136 proceedings: Washington academy of sciences 



and magnetic elements. The 11-year period is prominent. An oceanic- 

 type and a continental (Eurasian) type can be distinguished. The lat- 

 ter follows the sun-spot curve directly, whereas the former type follows 

 the sun spots inversely. There is also a third and very remarkable 

 type in which the curve changes more or less suddenly from direct to 

 inverse. This sudden inversion is brought out in many curves, com- 

 paring stations in different parts of the earth, and the inversion occurs 

 in very many cases at about the year 1896. 



When the temperature curves for different months of the year are 

 compared with the sun-spot curves, these three types of agreement 

 again appear in very puzzling and unexpected combinations. 



In addition to oceanic and atmospheric temperatures, other meteoro- 

 logical elements (air pressure, wind velocity, rainfall, cloudiness, mean 

 daily temperature-amplitude) show a relation to the sun spots, sun 

 prominences, and magnetic variations, and show not only the 11-year 

 period but also shorter periods of two, three, and five and one-half years. 



The fluctuations of the temperature at the earth's surface do not 

 follow directly the variations in the energy received from the sun as 

 determined by the measurements of Abbot and Fowle. The daily and 

 yearly temperature-amplitudes are believed to furnish sufficient I'efu- 

 tation of hypotheses based on supposed variations in the absorbing 

 and reflecting power of the atmosphere, as well as of Humphreys' hy- 

 potheses as to formation of ozone or effects of volcanic dust. Bland- 

 ford's hypothesis of the effect of increased evaporation in loweiing 

 continental temperatures at sun-spot maxima is also not supported by 

 the facts of tropical land and ocean stations. 



The mistake of most authors when they have discussed the causes of 

 temperature changes has been that they took for granted that the 

 average temperature at the earth's surface was directly dependent on 

 solar radiation, and would give a direct indication of heat received. 

 They have not considered sufficiently the fact that a ver}^ great pro- 

 portion of the sun's radiation is absorbed by the higher layers of our 

 atmosphere and that the distribution of heat in the atmosphere is of the 

 greatest importance for the temperatures at the earth's surface. They 

 seem very often to have forgotten that the variations in the sun's activ- 

 ity, and in the so-called ''solar constant," and also in the sun's electric 

 radiation, may primarily influence the higher layers of the atmosphere, 

 thus indirectly guiding the distribution of atmospheric pi*essure and the 

 circulation not only of these higher layers but also of the lower parts 

 of the atmosphere. In this manner the temperature of the higher 

 latitudes may be influenced more than that of the tropics, where the 

 conditions are so stable. 



The variation in pressure gradient seems much more closely related 

 to the temperatuie of land stations than is the variation in air pressure 

 itself. For instance, the Colombo-Hj^derabad gradient runs parallel 

 to the temperature in the Himalayas but opposite to the temperature 

 at Batavia, while Bombay forms an example of those strange reversals • 

 occurring about 1896. The Iceland-Azores gradient has exactly oppo- 



