neighbouring beds ; all the seeds in their pods are of the same size, that is to say, there is no little pea at the end as in the pod of the common pea, and they are very hardy and prolific. I procured a large number of seeds from the same bin, and selected seven weights, calling them K (the largest), L, M, N, 0, P, and Q (the smallest), forming an arithmetic series. Curiously, their lengths, found by measuring ten of a kind in a row, also formed an arithmetic series, owing, I suppose, to the larger and plumper seeds being more spherical and therefore taking less room for their weight than the others. Ten peas of each of these seven descriptions, seventy in all, formed what I called a "set."
I persuaded friends living in various parts of the country, each to plant a set for me. The uniform method to be followed was to prepare seven parallel beds, each i I feet wide and 5 feet long, to dibble ten holes in each at equal distances apart, and i inch in depth, and to put one seed in each hole. The beds were then to be bushed over to keep off the birds. As the seeds became ripe they were to be gathered and put into bags which I sent, lettered respectively from K to Q ; the same letters having been stuck at both ends of the beds. Finally, when the crop was coming to an end, the whole foliage of each row was to be torn up, tied together, and sent to me. All this was done, and further minute instructions, which I need not describe here, were attended to carefully. The result clearly proved Regression; the mean Filial deviation was only onethird that of the parental one, and the experiments all concurred. The formula that expresses the