Recognized HTML document


same mechanical problem had been solved still more comprehensively by a French mathematician. Professor Karl Pearson subsequently extended its application to variables not governed by the Gaussian Law, and the exact determination of the Index of Correlation by his refined method has now become the object of most biometric work.

I have received much help at various times from Mathematical friends. On one occasion, being impressed with the probability (owing to Weber's and Fechner's Laws) that the true mean value of many of the qualities with which I dealt would be the Geometric and not the Arithmetic Mean; I asked Mr. Donald Macalister, of whom I have already spoken, to work out the results. He, as a schoolboy, was the first to gain the prize medal of the Royal Geographical Society, then became the Senior Wrangler of his year at Cambridge, subsequently Chairman of the Medical Council, and is now Provost of Glasgow University. His memoir is supplementary to mine on the " Geometric Mean," Proceedings of f the Royal Society, 1879 [53].

My first serious interest in the Gaussian Law of Error was due to the inspiration of William Spottiswoode, who had used it long ago in a Geographical memoir for discussing the probability of the elevations of certain mountain chains being due to a'common cause. He explained to me the far-reaching application of that extraordinarily beautiful law, which I fully apprehended. I had also the pleasure of making the acquaintance of Quetelet, who was the first to apply it to human measurements, in its elementary binomial form, which I used in my Hereditary Genius.