TRANSACTIONS OF SECTION F. 825 



many persons who combine an agricultural with some other calling, and this in 

 an infinitely varying degree. The German and some other statistics endeavour 

 laboriously to give tables which take account of these persons with double 

 occupations and allot to them a place under more than one head. In England we 

 have no provision in our census for these cases, and a farmer and brewer or a 

 labourer engaged sometimes on a farm and at other times at other work may be 

 classed by the accident of the first entry in one or other categoi'y at random. 



By what is nearly a common consent, the attempted enumeration of the 

 agriculturally occupied population is connected rather with the general enquiries 

 of the census than with the crop returns of each year. Its value of necessity 

 depends on the coincident and relative record of the occupations, other than 

 agricultural, in which the inhabitants of any country are engaged. Such con- 

 siderations supply the answer to some of our less reflective writers on this 

 question, who would have a perennial investigation going on into the available 

 supply of agricultural labour — year by year, if not month by month. The move- 

 ment in the direction of concentration of growing numbers of the workers of a 

 nation in the urban districts, which is apparent iu so many countries besides our 

 own, and under the most opposite conditions of Governmental polity or agri- 

 cultural organisation, will no doubt form in a short time a very interesting topic 

 of statistical discussion. But the general figures cannot be handled with very 

 great advantage now at the distance of wellnigh a decade from the last enumera- 

 tions and at the moment when the taking of a new census is at hand. Until 

 that enquiry reveals its facts, the student of questions of relative rural population 

 may be referred to the mine of information collected by the Royal Commission 

 on Labour, and the late Mr. W. 0. Little's admirable and exhaustive analysis, 

 and to the most valuable statistical buff-book which the Board of Trade have 

 just issued from the pen of Mr. Wilson Fox. 



Equally or even more full of pitfalls for comparison are statistics of the size of 

 holdings, whether the comparison be made between one date and another in a 

 country like our own, or between one country and another. Not only will the 

 grades employed necessarily vary between country and country, but the starting- 

 point and detinition of what is a ' holding ' is usually entirely different. 



In one of the earliest meetings of the International Statistical Institute at 

 Rome I drew attention to the barrier thus offered to international comparisons on 

 the latter point. I then showed how occasionally it may happen that the recog- 

 nised 'holdings ' seem to have included every plot, however minute. Germany and 

 Belgium; and I may add Ireland, apparently made a beginning at zero. Great 

 Britain at one time regarded a quarter of an acre as a limit of statistical enquiry, 

 although since 1892 restricting the term ' agricultural holding ' to something over 

 an acre of land. Elsewhere, as in Holland and in the United States, refusals, 

 except under specially defined conditions, to take anything less than a plot of two 

 and a half or three acres in extent as a starting-point in the agricultural 

 enumerations are encountered. 



It is not always remembered that we ourselves have, even within the com- 

 paratively brief course of our official agricultural returns in Great Britain, held 

 more than one opinion as to what the siarting-poiut should be. At the first 

 collection of these statistics nothing under five acres was taken account of as agri- 

 cultural. The scope of the annual enquiry was subsequently extended to plots of 

 a quarter of an acre, and the limit was raised again eight years ago to the present 

 requirement, which refrains from requesting annual details of the acreage of their 

 crops from the occupiers of holdings of a single acre or less. As a matter of 

 administrative convenience there is very considerable advantage in the course now 

 pursued, and no real statistical loss is involved, for the land occupied by the 

 various petty crofts or gardens which escape annual record was found not to 

 reach one-tenth of one per cent, of the cultivated area, and such rare changes as 

 might occur in the crops raised on these minute sections of territory could in no 

 perceptible degree affect the value of the returns as affording a general view of 

 the curreut chfinge of agricultural practice. Changes, however, iu the unit of 



