Miscellanea 
533 
(ii) that the number of calories in the animal food is more influential than the number of 
calories in the vegetable food for the case of the boy, and the reverse is true for the case of the 
girl*. As a matter of fact the size of the probable errors shows us merely that it is very 
unsafe to draw any conclusions at all. On the basis of the scanty data provided by Miss 
Lindsay for the physical fitness of the individuals subjected to the various dietaries, no legiti- 
mate conclusions whatever can be drawn as to how wages or diet affect the individual ; still less 
can it be asserted that a return to porridge and milk, and an avoidance of the purchase of flesh, 
fish and eggs would save the situation. 
Indeed as far as any stress whatever can be laid on her slender material the conclusion to be 
drawn is entirely the other way, i.e. the greater the consumption of animal food relative to veget- 
able food the heavier will be the child. To test this the ratio of money spent on vegetable food 
( V) to the money spent on animal food (A) per individual was correlated with the weight of the 
child for constant age, i.e. the partial correlation a r v3,rjA was found. Further as the money 
spent on food might not have been spent to the best advantage the ratio of the number of 
calories in the vegetable food C v to the number in the animal food C A was also correlated with the 
weight for constant age, i.e. the partial correlation coefficient a r c^jc 4 , w was found. The following 
values were determined : 
Girls Boys 
«rr/A,» -"32 ±-14 --25 ±-16, 
• r W*« - -- 12 ±- 15 --23±-16. 
Now the correlations are not very large as compared with their probable errors, but 
they have all one sign, and this is negative. In other words the increase of expenditure 
on vegetable food relative to the expenditure on animal food, or the increase in the number 
of calories obtained from vegetable food as compared with the number obtained from animal 
food is associated in every case with decreased weight of the children. Thus we see that as 
far as any conclusion whatever can be drawn from Miss Lindsay's data, it is directly opposed 
to her statement that it is better to spend money on oatmeal, peas or beans than on flesh, fish, 
eggs, etc. She has started with the dogma that a calorie is of equal value whatever its origin, 
and not stayed to investigate whether it was even justified by her own material. Meanwhile 
her statement as to oatmeal has gone out as if it were a statistically demonstrated fact, whereas 
the only conclusion which would be justified in any measure by her data is that the proportion 
of animal to vegetable food should be kept as high as possible, if weight be taken as a test — 
although but a partial one — of the efficiency of a diet for a growing child. 
The elaborate dietaries carefully worked out by Miss Lindsay are of very small service 
indeed, because they have not been accompanied by any adequate anthropometric record of the 
families thus dieted. The present reviewer wrote at once to Miss Lindsay and later to Professor 
Noel Paton in the hope that it might still be possible to save the situation by some attempt to 
weigh at least the majority of the members of the families whose diets were recorded. This 
appears, however, to be no longer possible t, and the result is the not unfamiliar one — an 
elaborate piece of investigation has been carried out and practically no safe conclusions can be 
drawn from it. Those concerned in it have not studied beforehand what would be needful 
statistically to establish an inference, and the sections of the memoir (pp. 28 — 32) on " The 
possibility of Improving these Diets" and on "The Relationship of these Diets to Health" — 
which have been most widely quoted in the popular press — may express either correct or 
incorrect views ; the data provided are wholly insufficient to justify any conclusions at all. 
* A little consideration will show that this does not confute the result reached later, because 
in a r CAl „ and a r 0yw no attention is paid to the relative amounts of animal and vegetable food. 
t A final appeal to Dr Chalmers, since this paper was in type, has resulted in his most kindly 
attempting all that was possible under the circumstances. Some 04 children of from 5 to 14 years 
of age have been found and measured, and I hope in the next number of this Journal to revise the 
above numerical constants on this larger, but still sadly slender material. 
Biometrika ix 68 
