BE- -< 
 
 LIBRARY 
 
 CALIFORNIA 
 
Weather and Disease: 
 
 A CURVE HISTORY 
 
 OF THEIR 
 
 VARIATIONS IN RECENT YEARS. 
 
 BY 
 
 ALEX. B. MAcDOWALL, 
 
 M.A..F.R. Met. S. 
 
 LONDON : 
 
 THE GRAPHOTONE CO., i MACLEAN'S BUILDINGS, E.G. 
 
 I895- 
 
PREFACE. 
 
 THE influence of weather on health, a matter which concerns us 
 every moment of our lives, is yet one that is greatly in need of 
 elucidation. With growing attention to it, we become aware how 
 little it is understood, and how many unsolved problems it presents. 
 There are signs, here and there, of a deepening interest in this im- 
 portant subject ; thus, at Rome, e.g., a meteorological station has 
 been recently attached to the laboratories of the Public Health 
 Department, and students of the annexed school will be instructed 
 in the use of the instruments, and attend a course of lectures on 
 Meteorology applied to Hygiene. 
 
 The primary object of this book is to give an idea of the way in 
 which certain elements of our weather, and the mortality from some 
 well-known diseases, have varied in recent years. While some 
 direct attempt is made to trace the relations between weather and 
 disease, the work done is, generally, rather in the way of furnishing 
 data for comparison and study. Some of the facts presented, which 
 have been met with in going along rather unfrequented paths, may 
 be found, the author believes, both new and instructive. 
 
 The mode of exposition adopted is largely that of graphic curves ; 
 these, indeed, may be said to be the essence of the book. What is 
 written about the charts may be soon read. The reader is invited 
 to study them independently, and to follow out, with care, any 
 trains of thought which they may perhaps suggest to him. 
 
 It appears to the writer that the graphic method is far too little 
 utilised. In the best serials and books dealing with figures, one may 
 
find long and difficult accounts of how this or that item has varied 
 in a series of years, or other given time. A few simple lines on a 
 cross ruled space would at once make the whole thing clear. In the 
 actual case it is like groping for something in a dark room, where 
 one might first turn on the light of a glow lamp. 
 
 Columns and rows of figures are often, of course, indispensable. 
 But sometimes they are not. It may be desired, in surveying a wide 
 domain, merely to point out some salient features here and there ; 
 to indicate some noteworthy recurrence, the correspondence in 
 phase of two or more curves, or the like. Results may be sought, 
 rather than processes ; and a few figures may be given, while the 
 others may be left to be guessed approximately from the curves. 
 The graphic chart, indeed, is a multum in parvo mode of information, 
 which saves trouble in description, and presents the truth of figures 
 concisely, clearly, and impressively. It is remarkable, moreover 
 what new points of view are often gained when a column of figures 
 is converted into a curve. 
 
 A greatly extended use of the method, therefore, is here advocated, 
 and it is rendered practicable by the greatly enlarged facilities of 
 printing. Even a rough diagram may be regarded as better than 
 none. Those here offered make no pretention to fineness of quality 
 or mathematical precision. 
 
 Many of these curves are said to be smoothed. It may be well to 
 explain this. 
 
 The object of smoothing is to bring out those longer waves ot 
 variation which may be dimly perceived through the zigzags of the 
 actual curve, like a long ocean swell, obscured somewhat by the 
 minor waves and ripples. Suppose we have a series of annual or 
 other values, #, , c, d, e, /, g, &c., and we want to smooth these 
 with averages of 5. We first take the average of a to <?, and put it 
 down in place of c ; next the average of b to /j and put it down in 
 place of d ; and so on. The averages chiefly used in the charts are 
 
those of 5 and 10. In the latter case, the result is put down in the 
 fifth place. Sometimes the smoothed curves are given without those 
 of the actual figures ; but in many other cases both are given, so 
 that the nature of the relation between them need not be misunder- 
 stood. 
 
 The existence of demonstrable relations between weather and the 
 sunspot cycle of about eleven years, is now regarded by many 
 meteorologists as highly probable ; considering, especially, the proved 
 relations of sunspots to magnetic phenomena and to auroras, and the 
 obvious influences of solar radiation. The notion that a given sun- 
 spot phase (maximum or minimum), if it affect weather, must affect 
 it everywhere alike (giving everywhere excessive rainfall, excessive 
 heat, or other character), now appears to be untenable. If so, the 
 value of evidence from individual stations is enhanced, and it seems 
 desirable that any series of weather variations noticed to present 
 some fair amount of correspondence with the sunspot cycle through 
 a considerable number of its wave periods, should be brought forward 
 and discussed. It is in this spirit, and entirely without dogmatism, 
 that the four diagrams relating to sunspots and weather are offered 
 for consideration. 
 
 Some of the weather curves in this collection have previously 
 appeared in Nature, and are here reproduced by kind permission of 
 the Editor. The author's thanks are also due to Mr. F. J. BRODIE 
 and Mr. T. H. BAKER for the extension, to date, of useful tables 
 published by them ; one relating to fog, the other to harvests. In 
 preparing the section on disease, STEVENSON & MURPHY'S excellent 
 Treatise on Hygiene and Public Health has been found helpful. 
 
 It would be sanguine, perhaps, to imagine that in dealing with so 
 large a mass of figures, and making so many calculations, one has 
 never failed in accuracy. Should any mistake come to notice, the 
 author will be glad to have it pointed out. He Relieves the degree 
 of care exercised has been such as at least to avoid any grave error. 
 
CONTENTS. 
 
 L_WEATHER. 
 
 i. The rainfall of England. 
 
 2. The wettest and the driest month. 
 
 3. The best ten days in Summer. 
 
 4. The Winters of this century. 
 
 5. Another view of our Winters. 
 
 6. The occurrence of very cold days. 
 
 7. Long frosts. 
 
 8. A long wave of temperature. 
 
 9. A remarkable Winter (1890-91). 
 10. The Summers of this century, 
 ii. Spells of heat and their progression. 
 12. Bright sunshine in Spring, Summer, and Autumn. 
 13. Bright sunshine in the Winter half. 
 14. Sunspots and thunderstorms. 
 15. Sunspots and the relation of raindays to rainfall. 
 16. Sunspots and Spring rains at Geneva. 
 17. Sunspots and the temperature of the first quarter. 
 18. Early and late seasons ; harvest dates. 
 19. Early and late seasons ; first flowering of plants. 
 20. London fog ; East winds in Spring. 
 
IL DISEASE. 
 
 i. Measles. 
 
 2. Diphtheria. 
 
 3. Whooping cough. 
 
 4. Scarlet fever. 
 
 5. Enteric or Typhoid Fever. 
 
 6. Diarrhoea and Dysentery. 
 
 7. Small Pox since 1838. 
 
 8. Two centuries of Small Pox. 
 
 9. Progression in the Summer death wave. 
 10. Waves of Influenza. 
 
 ii. Quarterly deaths from Whooping Cough and Measles in London. 
 12. The incidence of disease in different towns. 
 13. A general view of London disease in 1893-94. 
 14. Some Zymotic diseases in London, 1893-94. 
 15. Respiratory disease in London in the first quarter. 
 16. Zymotic disease in severe and mild Winters. 
 
