46 First Results from the Oxford Anthropometric Laboratory 
the last three ages range from about 3 to 4^ lbs. no significance need be attached 
to this slight falling off. 
The lung capacity starting at its lowest in the 18 year old group with an 
average of 4184 c.c. also rises to its highest at 21 years where 4438 c.c. is reached, 
then follows a slight drop succeeded again by another rise. 
The weight is also at its lowest at 18 years of age when the average is 
10 st. 7 lbs. 9 ozs., this rises to 10 st. 12 lbs. at 19, and to 11 st. 2 lbs. at 21, then as 
in the case of the pull there is a drop of the average. Professor Bourne tells me 
that his experience with rowing men is something of the same kind, after 18 years 
there is a sudden rise of \ stone followed subsequently by a slight drop. 
These three measurements are fairly closely correlated one with another, the 
average correlation coefficients being : strength of pull and weight '46, strength 
of pull and lung capacity '37, lung capacity and weight "59. 
Knowing the correlation between each pair of three variables it is possible to 
find by means of a formula the correlation between any pair for a constant value 
of the third variable. In this case for instance it might perhaps be said that there 
is no independent correlation between strength of pull and lung capacity and that 
the apparent correlation is due to the fact that people with large lung capacities 
are also heavy and that their weight gives them an advantage at the pull test. 
That they do derive some advantage from their weight is perfectly true, but the 
application of the formula teaches us that for people of equal weight, the correlation 
between lung capacity and strength of pull is "14. 
To take for purposes of comparison the relations of weight, pull and stature. 
Here we have the correlation between stature and pull 21, between pull and 
weight '46 and between stature and weight "66. With a positive correlation 
between stature and pull it looks as if tall men had an advantage, but again 
applying the formula we find that for men of equal weight the correlation between 
stature and pull is actually negative, namely — "13, and that the tall men are at a 
distinct disadvantage. If we took from them besides the advantage of extra 
weight which their extra stature provides them with, the increased lung capacity 
which also goes with it, we would find that this disadvantage was greater still. 
The disadvantage itself may be either due to the fact that our apparatus is more 
difficult for a tall man properly to apply his strength at, or it may be that the tall 
men are more apt to be weedy and weak in the back than the short, at any rate 
during the early part of their lives. A certain amount of evidence in favour of 
the latter supposition is given by a comparison of the correlation coefficients of 
the different age groups, in the 18 year old group we find the correlation between 
stature and weight, lung capacity and strength of pull severally smaller that in 
the succeeding ages. The correlation between stature and weight is "50 for the 
18 year olds, "63 for the 19 year olds and goes up to "76 for those aged 21. As 
the p.e. in the latter case is '02 and in the former '04, this difference is probably 
significant. 
