Walter Frank Rai^hael Weldon. 1860 — 1906. 33 
in some cases parents of opposite deviation had been mated, so that a rathex' 
influential negative assortative mating resulted. But from other series of pedigree 
moth data that I have since seen, it seems to me probable that there is some 
special feature in heredity in moths, or possibly in those that breed twice in the 
year, and that the vast piece of work which Weldon and Warren undertook in 
1898 — 1901, may still have its lesson to teach us. At the time it formed another 
link in that chain of apparent failures which for a time, but only for a time, 
disheartened Weldon. 
In these three first years at Oxford, Weldon's intellectual activity was intense. 
The letters to the present writer, which in 1899 averaged one a week, in 1900 and 
1901 reached an average of two, and in some weeks there were almost daily 
letters. These letters not only teem with fruitful criticism and suggestion with 
regard to the recipient's own work, but contain veritable treatises — drawings, 
tables, calculations — on the writer's own experiments and observations. To the 
pedigree moth experiments was added in the summer of 1900 an elaborate series 
of Shirley Poppy growings, 1250 pedigree individuals being grown and tended in 
separate pots ; Weldon's records were the most perfect of those of any of the 
cooperators, and his energy and suggestions gave a new impetus to the whole 
investigation. They were ultimately published in Biometrika under the title, 
Cooperative Investigations on Plants, I. On Inheritance in the Shirley Poppy (22). 
As Weldon himself expressed it, the moths and poppies meant "a solid eight hours 
daily of stable-boy work through the whole summer, and through the Easter 
vacation, with decent statistical work between." The autumn of 1899 provided no 
proper holiday, but Christmas found the Weldons in Rome. After the Shirley 
Poppies were out of hand in the summer of 1900, the Weldons went to Hamburg 
and thence to Plon. The object of this visit was to collect Clausilia at Plon and 
Gremsmiihlen for comparison with the race at Risborough. The same aim — the 
comparison of local races— led Weldon at Christmas to collect land snails in 
Madeira. Thus he slowly built up a magnificent biometric collection of snail 
shells — i.e. one sufficiently large to show in the case of many local races of a 
number of species the type and variability by statistically ample samples. Of 
this part of Weldon's work only two fragments have been published, " A First 
Study of Natural Selection in Clausilia laminata (Montagu)" (23), and "Note on 
a Race of Clausilia itala (von Martens) " (24). In the first of these memoirs 
Weldon shows that two races of C. laminata exist, in localities so widely separated 
as Gremsmiihlen and Risborough, with sensibly identical spirals, although no 
crossing between their ancestors can have existed for an immense period of time, 
and although there are comparatively few common environmental conditions. At 
the same time, while no differential secular selection of the spiral appears to have 
taken place during this period, there yet seems to be a periodic selection of the 
younger individuals in each generation, the variability of the spirals of the young 
shells being sensibly greater than that of the corresponding whorls of adults. In 
other words stability to the type is preserved by selection in each new generation. 
Biometrika v 5 
