Statistics, repeatability and logic. Gaddum, Dale and one chicken

A story that used to be told in the coffee rooms concerned Sir John Gaddum FRS at a time when he was assistant to who was to be become Sir Henry Dale OM PRS, Director of the National Institute of Medical Research. Dale had a long interest in the hormones of the posterior pituitary.

The two mammalian hormones, vasopressin and oxytocin, had recently been partially (but not completely) separated from pituitary extract and Gaddum, appointed by Dale in 1927, had the job of comparing their effects on blood pressure. In contrast to mammals in which an injection of vasopressin causes and increase in blood pressure, hence the -pressin part of the name, Gaddum found in a chicken that vasopressin* caused a decrease in blood pressure. Thinking he ought to check the finding, he approached Dale for permission to buy another chicken or two.

Dale’s reply and point-blank refusal went along the lines of: Why on earth would you want another? Did you not do the experiment properly? Being assured by Gaddum that the experiment had been technically fine, Dale continued: If the result was observed and even if you had another dozen chickens in which you had the opposite result, you would still have to explain why you found a decrease in blood pressure in that chicken. So no, I cannot let you waste your time and the Medical Research Council’s money. No more chickens.

Had Dad's Army been on television on a still-to-be-launched television service I think we can envisage Dale adding, 'Stupid boy'.

This recording of blood pressure (shown then as now in mm of mercury)
before and after an intravenous injection of vasopressin in an
anaesthetised chicken was made on a smoked drum.
 From Gaddum's 1928 paper.

I suspect Gaddum, who became the expert on statistical treatment and design of bioassays, must have told the story himself to members of the Institute of Animal Physiology (now the Babraham Institute) when he was Director from 1958 until shortly before his death in 1965.

I shudder to think how many important findings are missed by the modern obsession with repeatability and statistical treatment of data before, during and after an experiment. If you cannot repeat a finding you have to be able find an explanation of why you cannot repeat it, and that explanation may itself be of biological significance. Perhaps that should be called the Dale-Gaddum rule?

Dale and Gaddum ca 1960
from here

*Vasopressin does not occur in birds. The avian version is arginine vasotocin

Gaddum JH. 1928. Some properties of the separated active principles of the pituitary (posterior lobe). Journal of Physiology 65, 434-440.