Chapter 70

You know how I work subjects: namely, if I stumble on any general remark, and if I find it confirmed in any other very distinct class, then I try to find out whether it is true,--if it has any bearing on my work. The following, perhaps, may be important to me. Dr. Wight remarks that Cucurbitaceae (55/1. Wight, "Remarks on the Fruit of the Natural Order Cucurbitaceae" ("Ann. Mag. Nat. Hist." VIII., page 261). R. Wight, F.R.S. (1796-1872) was Superintendent of the Madras Botanic Garden.) is a very isolated family, and has very diverging affinities. I find, strongly put and illustrated, the very same remark in the genera of hymenoptera. Now, it is not to me at first apparent why a very distinct and isolated group should be apt to have more divergent affinities than a less isolated group. I am aware that most genera have more affinities than in two ways, which latter, perhaps, is the commonest case. I see how infinitely vague all this is; but I should very much like to know what you and Mr. Bentham (if he will read this), who have attended so much to the principles of classification, think of this. Perhaps the best way would be to think of half a dozen most isolated groups of plants, and then consider whether the affinities point in an unusual number of directions. Very likely you may think the whole question too vague to be worth consideration.