STATISTICAL COMPARISONS. 475 



holdings of the Government debt, or the holdings of land, 

 which are the favourite in\'estraents of the masses of some 

 countries. But I do not know of any comparison of the kind 

 in which these conditions are complied with. French peasants 

 and working classes are often assumed to be much more saving 

 than the con esponding classes of England, but the statistical 

 proof seems to be wanting, and I am not sure that if the 

 accumulations of English unions, friendly societies, and 

 co-operative societies, were properly taken account of, as 

 well as Savings Banks, holdings of Government debt, and 

 investments in land, that the English working classes would 

 come so very badly out of the comparison. At any rate, the 

 comparison is more difficult than is often thought. 



Even if comparisons could be made there would remain the 

 question of the comparison of character. A working population 

 which feeds and clothes itself well and makes itself in all ways 

 efficient, provided it saves enough for security, may really be 

 making more of life than a population which starves itself in 

 the present through fear that it may starve in the future. _ The 

 proper proportion of saving for a working class community is 

 itself a subject which requires some study. 



These points are of special interest in new communities 

 where the working classes have large means. No good is done 

 by using unsound arguments eveu for so excellent an object as 

 the promotion of thrift. If examples are to be taken from 

 other countries such as France, the so-called example should 

 first of all be adequately explained, and a true comparison 

 made, and then an inquiry made as to whether and how far the 

 French example is sound and worthy of imitation. The fact 

 already brought out as to the larger proportion of old life in 

 France than there is in either Germany or the United Kingdom 

 may also render saving a greater necessity there in order that 

 as much may be got out of life as in the neighbouring countries. 

 The requirements as to saving may thus be essentially different. 



To sum up this branch of the discussion : what we may 

 say is that the rough comparison of communities as 

 regards moral characteristics based on statistics of educa- 

 tion, crime, insolvency, and the like, is entirely useless and 

 mischievous because the figures are of a kind that values can 

 only be assigned to them by the most careful study. To take 

 them haphazard from statistical abstracts and dictionaries, and 

 assume that figures called by the same names in different 

 countries have exactly the same values, is either foolish or dis- 

 honest. Dictionaries are for reference, and not intended to give 

 all the materials for discussion, and when they are used for pur- 

 poses for which they are not intended, all who are interested in 

 the subjects under discussion must look out. Some dictionaries, 

 however, might be mademoreuseful than they are by the addition 

 of a few notes to the figures, referring to such points as the 

 nature of the legislation applied to the subjects of the figures, the 

 mode of collecting the latter, and other vital qualifications of 

 the figures themselves. I may claim the ci'edit of privately 

 stifling many an argument which inquirers were going to use 

 by taking figures from books as they found them, because I 