SHEPHERD : " Weel, do ye ken, sir, that I never saw in a' my born 
 days what I could wi' a safe conscience hae ca'd bad weather ? The 
 warst had aye some redeemin' quality aboot it that enabled me to 
 thole it withoot yaumerin'. Though we may na be able to see, we 
 can aye think o' the clear blue lift. Weather, sir, aiblins no to speak 
 very scientially in the way o' meteorological observation, but 
 rather in a poetical, that is, a religious spirit, may be defined, I 
 jalouse, 'The expression o' the fluctuations and modifications of 
 feeling in the heart o' the heevens made audible and visible and 
 tangible on their face and bosom.' That's weather." Wilson's 
 "Noctes Ambrostance" 
 
I. WEATHER. 
 
 GENERAL DIRECTIONS. 
 
 (Applicable to both Sections) 
 
 THE plan adopted with each diagram is, first, to point out briefly 
 what the curves refer to, and then add a few remarks on their sense 
 and bearings. The scales, both vertical and horizontal, are varied 
 according to convenience. Sometimes curves are brought close 
 together, or superposed, having separate vertical scales ; in which 
 case a letter is used to connect each curve with its proper scale. A 
 frequently used method is to give the actual variations as a dotted 
 curve, and the result of smoothing as a continuous curve traversing 
 it. References to books, &c., are given at foot. 
 
10 
 
 i. THE RAINFALL OF ENGLAND. 
 
 DIAGRAM : Dotted curve, fluctuations of our annual rainfall since 
 1800. Continuous curve, the same smoothed with averages of 10. 
 
 The numbers here used are not so many inches of rain, but the 
 ratio numbers adopted by Mr. G. J. Symons, our chief rainfall 
 authority. The average yearly rainfall of a place being considered 
 as 100, then if a year, e.g., had rainfall 10 per cent, over the average 
 its number would be no ; if 10 per cent, under average, it would 
 be 90, and so on. The series for the whole country is worked out 
 from the data for a number of different stations. 
 
 The actual variations (in those ratios) show great irregularity ; 
 ranging from the very wet year 1852 (ratio 136) to the very dry one 
 1887 (ratio 69). Several other very wet and very dry years are 
 indicated. 
 
 It has been pointed out by Mr. Symons that the years 1834, ^44, 
 1854, J 864, 1874, an d 1884 were all dry years ; but 1894 probably 
 breaks the series. The years about 1878 are remarkable ; we have 
 there 9 successive wet years (1875-1883). Going back to the next 
 marked wave crest in the smoothed curve, 52 years before, we have 
 5 successive wet years (1827-1831). We are. able to go back still 
 further, to the early part of last century ; and our smoothed curve 
 has another notable crest about 1771 (i.e., 55 years back from 1826). 
 Then we have definite minima, or times of drought, about 1745, then 
 about 1804 (59 years interval), and again about 1853 (49 years); show- 
 ing a certain regularity in these long variations. In the noteworthy 
 group of years 1738-62 only one was above the average. At the begin- 
 ning of this century there were 10 dry years in succession (1800-9) a ^ 
 the last minimum only 5 (1854-58). Since 1878 the smoothed curve 
 has been in rapid descent, but its future course is not easy to predict. 
 Perhaps we may have a recurrence of prolonged drought like that in 
 last century. (See Symons' British Rainfall, 1870, 1881, 1892.) 
 
11 
 
12 
 
 2. THE WETTEST AND THE DRIEST MONTH. 
 
 DIAGRAM : Upper dotted curve, monthly rainfall at Camden Square, 
 London, in 1888-93 (in inches). Continuous curve, result of smooth- 
 ing with averages of 5 months, (a) below, smoothed curve (5 av.) of 
 monthly rainfall in S. W. Ireland (3 stations) ; (3) ditto, E. of England 
 (3 stations). 
 
 On an average, our wettest month in London is October ; our 
 rainfall is mainly autumnal. But we often, of course, find some 
 other month to be the wettest. In the upper part of the diagram 
 October comes out wettest in 4 of those years ; but July in 1888 and 
 1890. 
 
 The driest month, on an average, is February ; and there is one 
 very dry February in this series, that in 1891 ( - oi inch). But it is 
 the driest month in only 2 of the 6 cases (January, September, and 
 April being driest in the others). 
 
 In the west of Ireland and north of Scotland there is a different 
 regime. The bulk of the rainfall is generally later, in winter. The 
 continuous curve (a\ from averages of three stations in the Cork 
 region, (Dunmanway, Cork, and Kenmare), lags, in its crests and 
 hollows, behind the London curve, culminating somewhere from 
 October to January. The east of England curve (A), from averages 
 of 3 stations (Geldeston, Swaffham, and Holt), resembles the London 
 curve, with autumnal rainfall. On the Continent, north of the Alps, 
 we should find a preponderating summer rainfall. (Symons' British 
 Rainfall.) 
 
13 
 
 Ta..Jl\rt\j 5 Ja frlay 5. Ja.)tl<iy. S 
 
 S - JA TfiA *> 
 
14 
 
 3 . THE BEST TEN DAYS IN SUMMER. 
 DIAGRAM : Dotted curve, variation of number of years, out of 45, 
 each day of summer (considered separately) has had rain. Con- 
 tinuous curve, smoothed with averages of 10. 
 
 This diagram is the result of an examination of the last 45 years' 
 records of the weather at Greenwich in summer (June, July, August), 
 day by day, noting, in the case of each day, how often, that is, in 
 how many of those years, it had rain. By this mode of reckoning a 
 distinctly good time comes out in the end of June ; the continuous 
 curve going down to 147 on the 26th and 27th (10 days average 
 from 22nd or 23rd). 
 
 It need hardly be said, however, that no one who is u weather- 
 wise " would build immoderate hopes on this or any other part of 
 summer ! 
 
 The driest day thus reckoned is August nth, with the value 9 ; 
 while in 4 cases the curve rises to 24. (Greenwich Observations.} 
 
15 
 
16 
 
 4 . THE WINTERS OF THIS CENTURY. 
 
 DIAGRAM : Dotted curve, variation of mean temperature of winter 
 (December to February) at Greenwich since 1 800. Continuous curve, 
 smoothed with averages of 10. (Winters are designated by the year 
 in which they end ; thus 1814 means 1813-14.) 
 
 There are many ways of estimating winters in respect of severity. 
 That of mean temperature (just indicated) gives a fair measure. 
 
 It is difficult here to make out a decided periodicity, such as 
 might enable one to forecast coming winters. The severest winter 
 of the century, so far, appears to have been that of 1813-14 (M. T. 
 32*5), when a fair was held on the Thames. The cold was ushered 
 in by one of the worst of London fogs, lasting a week, and this was 
 followed by a snowstorm of 48 hours. There seems to have been a 
 general decrease of severity in winters (increase of mean tempera- 
 ture) till about the sixties, and an increase of severity since. This is 
 indicated by the smoothed curve, also by the series of severe winters 
 we have marked with dots, viz. : 1830, 1841, 1847, 1855, 1860 
 (ascending series) ; 1865, 1871, 1879, 1891 (descending series). The 
 M. T. of 1890-91 was 34*0, and that of last winter, 1894-95, is a little 
 higher (34*8). The winter of 1854-55 is remembered as the Crimean 
 winter, and that of 1870-71 in connection with the Franco-Prussian 
 war. In December, 1879, the absolutely lowest record of tempera- 
 ture in the United Kingdom was made, viz.: 23, or 55 below 
 freezing point, at Blackadder, in Berwickshire. Some very mild 
 winters are indicated. In 1868-69 the M. T. was 44*4 ; in 1833-34 
 and 1876-77 it was 43*4 ; in 1845-46 it was 43*2. We seem to have 
 come to a time of (at intervals) very severe winters. (See Table of 
 Monthly Temperatures, by Buchan, in Journal of the Scottish Mete- 
 orological Society, 3rd Series, No. 9.) 
 
