0.7 and the Sweet Pea Seeds

The sweet pea seeds from the packet marked 1875 arrived with a single number inked on the envelope: 0.7. I had weighed each seed in the laboratory at 42 Rutland Gate, then planted them, then weighed the daughters, and the ratio of the deviation of the daughter to the deviation of the parent came out at seven-tenths. A decimal place. One digit to the right of the point, and it has taken me a week to see what it does.

I confess I do not yet understand why the daughters of the heavy parents did not stay heavy. The heavy parent, chosen because it was heavy, produced daughters nearer the mean than itself. The light parent, chosen because it was light, produced daughters nearer the mean than itself. The mean held. My first reading was that the seed was somehow failing to transmit its own excess - that the parent’s peculiarity was being diluted in the act. But the dilution is not in the seed. The dilution is in the choosing.

Here is what I cannot yet name. When I select the heaviest parents, I am selecting two things at once: the part of their weight that will be passed on, and the part that was accident - weather, soil, the particular season of 1875. The second part does not pass. The daughter regresses because the accident does not repeat, and the seven-tenths is the fraction of the parent’s deviation that was heritable. The number is not a law of peas. It is a measurement of my own selection.

The correction, then. I began believing the decimal described the plant. I now believe it describes the relationship between parent and offspring within a population, and that the same seven-tenths would appear whether I studied peas or stature or anything else that varies and is partly inherited. The number is not in the seed. It is in the scatter.