616 Willis. — The Origin of Species by Large, rather than by 
and one would feel inclined to expect mutual intersterility as a matter of 
course. 
If so large a difference as having, or not having, endosperm, rumina¬ 
tion of endosperm, few or indefinite stamens, &c., &c., can occur, as it does 
occur, over and over again between genera which are obviously closely 
allied, we are evidently simply making difficulties for ourselves by supposing 
such differences to be gradually acquired. It must never be forgotten that 
gradual acquisition is an assumption of the theory of Natural Selection. 
Whether the differences were infinitesimal (or due to the universally occur¬ 
ring fluctuating variation), or whether they were more of the nature of 
sports, they were never supposed by Darwin and his followers to be 
anything but small, and evolution of new species was by their accumulation, 
whilst the larger groups were due to further accumulation and to destruction 
of the intermediate forms. Now the work which has been done to establish 
the theory of Age and Area goes to show that destruction of intermediates 
can no longer be invoked. There has been vast destruction of individuals, 
and probably of species which were only represented by a few individuals, 
but not of intermediates, unless these species which were destroyed were of 
intermediate type ; and in that case it is difficult to see how they could give 
rise to the later and more successful forms. Even in the earliest known 
geological horizons that contain the group there can be recognized many 
families of flowering plants that exist to-day, and that cover a very large part 
of the systematic range at present existing. They are as well and as widely 
separated as those now existing, and into families that now exist, and 
if these gaps were due to destruction, then Natural Selection must have 
operated with great rapidity and decision in the earlier ages of the flowering 
plants. If the earliest known flowering plants already show such differences 
as that between Monocotyledons and Dicotyledons, then evolution upon the 
Darwinian plan must have been going on previously (in flowering plants) 
for an enormously longer period than has since elapsed, or selection and 
destruction must have been much more rapid. 
The view that destruction of intermediate types was chiefly responsible 
for the differences between families might be more easily upheld were it not 
for the fact that one may find just as great differences occurring between closely 
allied genera, or even between species of the same genus, in which cases the 
time available, upon the Darwinian theory, must have been much less, and 
much less destruction was possible. Thus, for example, number of stamens 
is often a character of much importance in classification, yet, to take the first 
example that comes to hand, Lecythidaceae are marked by oo stamens, 
while the next family, Rhizophoraceae, show any number from 8 upwards 
to oo (and cf. below). Taking up Mr. Ridley’s ‘ Flora of the Malay Penin¬ 
sula ’, it opened at p. 380, and the first genus noticed was Trigonoch la my s , 
endemic to the region, with two species. T. Griffithii , with 6 stamens, 
