704 REPORT — 1889. 



modes of estimation are in use. Under the one system, the correspondent states 

 -whether, according to his observations in the field, each crop is ' over-average,' 

 ' averao-e,' or ' under-average.' The returns are tabulated, and a general verdict 

 for the whole country is arrived at according to whichever of the three heads just 

 specified secures the largest percentage of adherents. The error of this system lies 

 in its utter vagueness, inasmuch as all over-average estimates are treated as equal 

 amongst themselves, whilst the under-average estimates are similarly all regarded 

 as equivalent to each other. Moreover, it is inevitable that one over-average and 

 one under-average estimate must be regarded as equivalent to two average esti- 

 mates, and, therefore, counteract each other, no graduation of over-average or of 

 under-average estimates being attempted. The other system is free from these 

 errors. Its essential feature involves the taking of the number 100 to represent an 

 averao-e crop, and increasing or diminishing this according as the crop is above or 

 below average. Thus, 90 would indicate a crop estimated to give 10 per cent, 

 below an average yield, and 75 would denote a crop anticipated to fall one-fourth 

 below the usual yield. On the other hand, 150 would indicate a yield half as 

 large again as the average. This method, being more exact, is decidedly prefer- 

 able, particularly for the establishment of general numerical results applicable to 

 the whole country. 



But both systems are open to serious objection, in that they take no account of 

 area. Adopting the numerical system of estimate, it is suggested that every 

 number indicating the current valuation of the yield of a crop should be associated 

 with the area, expressed in acres, to which the estimate applies. Thus, an estimate 

 of 105 and one of 95 would, at present, be considered equivalent to two estimates 

 of 1 00, whereas one of the estimates might apply to hundreds of acres, and the 

 other onlv to a score or so. In each case the area in acres and the numerical 

 estimate should be regarded as factors, and the sum of all the resulting products 

 divided by the total number of acres would give a more reliable general forecast of 

 the probable yield per acre than has yet been obtained. In mathematical language, 

 the general forecast should be a function both of the numerical estimates of the 

 yield and of the area involved. The subject is of considerable commercial im- 

 portance, as market quotations are influenced by these forecasts. 



Figures and illustrations were given in support of the paper. 



3. Improved Dwellings for the Poor. By D. Gr. Hoet. 



It is manifest that, although the greatest improvements in vital conditions were 

 made by the cultured and intelligent, this would only serve to make still wider the 

 already wide gulf existing between the conditions of life and the death-rate of those 

 in comfort and those in poverty ; and in order to any marked effect being produced 

 on the general tables of mortality, the improved practice and improved conditions 

 must be brought down to and participated in by the masses ; hence the clamant 

 importance of the question of questions Lii economic science — improved dwellings 

 for the poor. 



The subject is a vast one, requiring great detail for its expansive treatment, but 

 this I reserve for a future paper, and limit myself to indicating, in broad lines (1) 

 the extent, depth, and urgency of the evils, and (2) the method proposed for their 

 cure. 



I. Any large city will suit for illustration. Take, first, Glasgow, of which I 

 know most. One-fourth of the inhabitants live in single apartments, nearly 70 per 

 cent, in houses of one and two apartments, often with lodgers. The death-rate in 

 such one- and two-roomed houses is 27"74 per thousand ; in three- and four-roomed 

 houses, 19'45 ; in houses of five rooms and upwards, 11"23. 



The great majority of the two-roomed houses, and a number of the single apart- 

 ments, are occupied by superior skilled workmen and others of a superior class, 

 whose mortality is much lower, and the high figure is made up by a comparatively 

 small proportion, in which the rate is enormously high. In one district, comprising 

 houses of all the classes, it is 42 per 1,000 ; what must it be in the worst class 



