2 74 



NATURE 



\Jan. 31, 1878 



The stations and the periods over which their registers of 

 summer and winter rainfall respectively extend are as follows : — 



Bengal . . . 



N.W.P. 



Calcutta 



Dacca 



Hazaribagh 



Patna 



Dehra Dun 

 Roorkee 

 Meerut 

 Benares 



" When the deviations from the local average seasonal 

 falls in each year are calculated for each of the above stations 

 separately, and the average taken for each year of the sun- 

 spot cycle, it is found, notwithstanding individual irregularities 

 which occur chiefly in the summer falls, (i) that the winter 

 rainfalls uniformly exhibit a marked tendency to vary inversely 

 with the sun-spots at all the stations, (2) that the summer rain- 

 falls show a corresponding tendency to vary directly with the 

 sun-spots, which, though strongly marked at the stations in the 

 N.W.P., is scarcely perceptible at the Bengal stations. The 

 result is best seen by combining several of the stations together, 

 and since, owing to the large differences between the actual 

 amount of rainfall at different stations, it is impossible to combine 

 the deviations from the local averages, registered in inches, I 

 have arranged the latter in the form of percentages of their 

 respective averages, then multiplied each percentage deviation 

 in each year of the sun-spot cycle by the number of years corre- 

 sponding to it at each station, added the several products for the 

 same year, and divided by the sum of the multipliers. By this 

 means each station contributes to the final result in proportion to 

 the extent of its register." Mr. Archibald then gives the tabu- 

 lated results of combining according to this method (i) the four 

 Bengal stations, (2) the four N.W.P. stations, and (3) all 

 together. 



From these tables it is seen that with very few excep- 

 tions the inverse relation between the two seasonal falls is 

 strongly manifested throughout, the] winter rainfall generally 

 tending to rise above the average in proportion as the summer 

 rainfall tends to fall below the same, and vice versd. The winter 

 rainfall moreover in every case tends to rise to a single maximum 

 exactly coinciding with the period of minimum sun-spot, descend- 

 ing thence to a single minimum which occurs a year or two after 

 the period of sun-spot maximum. The summer rainfall on the 

 other hand exhibits two maxima and minima, and though varying 

 more or less directly with the spots, this variation is principally 

 confined to the N.W.P. stations. 



The preceding peculiarities may be rendered still more apparent 

 if we take as a new mean for each year of the cycle the mean of 

 the mean percentage of the year itself together with half that of 

 the preceding and succeeding years. 



On the whole it is evident (i) that the winter rainfall through- 

 out Northern India as well as at Calcutta is subject to a periodic 

 variation amounting to nearly 50 per cent, of the average winter 

 fall and corresponding approximately with the inverse phases of 

 sun-spot frequency ; (2) that the variation in the summer rahifall, 

 though relatively much smaller, is of an almost exactly opposite 

 character, and that while well-marked in the N.W.P., it is 

 scarcely appreciable in Bengal ; (3) that from the last table the 

 cycle may be divided into two distinct portions, viz., the five 

 years preceding, and the six years succeeding, the year of sun- 

 spot maximum. In the former the winter rainfall is excessive 

 and the summer rainfall defective, while in the latter the inverse 

 relations hold, a fact somewhat analogous to the periodic change 

 in the direction of the wind at Oxford and Prague as recently 

 determined by Messrs. Main and Hornstein. 



It is scarcely possible at present to indicate the practical 

 deductions that may arise from a consideration of the preceding 

 data. One inference, however, would appear to be immediately 

 deducible, viz., that in any future comparison of the rainfalls of 

 Northern India and other countries similarly subject to distinct 

 seasonal rainfalls, due to distinct aerial currents— such as the 

 monsoon and the anti-trade winds—with the sun-spot period, 

 the summer and winter falls should be compared separately. 



otherwise it may be found that the combined effect of two oppo- 

 site seasonal variations renders the resultant variation in the 

 total annual fall very insignificant, or perhaps^obliterates it alto- 

 gether. 



