14 STATISTICS OF EXPENDITURE AND 



Ames, in a letter in answer to some queries made, has further 

 explained that the last sentence refers only to families with an 

 annual income of $1000 or less. " My experience, he adds, has 

 gone to prove that rental consumes from one-fifth to one-third of 

 the income of the very poor. Then the proportion grows gradu- 

 ally less as we reach the classes where the family income runs 

 from $8.00 to $12.00 per week. Those families receiving from 

 $12.00 to $15.00 seem to pay a smaller proportion of income as 

 rent, but classes receiving from $15.00 to $20.00 seem to grow 

 ambitious and desire to move into larger quarters. I am of the 

 opinion, although I have no facts to substantiate it, that if we 

 were to take classes receiving annually $1000 a year and over, 

 we would find the rental proportionately diminishing the higher 

 we go" Thus, Mr. Ames's results hardly bear out Engel's law, 

 that the percentage expended on rent is invarirbly the same 

 whatever the income ; and it appears necessary to modify the 

 law, at least, if we admit subdivisions of the working classes. 

 The proportion is highest for the very poor, varying from 25 to 

 30 per cent. ; for the ' real industrial ' classes it falls to 16 per 

 cent. ; and then rises to 25 per cent, for highly skilled mechanics, 

 and then gradually falls for families whose income exceeds one 

 thousand dollars. 



Before we go on to enquire what sort of accommodation is 

 obtained for this expenditure of income, it is necessary to verify 

 the assumption made in last paragraph that there is compara- 

 tively little distinction of classes in Canada. In a sense this is 

 an obvious fact, a matter of ordinary remark by every observer. 

 While there are few in Canada who are very wealthy, there are 

 probably as few who are in actual distress. The Census Report, 

 1891, enables us to verify to a certain extent this common obser- 

 vation. For each census district we have given, in a series of 

 tables, the number of houses and the number of rooms in each 

 house. It would be obviously impossible, having regard to the 

 limits of time and space, to analyze the returns for the whole 

 Dominion ; and since in the country districts there is little differ- 

 ence of class, it is necessary only to examine the returns for the 



