202 
Willis . — The Relative Age of Endemic 
body and the extraordinarily modified flowers, were older than such forms 
as Tristicha or Podostemon , which are almost radially symmetrical, and 
come near to the ordinary type of submerged water plant. Yet the latter 
are widespread and almost universal, covering the whole range of distribu- 
tion of the families, while the violently dorsiventral forms are all endemic to 
comparatively small areas, Lawia , for example, occurring from Ceylon 
to Bombay, Castelnavia in the Araguaya and one other river in Brazil. It is 
impossible to talk of local adaptation in these plants, as I have elsewhere 
pointed out ( 18 ) ; there is nothing to be adapted to. The non-dorsiventral 
forms are just as common as the dorsiventral, whether in slowly or in swiftly 
moving water. This family is perhaps the most promising in all the flower- 
ing plants in which to study evolution or anything connected therewith ; the 
moment that I saw them growing in the river at Peradeniya, I realized that 
there was an unrivalled group for the study of Adaptation, in which at that 
time I was a whole-hearted believer. 
Mutation. 
Later, Mr. Ridley objects to the mutation theory, and quotes against it 
numerous examples which show that he confuses mutations with varieties, 
and with the direct effect of changed conditions, and that he does not clearly 
distinguish between infinitesimal variations and large changes. He says 
that c an organism . . . produces . . . varieties, which if more suitable to the 
surrounding conditions than the parent form are selected . . .’ Does an 
infinitesimal variation at once produce a variety ? 
Mr. Ridley does not seem to be quite sure whether he will have 
Natural Selection with infinitesimal variations or with large changes ; he 
evidently has an uneasy feeling that if he adopt the latter he rules Natural 
Selection out of court (10, 11), for if it cannot act upon a small beginning, 
nor determine that a large variation in one direction shall be followed 
by another in the same direction, it cannot be the determining factor in the 
production of the finished article, nor can it be its explanation. 
So long as we keep to infinitesimal variations, in the literal sense 
in which they were understood until the coming of the mutation theory, it 
is quite simple to evolve anything, provided (i) that the variations are fully 
hereditary without regression, which we now know them not to be, (2) that 
they are differentiating, and not simply linear, (3) that the necessary varia- 
tions appear, and (4) that Natural Selection can act — that their appearance 
gives sufficient advantage to the plant to ensure their survival in at least a 
majority of cases. But, if we once adopt large changes, the whole case is 
altered. We know no reason why a large change in one direction should 
be followed by further large changes in the same direction. There is 
nothing for it but to admit that the whole of a specific or perhaps even 
generic change may appear, but Natural Selection has nothing to do with its 
