NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 295 



If the daily number of solar prominences be taken as a measure of 

 solar activity, we obtain the same result for the eleven-year period 

 (cf. figs. 69 and 70). But in the shorter periods of a few years, 

 the atmospheric circulation seems to fluctuate inversely as the num- 

 ber of prominences, according to the observations at the Roman 

 observatory. This may be seen in figures 107 and no, where the 

 inverted curves R and V, respectively, represent the departures 

 of the daily number of prominences observed at Rome (cf. also 

 fig. 70, B and RC). The same thing was practically found by the 

 two Lockyers [1902, 1904] and by Bigelow [1908], that, e. g., the 

 pressure of Bombay should fluctuate directly as the solar promi- 

 nences, and the pressure at Cordoba inversely. 



Upon closer examination of the curves in figures 107 and no 

 we find, however, that the barometric variations demonstrated by 

 these curves frequently occur earlier than the corresponding varia- 

 tions exhibited by the curve of prominences, and besides it is only 

 the middle part of the inverted curve R, figure 107, between the 

 years 1885 and 1895, that agrees with the barometric curves. 



It has also to be considered that the observations of prominences 

 made at Palermo and Catania dififer much from the Roman obser- 

 vatio'ns. 



It is the Roman observations that have been used by the Lock3^ers 

 (and also chiefly by Bigelow) ^ and they have paid most attention 

 to the above mentioned years where there seems to be a remarkable 

 agreement between the curve of prominences and the barometric 

 curves. This explains their unexpected results. 



By special treatment of the relative numbers we have been able 

 to demonstrate a few-years period in the sioi spots, similar to those 

 found by the Lockyers, and by Bigelow in the prominences. We 

 consider it probable that the sun spots give a more trustworthy 

 measure of the solar activity, especially as the variations of sun 

 spots agree remarkably well with the variations in terrestrial mag- 

 netism, which is probably a sensitive measure of the changes in the 

 amount of solar energy received by our globe. 



The curves 6" and 71/, in figure 107, showing the fluctuations in 

 sun spots and in the daily variations of the magnetic declination 

 at Christiania, have been formed by plotting, on squared paper, the 

 consecutive twelve-month means, from which have been subtracted 



^As was mentioned before, Mr. Bigelow made, however, a serious mistake 

 in his computations of the mean daily number of prominences, by not dividing 

 the numbers observed in each month by the numbers of days of observation. 



