Dec. iS, 1879] 



NATURE 



167 



Although in both tables the rainfall variation was not so 

 regular as the sun-spot variation, yet there was a remarkable 

 parallelism. At all events, both tables showed that the rainfall 

 uas greatest when the spots were most numerous, and on the 

 whole, least when they were fewest. 



From the several results now obtained it was concluded that 

 any relation that might subsist between the rainfall of Paris and 

 sun-spots was direct, instead of inverse, and that excessive rain- 

 fall in the present minimum period, or year, would be merely an 

 exception to the general rule. 



If. Flammarion was careful to state that the recent wet 

 weather and paucity of sun-spots might have been only a 

 coincidence. Were the far more frequent cases (during nearly 

 two centuries) of comparatively wet years and seasons when solar 

 maculation was greatest, and of comparatively dry years and 

 seasons when it was least, also mere coincidence*? Various 

 considerations had led to the conclusion that they were not. 



The circumstance that the sun's physical state, as indicated by 

 the changes that took place in and above the photosphere, was 

 subject to periodic variations, afforded ground for supposing that 

 corresponding variations took place on and near the earth's surface. 

 As a matter of fact, it was now universally admitted, although long 

 contested, that terrestrial magnetism and auroras were subject to 

 variations corresponding directly with those of the sun. Would it, 

 then, as M. Flammarion had asked, te at all surprising to find 

 that meteorological phenomena were subject to similar variations ? 



Supposing it were fully proved that the rainfall of the whole 

 globe varied directly as the amount of sun-spots, it could not be 

 expected that the law would invariably hold good everywhere. 

 At any given place there were exceptions to every meteorological 

 cycle. For example, on an average, the diurnal temperature in- 

 creased from near sunrise to an hour or two after noon, and then 

 decreased ; but in many parts of the world there were frequent 

 exceptions, and these were so great that the coldest and warmest 

 hours might respectively occur at any time of the day or night. 

 Yet there was a daily cycle of temperature corresponding with 

 the position of the sun. Again, there was a diumal oscillation 

 of the atmospheric pressure, which, within the tropics, was very 

 regular, though now and then disturbed or entirely masked, but 

 which, in many extra-tropical countries, could not be determined 

 except by taking means of hourly observations carried on for 

 many days. Hourly barometric observations made on ten or 

 more successive days, or cycles, in high latitudes, might not show a 

 trace of the mean diurnal oscillation, which nevertheless existed. 



It could not be said, then, that because the rainfall of a 

 place did not invariably increase and decrease as the sun spots 

 did, there was no rainfall cycle c ^responding with the sun-spot 

 cycle. On the contrary, considering the capriciousness of the 

 climate of Paris, it was somewhat surprising that a mean of the 

 rainfall for only five cycles gave such results. There were many 

 five consecutive days on which hourly observations would not 

 give more favourable results for determining the daily march of 

 the pressure of the atmosphere. But while there were 365 cycles 

 of the diurnal oscillation of the barometer in one year, the 

 same number of eleven-year rainfall cycles, if such cycles 

 existed, extended over 4,051 years ; so that it was easier to dis- 

 cover the former than the latter by observations at a single 

 station. If the rainfall of Taris and the sun-spots had been 

 observed and compared fcr as many sun-spot cycles as there 

 were cycles of the diurnal oscillation of the barometer in one 

 jcar, an eleven-year cycle of the rainfall might now be as well 

 established as the diurnal cycle of the atmospheric pressure or 

 llie diurnal cycle of the temperature of the air. But the number 

 of sun-spot cycles during which ob-ervations of the rainfall had 

 as yet been made were few. Was it necessary, then, to Wait 



thousands of years before it could be known whether or not there 

 was an eleven-year rainfall cycle ? It w as believed that such was not 

 the case. The problem might be solved in a much shorter time. 



If the total annual precipitation over the whole globe were 

 accurately known for eleven years, and if it were found that it 

 was not a constant quantity, but increased from a minimum in 

 the first year to a maximum in the fourth or fifth, and then 

 decreased to a minimum in the eleventh, there would be a strong 

 probability that this variation was due to some cause operating 

 from without, and that that cause resided in the sun. For to what 

 could such a phenomenon be attributed but to a variation in the 

 action of the great central luminary upon which the production and 

 condensation of aqueous vapour depended ? And if continued 

 observation of the total precipitation over the globe showed 

 repetitions of the same variation during several periods of 

 which the mean length was about eleven years, it would be 

 somewhat difficult to avoid the conclusion that the sun's radiant 

 energy was subject to a corresponding variation, even if no 

 trace of such variation had as yet been discovered. 



Suppose, now, that in the course of time it were found that there 

 was a periodic variation of the physical state of the sun, and 

 that this variation had the same duration and characteristics as 

 the previously known variation in the amount of aqueous pre- 

 cipitation, would it not be concluded that the latter variation 

 was intimately connected with the solar variation, although the 

 nature of the connection might be a mystery ? 



Similarly, it might be argued that if the sun, upon which 

 aqueous precipitation depended, was subject to variation, pre- 

 cipitation would be subject to a corresponding variation. 



Since, then, it was an established fact that the sun, as shown 

 by a periodic increase and decrease of spots, faculse, and erup- 

 tions, extending over a period of about eleven years, was subject 

 to variation, it might reasonably be inferred that there was a 

 similar variation in aqueous precipitation. And if actual 

 measurements of the total annual amount of precipitation over 

 the globe during one sun-spot cycle showed a variation similar in 

 every respect to the solar variation, it would be concluded, not 

 only that there was a rainfall variation, but that probably it was 

 intimately connected with the sun-spot variation. 



Theoretically, then, the object should be to ascertain the 

 annual rainfall of the globe. If this could be done for a few 

 sun-spot cycles, the question of a corresponding rainfall cycle 

 would be settled. But the total annual precipitation could not 

 be ascertained ; for, in addition to other obstacles, some parts 

 of the earth's surface were inaccessible. 



It was more than probable, however, that, supposing a rainfall 

 cycle existed, observations made at numerous points, in both 

 hemispheres, would detect it in a comparatively short time. If, 

 for example, in the course of a few sun-spot cycles, the rainfall 

 in many remote parts of the world, and under every variety of 

 climate, afforded strong evidence of corresponding cycles, if the 

 evidence became stronger as the number! of", observing stations 

 increased, and if a mean of all the results showed a rainfall 

 variation closely agreeing with the sun-spot variation, it would 

 be difficult to resist the conclusion that the rainfall had a periodi- 

 city connected in some way or other with the solar periodicity. 



Now, taking only the four sun-spot cycles from 1824 to 1867, 

 so as to avoid objections to earlier observations, it had been 

 found, as shown at former meetings of the society, not only that 

 the rainfalls of the British Islands, the Continent of Europe, 

 America, India, Mauritius, the Cape, and Australia, had, as far 

 as could be ascertained, been greatest when the sun-spots were 

 most numerous, and vice versd, but that a mean of all the obser- 

 vations taken at 138 stations gave the following results, when 

 compared with the sun spots for the same four cycles : 



