138 Discussion 



After about four weeks there was an increase of 5-7 per cent in total body 

 water, but the interesting thing is that the plasma potassium and plasma 

 sodium concentrations remained unchanged. 



Milne : Prof. Wallace, the main change in calcium with these diets was 

 in the skeletal calcium. Have you any information on changes in soft 

 tissue calcium, particularly kidney calcium? In my experience it varies 

 tremendously in rats on different calcium diets. 



Wallace : The calcium in the body is almost entirely skeletal and with 

 this kind of data it is impossible to say just where this calcium is. You 

 have to study the individual tissues. 



Fourman : Do you think that the increase in bone which you suggested 

 took place is an increase in trabecular bone — so-called freely available, 

 mobilizable, bone tissue? 



Wallace : We are not certain but think it is probably both cortical and 

 trabecular. We would like to know if the large animals on the high 

 electrolj^e intakes have more easily mobilizable bone tissue under 

 conditions of stress. 



McCance : You began by putting up charts of balances showing that if 

 the diet contained more sodium and potassium, the child absorbed and 

 retained more. Yet you find by experiment that you do not alter the 

 composition of the body. Can you reconcile those observations? 



Wallace: This is a purely technical matter on which I have strong 

 feelings. In a balance experiment the quantity of food entering the 

 body and the excreta recovered are always slightly less than the measure- 

 ments indicate. The more refined the technique the smaller this error is. 

 Also, the greater the concentration of a nutriment in the intake the 

 greater will be the error when compared with intakes of lower concen- 

 tration but of equivalent caloric value. When subtraction is used to 

 calculate the balance these errors accumulate. The errors in doing a 

 balance are not randomly plus or minus as is generally believed, but 

 systematically positive. Much of the arithmetical difficulty arises because 

 one must subtract two quite large numbers to obtain the usually very 

 small balance value. At zero intake the balance method becomes more 

 accurate. Body composition estimates such as can be made from Bene- 

 dict's and Gamble's fasting data agree with direct analysis data quite 

 well. However, body composition estimates made from balance data 

 with infants fed with cow's milk and human milk are always widely diver- 

 gent, even when weight gains are equivalent. The higher the intake of a 

 constituent the greater the apparent retention. Eventually the retention 

 becomes patently absurd. 



