August 9, 1906] 



NA TURE 



553 



values, has decreased considerably of recent years, 

 and has nearly obliterated the eleven-year variation. 



Going further afield, the curve for Batavia (East 

 Indifs) is very similar to that of Bombay. Coming- 



IB700 



2500 

 5000 



SUNSPOTCURVE'^oo 



1000 



I 



19000 



18800 ^ 

 Flc, 2. — Long bari 





to the .\ustralian continent, it will be noticed that 

 the eleven-)ear variation is well indicated in the 

 .Adelaide curve, but the amplitudes are much greater. 

 Particular attention is directed to 

 the maximum about the years 1876- 

 1S78, because in the curves for Mel- 

 bourne, Sydney, and Perth this inche 

 becomes quite insignificant. In 

 fact, it is the dropping out of 

 this maximum which gives the S™AMERICA 

 Australian curves quite a different 

 appearance from those of India, 

 although in many other respects 

 they closely resemble the Indian 

 changes. 



The .Australian curves thus in- AiioTnAi ia 

 dicatc two principal maxima about AJolKnLI/l 

 the epochs 1S68 and 18S7, with an 

 intermediate subsidiary maximum 

 about 1878 ; the principal maxima 

 are thus nineteen years apart. Tlie 

 curve given at the bottom of Fig. i 

 is drawn to represent in a general 

 manner this variation, and to serve S™AMERICA 

 as a comparison to the other curves '"vcBTto 



which follow. 



.An examination of the .South 

 .American pressures was next under- 

 taken. Here, as I have said before, 

 the data are not too numerous, but 

 I think they are sufficient to demon- ^"■- .' — Tl 



strate a long variation that is in 

 operation and the epochs of the maxima and 

 minima. 



The same method of four-year means was employed 



NO. 191 9, VOL. 74] 



as before, and curves drawn for five stations. Three 

 series of observations represent the .Argentine Republic 

 stations Cordoba, Goya, and .San Juan, while 

 .Santiago (Chili) and Curityba (Brazil) are also used, 

 as they are stations situated more 

 westerly and easterly respectively. 

 Curves representing barometric 

 changes at these places are all 

 given in Fig. 2, and arc drawn on 

 the same scale. .Although they ex- 

 tend over different periods of time, 

 there is sufficient overlapping in all 

 cases to allow one to draw con- 

 clusions as to the general kind of 

 variation over this area. 



.As was done in the case of 

 Fig. I, a curve is here drawn at the 

 foot of the South .American curves 

 to show the general nature of the 

 variation in this region. Two prin- 

 cipal maxima are very obvious 

 about the 3'ears 1874 and 1893, 

 while there seems to be an indica- 

 tion of a subsidiary maximum the 

 mean of which is about the year 

 1S83. We are here in the presence 

 of a barometric change of long 

 duration the principal maxima of 

 w hich are also about nineteen years 

 apart, so far as these observations 

 inform us. 



The question now arises. How 

 does this South American variation 

 compare with those shown to exist 

 in India and Australia? This can 

 be easily answered by comparing 

 the curves brought together in 



F'R- 3- 



The first point of importance is 

 that the South .American and Australian curves have 

 principal maxima about nineteen years apart, while 

 situated between them is another maximum of a sub- 



pared \ 



sidiary nature. The second is that the epochs of 

 these maxima in these two widely separated areas 

 are not coincident. Further, we are not here in the 



