tural dissimilarity necessarily going together, except indirectly
and by no means strictly. Look at the case of pigeons, fowls, and
cabbages.

I overlooked the advantage of the half-sterility of reciprocal crosses;
yet, perhaps from novelty, I do not feel inclined to admit the
probability of Natural Selection having done its work so clearly.

I will not discuss the second case of utter sterility; but your
assumptions in par. 13 seem to me much too complicated. I cannot believe
so universal an attribute as utter sterility between remote species was
acquired in so complex a manner. I do not agree with your rejoinder on
grafting; I fully admit that it is not so closely restricted as
crossing; but this does not seem to me to weaken the case as one of
analogy. The incapacity of grafting is likewise an invariable attribute
of plants sufficiently remote from each other, and sometimes of plants
pretty closely allied.

The difficulty of increasing the sterility, through Natural Selection,
of two already sterile species seems to me best brought home by
considering an actual case. The cowslip and primrose are moderately
sterile, yet occasionally produce hybrids: now these hybrids, two or
three or a dozen in a whole parish, occupy ground which _might_ have
been occupied by either pure species, and no doubt the latter suffer to
this small extent. But can you conceive that any individual plants of
the primrose and cowslip, which happened to be mutually rather more
sterile (i.e. which when crossed yielded a few less seeds) than usual,
would profit to such a degree as to increase in number to the ultimate
exclusion of the present primrose and cowslip? I cannot.

My son, I am sorry to say, cannot see the full force of your rejoinder
in regard to the second head of continually augmented sterility. You
speak in this rejoinder, and in par. 5, of all the individuals becoming
in some slight degree sterile in certain districts; if you were to admit
that by continued exposure to these same conditions the sterility would
inevitably increase, there would be no need of Natural Selection. But I
suspect that the sterility is not caused so much by any particular
conditions, as by long habituation to conditions of any kind. To speak
according to pangenesis, the gemmules of hybrids are not injured, for
hybrids propagate freely by buds; but their reproductive organs are
somehow affected, so that they cannot accumulate the proper gemmules, in
nearly