17 
 
 7800 
 
18 
 
 5. ANOTHER VIEW OF OUR WINTERS. 
 
 DIAGRAM : Dotted curve, variations in the number of frost days in 
 October to April, since 1842. Continuous curve, smoothed with 
 averages of 5. (1842 means 1841-42, &c.) The curves are inverted. 
 
 We have considered the mean temperature of the three months, 
 December February ; we here consider the number of frost days in 
 the seven months, October April. The vertical scale here increases 
 downwards, so that low points mean severe winters as in the former 
 diagram. 
 
 This way of looking at winters leads to some rearrangement ot 
 them as regards intensity. Thus the winter season 1887-88 comes 
 into prominence, exceeding both 1878-79 and 1890-91. While it 
 had fewer frosts (and higher mean temperature) than these in the 
 three winter months, it had considerably more frost in the four 
 other months ; March and April were both very cold. Several other 
 cases may be noticed. The number of frost days ranges from 90 in 
 1887-88 to 24 in 1883-84. (See Harding's Table, Quarterly Journal 
 of the Royal Meteorological Society, 1891, p. 113.) 
 
19 
 
20 
 
 6. THE OCCURRENCE OF VERY COLD DAYS. 
 
 DIAGRAM : Upper curve, variation in number of days in each winter 
 season (at Greenwich, since 1842) with maximum temperature not 
 over 32. Lower curve, variation in number of days in which the 
 minimum fell below 20. 
 
 A day may fairly be considered very cold, in which the maximum 
 temperature does not rise above freezing point. In 10 of the winters 
 considered there were no such days, and in one case (1890-91) the 
 number rose to 27. The average is about 5. Conspicuous points of 
 the curve occur as follows : 
 
 1844-45 I 854-55 1866-67 1878-79 1890-91 
 13 15 12 18 27 
 
 which shows a certain regularity of recurrence at 10 or 12 years' 
 interval. But we cannot say whether this will continue. Last 
 winter (1894-95) had 17 of those days. 
 
 The lower curve is drawn from a table by Mr. Harding (extended) ; 
 it relates to days with minimum under 20. The Crimean winter 
 (1854-55) here comes into prominence. The last few winters present 
 an unusual succession of high numbers. In 1894-5 there were n of 
 those days, all in February. (See Nature, vol. 51, p. 416 ; Greenwich 
 Observations ; and Harding, loc. cit.) 
 
21 
 
22 
 
 y. LONG FROSTS. 
 
 DIAGRAM : Lines showing the longest continuous frost in each 
 winter, at Greenwich, since 1850-51. 
 
 This is in five sections, each representing a month (November- 
 March). The longest continuous frost is that of last winter 1894-95, 
 extending from January 2ist to February 2Oth, *>., 31 days. There 
 was one of 28 days in February March, 1886, and one of 26 days in 
 December January, 1890-91. None in the earlier part of the period 
 come up to these. In the mild winter 1877-78 no frost lasted more 
 than 4 days, and there were 5 groups of that length. (The data are 
 from a table by Mr. Harding, extended.) ( Quarterly Journal of the 
 Royal Meteorological Society, 1886, p. 235 ; Greenwich Observations?) 
 
'So-i 
 
 55- 
 
 L 1 i i 
 
 29 
 
 1 1 1 I 
 
 '7 23 
 
24 
 
 8. A LONG WAVE OF TEMPERATURE. 
 
 DIAGRAM : (a) Dotted curve, a smoothed one (5 av.) of the mean 
 temperature of the first, plus the last, quarter of each year, at 
 Greenwich, since 1830. Continuous curve, averages smoothed in 
 the same way. () Decade averages of the same mean temperature, 
 beginning with 1830-39 (the last point only 1890-93). 
 
 To take another standpoint with regard to the fluctuations of 
 cold, we here consider the mean temperature of the first plus the last 
 quarter in each year. Note, in this dotted curve (a), the long rise 
 from a low point in 1838 (40*5), and, after some fluctuations at a 
 higher level, the return to a low point in 1887 (40*6). The twice- 
 smoothed curve presents crests at about 10 years' intervals (from 
 1850). 
 
 The lower curve (3), giving averages in decades on a larger scale , 
 rises to a maximum in the sixties, and then falls to a point lower 
 than the initial one. (See Buchan's Table, loc. ctt.) 
 
636 
 
 42-4 
 4r8 
 
 &*- 
 
 WO 
 
 M$ 
 
 fa '6 - 
 
 L I I \ I i I i I ( I I I I 
 /8dO &" to fa *o '66' '60 '6* 70 7* '80 '$$ '?<> 9* 
 
9 . A REMARKABLE WINTER (1890-91). 
 
 DIAGRAM : (a) Smoothed curve (5 av.) of daily minimum tempera- 
 ture in winter months at Greenwich, (p) Ditto at Nairn in the 
 north of Scotland. 
 
 The severity of the winter 1890-91 is well remembered in London. 
 Among other effects nearly eight weeks skating was enjoyed. 
 
 Were these two smoothed curves of daily minimum temperature 
 put before someone to say which was the London curve and which 
 the Nairn one, he would probably assign the milder to London. 
 This winter, however, was considerably colder in London than in 
 the North. 
 
 This (smoothed) London curve goes below freezing point on 
 December 9th, and does not return to it till January 23rd. 
 
 An interesting feature appears in those minor waves of variation. 
 They largely correspond in the two curves, those of the London 
 curve lagging, however, a few days behind those of the Nairn curve. 
 This extends through most of the winter. Such lagging corres- 
 pondence between southern and northern stations may be traced in 
 other cases of severe weather, but not always. (Daz'ly Weather 
 Report!) 
 
27 
 
 /v- /Q . 2*. 2a Tebj 8 10 /9 23 2S 
 
io. THE SUMMERS OF THIS CENTURY. 
 
 DIAGRAM : Dotted curve, variations in mean temperature of summer 
 (June, July, August) at Greenwich since 1800. Continuous curve, 
 twice smoothed with averages of 5. 
 
 This chart is similar in design to that of winters, except that in 
 smoothing the actual values are first smoothed with averages of 5, 
 then the averages are similarly smoothed. 
 
 The distinctly cold time about 1814 seems to have affected our 
 summers as well as our winters. In the smoothed curve we may 
 detect a certain regularity, with intervals of about io years, for a 
 considerable period. But it must be admitted that, as in the case of 
 winters, the actual variations are very irregular. 
 
 A few of the very hot summers are indicated. 1846 comes out 
 the highest by this reckoning, with M. T. 66'8. It was preceded 
 by a backward spring, and was very dry (" hottest June on record"). 
 The next, 1826 (M. T. 66.3), was also very dry throughout, with a 
 remarkably early and good harvest. 1859 an d 1868 may also be 
 noted. No recent summers come up to these. On the other hand 
 we have the very cold and wet summer of 1816, with one of the 
 worst harvests known. " After St. Swithin's day it rained 25 days 
 out of 30." 1860 was another disastrous summer, June wet through- 
 out, also August ; "a wetter season than 1799 and 1816, but not 
 quite so late." 
 
