622 Willis .— The Origin of Species by Large , rather than by 
But space will not permit of giving long lists of families and genera that 
illustrate what we have said about the well-known (but little considered) fact 
that a character may at times be family, at other times only generic 
or specific. Whether leaves are alternate or opposite is a character that is 
frequently of great family or generic importance, yet in at least 87 families, 
and in numerous genera, both may be found. Intermediates are not 
possible, and neither arrangement has any use-value as against the other. 
When the character is of great systematic importance it has in general 
no conceivable use-value as. against the contrasted character, and cannot 
have been the subject of Natural Selection. It is impossible to conceive that 
it can matter in the struggle for existence whether the corolla is valvate or 
imbricate, the leaves alternate or opposite, the endosperm ruminate or 
equable, and so on, or whether, to take even more widespread characters, 
the embryo is mono- or di-cotyledonous, or the leaves net- or parallel-veined. 
As Guppy says ( £ Age and Area p. 102), the Darwinian theory ‘ implies that 
the simpler, least mutable, and least adaptive characters that distinguish the 
great families are the last developed. This could never have been.’ 
Further than this, they are not capable of change from one to the other, 
whether by gradual variation or by destruction of intermediates, or both. 
One cannot conceive of alternate leaves becoming gradually opposite, nor in 
fact does one find any fossil relics showing intermediate stages. The same 
may be said of the various aestivations of the corolla, of trimery, of porous 
opening of anthers, number and arrangement of the ovules, &c., &c. It is 
impossible to conceive, as I have maintained for twenty years, that Natural 
Selection can have produced such characters, and to argue that they are 
correlated with important characters due to Natural Selection is simply to 
invoke incomprehensibility, as did the special creationists. These latter 
characters cannot be external, or they would have been noticed and utilized 
by the taxonomists, and to argue that they are all internal is to ask too 
much of credulity ; nor does it get over the chemical difficulty which was 
pointed out so long ago as 1877 (p. 6c6). If internal characters proceed by 
gradual selection, why does their external manifestation go in jumps, as 
must of necessity be the case with many of the most important characters? 
What intermediate types are conceivable between porous, valvular, 
longitudinal, and transverse opening of anthers—all at times family 
characters ? 
Or take ‘smaller’ variations, i. e. variations which we do not know to 
be in reality any smaller, but variations which experience has proved to be 
only available as specific or generic. In Abrus (Leguminosae, Papilionatae- 
Vicieae), in Adenanthera (do. Mimosoideae-Adenanthereae), in Ormosia 
(do. Papilionatae-Sophoreae), and in Rhynchosia (do. Papilionatae-Phaseo- 
leae) some of the species have seeds which are sharply divided into a red 
end and a black end by a difference in the colour of the testa, which is 
