'ov. 25, 1880] 



NA TURE 



of (he Bombay Presidency, and that the proportionate increase 

 or decrease of the abuormal rainfall, corresponding to a fall or 

 rise in the abnormal pressure of a tenth of an inch of mercury, 

 amonnts to more than one hundred per cent, of the normall fall ; 

 but ttat the variations of the ordinary monsoon gradients pro- 

 duce very different effects on the rainfall of different districts, 

 depending on the geographical peculiarities of the particular 

 locality. 



From all these facts it is clear that there is some intimate rela- 

 tion between the variations of sun-spots, barometric pressure, 

 and rainfall ; and as famines in general are induced by a defi- 

 ciency of rain, it is probable that they also may be added to the 

 above list of connected phenomena. What is required in order 

 to gain an insight into the causal relation of these variations is 



that they should each and all be studied in greater detail than 

 has hitherto been attempted. Accordingly I commenced, more 

 than a year ago, a detailed investigation into the nature of the 

 abnormal variations of barometric pressure, and have been led 

 to the discovery of some new facts which appear to me to be of 

 sufficient importance to render it desirable that they should be 

 published in anticipation of the theoretical conclusions deducible 

 therefrom. 



Commencing with the daily abnormal barometric variations 

 observed at several stations in Western India, it was soon found 

 that as the time over which an abnormal barometric fluctuation 

 extended became longer and longer, the range of the fluctuation 

 became more and more uniform at the various stations, thus 

 leading to the conclusion that the abnormal variations of long 



■duration affect a very wide area. To test this inference it became 

 necessary to compare the observations recorded at Bombay with 

 those of some distant tropical station. Batavia was chosen, and 

 on curving the daily observations side by side with those of 

 Bombay, the degree of accordance between them was found to 

 be truly surprising, considering how far the two stations are 

 apart. The next step was to compare the monthly abnormal 

 variations of these two stations, and finding that they presented 

 many similar features, as well as some differences, to smooth the 

 variations by taking three-monthly means. The degree of ac- 

 ■cordance was now found to be still greater, many of the discord- 

 ances having been eliminated in the process of smoothing ; but 

 as some differences were still observable the process was re- 

 peated, giving nine-monthly means of abnormal pressure corre- 



sponding to the middle of the months January, April, July, and 

 October of each year. The carves obtained in thii way for 

 Bombay and Batavia were then found to be almost identical in 

 form, but with this very remarkable difference : the curve far 

 Batavia was seen to lag very persistently about one month behind 

 the Bombay curve. Similar results were then worked out from 

 all the available data for the following tropical stations : St. 

 Helena, Mauritius, Madras, Calcutta, and Zi-ka-wei, and for 

 comparison with them the monthly sun-spot areas ^ were treated 

 in exactly the same manner. The results are given m the fol 

 lowing table, and graphically represented by the contmuous 

 curves on the annexed plate : — 



' Taken from the paper by Messrs. De La Rue, Stewart, and Loewy, 

 published in the Pldlasophkal Transactions for 1870, p. 122. 