 The long wave of variations extending from 1860 to (say) 1879 
 (another cold summer), may be noted as interesting. (See Buchan's 
 Table, loc. cit., Baker's Records of the Seasons). 
 
30 
 
 ii. SPELLS OF HEAT AND THEIR PROGRESSION. 
 DIAGRAM : (a) Dotted curve, variations of daily maximum tempera- 
 ture at Parsonstown in August, 1893. Continuous curve, smoothed 
 with averages of 5. (b) Ditto for Paris, (e) Ditto for Munich. 
 
 The spring and summer of 1893 were remarkable in many ways. 
 We here consider the intense heat in August. 
 
 These three pairs of curves of maximum daily temperature for 
 Parsonstown (in the heart of Ireland), Paris, and Munich, culminate 
 one <after the other in the order given. Thus, the absolute maxima 
 were reached at Parsonstown on the 1 4th- 1 6th, at Paris on the i8th, 
 and at Munich on the 23rd ; the maxima of the smoothed curves 
 on the I4th, i6th, and 2ist respectively. 
 
 The fall of temperature in this month at some stations was very 
 great and sudden. Thus in Paris we find a maximum of 96 on the 
 1 8th, and one of 71 next day. 
 
 The kind of progression indicated may often be noticed in com- 
 paring spells of heat at various Continental stations and those at our 
 own, and spells of cold show it also. (Daily Weather Report^) 
 
32 
 
 12. BRIGHT SUNSHINE IN SPRING, SUMMER, AND 
 
 AUTUMN. 
 
 DIAGRAM : (a) Dotted curve, variations of bright sunshine at Green- 
 wich in spring. Continuous curve, smoothed with averages of 5. 
 (b) Ditto for summer, (c) Ditto for autumn (with larger scale). 
 
 The records of sunshine at Greenwich, now extending over 18 
 years, present various points of interest. In these three pairs of 
 seasonal curves for spring (March, April, May), summer (June, July, 
 August), and autumn (September, October, November), we may 
 note this relation : We have first a wave crest in the spring, then a 
 few years later, one in the summer, then a few years later, one in the 
 autumn. The actual maximum years are spring 1882, summer 1887, 
 and autumn 1891. 
 
 The time is too short, of course, to enable us to say whether there 
 is anything like a cycle in bright sunshine ; but it may be well to 
 notice whether this succession is repeated in future years, and 
 whether the crests of a seasonal wave recur at about 10 years' 
 interval. 
 
 The sunshine of the summer of 1887 (Jubilee year) was in sharp 
 contrast to that of the following summer. One was the highest 
 amount in this series (715 hours), and the other the lowest (372 
 hours). The brightest spring (1893) had 572 hours, the brightest 
 autumn (1891) had 295 hours. 
 
 Sunshine observations at Greenwich are made with the Campbell 
 Stokes instrument, in which the sun's heat is focussed on a strip of 
 millboard by a glass globe. (Greenwich Observations.} 
 
33 
 
 7377 7 8 79 
 
34 
 
 1 3. BRIGHT SUNSHINE IN THE WINTER HALF. 
 
 DIAGRAM : (a) Variations in amount of bright sunshine in October- 
 March at Hillington, Norfolk (from 1881-82). (b) Ditto at Cullomp- 
 ton, Devon. (<;) Ditto at Eastbourne (from 1886-87). (X) Ditto at 
 Newton Reigny, Cumberland (from 1884-85). (In chart, 1882 means 
 1881-82, &c.) 
 
 This diagram shows how sunshine in the winter half has recently 
 varied at 4 stations widely apart. It reveals a distinct upward 
 tendency in the amounts of late years. Thus at Eastbourne, the 
 sunshine continuously increased year by year from 381 hours in 
 1887-88 to 625 hours in 1893-94. Last winter breaks the succession. 
 We may give the Cullompton figures since 1883-84. 
 
 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 
 345 293 351 434 390 384 373 489 494 534 555 
 S 7 3 a^) ed 330 359 392 403 382 415 452 506 528 
 
 Important effects are doubtless associated with these long varia- 
 tions. (Meteorological Record.) 
 
36 
 
 I4- SUNSPOTS AND THUNDERSTORMS. 
 DIAGRAM : (a) Dotted curve, variation in thunder-days at Green- 
 wich (April-October of each year since 1850). Continuous curve, 
 smoothed with averages of 5. (b) Similarly smoothed curve of days 
 with thunder at Berlin (in year), (c) Ditto for Geneva, (d) Sun- 
 spot curve (inverted). 
 
 We may now proceed to consider a few cases of apparent corres- 
 pondence between certain weather phenomena and the sunspot 
 curve. Whether they indicate some causal relation, or are merely 
 of fortuitous nature, is another question. In any case it will be 
 interesting to observe how far, in the future, the correspondence is 
 maintained. 
 
 We begin with thunderstorms. Our records of this class of 
 phenomena seem rather inadequate. The Greenwich curve (dotted 
 one, a) is arrived at by picking out and counting the number of days, 
 in April to October of each year, on which thunder was heard. In 
 the inverted sunspot curve, at foot, there are minima in the years 
 1856, 1867, 1878, and 1889. Corresponding fairly with three of 
 these we have in the Greenwich smoothed curve maxima in 1857, 
 1878, and 1890, but no well-marked wave for the sunspot minimum 
 of 1867. 
 
 In the similarly smoothed curves for Berlin and Geneva (b and c), 
 the reader will easily trace by the figures a considerable corres- 
 pondence of the same kind with the sunspot curve ; maxima of 
 thunderstorms in general following minima of sunspots, and minima 
 of thunderstorms maxima of sunspots. (See Nature, Vol. 46, p. 488.") 
 
38 
 
 I5- SUNSPOTS AND THE RELATION OF RAINDAYS 
 TO RAINFALL. 
 
 DIAGRAM : (a) Dotted curve, variation of the ratio of raindays to 
 rainfall at Greenwich each year since 1835. Continuous curve, 
 smoothed with averages of 5. (fr) Sunspot curve (normal). 
 
 If we take the total number of days with rain in a year, and 
 divide by the rainfall of the year, we get the number of raindays per 
 inch of rain. The smoothed curve (for Greenwich) since 1835 
 presents maxima as follows : 
 
 1839 l8 49 l86 2 l8 7 2 l88 4 
 
 while the sunspot curve shows maxima : 
 
 1837 1848 1860 1870 1883 
 
 The former maxima, therefore, slightly lag behind the latter (which 
 might be considered natural if it were a case of cause and effect). 
 Similar relations may be noticed between the minima. 
 
 According to this view there would appear to be, on the whole, 
 most rain per rainday (or fewest raindays per inch) near or soon after 
 minimum sunspots, and the opposite near or soon after maxima. 
 (Annual Summary.) 
 
39 
 
40 
 
 1 6. SUNSPOTS AND SPRING RAINS AT GENEVA. 
 
 DIAGRAM : () Smoothed curve (5 av.) of rainfall in spring at 
 Geneva from 1836. (b) Dotted curve, variation in number of rain- 
 days in spring at Geneva. Continuous curve, smothed with averages 
 of 5. (c) Sunspot curve (inverted). 
 
