REGENERATION OF .LOST PARTS IN ANIMALS. 43 
Even in ordinary development things may go wrong, and a 
headless creature may be formed. 
If we believe that this is the best of all possible worlds— 
i.¢., that adaptations are perfect—then the fact that regenera- 
tion does not always work rightly may prove a stumbling- 
block. But there seems little reason for accepting the 
belief. 
Let us consider a few more instances. Werner points out 
that when a lizard regrows its tail, it does not always adhere 
to the pattern (8. B. Akad. Wiss. Wien, v. (1896), pp. 34-35). 
When the scales are comparatively simple, the regeneration 
is almost perfect, but when the scales are complex and there 
is much ornamentation, the regenerated tail is simpler than 
that which has been lost; it tends to be an ancestor’s taal 
that is regrown. Or to put it in another way, the orna- 
ments and elaborations which are present in the adult, 
though absent in the embryo, are absent also in the regen- 
erated tail. This does not seem surprising, if regeneration be 
regarded as due to a local persistence and re-awakening of 
embryonic growing powers. It is interesting also to notice 
that if the upper or distal part of the tail be different from 
the proximal part nearer the tip, then the regenerated tail 
follows the distal pattern. 
There seems, indeed, to be a widespread tendency towards 
the reproduction of a simpler or ancestral form, or, in some 
cases, of a simpler, more embryonic form. Thus in cock- 
roaches and walking-stick insects (Phasmide), where there 
are normally five tarsal joints at the end of the leg, the re- 
generated limb has only four tarsal joints, which is believed to 
be the ancestral number. Weismann cites the observation of 
Fritz Miiller, that in a Brazilian shrimp, Atyoida potimorim, 
the long clawed forceps are replaced in regeneration by the 
older short-fingered type of forceps, seen in the allied genus 
Caridina, That this cannot be due to mere lack of material 
is evident, he says, from the fact noted by Barfurth that the 
four-fingered hand of the Axolotl is replaced after amputa- 
tion by a more typical five-fingered hand. It might also be 
noted that regeneration will occur perfectly in half-starved 
animals. Weismann’s particular theory is that the hypo- 
thetical regeneration-germs which he supposes to reside in 
