ON OFFICIAL STATISTICS. 419 



allowed for the taking of the census, but much more time is often 

 taken in the operation. The reason of this is said to be that the 

 appointment of the census collectors is a political job, and it is 

 considered essential that each collector should be long enough at 

 work to enable him to earn a good round sum ; and thus the 

 accuracy of the census is sacrificed to the exigencies of politicians. 

 The United States census of 1870 was aa admitted failure in 

 consequence of this, the returns giving several millions short of 

 the actual population ; and although the census of 1880 was less 

 defective, or at any rate did not err in the same direction, it is 

 questionable if the population was not largely over stated in its 

 returns ; though not perhaps to the same extent as the population 

 of 1870 was understated in the returns of that year. 



The census of ancient Rome was taken once in a lustrum, or 

 every five years. We in all but two of the colonies are behind 

 the ancients in this respect as we take ours only once in ten years. 

 It is my opinion that in countries like these, with rapidly increas- 

 ing populations, not more than five years ought to intervene 

 between two censuses. It will be remembered that at the census 

 of 1881 the population estimates which had been previously 

 formed in all the colonies were wide of the truth, the error in the 

 three eastern colonies alone amounting to as much as 112,000. 

 And if the error was so great in the mere numbers of the popula- 

 tion what chance could there be of our possessing a correct idea 

 of the birth places, religions, ages, occupations, conjugal condition, 

 or education of the population 1 An Httempt is certainly being 

 made now in two of the colonies to avoid such a discrepancy in 

 future by making an allowance to correct certain supposed errors 

 in the returns of immigration and emigration ; but we who make 

 the estimates know we are working in the dark and we have much 

 need of a census to enlighten us. 



Another reason for taking a census often is that the experience 

 gained by those engaged in the work is to a great extent lost in 

 so long a period as ten years. Every one, after superintending 

 the operations of a census, feels how much more effectively he 

 could do so if had to begin again. But after the expiration of ten 

 years, the lapse of time and attention to the details of quite 

 ditierent duties, have probably caused him to forget mucli of 

 what he had learned, and even if this should not be tlie case, the 

 start" he had organised and trained with care to a full knowledge 

 of their work is scattered, and he has to begin again the harassing 

 task with a new set of men. There is no doubt in my mind that 

 if the census were taken at shorter intervals, a better result 

 would be obtained. 



I am not in favour of burdening the census collectors with 

 work outside the usual legitimate inquiries, connected with the 

 enumeration of the population, namely, the numbers of the 



