vertebrae can 
be produced in the total absence of a notochord), it may be as 
well to quote, first of all, the results of my experiments on 
caudal regeneration from the posterior surfaces of caudal seg- 
ments which were cut in half,i.e. cut transversely between 
anytwo autotomy planes, ie.intervertebrally. 
These experiments were successful on four occasions (Text-fig. 
5, B, B’) and in each case, though the notochord was well 