With reference' to Mr. Hill's letter in Nature, vol. xvi. p. 505, 

 Mr. Blanford writes that he learns that Mr. Hill was not aware 

 of the existence of his (Mr. Blanford's) paper in the forty-fifth 

 volume of the yoiirnal of the Asiatic Society of Bengal 

 (1875), *' hence, perhaps, what I cannot but regard as 

 his under-estimate of the extent and validity of the evidence 

 opposed to his view. He discusses the registers of three sta- 

 tions, one in the North-west Himalaya, and two on the dry 

 plains of the Upper Gangetic Valley. My conclusions were 

 based on the data of eleven stations altogether, viz., one in 

 Roorkee, which is also selected by Mr. Hill, and one in Behar ; 

 one on the Eastern Himalaya, one on the plateau of Western 

 Bengal, one in Orissa, one in the Andaman Islands, and the 

 remainder in Lower Bengal and Cachar, Moreover, I was care- 

 ful to eliminate all errors arising from the use of different un- 

 compared instruments ; and how necessary such a proceeding i?, 

 I illustrated by the remark that I have known sun-thermometers 

 bearing the names of the best London mikers differing 10° and 

 1 5° in their indications when exposed side by side under similar 

 conditions to the sun. This precaution Mr. Hill has not taken, 

 and I think his results are probably in a great measure due to that 

 fact." Mr. Blanford thinks the sudden changes in the Roorkee 

 register may be accounted for by the fact that the thermometer 

 was twice changed, and the apparent increase in the wind's 

 velocity by the shifting of the anemometers at Benares and 

 Hazaribagh. 



With respect to what Mr. Hill has said of the elements of 

 error probably inherent to the method of discussion which Mr. 

 Blanford adopted, while he admits the great difficulty there is in 

 eliminating the effect of disturbing causes, he cannot admit that 

 any systematic error was introduced, in the way suggested by 

 Mr. Hill. 



Mr. Blanford concludes : — "While on this subject I would direct 

 attention to the importance of regular actinometric observations, of 

 an absolute not merely relative character (such as are shown by 

 the ordinary sun thermometers). The importance of making the 

 solar changes a part of meteorological study is now fully recog- 

 nised, and it is understood that a trained photographer is about 

 to be sent to India to take photographs of the sun, but if this is 

 so, regular actinometric observation should certainly form a part 

 of the work. The best place perhaps would be Leh, where the 

 atmosphere is remarkably free from haze and dust, which is not 

 the case on the plains of Upper India ; nor indeed, in dry 

 weather, on the north-west Himalaya. At Leh, 11,500 feet 

 above the sea, the radiation is most intense. Regular observa- 

 tions with the actinometer carried on for a few years at this 

 place should satisfactorily decide the question of the variation of 

 the sun's heat." 



Mr, Blanford also sends a reply to the letter of "Old Ma- 

 drassee " in Nature, vol. xvi. p. 519. Mr. Blanford believes 

 that to anyone who has seen or can readily refer to the report on 

 the question of the periodical variation of the rainfall of Madras, 

 it will be abundantly obvious either that " Old Madrassee " can 

 never have seen that report or that he must have misinterpreted its 

 whole purport and argument, and in his references to it, must 

 have trusted to a somewhat unusually treacherous memory. 



On the subject of solar radiation and sun-spots Mr. Hill writes 

 that since his article (vol. xvi. p. 505) was written he has gone 

 over the registers of four other stations at which solar thermo- 

 meters have been in use for five or six years. The former 

 method of treatment is not applicable to these on account of 

 breaks in the registers and changes in the instruments ; but 

 adopting suitable methods to compensate for this, the results are 

 as given in the following table, which shows the variation of each 

 year from the preceding one : — 



1876. 



+ 2'0 



;■" I '5 

 + 2-9 

 - 0-3 



+ oi 



Owing to inequalities in the number of months combining tc 



