Ethel M. Elderton 
59 
then sub-divided the patterns under 53 lieads and an experiment was made with 
the fore, middle and ring fingers of 150 paternal couples by Mr F. Howard Collins. 
The resulting table shows the same thing but the multiplicity of sub- divisions, 
especially when the number of pairs used is comparatively small, makes it even 
more difficult to assign any degree of correlation though some undoubtedly exists. 
Seventeen sets of twins show an even closer resemblance, two sets agreeing in all 
their three couplets, four agreeing in two and five in one. There ai^e instances of 
partial agreement in five others. 
The knowledge of the inheritance of type of finger-print was at this stage in 
1903 when Galton began to collect the material which forms the chief part of the 
data used in this paper. The total number of persons whose finger-prints were 
taken was 2,300 but many of these are isolated individuals and of no use for the 
study of inheritance, but when these are omitted we have over 400 cases in the 
tables dealing with maternal inheritance, over 300 in the paternal tables and 
between 800 and 900 in the fraternal tables. The most complete family schedules 
were those collected by Mr Perrycoste from the iidiabitants of Polpeno, Cornwall, 
and in this material we have an extraordinarily interesting series. In some cases 
we have the finger-print of some old man and his wife, the finger-prints of all 
their living children, of the husbands and wives of these children and of many of 
the old folks' grandchildren. The people of Polperro evidently in 1900 had not 
started restricting their birth-rate and we have families of eight and nine brothers 
and sisters with as many children again of the married ones. In Galton's later 
collection the finger-prints of the first fingers only were taken and the reason 
for this limitation is given in the schedule sent out from which I quote the 
following. 
"It is proposed to confine the collection to prints of the two forefingers of many 
persons, to rolled impressions of them, and to ask fox four prints of each. 
The two forefingers are selected because their patterns are more varied than 
the rest in respect to ' radial ' dii-ection, and not less vai'ied in other respects. 
Also because rolled inipressions can be obtained most easily from them, requiring 
as they do the extension of the finger in use, while the others are tightly closed to 
be out of the way. Lastly, the degree of closeness of correlation between the two 
forefingers has been found to be veiy suitable to the inquiry." 
Unluckily experience has shown us that it is more satisfactory to work with the 
prints of all the fingers of both hands, so I also considei'ed the larger and earlier 
general finger-print collection made by Galton and I have found therein about 350 
individuals who had a brother or a sister recorded ; in this case all their fincrer- 
prints had been taken. It was on this collection that most of the published work 
done by Sir Francis on finger-prints was based, but I am not clear whether the 
special material he used for the early consideration of inheritance was taken from 
this data or not. If it were then, I think, only a selection of the available pairs 
of brothers was extracted nnless the general collection was added to in more 
recent years after the publication of his book in 1892. 
