696 SCIENCE PROGRESS 



to which they have particular reference. In addition, all local records* 

 published in the many British ornithological and natural-history journals 

 up to 191 8, are systematically given under their particular counties. It is 

 therefore possible for an ornithologist visiting a part of England new to him 

 to have before him a complete list of all published references to the birds of the 

 district. The extraordinary value of this to the serious student is self-evident. 



To those who do not possess the first volume, the long list of general works 

 of reference given in Part I, now before us, will also prove useful. 



The book is well produced, but lovers of their private libraries will be 

 sorry that it is not printed on the same paper as the companion volume, 

 though the two will match in size and approximately in thickness. There 

 are several printers' errors, e.g. on p. 12 appears Atchison, G.T., while on p. 13 

 appears Aitchison, G.T. The first is correct. 



We can find no reference under Cumberland to The Home Life of the 

 Terns, by W. Bickerton, although it contains a lengthy account of the famous 

 Ravenglass Gullery. But omissions are so few that criticism on that score is 

 unmerited, and we can but congratulate the authors on the successful 

 beginning of an undertaking for which all British ornithologists will be heartily 

 grateful. 



Wm. Rowan. 



MISCELLANEOUS 



Vital Statistics. An Introduction to the Science of Demography. By George 

 Chandler Whipple. [Pp. xii+517.] (New York: John Wiley & 

 Sons ; London : Chapman & Hall, 1919. Price 18s. 6d. net.) 



This book is intended primarily for students in Schools of Public Health, 

 and with this end in view it deals with the collection of data, with tabulation, 

 and with graphical representation rather than with the working out of the 

 data when obtained. Probability and correlation are touched upon, but we 

 are inclined to think that this part of the book will not prove so enlightening 

 as the earlier portion. In the first chapters there is most useful information 

 for those collecting and tabulating data, and warnings as to the importance 

 of stating exactly what limits are used in grouping, how ages are reckoned 

 (whether age last birthday or age to the nearest year), and as to various other 

 matters all of vital importance if good use is to be made of the data collected. 



Throughout the book there are hints as to the intrepretation of results 

 obtained from statistics which are very valuable. We would especially 

 recommend the chapter on the Statistics of Particular Diseases, as it gives 

 many examples of the danger of drawing conclusions from mortality rates 

 without a most careful study of those rates. For example, a much higher 

 death-rate from tuberculosis is found in Richmond than in New York City, 

 but when a further analysis of the rates is made Professor Whipple points out 

 that it is due to the far larger proportion of coloured persons in Richmond 

 than in New York, and if the population be sub-divided into white and coloured, 

 New York is found in both cases to have a higher death-rate from tuberculosis 

 than Richmond. In this chapter, Prof. Whipple gives a diagram showing 

 the growth of water filtration and the decrease in typhoid fever death-rates ; 

 such a diagram proves nothing, both have gone down with time, but that does 

 not necessarily mean that one depends on the other. 



The chapter on correlation does not appear to be so adequate. It is 

 difficult to understand what is meant by the statement that " Pearson's 

 coefficient (of correlation) is not quite the same " as Galton's, when Galton's 



coefficient is given as — — . This is the product moment coefficient used in all 



n<T x <r v 



cases when the correlation table is quantitative both ways, and the means 



of the arrays lie on a straight line. It is hardly fair to call it either Galton's 



