136 THE CANADIAN NATURALIST. [Dec. 



taken as a very favourable sign of the sanitary conditions. Column 

 E gives the percentage of the total deaths which took place under 

 five years of age. If accurate, unless there were some special 

 infantile epidemic, the high or low percentage in this column 

 ought to be a sure test of sanitary condition ; but the high 

 rate in healthy Upper Canada, never falling below 35 p. c, and in 

 even the country districts of Lower Canada (with the exception of 

 Soulanges), needs some explanation not yet given. Column F 

 gives the number, out of every hundred deaths, which were of 

 people above the allotted term of 70 years of age. Contrary to 

 the previous columns, it ought to be highest in the most healthy 

 districts; but the numbers are so low that they could only be 

 trusted on an average of years, or for a large population. Thus 

 the low rate for Three Rivers, and the very high rate for Soulanges 

 (iiearly five times that of Montreal) are probably accidental. 

 Column CI exhibits the proportion between the births and deaths 

 in the year ; the figures representing the deaths in each district to 

 every hundred births. If accurate, these ought to be lowest in the 

 most healthy districts, as we see in the case of Vercheres which 

 presents only half the death-rate of Montreal. 



The last column, H, representing the number of Catholics out 

 of every hundred in the population, has been added to test the 

 value of a suggestion made in certain quarters that the religious 

 customs of the French Canadians, who bring their infants to be 

 baptized in the church, even in the coldest weather, was a main 

 cause of the excessive infantile mortality of Montreal. It will be 

 seen that the proportion of Catholics is less in Montreal than in 

 any other quoted district of Lower Canada, except Sherbrooke. 



The returns may be regarded (subject to exceptions) as suffi- 

 ciently correct to show the comparative mortalities of cities and 

 adjacent counties, and to compare these with the ratios worked-out 

 from the preceding census. It is presumed that the causes of in- 

 accuracy will affect the different returns in somewhat of the same 

 ratio. They must also be taken (whether accurate or not) as our 

 only data for the actual population ; and, by comparison with the 

 census of 1851, for the yearly average rate of increase. There 

 was no temptation to * ; cook the figures'' in this, the easiest part 

 of the work ; least of all, to reduce the population below its actual 

 extent. 



In all the columns which include Quebec city, two sets of figures 

 are bracketed together for the reason stated above. Analogy proves 



