With respect to trees on islands, I collected some cases, but took the main facts from Alph. De Candolle, and thought they might be trusted. My explanation may be grossly wrong; but I am not convinced it is so, and I do not see the full force of your argument of certain herbaceous orders having been developed into trees in certain rare cases on continents. The case seems to me to turn altogether on the question whether generally herbaceous orders more frequently afford trees and bushes on islands than on continents, relatively to their areas. (110/5. In the "Origin," Edition I., page 392, the author points out that in the presence of competing trees an herbaceous plant would have little chance of becoming arborescent; but on an island, with only other herbaceous plants as competitors, it might gain an advantage by overtopping its fellows, and become tree-like. Harvey writes: "What you say (page 392) of insular trees belonging to orders which elsewhere include only herbaceous species seems to me to be unsupported by sufficient evidence. You cite no particular trees, and I may therefore be wrong in guessing that the orders you allude to are Scrophularineae and Compositae; and the insular trees the Antarctic Veronicas and the arborescent Compositae of St. Helena, Tasmania, etc. But in South Africa Halleria (Scrophularineae) is often as large and woody as an apple tree; and there are several South African arborescent Compositae (Senecio and Oldenburgia). Besides, in Tasmania at least, the arborescent Composites are not found competing with herbaceous plants alone, and growing taller and taller by overtopping them...; for the most arborescent of them all (Eurybia argophylla, the Musk tree) grows...in Eucalyptus forests. And so of the South African Halleria, which is a tree among trees. What the conditions of the arborescent Gerania of the Sandwich Islands may be I am unable to say...I cannot remember any other instances, nor can I accept your explanation in any other of the cases I have cited.") In page 4 of your letter you say you give up many book-species as separate creations: I give up all, and you infer that our difference is only in degree and not in kind. I dissent from this; for I give a distinct reason how far I go in giving up species. I look at all forms, which resemble each other homologically or embryologically, as certainly descended from the same species.