s possible to infer,
taking other evidence also into consideration, that the ancestors of the
starfish were at one stage of their existence stalked and sessile
organisms. But this leaves unanswered the question as to how and why the
starfish does still repeat after so many millions of years part of the
organisation of one of its remote ancestors. Why is this feature
retained, and by what means has it been conserved through countless
generations? It is clear that the answer can be given only by a science
of the causes of the production and retention of form, by a causal
morphology, based upon a study of heredity and development.

From the point of view of the pure morphologist the recapitulation
theory is an instrument of research enabling him to reconstruct probable
lines of descent; from the standpoint of the student of development and
heredity the fact of recapitulation is a difficult problem whose
solution would perhaps give the key to a true understanding of the real
nature of heredity.

To make full use of the conception of the organism as an historical
being it is necessary then to understand the causal nexus between
ontogeny and phylogeny.

We shall see in the next chapter that the transformation of morphology
from a comparative to a causal science did take place towards the end of
the century, and that some progress was made towards an understanding of
the relation between individual development and ancestral history,
particularly by Roux and Samuel Butler, working with the fruitful
Lamarckian conception of the transforming power of function.

[456] The importance of convergence came to be realised
after the vogue of phylogenetic speculation had
passed--see Friedmann, _Die Konvergenz der Organismen_,
Berlin, 1904, and A. Willey, _Convergence in Evolution_,
London, 1911. Also L. Vialleton, _Elements de
morphologie des Vertebres_, Paris, 1912.

[457] From this point of view there is a very profound
analogy between artificial and natural selection. Upon
the theory of natural selection organisms are lifeless
constructs which are mechanically perfected by external
agency, just as machines are improved by a process of
conscious selection of the most successful among a
number of competing models. (_Cf._ passage quoted below,
on p. 308.)

[458] _Arch. f. mikr. Anat._, xi. (suppl.), 1874; _Morph.
Jahrb._, ii., 1876, v. 1879, and vii., 1882.

[