I have received your very kind note of April 8th. In truth it is preposterous in me to give you hints; but it will give me real pleasure to write to you just as I talk to Hooker, who says my questions are sometimes suggestive owing to my comparing the ranges, etc., in different kingdoms of Nature. I will make no further apologies about my presumption; but will just tell you (though I am certain there will be VERY little new in what I suggest and ask) the points on which I am very anxious to hear about. I forget whether you include Arctic America, but if so, for comparison with other parts of world, I would exclude the Arctic and Alpine-Arctic, as belonging to a quite distinct category. When excluding the naturalised, I think De Candolle must be right in advising the exclusion (giving list) of plants exclusively found in cultivated land, even when it is not known that they have been introduced by man. I would give list of temperate plants (if any) found in Eastern Asia, China, and Japan, and not elsewhere. Nothing would give me a better idea of the flora of United States than the proportion of its genera to all the genera which are confined to America; and the proportion of genera confined to America and Eastern Asia with Japan; the remaining genera would be common to America and Europe and the rest of world; I presume it would be impossible to show any especial affinity in genera, if ever so few, between America and Western Europe. America might be related to Eastern Asia (always excluding Arctic forms) by a genus having the same species confined to these two regions; or it might be related by the genus having different species, the genus itself not being found elsewhere. The relation of the genera (excluding identical species) seems to me a most important element in geographical distribution often ignored, and I presume of more difficult application in plants than in animals, owing to the wider ranges of plants; but I find in New Zealand (from Hooker) that the consideration of genera with representative species tells the story of relationship even plainer than the identity of the species with the different parts of the world. I should like to see the genera of the United States, say 500 (excluding Arctic and Alpine) divided into three classes, with the proportions given thus:—
100/500 American genera;
200/500 Old World genera, but not having any identical species in common;
200/500 Old World genera, but having some identical species in common;
Supposing that these 200 genera included 600 U.S. plants, then the 600 would be the denominator to the fraction of the species common to the Old World. But I am running on at a foolish length.
There is an interesting discussion in De Candolle (about pages 503-514) on the relation of the size of families to the average range of the individual species; I cannot but think, from some facts which I collected long before De Candolle appeared, that he is on wrong scent in having taken families (owing to their including too great a diversity in the constitution of the species), but that if he had taken genera, he would have found that the individual species in large genera range over a greater area than do the species in small genera: I think if you have materials that this would be well worth working out, for it is a very singular relation.
With respect to naturalised plants: are any social with you, which are not so in their parent country? I am surprised that the importance of this has not more struck De Candolle. Of these naturalised plants are any or many more variable in your opinion than the average of your United States plants? I am aware how very vague this must be; but De Candolle has stated that the naturalised plants do not present varieties; but being very variable and presenting distinct varieties seems to me rather a different case: if you would kindly take the trouble to answer this question I should be very much obliged, whether or no you will enter on such points in your essay.
With respect to such plants, which have their southern limits within your area, are the individuals ever or often stunted in their growth or unhealthy? I have in vain endeavoured to find any botanist who has observed this point; but I have seen some remarks by Barton on the trees in United States. Trees seem in this respect to behave rather differently from other plants.
It would be a very curious point, but I fear you would think it out of your essay, to compare the list of European plants in Tierra del Fuego (in Hooker) with those in North America; for, without multiple creation, I think we must admit that all now in T. del Fuego must have travelled through North America, and so far they do concern you.
The discussion on social plants (vague as the terms and facts are) in De Candolle strikes me as the best which I have ever seen: two points strike me as eminently remarkable in them; that they should ever be social close to their extreme limits; and secondly, that species having an extremely confined range, yet should be social where they do occur: I should be infinitely obliged for any cases either by letter or publicly on these heads, more especially in regard to a species remaining or ceasing to be social on the confines of its range.