THE SOLAR HYPOTHESIS. 



247 



comparable figures I made monthly means of the 121 observations taken at Mount Wilson from 

 June to October 1905 and from May to October 1906. The differences between these figures, 

 compared with the differences between the departures of the corresponding months, led me to 

 the supposition that a departure in temperature of 1° F., in a monthly mean observed at Arequipa, 

 is due to a departure of about 0.015 of the 'solar constant' from its normal value. If this is the 

 case, we may admit that a comparably [comparatively?] small lowering of the 'solar constant,' 

 if permanent, could produce climatical changes such as those which have really existed during the 

 Pleistocene ice age. The required diminution would indeed fall entirely within the range of the 

 momentary changes observed at Mount Wilson, the extreme values being 1.93 and 2.14 calories. 

 But it is useless to make more far-reaching speculations, the acquired facts being sufficient to 

 show the cause of the formation of pleions in tropical regions." 



Mauritius Unflian Ocean) 



Tananarive (Madagascar) 



Bulauayo(.S. Africa) 



Fig. 82. — Monthly Departures of Temperature in South Equatorial 

 Regions, showing Agreement, after Aretowski. 



Ai-ctowski's conclusion that the pleions of equatorial regions are really due to changes 

 in the solar constant receives support from the fact that other places in the same zone of 

 cUmate are characterized by similar variations. Thus figure 82 shows the smoothed 

 curves of monthly departures for four stations having a latitude of from 16° to 20° S., 

 but distributed well around the world in longitude. The agreement of the four curves is 

 unmistakable. If seasonal variations played a part in the matter this agreement would 

 possess no significance, but such is by no means the case. Each point on the curves repre- 

 sents the mean of 12 months, the middle point of a year representing the mean of one 

 January to the following December, the next point February to the succeeding January, 

 then March to February, and so forth. 



The next diagram, figure 83, represents the similarly smoothed mean departures of 

 four stations ranging across the torrid zone from Arequipa, latitude 16° 23' S., through 

 Batavia, 6° 10' S., and Colombo, 6° 59' N., to Bombay, 18° 54' N. The upper three curves 

 agree fairly well. Batavia, which has a typically equatorial climate of the simplest sort, 

 has a curve like that of Arequipa except that it is less sinuous and the maxima of 1900 and 

 the end of 1907 almost flatten out. Colombo in Ceylon has a cHmate similar to that of 

 Batavia except that it is not quite so simple, being influenced somewhat by monsoon winds 

 due to the great size of Asia. Its curve, however, resembles that of Batavia, except that 

 it is decidedly more irregular and lags a Uttle behind that of Arequipa. Bombay, the most 

 northern of our stations, is quite different from the rest. Probably this is because its chmate 

 is highly complex by reason of the strong contrast between the northeast monsoons or trades 

 of the dry winter and the southwest or true monsoons of the rainy summer. Its curve, then, 

 as might be expected, shows the same periodicity as that of the other stations, but with a lag 

 of a year or more in the main crests and with the curious addition of minor crests corre- 

 sponding to the Ai-equipa crests, as occurs in the years 1905 and 1907. It is as if the 



