346 SCIENCE PROGRESS 



p. 160, and on the same page the name of Hiss is spelt with one s. On p. 162 

 Ghon's name is twice misspelt. 



We have already alluded to the excellence of the coloured illustrations, which 

 exhibit the cutaneous rashes, the morbid anatomy, the histology and the bacteri- 

 ology of cerebro-spinal fever. A full bibliography and a good index are added, 

 completing a volume which should prove of great value for many years. 



F. W. Andrewes. 



An Inquiry into the Statistics of Deaths from Violence and Unnatural 

 Causes in the United Kingdom. By William A. Brend, M.A., M.D., 

 B.Sc. [Pp. 80.] (London : C. Griffin & Co., 191 5. Price 3s. 6d. net.) 



The common saying that there are three kinds of lies — ordinary lies, damned lies, 

 and statistics— is unfortunately a witticism which contains a large amount of truth. 

 This is, however, no reflection on the statistician, whose services are not usually 

 employed in recording the data. The erratic way in which statistics are supplied 

 is well illustrated by the author of the work under review. There is nothing in the 

 whole work — in fact, there could be nothing, with the returns at the disposal of 

 the author — to enable the reader to say whether or no we had attained or were 

 approaching the irreducible minimum of deaths from violence. There are indica- 

 tions, but indications only, that the number of deaths from unnatural causes are 

 far in excess of what we might reasonably hope, with average care, to reduce very 

 considerably. 



The chief aim of the work would seem to be to show the chaotic state of our 

 records. The records of unnatural deaths given in the Annual Report of the 

 Registrar-General differ from those supplied by the Local Government Board, and 

 also from those supplied by the coroners' returns ; the difference being probably 

 due to the various systems of classification employed. There seems to have been 

 no attempt to bring uniformity into the methods of recording, even when such 

 uniformity could easily be attained, such as the methods of geographical distribu- 

 tion, and arrangement in the same age groups. 



The classification of the cause of death and the difficulties of nomenclature 

 might certainly be considered more difficult problems ; but a useful working basis 

 might be arrived at if medical men would only consider the possibilities of statistics 

 more fully. They would see that, while a medical practitioner might be justified 

 in considering his own method preferable to those of his confreres, a persistence 

 in his own classification and refusal to join in the general method, while not per- 

 haps reflecting on his ability as a doctor, does certainly point to his incompetence 

 as a Public Health officer. 



It is curious to hear that officials, such as coroners, should ever consider their 

 returns as their own personal property; and although they may not refuse informa- 

 tion to those requiring it, the fact that they have the power to refuse, and are not 

 obliged to send their records in a prescribed :form, is a scandal. The figures 

 given in the book are the records for the year 191 2, taken from the sources men- 

 tioned in the introduction ; but there is no indication as to whether they are larger 

 or smaller than the average for the last few years, obtained from the same sources. 



The author clearly points out the unreliability of the data, but does not help us 

 to estimate the amount of that unreliability. As to whether the figures are 5, 5°> 

 or even 500 per cent, out, is left to our instinct. Now on purpose that statistics 

 should be of value it is necessary to consider the following points : (a) the relia- 

 bility of the sources ; (b) their completeness or incompleteness ; {c) the manner of 

 grouping with reference to the taking of an average, or for correlation. Associated 



