IS89-] Transmission of Acquired Characters. 565 
^- — To the objection that this proves too much, — that the 
cusps thus formed would keep on growing, it may be said [a), 
that in the organism itself these reactions occur least in the best 
adapted structures. This proposition is difficult to demonstrate 
in the case of the teeth, but may be readily demonstrated in what 
are known as the phenomena of displacement in the carpals and 
tarsals where growth has a direct ratio to impact and strain. {b\ 
In the organism itself growth does not take place beyond the 
limits of adaptation ; there is, therefore, no ground for the suppo- 
sition that overgrowth will take place by transmission, {c), 
Either by the selection or Lamarckian theory development is 
held in check by competition between the parts ; there is a limit to 
the nutritive supply ; in the teeth, as elsewhere, the hypertrophy 
of one part necessitates atrophy of another. 
C. — A general objection of considerable force is that we find 
other adaptations, equally perfect, in which the Lamarckian prin- 
ciple does not apply ; why then invoke it here ? To this it may 
be said that there is no theoretical difficulty in supposing that 
while natural selection is operating directly upon variations of the 
first class, the Lamarckian principle is producing variations of the 
second class, and while selection does explain the former, it falls 
far short of explaining the latter. 
T). — Finally, if Weissmann succeeds in invalidating the sup- 
posed proofs of the Lamarckian principle derived from pathology 
and mutilations, this will not affect the argument from palaeon- 
tology and comparative anatomy, for these proofs involve two 
elements which are not in our theorem : {a), immediate trans- 
mission of characters ; {b), transmission of characters impressed 
upon the organism and not self-acquired. 
r/iird.—As regards the adequacy of the selection principle to 
explain these variation phenomena. It is not necessary io repeat 
here the well-known current theoretical objections to this prin- 
ciple, but simply to point out the bearing of this paL-Eontological 
evidence. In Weissmann's variation theory the preponderating in- 
fluence must be conservative ; however it may explain progressive 
modification, or even correlation of old characters, it does not 
admit that the genesis of new characters should follow definite 
