I should be very much obliged for your opinion on the enclosed. You may remember in the three first volumes tabulated, all orders went right except Labiatae. By the way, if by any extraordinary chance you have not thrown away the scrap of paper with former results, I wish you would return it, for I have lost my copy, and I shall have all the division to do again; but DO NOT hunt for it, for in any case I should have gone over the calculation again. Now I have done the three other volumes. You will see that all species in the six volumes together go right, and likewise all orders in the three last volumes, except Verbenaceae. Is not Verbenaceae very closely allied to Labiatae? If so, one would think that it was not mere chance, this coincidence. The species in Labiatae and Verbenaceae together are between 1/5 and 1/6 of all the species (15,645), which I have now tabulated. Now, bearing in mind the many local Floras which I have tabulated (belting the whole northern hemisphere), and considering that they (and authors of D.C. Prodromus) would probably take different degrees of care in recording varieties, and the genera would be divided on different principles by different men, etc., I am much surprised at the uniformity of the result, and I am satisfied that there must be truth in the rule that the small genera vary less than the large. What do you think? Hypothetically I can conjecture how the Labiatae might fail--namely, if some small divisions of the Order were now coming into importance in the world and varying much and making species. This makes me want to know whether you could divide the Labiatae into a few great natural divisions, and then I would tabulate them separately as sub-orders. I see Lindley makes so many divisions that there would not be enough in each for an average. I send the table of the Labiatae for the chance of your being able to do this for me. You might draw oblique lines including and separating both large and small genera. I have also divided all the species into two equal masses, and my rule holds good for all the species in a mass in the six volumes; but it fails in several (four) large Orders--viz. Labiatae, Scrophulariaceae, Acanthaceae, and Proteaceae. But, then, when the species are divided into two almost exactly equal divisions, the divisions with large genera are so very few: for instance, in Solanaceae, Solanum balances all others. In Labiatae seven gigantic genera balance all others (viz. 113), and in Proteaceae five genera balance all others. Now, according to my hypothetical notions, I am far from supposing that all genera go on increasing forever, and therefore I am not surprised at this result, when the division is so made that only a very few genera are on one side. But, according to my notions, the sections or sub-genera of the gigantic genera ought to obey my rule (i.e., supposing a gigantic genus had come to its maximum, whatever increase was still going on ought to