On pp. 220-226 of your book, you state with extreme felicity, and much better than he does, Weismann's theory of the causes of variation. But it does not occur to him, and does not seem to have occurred to you, that there is a curious and unaccountable interruption in the ascending grades of sexual differentiation, for in the vegetable kingdom these do _not_ follow the grades of taxonomic ascent; but, on the contrary, and as a general rule, the _lower_ the order of evolution, the _greater_ is the tendency to bi-sexualism. Diœcious species (_i.e._ male and female organs on different plants) occur in largest proportion among the lower Cryptogams, less frequently among the higher, and more rarely still among Phanerogams. Monœcious species (_i.e._ male and female organs on the same plant, but locally distinct) occur chiefly among the higher Cryptogams and lower Phanerogams; Hermaphrodite species (_i.e._ male and female organs in the same flower) occur much more frequently among higher Phanerogams. There is, besides, another difficulty. According to Weismann and yourself, it is natural selection that has brought about sexuality 'for the sake of better results in the offspring,' by making them more variable or plastic. But how can natural selection act prophetically? Unless the variability is of use to the individuals at each stage of its advance, it cannot come under the sway of natural selection, however advantageous it may _eventually_ prove to the type. But, if one thinks about it, how can such variability be of any use to the individual? Observe, beneficial variability is quite different from beneficial variation. It is the _tendency to vary_ that is in question, not the occurrence of this, that, and the other display of it. Now, I do not see how sexuality can have been evolved by natural selection for the purpose of securing their tendency _in the future_, when it can never be of any use to individuals _of the present_. Each individual of the present is an accomplished fact; the tendency to produce variable offspring is, therefore, of no use to it individually, and so natural selection would have no reason to pick it out for living and propagating. Such is my difficulty touching this point. Another is, why do we meet with such great differences between (sometimes) allied natural genera, and even whole natural orders, as to the facility with which their constituent species hybridise? For example, species of genus Geranium will hybridise almost better than any other, those of the Pelargonium scarcely at all. I hope that at some time you will be able to get sent to me seeds of species peculiar to oceanic islands, should you hear of any botanists who are visiting such islands. G. J. ROMANES.