Brass calipers at the South Kensington Museum anthropometric

The brass calipers at the South Kensington Museum anthropometric laboratory recorded a head width of 15.4 centimeters for the subject numbered 742 on the afternoon of May 12, 1884. This single datum, when placed against the accumulating records of several thousand Londoners, occupies a position so far removed from the mean that it threatens the symmetry of the binomial distribution. The bell curve requires a smooth tapering toward the extremities, yet this specific measurement refuses to reconcile with the expected frequency of the population. It persists as a jagged protrusion on the chart, a point of resistance that the quincunx cannot immediately absorb.

A single deviation of this magnitude suggests a failure in the instrument or a clerical error in the ledger. However, a re-examination of the subject confirmed the measurement. The calipers are true; the subject is authentic. When such a fact fails to fit the curve, the initial impulse is to treat it as a sport or a monster, an observation outside the jurisdiction of natural law. This is a mistake in method. The individual is merely a sample of one from an underlying distribution that we have not yet fully mapped. If the curve cannot accommodate the point, the fault lies in the insufficiency of the sample size or the misidentification of the population.

The regression toward mediocrity dictates that the offspring of such an extreme outlier will inevitably move back toward the mean of the race. The stubbornness of this 15.4-centimeter measurement is temporary when viewed through the lens of successive generations. The individual case is a transient disturbance, whereas the distribution is the permanent reality. The most anomalous point in a scatter plot is still governed by the law of error; it simply represents the rare occasion where the multiple independent causes of variation have all acted in the same direction simultaneously. The observation remains in the ledger as a valid entry. I shall increase the total sample to ten thousand to see if the gap between this outlier and the shoulder of the curve begins to fill with intervening cases.