 The spring rains of Geneva since 1836 seem to have varied in a 
 way suggestive of a relation to the sunspot cycle. Both raindays 
 and rainfall exhibit the phenomenon, in which maxima of the 
 smoothed curves are found coinciding with or closely following 
 minima of the sunspot curve ; while the minima of the former have 
 a similar relation to the maxima of the latter. (See Nature, Vol. 50, 
 p. 475, and Archives des Sciences.) 
 
41 
 
 #36 fy 'Wte"Jte Si 'fa '7 '6* '62 ** fy '7*- '7* '79 '81 '*V 
 
I7 ._SUNSPOTS AND THE MEAN TEMPERATURE OF 
 THE FIRST QUARTER. 
 
 DIAGRAM : (a) Dotted curve, variation in mean temperature of the 
 first quarter at Greenwich since 1841. Continuous curve, smoothed 
 with averages of 5. (3) Sunspot curve (normal). 
 
 Our fourth and last example relates to the amount of cold in the 
 first quarter of the year, the variations of this being measured by 
 mean temperature. 
 
 The sunspot maxima being in 
 
 1848 1860 1870 and 1883 
 
 we find maxima in the smoothed temperature curve in 
 
 1850 1861 1870 and 1883; 
 
 mild first quarters, therefore, in general, accompanying or following 
 many sunspots, and cold first quarters, few sunspots. The actual 
 variations in this and other cases, are of course very zigzag in 
 character ; and it will not be overlooked that we may find among 
 the groups at the sunspot maxima first quarters with a low tempera- 
 ture, or among the groups at minima first quarters with a high 
 temperature. 
 
 The oscillations of temperature here considered lie between 34'7 
 in 1855 and 44-2 in 1846 and 1872. (Buchan's Table, loc. ctf.) 
 
43 
 
1 8. EARLY AND LATE SEASONS. HARVEST DATES. 
 
 DIAGRAM : Dotted curve, variations in date (number of day in year) 
 when wheat cutting was begun at Mere, in Wilts. Continuous 
 curve, smoothed with averages of 10. 
 
 Some light on the general character of a season may be had from 
 the dates at which various harvest operations commence in a locality. 
 In a useful table Mr. T. H. Baker has given the dates when wheat 
 cutting commenced in his district in Wiltshire, in a long series of 
 years. The ordinary dates are each converted into the correspond- 
 ing number of the day in the year. The upper parts of this dotted 
 curve, then, it will be understood, represent late seasons (harvests), 
 and the low points early ones. 
 
 These variations do not present much regularity. We might 
 note that a very late harvest apparently comes after a lapse of about 
 19 years. Thus 1821 was very late ; then 1838 (17 years) ; then 
 1860 (22 years) ; then 1879 ( J 9 years.) Again, we find very early 
 harvests in 1826 ; in 1846 (20 years) ; in 1868 (22 years) ; and in 
 I ^93 (25 years) ; an average interval of 22 years. The smoothed 
 curve descends to low points about 1829 and 1868, which are 39 
 years apart. It may be here mentioned that Brueckner supposes 
 the weather of a great part of our globe to be subject to a cycle of 
 about 35 years, cold and wet periods alternating with hot and dry 
 ones ; and that 1830 and 1860 were approximately centres of hot and 
 dry periods, while 1815, 1850, and 1880 were centres of cold and 
 wet ones. Perhaps the smoothed curve may be said to harmonise 
 so far with this view. (Baker's Records of the Seasons). 
 
4G 
 
 1 9. EARLY AND LATE SEASONS. FIRST FLOWER- 
 ING OF PLANTS. 
 
 DIAGRAM : (a-e) Five curves showing variations in date (day of 
 year) of first flowering of (a) dog rose, (b) oxeye, (c) black thorn, 
 (d) wood anemone, (e) coltsfoot, at Strathfield Turgiss, N.E. Hants. 
 (/") Dotted curve, average of above. Continuous curve, smoothed 
 with averages of 5. (^) Dotted curve, variations in date of first 
 flowering of Ribes Sangnineum at Edinburgh (1850-1874). Con- 
 tinuous curve, smoothed with averages of 5. 
 
 We here consider early and late seasons from another point of 
 view, that of the first flowering of plants, a branch of the science of 
 phenology, (on which a valuable report is annually presented to the 
 Royal Meteorological Society). The high points of those five curves 
 (a-e) are again late dates, the low ones early dates. In the smoothed 
 average curve below (/) we find a wave rising from a hollow about 
 1882 to a crest in 1887-1889. There was probably another crest 
 about 1878. In the smoothed curve of Ribes Sanguineum at Edin- 
 burgh (g), to the left (with separate scale), we find two waves 
 culminating in 1857 and 1866. It may not be amiss to note (with- 
 out laying stress on it) that all those crests in the English and Scotch 
 curves are near sunspot minima. 
 
 It should be borne in mind that all dates here considered are a 
 great deal earlier than those of wheat cutting. It is right to state 
 that those series for Strathfield Turgiss are not perfectly homogeneous 
 throughout, but they cannot be far from the truth. (See Annual 
 Reports on Phenology to the Royal Meteorological Society, by Rev. 
 Mr. Preston and Mr. Mawley, in the Quarterly Journal of the Royal 
 Meteorological Society.) 
 
90 
 
48 
 
 20. LONDON FOG. EAST WINDS IN SPRING. 
 
 DIAGRAM : Upper curve, variation in number of days with fog in 
 London in the winter half, since 1872. Lower curve, variation in 
 number of days of easterly wind at Greenwich, since 1860. 
 
 LONDON FOG. A definite and uniform system of recording fog is 
 very desirable, and some uncertainty is at present felt in comparison 
 of existing records, through want of agreement as to what shall be 
 called a fog. The tables of London fog, prepared by Mr. F. J. 
 Brodie, may, however, be taken as adequately showing how this 
 unwelcome visitant has behaved in recent years. Confining our 
 attention to the winter half (October-March), we find as conspicu- 
 ously bad 'seasons 1873-74, 1879-80, 1886-87 (i.e. at 6 and 7 years' 
 interval) , the respective number of fog days being 59, 68, and 75. 
 But the bad season 1890-91 rather breaks this regularity ; it had 69 
 fog days. Our winters of late have been much freer of fog. 
 
 EAST WINDS IN SPRING. One of the annual tables issued by 
 Greenwich Observatory gives, for each month of the year, the 
 number of days of N. S. E. and W. winds (all directions reduced to 
 these four). Adding the numbers for E. wind in the three spring 
 months (March, April, May) since 1860, we get a series yielding the 
 lower curve, from which it appears that the spring of 1880 was the 
 most conspicuous for this character (38 days), while 1886 and 1893 
 come second and third (34 and 33). The earlier maxima seem to 
 have been lower. (See Brodie's Paper on London Fog, Quarterly 
 Journal, October, 1893, p. 233 ; Annual Summary^) 
 
// 7 t VA '76 Vs to s*. '$/, 'zb 
 
 fo '<}2- 
 
II. DISEASE. 
 
 Like crowded forest trees we stand, 
 And some are marked to fall." 
 
 Cowper. 
 
52 
 
 i. MEASLES. 
 
 DIAGRAM : (a) Dotted curve, variations of measles death rate in 
 London since 1838. Continuous curve, smoothed with averages of 
 10. (K) Similarly smoothed curve, with separate scale, for England 
 and Wales. 
 
