1 7 6 



NA TURE 



[December 2 1, 1905 



merely tantalising', but I should like to present the 

 cisc as it occurred. When the figures for the aver- 

 age wind velocity were being put together, I inquired 

 about the variation from year to year. The monthly 

 values had not been combined, and a glance showed 

 lb - last year (1903) to be one of exceptionally high 

 velocity. Fur the complete year, since calculated, it is 

 twi nty-one miles per hour; the average for the twelve 

 years is eighteen miles per hour. I noted 1903 as the 

 year of heavy rainfall in this country, and asked 

 about 1893, the year of drought, especially in thespring 

 months. I found the wind velocities at St. Helena 

 were for the first half 



Jan. 

 16 



as against 



Jan. 



Feb. 

 14 



May Ju 



Feb Mar. Ap. May Jun 

 20 l6 20 16 10 



in 1S93, 



in 1903. 



without hope that the evidence for organic connection 

 would develop with further investigation. When 

 plotting the curves of wind velocity for individual 

 years, I noted that 1898 was an exceptional year, be- 

 cause it had two maxima of wind velocity, one in 

 March and one in October, instead of the usual single 

 one in September. Some information that I had for 

 Southampton seemed to indicate a similar state of 

 things for rain in England (south) in that year. 1 

 had the monthly rainfall figures for England* (south) 

 computed for each year, and looked at once to those 

 for 1898. Here are the figures for the two variables 

 compared for that year. 



St. Helena wind velocity — 



Jan 



The first two are the lowest velocities of those months 

 0,1 record; the others are low, but not the lowest. 

 The blanks mean that the instrument was not work- 

 ing properly. This suggested some sort of connec- 

 tion, a stronger trad'- wind being associated with a 

 heavier rainfall in this country. 1 obtained the 



Feb. Mar. Ap. May Jun. 

 19 19 22 2C 19 16 



South of England rainfall — 



Inch 



July Aug. Sepl. 

 10 15 22 



monthly values and plotted the several years' vari- 

 ation. There was unmistakable evidence of a large 

 seasonal variation with a maximum in September and 

 a minimum in May. I plotted the average 51 

 variation of the St. Helena wind for the twelve years, 

 and against it the seasonal rainfall in the south of 

 : d for thirty-five years, which I happened to 



have .it hand. The curves are reproduced in the 

 figure (Fig. 11. The similarity is surprising. Of 

 course, the seasonal rainfall is nut the same in all 

 localities, even in the British Isles. Somewhat similar 

 curves are, however, in he found for Stykkisholm, in 

 Iceland, and for Hakodate, in Japan, -... that the case 

 ivai not quite an isolated one. I was, therefore, not 



NO. l886, VOL. 73] 



Jar. Feb. Mar. Ap. May June fuly Aug. Sept. Oct. Nov. Dec. 



071 15S 112 i 39 3-59 i - 4 6 0-49 1-37 0-99 3- 4 S 3-67 2-86 

 There is unmistakably the second maximum of 



rainfall. It is in May, generally the driest month, 

 two months after the unusual 

 second maximum of wind velocity 

 at St. Helena. The ordinary 

 autumnal maximum of rainfall is 

 delayed a month until November, 

 just as the wind maximum is de- 

 layed a month until October. 



As a test case this seemed to be 

 almost conclusive and the connec- 

 tion to be put beyond doubt, but in 

 meteorological matters there are 

 many disappointments. Some gob- 

 lin seems to be in possession of this 

 castle in the air; we see a glimpse 

 of light; knock at the door; the 

 goblin opens it almost wide enough 

 to lei us in, and then he slams it 

 in our faces with a laugh. One 

 can almost hear the. mischievous 

 Puck crowing to the 



" Captain of our fairy band, 

 Helena is here at hand, 

 Anil those things do best please me 

 I li.ii befall preposterously. " 



There is even a faint echo of the 

 wicked exclamation 



" I.i id, what fools tli.-. mortals be! " 



When one turns from the average 

 11I years 10 the individual years, 

 after the curious test ease of [898 

 one must confess that while the 

 seasonal variation is maintained 

 fairly well in the trade wind, year by 

 year, one cannot recognise it in the rainfall. There 

 appears, perhaps ii is hardly necessarj to saj 50, 

 to be no regular seasonal variation in a single year 

 of English rainfall. Any month may he the wettesl 

 tin nth or perhaps the driest, and so a fitful 

 parallelism is rudely interrupted by a wet July or 

 some unaccountable abnormality. The phenomen- 

 ally wet vear, 1003, is truly the year of greatest 

 trade wind velocity", but the order of wind velocity 

 is not fegularl) the order of rainfall values; one. 

 wonders whether the recorder has always been work- 

 ing as one would wish ; and when the monthly 

 rainfall average is taken for the twelve correspond- 

 ing years instead of the thirty-five years, the curious 



