236 



KNOWLEDGE 



[December 1, 1898. 



in the Lake District (from 1840) ; and further down, that 

 for Eothesay, in Bute. These are distinctly alike, and I 

 find that the curve for Portlaw, in Waterford (added in 

 the lower division), may fairly be ranked with them. The 

 Rothesay curve, from 1856, might be roughly described as 

 forming a long three-crested wave. We are in a position 

 to carry back the curve to the beginning of this century, 

 and the first part here given {before 1856) proves to be 

 the third part of another long triple-crested wave, com- 

 mencing in 1822. 



Now there is an interesting correspondence, in these 

 Keswick and Rothesay curves, to the sunspot curve (which 

 is shown in the upper division). The sunspot maxima 

 occurred in 1848, 1860, 1870, and 1883 ; and, taking the 

 Keswick curve, we have relative maxima in 1850, 1801, 

 1870, and 1884. The general course of the curves is 

 similar. 



This correspondence, at least in the case of Rothesay 

 (which has been previously noticed*), can be traced back 

 to early in the century I in the first decade it seems to 

 fail). Our interest is naturally roused by it ; but to affirm 

 a causal relation between the phenomena may be con- 



-i6 



4-i 



48 42 



44 40 



40 38 



36 36 



1840 '4 '8 '52 '6 '60 '4 '8 '72 '6 '80 '4 '8 '92 



35 



33 



31 



29 



27 



25 



23 



PlO. 2. 



sidered at least premature. Certain anomalies met with 

 in the actual values, and pointed out by Mr. Abercromby, 

 call for consideration, though we may not quite accept his 

 view about smoothing processes. It might seem to some 



* Sec Abercromby's Weather, p. 325. 



a sufficient condemnation of the sunspot theory, that we 

 have not the same kind of curve everywhere else. But 

 with the meteorological Imowledge of to-day, it may fairly, 

 I think, be argued, that opposite weather in two regions 

 {e.r/., great wetness in one, and great drought in another) 

 might both be traceable to the condition of the sun, as 

 revealed by spots. On this point, however, I must not 

 linger. 



Over what extent of region, precisely, this kind of 

 curve is obtainable, I cannot say. If we go up to Cape 

 Wrath, however, or down to Truro, we find in both cases 

 a different rcijime. The curves for these places are shown 

 in Fig. 2 (upper half). 



The curve for Cape Wrath, indeed, culminates about the 

 same years, but in the general relation of different parts 

 of the curve there is considerable change. The Truro 

 curve might perhaps be described as forming three long 

 waves. 



Another three-wave curve is that of Ventnor (lower 

 division of Fig. 2), and one meets with several curves of 

 this type in different parts of the country, even widely 

 apart. Speaking generally, there is first a prominent 

 wave-crest in one of the years 1846-50, then a crest 

 about 1861, and the third, most conspicuous of all, some- 

 where in 1874-80 ; from which point the curve descends 

 rapidly and with little break to a low point about 1888. 

 With the Ventnor curve I give that for Bothwell Castle, 

 in Lanarkshire, N.B., which has its third crest in 1874, 

 while that of Ventnor is in 1878. The curve for Brechm, 

 east coast of Scotland (Fig. 3), has also three waves, 

 with the third culminating in 1877. Other cases might 

 be given. 



The Sheffield curve (Fig. 8) might perhaps be regarded 

 as a kind of transition form to what we find in the east of 

 England (shown below). This East Anglian curve is one 

 for Dickleburgh, in Norfolk, and (last eighteen years) for 

 Norwich. From a high point in 1843 the curve (roughly 

 speaking) takes a long downward sweep, and rises again to 

 a high crest in 1877. (The Oxford curve is very simUar, 

 only the earlier culmination is in 1850.) In a curve for 

 Boston, Lincolnshire, the two chief crests are in 1847 and 

 1881. 



In the Greenwich curve (not given) a high point is 

 reached at 1867 (compare Truro and Ventnor), in addition 

 to crests at 1850, 1861 and 1879, which latter is tlie 

 highest. 



In a remarkable work published a few years ago, Prof. 

 Brueckner, of Berne, has traced variations of climate in 

 different parts of the globe since 1700 ; and from a large 

 induction of facts, he infers the existence of a weather 

 cycle of about thirty-five years in the greater part of the 

 land surface of the earth. It is interesting in this con- 

 nection to note that Lord Bacon, in one of his essays 

 (" On the Vicissitudes of Things"), refers to such a cycle as 

 being spoken of in the Low Countries. 



Taking recent years, Brueckner finds that there was a 

 cold and wet period from about 1841 to 1855, and again 

 from 1871 to 1885, while the period between these 

 (1856-70) was warm and dry. The east of England is 

 included by him among the regions for which this rule 

 holds, and our East Anglian curves are quite in accordance 

 with this view. In certain regions (comprising Ireland, 

 Scotland, West England, the Atlantic islands, &c.) the 

 variations are stated to be exceptional. 



The following years, according to Brueckner, are, 

 approximately, centres of cold and wet periods — 1700, 

 1740,1780, 1815, 1850, and 1880; while the following 

 are centres of warm and dry periods — 1720, 17C0, 1795, 

 1830, and 1860. 