 In the study of diseases we may, in a general way, discriminate 
 between three groups. There are first the diseases which medical 
 and sanitary science appears to be overcoming, (unless we are to 
 interpret the recent decline in their fatality as merely the retiring 
 slope of a long cyclical wave). Next, there are diseases which seem 
 to be growing amongst us, spite of all applied knowledge (and the 
 application of knowledge is of course by no means co-extensive with 
 the knowledge itself). Lastly, there are diseases of which one can 
 hardly affirm either permanent growth or decrease, they keep much 
 about the same. 
 
 Our concern here will be chiefly with diseases of the zymotic 
 class, and we may begin with measles. 
 
 The smoothed London curve shows several long waves, with crests 
 about 1843 (?), 1862, and 1885 (/>., at 19 and 23 years' interval), 
 and hollows about 1852 and 1871 (19 years). But on the whole we 
 seem to remain in statu quo. The epidemics of 1839 and 1845, 
 indeed, appear to have been much more severe than any since, 
 reaching the figures 1,133 an d 1,122 per million living. The rate 
 goes down on the other hand to 246 in 1852. The smoothed curve 
 for England and Wales resembles that for London. 
 
 Measles, in this country, has two maxima in the year ; the larger 
 in November-January, the smaller in May-June. It must be re- 
 membered that here and elsewhere we deal only with mortality, and 
 the prior time of incubation and illness should be taken into account 
 in considering seasonal influences. It is probable that measles is due 
 to some microbe, but this has not yet been identified. (Annual 
 Summary; Annual Report of Registrar General^) 
 
53 
 
 1838 Uf '!&- '*/ 30 '$3 
 
54 
 
 2 DIPHTHERIA. 
 
 DIAGRAM : (a) Dotted curve, variations of diphtheria death rate in 
 London since 1859. Continuous curve, smoothed with averages of 
 10. (b) Similarly smoothed curve, with separate scale, for England 
 and Wales. 
 
 The increase of diphtheria in recent years is notorious. Our 
 curves begin at 1859, when the disease began to be registered 
 separately (from scarlet fever). The London death rate, which was 
 rather high in some of the fifties and sixties (284 in 1859, 275 in 
 1863), went down to a minimum in 1872 (80) ; since which year 
 it has x been rising on the whole, till in 1893 the enormous rate of 760 
 was reached. In 1894 there is some improvement. The rise in 
 England and Wales has been less rapid. 
 
 Some allowance is probably to be made for improved diagnosis 
 and registration ; but even so, a large increase (in London) is 
 unmistakeable. 
 
 The alarming spread of diphtheria in recent years has been attri- 
 buted to the growing aggregation of children in schools, &c. ; but 
 there is some reason to think the disease is subject to a long cycle, 
 a wave crest of which we have been lately approaching. 
 
 Diphtheria is far more common in temperate and cold regions than 
 in the tropics. In this country it is most common in the last 
 quarter, and lowest in the summer months. Cold and damp 
 apparently increase the susceptibility to infection. 
 
 The microbe of diphtheria is that known as Loffler's bacillus. 
 {Annual Summary ; Annual Report of Registrar General?) 
 
56 
 
 3 WHOOPING COUGH. 
 
 DIAGRAM : Method as in i and 2. 
 
 This widespread disease, so fatal to children, appears to have been 
 declining, on the whole, in recent years (since the sixties). The 10 
 years' average for London in 1857 was 921 (per million living) ; but 
 in 1889 it had got down to 628. 
 
 The London rate was very high in 1861 (1,260), also in 1841, 1869, 
 1878, and 1882, (all over 1,180). Its lowest point was reached in 1883 
 (410). A relatively high point is reached, it will be seen, every 3 to 
 ; years. 
 
 In London, whooping cough increases steadily from its lowest in 
 September to a maximum in April, from which it declines through 
 the warmer months. It is often observed to become worse when the 
 wind turns cold or suddenly changes. Its specific microbe (if it have 
 one) is not yet identified. (Annual Summary ; Annual Report of 
 Registrar General.) 
 
57 
 
58 
 
 4- SCARLET FEVER. 
 
 DIAGRAM : Method as in i and 2. 
 
 This disease presents a still more evident decline in recent years, 
 (since the sixties). From a 10 years' average of 1,133 in 1865, the 
 smoothed London curve descends to one of 240 in 1889. The 
 extreme variations of the actual death rate in London are 2,131 in 
 1848, and 142 in 1891. 
 
 The variations in the death rate of this disease have been pretty 
 regular in their recurrence. The dotted curve reaches wave crests 
 every 4-7 years. The prevalence and mortality of scarlet fever are 
 greatest in the autumn and least in the spring, and its seasonal curve 
 is thus, roughly, opposite to that of whooping cough. The maximum 
 is in October. Curiously, in New York, scarlet fever is most 
 prevalent in spring and least in autumn. 
 
 According to Dr. Ballard, a temperature above the average for the 
 season, and a dry atmosphere, with little rain, favours the prevalence 
 of scarlet fever more than the reverse conditions. The disease has 
 been connected with a certain streptococcus as its producing cause. 
 
 One is disposed to ask whether the long wave of temperature 
 indicated in I. 8 may not have something to do with the course of 
 this and some other diseases (whooping cough, diphtheria) ? {Annual 
 Summary ; Annual Report of Registrar General?) 
 
60 
 
 S.-ENTERIC OR TYPHOID FEVER. 
 
 DIAGRAM : Method as in i and 2. 
 
 Our curves here extend only from 1869, the time when enteric 
 fever began to be registered separately from typhus and simple or 
 ill defined fever. Note the steady decline in both curves ; from a 
 death rate of 337 in 1869, the London curve goes down to 102 in 
 
 1892. 
 
 The prevalence and mortality of this disease are greatest in the 
 autumn and least in the spring and early summer. The bacilli 
 known as Gaffky's are believed by many pathologists to be the cause 
 of enteric fever. 
 
 This fever, it is well known, is distinguished from typhus fever by 
 the intestinal lesions present. The mortality from typhus fever 
 has also greatly declined of late years. (Annual Summary ; Annual 
 Report of Registrar General.} 
 
4. - 
 
 3m 
 
 220 
 
 /So 
 
 /to 
 
 220. 
 
 1*0 
 
 V* 
 
 
 7S- 7? 7? 
 
 1 1 
 
 93 
 
62 
 
 6. DIARRHCEA AND DYSENTERY. 
 
 DIAGRAM : (a) Dotted curve, variations of the death rate in 
 London since 1838. Continuous curve, smoothed with averages of 
 5. (b) Similarly smoothed curve for England and Wales. (c) 
 Similarly smoothed curve of the mean temperature of July at 
 Greenwich. 
 
 In the London smoothed curve we may notice three long waves 
 culminating in 1851, 1870, and 1885, (19 and 15 years apart). The 
 range of the death rate is from 1,683 m I ^49 to 253 in 1839. Recent 
 values are still above those at the beginning of the curve. 
 
