1. The weekly budgets of eighty out of ninety-nine families in four rural schools (containing 165 school children) and ninety-seven out of 200 families in one urban school (containing 183 school children) have been analysed.

2. Evidence is given that housewives, despite varying social status, spend between 9 and 11s. per man per week on food when the income is sufficient to allow this.

3. The lowest level of sufficiency in diet according to the instincts of the housewife is seen to correspond, at prices ruling in this district in the winter of 1936–7, with the cost of the diet of the British Medical Association Nutrition Report (1933) (7s. 8¾d.). This has been called standard A and compared with standard B costing 5s. 10¾d. which contains 1000 cal. and 13 g. of first-class protein a day less. Standard B is probably the lowest standard of diet which any authority would be prepared to recommend.

4. In the school in a prosperous urban district, in working class families (income not exceeding £4) where the children were fully embarked on the school period (i.e. the eldest child over 7 years but no child yet earning more than 10s. weekly), and which contained more than one child, 60% of children fell below standard A and 47·6% below standard B.

5. In the rural schools taking all the children, 72·2% were below standard A; where there was an income below 55s., 78% fell below standard A; below 40s. 85%; and below 35s. 91·7%. The figures below standard B were 40·6, 52·4, 60, and 75% respectively.

6. Both in urban and rural families, with the exception of high income groups (e.g. farmers), increase in the size of the family steadily lowers the allowance of food per unit. In medium wage earners the allowance of 9–11s. per unit per week when the family is small, is reduced to about 3s. when the family grows to a large size and before the first children begin to earn.

7. Most significant is the decline in the provision of first-class protein foods both in town and country in proportion as the family grows. Evidence is given that this contrasts markedly with the recognized needs of growth in childhood, and it is clear that in both town and country the large family in low wage groups produces a condition of extreme protein deficiency. Accepting 2s. per unit per week as a conservative estimate of the minimum expenditure upon first-class proteins, 13·2% of the school children in the urban school and 32·7% in the rural schools were subsisting on amounts suggesting protein starvation.