142 READINGS IN EVOLUTION, GENETICS, AND EUGENICS 
nearly all the numerous species of snakes, he should have added a tiny 
rudiment in the case of the Python—and even in that case should 
have maintained his ideal very inefficiently, inasmuch as only two 
limbs, instead of four, are represented ? How much more reasonable 
is the naturalistic interpretation; for here the very irregularity of 
their appearance in different species, which constitutes rudimentary 
structures one of the crowning difficulties to the theory of special 
design, furnishes the best possible evidence in favour of hereditary 
Fic. 18,—A pteryx australis. Drawn from life in the Zodlogical Gardens, 
% Dat. size. _The external wing is drawn to a scale in the upper part of the cut. 
The surroundings are supplied from the most recent descriptions. (From 
Romanes.) 
descent; seeing that this irregularity then becomes what may be 
termed the anticipated expression of progressive dwindling due to 
inutility. Thus, for example, to return to the case of wings, we have 
already seen that in an extinct genus of bird, Dinornis, these organs 
were reduced to such an extent as to leave it still doubtful whether so 
much as the tiny rudiment hypothetically supplied to Figure 15 was 
present in all the species. And here is another well-known case of 
another genus of still existing bird, which, as was the case with 
Dinornis, occurs only in New Zealand (Fig. 18). Upon this island 
there are no four-footed enemies—either existing or extinct—to escape 
from which the wings of birds would be of any service. Conse- 
