NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 20/ 



finds a longer period of eleven years and one of thirty-three years 

 and even one of a hundred and ten years. The eleven-year period 

 he found divided into two parts with two maxima and two minima. 



In figures ^6 and 'jy we give a comparison of the variations of 

 the air temperature in Stockholm and the sun spot variations. The 

 upper curve designated by t^ shows the temperature variations as 

 they are exposed by consecutive smoothed means of twelve months. 

 The second curve underneath designated by t, shows the tempera- 

 ture variations according to consecutive means of twenty-four 

 months. At the bottom is given the sun spot curve with increas- 

 ing values downward. In the t^ curve there appear the biennial 

 variations with a considerable part of their amplitude. Com- 

 paring this curve with the smoothed curves in which the two- 

 year period is entirely eliminated one sees that a large number 

 of the short variations have disappeared and in single cases one 

 can see quite distinctly the great strength which the biennial 

 period attained. Such observations may be made for the period 

 from 1810 to 1820 or for that from 1846 to 1856 or concerning 

 the relations of the middle and end of the nineties. 



In relation to the longer interval variations we will note particu- 

 larly the tg curves. In several cases one sees well-marked tempera- 

 ture minima in these curves at times when the sun spot minima 

 occurred, as for example 1844, 1855, and 1867. However the sun 

 spot minimum is often long continued ; that is to say, the inflection 

 in the curve which comes at the place of minimum is not particu- 

 larly well-marked and not so sharp as in other cases. In these 

 long-stretched-out minima one finds the temperature minimum not 

 at the lowest point of the curve, but in the transition time of the 

 bending of the curve towards the long minimum. This is, for 

 example, the case in the years 1808, 1820, 1876, 1888, and 1899. The 

 temperature maximum that follows such a temperature minimum 

 is apt to fall while the sun spots are yet few and have hardly de- 

 parted from the minimum value. In the other cases where the sun 

 spot minimum is more definite and is restricted to a shorter time 

 interval the temperature maximum during the rise of the sun spot 

 numbers occurs between minimum and maximum of sun spots. 

 This relation is conspicuous in the years 1846, 1858, and 1868-69, 

 and besides that in a couple of cases more. Generally near the time 

 of sun spot maximum there is found a new temperature minimum, 

 and this is indeed the case in all of the series of years which we have 

 investigated, that is, from 1800 to 1910. In individual cases it hap- 