 Diarrhoea is now reckoned as largely of epidemic character. With 
 regard to its association with heat, it would appear from recent 
 researches that atmospheric temperature is of less importance in 
 relation to it than earth temperature. According to Dr. Ballard, the 
 summer rise of diarrhoeal mortality does not commence till the 
 mean temperature of a 4 ft. deep earth thermometer reaches about 
 56, no matter what temperature has previously been reached by 
 the atmosphere. The two (death rate and earth temperature) reach 
 their maximum in the same week, and the decline of both is 
 similarly slow. The average summer wave of diarrhoea in London 
 rises to a maximum about the end of July : it is mainly due to 
 infantile diarrhoea. We may trace a considerable amount of corres- 
 pondence, (from the nature of the case it could not be exact), between 
 the smoothed July temperature curve (^ and the London diarrhoea 
 curves (a}. Calm weather promotes diarrhoea mortality, while wind 
 lessens it. Rainfall affects it indirectly through the temperature of 
 the ground. Moderate dampness of soil favours the disease, but 
 excessive wetness or dryness are thought to lessen it. 
 
 Several kinds of bacilli have been connected with the disease. 
 {Annual Summary ; Annual Report of Registrar General, &c.) 
 
64 
 
 7. SMALL POX SINCE 1838. 
 
 DIAGRAM : (a) Actual variations of the small pox death rate in 
 London. (3) Ditto, in England and Wales (data lacking 1843-1846). 
 
 Excluding the great epidemics of 1838 and 1871-72 a gradual 
 decline is apparent, especially in the lower curve. In the years 1838 
 and 1871 the London death rate rose to 2,168 and 2,422 respectively 
 (per million living). In 1889 the return is nil. In the next 4 years 
 there is a gradual rise (represented by the figures i, 2, 10, 48), 
 checked in 1894, however. In the London curve relative maxima 
 are reached every 3-6 years. 
 
 In Oriental countries the spread of small pox appears to be 
 favoured by cold and retarded by heat. In Europe the mortality is 
 greatest in spring and autumn. {Annual Summary; Annual 
 Report of Registrar General?) 
 
65 
 
8. TWO CENTURIES OF SMALL POX. 
 DIAGRAM : Ratio of deaths from small pox to 1,000 deaths from all 
 causes in London, averaged in decades (1701-10, &c.). 
 
 In a valuable paper contributed to the Statistical Society, Dr. Guy 
 has given statistics of 250 years of small pox in London, showing, 
 year by year, the ratio of deaths from small pox to i ,000 deaths from 
 all causes. Extending the table to 1891 we have made decadal 
 averages of those ratios since 1701. The data between 1830 and 
 
 1840 are defective. 
 
 The curve rises high in the latter half of last century (average for 
 1861-1870, 102-5), but in this century there has been a rapid drop> 
 and in the last decade we come to an average of 6-9. 
 
 With regard to the vexed question of vaccination, some dates may 
 here be given. The practice was introduced by Jenner about the 
 end of last century (1708). Vaccination was optional before 1853 ;. 
 then from 1854-1871 it was obligatory, but not efficiently enforced ; 
 and since 1871 it has been more efficiently enforced. (See Dr. Guy's 
 Paper to Royal Statistical Society, Vol. XLV.) 
 
67 
 
 /w - 
 
 80 
 
 6-0- 
 
 Ao- 
 
 30 
 
 lo - 
 
 - 
 
 .-*. .' 
 
 ,' i 
 
 
 
 4 
 
 
 \ 
 
 
 - 
 
 
 
 
 
 - 
 
 
 
 -., 
 
 **.-.-' 
 
 I ! I I 
 
 
 l 1 1 1 
 
 i I 1 
 
 ffp 'lo "to 'do '&o o '60 '70 'do '<fo /8oo '/o '?o 30 '44 3~o &o '70 'So V 
 
68 
 
 9. PROGRESSION IN THE SUMMER DEATH RATE 
 
 WAVE. 
 
 DIAGRAM : (a} Dotted curve, weekly death rate, Plymouth, from 
 22nd to 39th week of 1893. Continuous curve, smoothed with 
 averages of 5. (3) Ditto for Cardiff, (c} Portsmouth. (#0 Brighton. 
 00 Hull. 
 
 There is generally a well-marked wave in the summer death rate, 
 and if we compare these waves for different places, some interesting 
 relations are met with. Here we take the hot summer of 1893, and 
 show the death rate variations from the 22nd to the 39th week (/>., 
 say, June to September) at five places. The smoothed curves bring 
 out more plainly the retarded culmination of these waves in an 
 easterly direction (which is probably a temperature effect). Thus 
 we have the Plymouth wave crest about the 25th week ; Cardiff, the 
 26th ; Portsmouth, the 29th ; Brighton, the 34th. In the north, 
 Hull is still later, 35th week. This kind of relation may be observed 
 in other years, but is not invariable. (Annual Summary.} 
 
<t 
 
 3 
 
 r 72 
 
 22. 
 
 e 
 
 32 
 
 \ \ 
 
 Wym. 
 
 Card. 
 
 ; Hub 
 
 26 'J8 30 3z 34- 36 28 
 
 no. 
 
70 
 
 io. WAVES OF INFLUENZA. 
 
 DIAGRAM : Weekly deaths in London in the five epidemics. 
 
 The five recent successive visitations of influenza are here 
 depicted. The worst of the series, as regards highest mortality in a 
 week, was that in the winter of 1891-92 ; there were as many as 
 506 deaths in the week ending January 23rd. The last epidemic 
 (1895) comes very near this however, with 473 deaths in the week 
 ending 9th March. Both of these waves are sharper in their 
 culmination than the wave of 1891. 
 
 We may note the dates of maxima in those 5 curves : week end- 
 ing January i8th, May 23rd, January 23rd, December i6th, March 
 9th. 
 
 Hirsch considers that influenza is independent of seasons, and of 
 the influences of weather. 
 
 The last previous group of influenza epidemics was about the 
 forties ; 1 847 and 1 848 were very bad years. The chief remaining 
 epidemics of the century were in 1803, 1833, and 1837-38. The 
 disease apparently alters in type from time to time. There is 
 reason to believe that a particular bacillus is in causal relation to it. 
 (Weekly Return^ 
 
// /8 26 1-4/6 2$ 30 6 '3 1 ^7 
 
 Tzb 
 
 O /6 23 
 
72 
 
 ii. QUARTERLY DEATHS FROM WHOOPING COUGH 
 AND MEASLES IN LONDON. 
 
 DIAGRAM : (a) Whooping cough. (6) Measles. (1/76 means first 
 quarter of 1876, &c.) 
 
 It is instructive to put into curve form the numbers of deaths from 
 a given disease, in quarters, in a series of years. The diagram gives 
 those for whooping cough and measles in London since 1875. Each 
 space between two marks at the foot indicates a year, and is divided 
 into four. The whooping cough curve, it will be seen, reaches a 
 maximum with great regularity about every two years ; the first 
 quarter of 1876, the second of 1878, the first of 1880, and so on. 
 The worst quarter was the first of 1882, when there were 2,224 
 deaths. The measles curve also reaches maxima at about two years' 
 intervals, and these maxima evidently tend to increase in height of 
 late: the highest figure is in the second quarter of 1894 (1,745 
 deaths). 
 
 Comparing the maxima of the two curves, it will be noticed that 
 through a great part of the period they alternate ; measles one year, 
 whooping cough the next, and so on. (Quarterly Reports of Registrar 
 General?) 
 
23 
 
 I II I I I I 
 
74 
 
 12. THE INCIDENCE OF DISEASE IN DIFFERENT 
 
 TOWNS. 
 
 DIAGRAM : (a) Quarterly deaths from whooping cough in Bristol 
 since 1875. (b) Ditto, in Portsmouth, (c) Quarterly deaths from 
 scarlet fever in Bristol, (d) Ditto, in Leeds. 
 
