Howard's Cycle of Eighteen Years in the Seasons of Britain. 71 



of ike climate (or average of all the observations of these eighteen years) 

 being 48°*126*. The nine years from 1824 to 1832 average 48°*879; the 

 nine years from 1833 to 1841 give 47 0, 374. The difference of 1-405 is 

 about equal to the difference in warmth between Ackworth, N. lat. 53° 

 38' 57", and London. I therefore call the former nine the warm, and the 

 latter nine the cold years of the cycle. The curve shows palpably the 

 bulk of the years of high temperature on the right, and of those of low 

 temperature on the left of the dividing line, but with two striking excep- 

 tions. There is a very cold year, 1829, among the warm, and a very warm 

 year, 1834, among the cold; and these considerably reduce the difference 

 between the two averages : the comparison or contrast holds best, there- 

 fore, among the years in detail. 



" The full flexuous line in fig. 1. shows the variation from year to year 

 of the total rain collected by the gauge in each. It is not here as with the 

 temperatures ; the amount of rain is balanced, or nearly so, in each nine 

 years. Thus out of 472-93 inches fallen in the whole cycle, 238-60 inches 

 appear to have fallen on the warm,an& 234-33 inches on the cold side, making 

 the ann-ual averages respectively 26*51 and 26*04 inches nearly ; which is 

 about an inch more on the whole per annum than is found to fall near Lon- 

 don — the level being at the ground, in both. If we .now look through the 

 curve (I beg pardon of mathematicians for applying the term to such a line), 

 we shall probably be first struck with an extreme of dryness (1826) fol- 

 lowed by an extreme of wetness (1828) on the warm side; then, with a 

 gradation from very wet (again following very dry) in 1830 to very dry in 

 1835 ; and this again mounting by steps to extreme wet again in 1839. In 

 fact, ten years, from 1830 to 1839, show a gradual decrease, and again an 

 increase of rain, protracted through the half-cycle, while eight years, from 

 1840 to 1829 (passing thus back to make the cycle), show repeated and 

 more extended oscillations performed in shorter times ; yet with results 

 so nearly the same, that the first set of years here specified show an average 

 rain of 26*36 inches, while the second set average 26*16 inches. Again, on 

 comparing rain with temperature, we find 1826 in the extreme at once of 

 xvarmth and dryness, and 1839 in those of wet and coldness ; but 1828 (in 

 the extreme of wetness) is equal in heat to the dry 1826; and 1829 is both 

 dry and very cold. The quantity of rain therefore is not regulated by the 

 temperature of the year : we may get it with heat, brought by winds highly 

 vaporized from the tropic ; or with cold, from the condensation effected 

 by the approach of northern air to our own atmosphere, previously charged 

 with vapour to the full ; and the dryness of 1829, with so much of cold, 

 may have been the result of the great deposition of rain in the previous 

 season. The only rule then that prevails throughout seems to be compen- 

 sation ; a wet year against a dry one, &c, and so of whole runs of seasons ; 

 and we must examine the winds for the cause." 



The author next proceeds, from the review of the rain and tem- 

 perature of whole years, as above, to an analysis of the distribution 

 of these through the several months of the year, with the view of 

 ascertaining the " difference, under equal quantities of rain, of the 

 warm from the cold side of the cycle, as regards the most important 

 of its effects, the fruitful or unfruitful character of our seasons." 



A full flexuous line in fig. 2.. presents the monthly rain, in its 

 total amounts under each month, for the nine years 1824 to 1832, 



* The numbers are here stated as given in the Corrigenda at the end of 

 the work. 



