IMPORTANCE OF STATISTICAL IDEAS. 121 



to elucidate the phenomena, so that this particular explanation cannot 

 be overlooked. 



4. There remains, however, the question which many people have 

 rushed in to discuss — viz., whether the reproductive power of the popu- 

 lations in question is quite as great as it was fifty or sixty years ago. 

 We have already heard in some quarters, not merely that the repro- 

 ductive energy has diminished, but suggestions that the populations in 

 question are following the example of the French, where the rate of 

 increase of the population has almost come to an end. Apart, however, 

 from the suggestions above made as to the abnormality of the increase 

 fifty or sixty years ago, so that some decline now is rather to be ex- 

 pected than not, I would point out that the subject is about as full of 

 pitfalls as any statistical problem can be, for the simple reason that it 

 can only be approached indirectly, as there have been no statistical 

 records over a long series of years showing the proportion of births to 

 married women at the child-producing ages, distinguishing the ages, 

 and showing at the same time the proportion of the married women to 

 the total at those ages. Unless there are some such statistics, direct 

 comparisons are impossible, and a good many of the indirect methods 

 of approaching the subject which I have studied a little appear, to say 

 the least, to leave much to be desired. We find, for instance, that a 

 comparison has been made in Australasia between the number of mar- 

 riages in a given year or years and the number of births in the five or 

 six years following, which show, it is said, a remarkable decline in the 

 proportion of births to marriages in recent years as compared with 

 twenty or thirty years ago. It is forgotten, however, that at the 

 earlier dates in Australasia, when a large immigration was taking 

 place, a good many of the children born were the children of parents 

 who had been married before they entered the country, while there are 

 hardly any children of such parents at a time when immigration has 

 almost ceased. The answer to such questions is in truth not to be 

 rushed, and the question with statisticians should rather be how the 

 statistics are to be improved in future, so that, although the past 

 cannot be fully explained, the regular statistics themselves will in 

 future give a ready answer. 



5. One more remark may, perhaps, be allowed to me on account of 

 the delicacy and interest of the subject. To a certain extent the 

 causes of a decline in reproductive energy may be part and parcel of 

 the improved condition of the population, which leads in turn to an 

 increase of the age at marriage, and an increase of celibacy generally 

 through the indisposition of individual members of the community to 

 run any risk of sinking in the scale of living which they may run by 

 premature mai-riage. These causes, however, may operate to a great 

 extent upon the birth-rate itself without diminishing the growth of 



