18 
Walter Frank Raphael Weldon. 1860—1906. 
criticise extremely complicated algebraical investigations, and to reproduce and 
often simplify them for the use of his own students. He had, however, a touch 
with observation and experiment rare in mathematicians. In problems of 
probability he would start experimentally and often reach results of great com- 
plexity by induction. Thus he was able to find out a number of problems 
relating to the correlation between a throw of n dice, and the result obtained 
when a re-throw of m out of the number n was made, and others relating to the 
mixture of n packs of cards and the throwing out of random portions*. In 
all these cases Weldon was illustrating by a game of chance a definite biological 
process. 
From 1890 onwards, Weldon's knowledge, theoretical and experimental, of the 
theory of chance increased by bounds. Weldon and the present writer both 
lectured from 1 to 2, and the lunch table, between 12 and 1, was the scene of 
many a friendly battle, the time when problems were suggested, solutions brought, 
and even worked out on the back of the menu or by aid of pellets of bread. 
Weldon, always luminous, full of suggestions, teeming with vigour and apparent 
health, gave such an impression to the onlookers of the urgency and importance 
of his topic that he was rarely, if ever, reprimanded for talking ' shop.' 
It is difficult now, after fifteen years of common work and continuous inter- 
change of ideas, to distinguish where one or other idea had its source, but of this 
the writer feels sure, that his earliest contributions to biometry were the direct 
results of Weldon's suggestions and would never have been carried out without 
his inspiration and enthusiasm. Both were drawn independently by Galton's 
Natural Inheritance to these problems, but the papers on variation and correlation 
in shrimps — which in rough outline are types of all later biometry — were published 
before their friendship had begun. 
Weldon's work at University College commenced in 1891. The house in Wimpole 
Street was taken and, if possible, life became more intense. Easter was spent at 
Chartres. In the summer came the annual visit to Plymouth, where work on 
crabs was now to replace that on shrimps. September gave some rest with a sea 
trip to Malta. In October came the college inaugural lecture for the session, 
Weldon taking as his subject the statistical treatment of variation. At Christmas 
there was a break for opera in Munich and Dresden. This year and the next 
were strenuous years in calculating. The Brunsviga was yet unknown to the 
youthful biometric school ; the card system of correlating variables was still un- 
developed, we trusted for multiplication to logarithms and Crelle, and computors 
trained to biometric work had to be created. The Weldons toiled away at masses 
of figures, doing all in duplicate. At Easter, 1892, they went to Malta and Naples, 
and the summer was spent over crab-measurements at the zoological station in the 
* In the summer of 1905 a great deal of work was done by the present writer in conjunction with 
Weldou on mixtures of card packs, tlie main features of the work having been already outlined by 
Weldon. The results are summed up in a theory of determinantal inheritance which, it is hoped, 
will be eventually published. 