 Our large towns, even when not far apart, may be quite differently 
 affected by disease. In illustration of this, two comparisons are here 
 given. In the case of whooping cough at Bristol and Portsmouth, 
 these quarterly curves show (during much of the period considered) 
 that when the Bristol curve has culminated, the Portsmouth curve 
 has been low ; and when the latter has culminated, the Bristol curve 
 has been low. This is not, however, an unvarying rule. Note, 
 further, that the intervals between crests in both cases tend to be a 
 little wider than in London. 
 
 The other case is that of scarlet fever in Bristol and Leeds. While 
 the Bristol curve has maxima in 1875, 1880, and 1887, Leeds has 
 them in 1875, 1878, and 1883. (Quarterly Reports of Registrar 
 General.} 
 
75 
 
 ' > '77 'J9 '7? '80 '91 '92. '83 
 
76 
 
 13. A GENERAL VIEW OF DISEASE IN LONDON IN 
 
 1893-94. 
 
 DIAGRAM : (a) Smoothed curve (5 av.) of weekly deaths from 
 diseases of the respiratory system in London, (b) Ditto, from the 
 chief zymotic diseases, (c) Ditto, from phthisis, (d) Ditto, from 
 diseases of the circulatory system. 
 
 The years 1893 and 1894 were very opposite with regard to 
 weather, which may be considered an advantage from our present 
 standpoint. 
 
 In this group of smoothed curves we may first note the prepon- 
 derance of diseases of the respiratory system, and the high point 
 reached by this curve in winter, (719 in the 49th week of 1893, />., 
 early in December). In both descents of the curve from the maxima 
 there is a curious reversal. Next we have the curve of zymotic 
 diseases (including small pox, measles, scarlet fever, diphtheria, 
 whooping cough, typhus, enteric fever, simple continued fever, 
 diarrhoea and dysentery, and cholera), with its striking summer 
 wave, (diarrhoea chiefly), in the very hot summer of 1893 ; in the 
 2yth week the average reached 461. In 1894 we find a different 
 state of things, two much smaller waves culminating in the I9th 
 (310) and 3 ist (303) weeks respectively, the first due to measles 
 chiefly, the second to diarrhoea. 
 
 Lastly, we have curves for phthisis and for diseases of the circula- 
 tory system, both culminating in weeks of winter, (phthisis, 186 
 deaths in the 52nd week of 1893, circulatory system 160 in the 49th 
 week of 1893). This classification of disease is of course not 
 exhaustive. ( Weekly Return.} 
 
77 
 
 Kftf'no 
 
 /3 IS 23 2 33 ^s 4-3 4S < 
 
78 
 
 14. SOME ZYMOTIC DISEASES IN LONDON IN 
 
 1893-94. 
 
 DIAGRAM : (a) Smoothed curve (5 av.) of weekly deaths from 
 measles in London, (b] Ditto, from diphtheria, (c) Ditto, from 
 whooping cough, (d) Ditto (with separate scale), from scarlet fever. 
 
 We have just looked at zymotic diseases as a whole ; we may now 
 consider some members of the group in the same way. 
 
 Here we at once notice the high measles wave of 1894, going up 
 to 162 in the 2ist week, a great contrast to that of the previous 
 year. Diphtheria reaches a pronounced maximum in the 45th week 
 of 1893, and whooping cough in the beginning of 1894. 
 
 The scarlet fever curve (given with separate scale below to avoid 
 confusion) also culminates in the last quarter of 1893, and in general 
 the values for that year are above those for the second year. 
 ( Weekly Return.} 
 
79 
 
 10 1$ 2o 2$ 30 3$ W 46 
 
 3 8 13 /3 23 & 33 38 43 trf .$3 
 
80 
 
 15. RESPIRATORY DISEASE IN LONDON IN THE 
 FIRST QUARTER. 
 
 DIAGRAM : (a) Dotted curve, variation in first quarter death rate 
 (per i ,000) from diseases of the respiratory organs, in London, since 
 1855 (inverted}. Continuous curve, smoothed with averages of 5. 
 (3) Smoothed curve (5 av.) of mean temperature in first quarter at 
 Greenwich. 
 
 This smoothed and inverted curve of respiratory disease in the 
 first quarter exhibits two well-marked waves with crests (low death 
 rate) in 1870 and 1883. 
 
 Respiratory disease is of course closely associated with cold 
 weather. A spell of cold weather, in winter, is generally followed 
 by a rise in the deaths from bronchitis, &c. We have already con- 
 sidered the mean temperature of the first quarter (No. 17 in first 
 part), and we repeat the smoothed curve of it since 1755. There is 
 a fair amount of correspondence with the respiratory curve (from 
 the nature of the case we could not expect it to be very exact), the 
 temperature wave showing maxima in 1870 and 1883, and the 
 average death rate being then low. Now the temperature curve we 
 found to correspond fairly with the sunspot curve. But to say that 
 we cough more or less, according as the sun's face is less or more 
 spotted, might argue some fitness for the attention of Commissioners 
 in Lunacy. (Anmtal Summary ; Buchan's Table, loc. ctt.} 
 
82 
 
 1 6. ZYMOTIC DISEASE IN SEVERE AND MILD 
 WINTERS. 
 
 DIAGRAM : Six smoothed curves (3 av.) of weekly deaths from zymotic 
 disease in London, during first quarter of six years named. Con- 
 tinuous curves, severe winters ; dotted curves, mild winters. 
 
 The relation of zymotic diseases to cold is a somewhat complex 
 subject. It may be well to ask, What kind of mortality has there 
 been from such disease in recent severe winters ? It will be remem- 
 bered that the winters 1878-79, 1890-91, and 1894-95 were a ll notably 
 severe. Allowing for development of disease, we note the weekly 
 deaths from zymotic diseases (as specified under 13), not in the 3 
 winter months, but in the first quarter, and smoothing the figures 
 with averages of 3, we obtain the three continuous curves. Then 
 we do the same with the three mild winters 1882-83, 1883-84, 
 1884-85 (dotted curves). 
 
 It may be noted that the three mild winter curves all show more 
 deaths from zymotic disease than the curves of the very severe 
 winters 1890-91 and 1894-95, notwithstanding growth of population 
 and growth of at least diphtheria. The 1879 curve shows a higher 
 mortality than those two, coming second to that of 1885; (there was 
 a great deal of whooping cough). On the other hand, we must 
 remember the progress that has taken place in overcoming zymotic 
 disease generally, and of this the relative position of the curves 
 1879, I ^9 I an d 1895 is no doubt partly an expression. Still, it 
 appears from various evidence that intense winter cold is on the 
 whole unfavourable to the mortality from zymotic disease, and mild 
 winter weather favourable ; and in this we may find some justifica- 
 tion of the saying. u A green Yule makes a fat churchyard," which, 
 if we take the large mortality from respiratory diseases into account, 
 and its rise in severe cold, is rather falsified. ( Weekly Return?) 
 
730 
 
 WO 
 
 160 
 
 /<fO 
 
 tto 
 
 1-2 Q 
 
 J 1 I 
 
 J 1 L 
 
 2 3 
 
 S 6 7 S 
 
 ft /2- 
 
PRINTED BY 
 
 The (ipctphevtone Co. 
 
 BUSH HILL PARK, 
 ENFIELD.