366 



NA TURE 



[February 15, 1900 



the New York City Station of the United States Weather 

 Bureau. There were copied the mean temperature, barometer 

 and humidity, the total movement of the wind, the character 

 of the day and the precipitation for each of the days of those 

 same three years. Then, by a somewhat laborious process of 

 tabulation, excesses or deficiencies in the occurrence of arrest 

 for drunkenness were determined. In the accompanying 

 diagrams these are shown for the different months of the year, 

 and for definite conditions of temperature. In each, the heavy 

 horizontal line represents the normal occurrence, distances along 

 the abscissa line the months of the year and definite temperature 

 groups, while ordinate distances show excesses or deficiencies 

 in percentages of the expectancy. The extremes of the temper- 



JAN. FEB. MAR. APR. MAY. JUKI. JUL. AUfr, SEP. OCT. NOV. DEC. 



SOI. 



aofo 



+ 

 



ature curve are omitted where the probable error equalled^or 

 exceeded the discrepancy in occurrence. 



As may be seen from both of the diagrams, the surmise o the 

 reviewer with respect to the use of intoxicants is erroneous. In 

 fact, their showing is somewhat surprising. As shown by Fig. i, 

 the prevalence of intoxication during the cold months is much 

 in excess of that for the warm ones, December giving the police- 

 court 47 per cent, more business from its use than did July. 

 The physiological problem which this fact might suggest, as to 

 the effect of equal quantities of alcohol taken under different 

 conditions of temperature, I do not here attempt. If there is 

 not a marked difference in this respect, our figures would indi- 

 cate that much more liquor was drunk in 

 the City of New York during the colder 

 months of the years studied than during 

 those of the other extreme of temperature. 

 We claim no broader bearing for the 

 problem ; but even this is interesting. 



The difference might l)e due either to 

 social or meteorological influences. Under 

 the first we may consider the effect which 

 certain holidays might have upon the pre- 

 valence of drunkenness. Undoubtedly 

 some days of the year are made the occa- 

 sion of a drunken debauch by persons so 

 inclined, and Christmas is one of them. 

 But the 4th of July is perhaps just as much 

 of a favourite for such diversion to us in 

 America, a fact which would swell the 

 numbers for that month. This, however, 

 fails to show any such effect. In fact, a 



careful inspection of the record, although showing a slight 

 increase of drunkenness for the festivals mentioned, proves it 

 to be too small to account for the monthly showing. The ex- 

 cesses for the cold months are due to a large daily occurrence, 

 and the deficiencies for the cold ones to the reverse conditions. 



Another social condition which might effect the results is the 

 exit from the city for the summer of some who are brought 

 with some regularity before the bar of the police-court during 

 the rest of the year. My study of assault and battery would, 

 however, lead me to believe that the influence of this exodus is 

 not great. It would be reasonable to infer that arrests for these 

 crimes and for drunkenness would, for the most part, be made 

 from the same social stratum, and that social conditions which 



1 would influence the prevalence of one would have the same 



i effect upon the other. That is, if public drunkards were gone 



j in any numbers from the city, public brawlers would be also. 



i Yet this is precisely the reverse of what our study of assault has 



shown. Upon Fig. i I have shown, by means of a dotted curve, 



I the arrests for this crime for the same years. It shows as 



I marked excesses for the warm months as we have deficiencies 



for drunkenness for that season, a fact which would lessen the 



! validity, if not entirely negative the weight of any migration 



theory which might be brought to bear upon the problem. 



The third hypothesis is that of the direct effect of the 

 peculiar meteorological conduions, and it seems to be the 

 most plausible. Of these, temperature is the only one which 

 we shall here consider. As shown by Fig. 2, 

 the relation between expectancy and oc- 

 currence was worked out for each of the 

 temperature groups indicated at the top, 

 and represented by the curve. Low tem- 

 peratures made business for the police 

 judge, and high ones lessened his labours. 

 Of coarse, if our conclusions in the pre- 

 ceding paragraph on occurrence were 

 erroneous, those from this figure would be 

 also. In that case, deficiencies for high 

 temperatures shown here would be but 

 concomitant variations. 



The summer is hot. If there be but few 

 arrests for drunkenness during the summer, 

 there can be but few during high tem- 

 peratures. On the other hand, if high 

 temperatures so affect the individual that 

 less stimulant is demanded than during 

 those which are lower, we have here the 

 cause of the peculiarities shown in Fig. i. 

 There are many reasons for believing 

 that this is the case. In the first place, 

 there is every reason to believe that 

 the vitality of the body is lower in cold weather than 

 during that which is moderately warm. This in itself would 

 influence the demand for stimulant. A "wee drappie" 

 is taken when 7ieeded, and for many this means a drunk. No 

 doubt many of the habttiuh of the police-court as prisoners^ 

 struggle against their tendencies to drink, knowing the conse- 

 quences. When vitality is great, they do so with success. For 

 days, and perhaps weeks, they are winners, but the time comes 

 when the fight is too severe, and they succumb. That was on 

 the day when vitality was at its lowest ebb, and the cold con- 

 tributed to that condition. The poor fellow was cold ',. 

 he was weak. The stimulant would give him immediate 



2,0' IS' 30* ??'- 40* 45° 



55* feC t,S° W IS' 80 8S- 



though temporary relief. He took it, and our figure shows the 

 result. 

 I In conclusion, I would say that I recognise the limitations 

 ' of this method of study. By its very nature, each meteoro- 

 logical condition is treated as if the others were not at the same 

 time present. This would, liowevei;, introduce no error unless 

 two or three tended to vary concomitantly. In that case, the 

 j effects of one might be imputed to another. We recognise, too, 

 ' that a study of drunkenness does not have quite the bearing 



NO. I581, VOL. 61] 



