228 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. JO 



period from 1878 to 1889 the temperature variations in the months 

 December and January (see also the winter curve W) goes directly 

 with the curve for the sun spots and the prominences. In the curves 

 for December and January (probably in February, and see also the 

 curve for October) we find even the three shorter periods in the 

 prominence curve with small maxima in the years 1881 to 1882, 

 1884, and 1887. In the same sun spot period, 1878 to 1889, the 

 run of the temperature curves for the summer months goes approxi- 

 mately inverted to the sun spot curve and the prominence curve. 

 For the next sun spot period, 1889 to 1901, on the other hand, the 

 agreement between the temperature curves and the sun spot and 

 prominence curves is much less regular. The curve for the winter 

 months shows a minimum corresponding to the maximum of sun 

 spots and prominences in the year 1893 and the curves for Novem- 

 ber, December, and January have furthermore a remarkable maxi- 

 mum in the years 1897 to 1898 that is particularly well marked in 

 the January curve and that has nothing corresponding to it in the 

 prominences and sun spot curves. The temperature curve SO 

 for the three summer months runs generally reverse to the sun spot 

 curve and the prominence curve. It is very noticeable that this 

 summer curve is much more similar to the yearly curve than the 

 winter curve W is. The cause of this is that the temperatures in 

 March-April and November with their great variations are for 

 the most part inverted to the winter temperature. 



THE TEMPERATURE VARIATIONS IN DIFFERENT MONTHS OF THE 

 YEAR AT LIEPE's STATION ONE 



In figure 84 we give the curves for each month of the anomalies 

 of the surface temperature after combining by two- and three-years 

 smoothing of Liepe's station I. We see that the variations in 

 the surface temperature in this most easterly part of the Atlantic 

 Ocean run almost exactly in the same direction in the different months 

 of the year. Besides, we have a great similarity with the variations 

 of the sun spots, particularly in the spring months up to June, when 

 the temperature minimum, as the reader will see, goes almost exactly 

 coincident with the minimum of the sun spots, while the maximum is 

 in part one or two years after the sun spot maximum. In the other 

 months the temperature minimum also falls with the sun spot mini- 

 mum fairly well, while the temperature maximum in part is several 

 years after the maximum of spots. Particularly in the months, June 

 to November, there is such a strongly marked tendency to a divi- 



