JUNE 15, 1905_ 



NA TURE 



147 



WE A THER J NFL UENCES. 

 Weather Influences : an Empirical Study of the 

 Mental and Physiological Effects of Definite Meteor- 

 ological Conditions. By Dr. E. G. Dexter. Pp. 

 xxxi + 286. (New York: The Macmillan Company; 

 London : Macmillan and Co., Ltd.) Price Ss. 6d. 

 net. 



THE effect of changes of weather on human 

 activities has been the subject of much discus- 

 sion, and each of us has no doubt formed an opinion 

 on how he individually is affected by different meteor- 

 ological conditions. The problem as affecting the 

 behaviour of humanity in the mass has, however, 

 received but scant attention hitherto. The statistical 

 method affords the means of obtaining numerical re- 

 sults which enable us to estimate the importance of 

 such effects. 



Meteorological statistics are nowadays available 

 from most centres of population ; social statistics are 

 also plentiful, yet of these only a limited number can 

 be made to yield information on the general conduct 

 or the working capacity of the community as a whole. 

 In the book before us Dr. Dexter has collected and 

 discussed sixteen classes of data culled from school 

 records, covering both questions of attendance and 

 conduct, police records dealing with cases of assault, 

 drunkenness, murder, suicide, arrests for insanity, 

 discipline in penitentiaries and the health of the force, 

 the death register, registers of attendance in the 

 out-patient departments of hospitals, and records of 

 the number of clerical errors discovered in the books 

 of certain banking establishments. The latter are the 

 onlv data studied which deal exclusively with mental 

 activities. All the records refer to New York City 

 or to Denver, Colorado. The meteorological statistics 

 with which thev have been compared were obtained 

 from the V.S. Weather Bureau. 



The effects of seasonal changes are first discussed, 

 and then the influence of each of the meteorological 

 elements is considered separately. The general 

 method of arranging- the material for this purpose 

 will be clear from the following description of that of 

 dealing with the connection between temperature and 

 assault. The davs falling within the period con- 

 sidered were arranged in groups according to their 

 mean temperatures, each group extending over a 

 range of 5° F. On the assumption that temperature 

 has no effect on assault, the number of days in each 

 group is proportional to the " expectancy " of assault 

 for that group. The actual number of occurrences of 

 assault on the days of the group is computed as a 

 percentage of the " expectancy," and curves have 

 been drawn using the " occurrences " as ordinates 

 and temperatures as abscissae. 



In dealing witli the element rainfall the usual 

 meteorological distinction has been drawn between 

 days of rainfall, on which o-oi inch of rain or more 

 was measured, and dry days. It seems a pitv that a 

 further subdivision was not made. Most of us would 

 be inclined to draw a wide distinction between showery 

 days with only a few hundredths of an inch of rain- 

 fall and days of steady downpour. Even if such a 

 further subdivision had been adopted, days with a 

 NO. 1859, VOL. 72] 



few heavv showers would not be distinguishable from 

 days of continuous fall ; probably a classification on 

 the basis of duration rather than amount of rainfall 

 would yield results which would repay the labour 

 involved in tabulating the records of self-registering 

 rain gauges. 



The majority of the curves show fluctuations which 

 are greatly in e.xcess of any whicli could be due merely 

 to accidental variations. The number of data is in 

 some cases extremely large (about 40,000 cases of 

 assault), and there can be no doubt about the genuine- 

 ness of the effects of meteorological changes. 



The interpretation of the results is, however, a 

 matter of considerable difficulty, and the possible 

 influence of other than meteorological causes has to 

 be steadily borne in mind. The general line of argu- 

 ment adopted regards the curves as compound func- 

 tions of " irritability " or " emotional state " and 

 "available" or "reserve energy." Thus, to return 

 to the temperature-assault curve, we find a marked 

 deficiency of occurrences at low temperatures. This 

 has been taken to mean that under these conditions 

 so large a portion of the vital energy is used up in 

 supporting normal metabolic processes that the sur- 

 plus available for active disorder is small. Under 

 warmer conditions our pugilist, in addition to being 

 more out of doors and thus seeing more of his 

 neighbour, has more reserve energy available for 

 active warfare, and the work of the police is propor- 

 tionately increased. Above 65° the curve commences 

 to rise with increased rapidity. Fighting energy is 

 now at its prime, and at the same time " irritability " 

 or quarrelsomeness is rapidly increasing. The tem- 

 perature group 8o°-85° shows a conspicuous maxi- 

 mum in the relative frequency of assaults. In the next 

 group, 85°-90°, the curve exhibits a sudden drop. 

 Irritability may very possibly be at a maximum, but 

 the energy necessary to commence war is lacking, and 

 a mere desire to fight is not a punishable offence. It 

 is an interesting fact that the curve for women shows 

 the above effects even more conspicuously than the one 

 for men. A similar accentuation of the general charac- 

 teristics is shown in all cases in which the number of 

 data is sufficiently large to justify a separation of the 

 sexes, so that it would appear that women are, on the 

 whole, more susceptible to weather influences than men. 



Some of the most interesting and at the same time 

 most inexplicable curves are those which show the 

 effect of the height of the barometer on human 

 activities. With a few exceptions all classes of data 

 show a marked e.xcess of occurrences for periods of 

 low barometer and a corresponding deficiencv when 

 readings are high. We cannot set this down to the 

 direct effects of the diminution of pressure on the 

 human organism; crime, &c., does not increase with 

 altitude. .Attempts at explanation by calling to our 

 aid the usual accompaniments of a low barometer, 

 viz. wind, rain, or cloud also fail, for when the 

 effects of these elements are considered separately we 

 find that in a number of cases the results contradict 

 the hypothesis. Dr. Dexter directs attention to the 

 peculiar " feel " which some people have for the 

 approach of a storm, but this hardly amounts to an 



