42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 76 



both of grade 2. This suggests a similar comparison, II, of the eclipses 

 in and out of depressions, excluding all southern eclipses, and III, 

 including only southern eclipses. 



The results of comparison II are, mean grade of eclipses in depres- 

 sions, 0.50, without depressions, 1.18; of comparison III, mean grade 

 of eclipses in depressions, 2.00, without, 1.73. So that the effect of 

 volcanic haze has been practically confined to the northern zone of 

 the shadow. 



As to the suggested eft'ect of the solar cycle of sunspots, etc. ; 

 during the period considered there is no obvious way in which such 

 an effect could be disentangled from the effects of dust. 



Dust is continuously blowing into the air from the deserts and 

 seasonal drought areas of Asia, Africa, North and South America 

 and Australia, whence it goes far, as in the trade wind belts of the 

 North Atlantic and on the Pacific to leeward of the loess areas of 

 China. 



Then, the number of known volcanic dust eruptions, big enough 

 to send clouds well up toward the stratosphere, is considerable, and 

 in the past there have doubtless been very many great unknown 

 eruptions; an illustration is the great volcanic dust fall of 1907, over 

 the Seward Peninsula and the Yukon Valley, but of unknown origin ; 

 this coincides well enough with one of the minor depressions in 

 Kimball's curve. The details of many known eruptions, the height 

 of the cloud and the spread of the ash, are ill-reported. Thus, an 

 eruption ascribed at first to one of the craters of Skaptar Jokull was 

 seen at Reykjavik, Jan. 9, 1873, and " for days thereafter the cloud 

 stood high in the sky." Thoroddsen assigns this to the Kverkfall, 

 which is 162 miles from Reykjavik ; whence the cloud must have 

 been over 15,000 feet high to be seen there at all; how much higher 

 it was is unl^nown ; and the spread of the ash outside of Iceland is 

 unknown. This must have been a very important eruption. 



The pyrheliometric observations at Arequipa, Peru, showed no 

 effect of the great Katmai depression. Whence we may conclude 

 that probably volcanic dust from either temperate zone has little effect 

 on atmospheric transparency the other side of the equator. As the 

 greatest deserts and a very large number of the dust-producing 

 volcanoes are in the northern hemisphere, it may be that the darkness 

 of the northern part of the earth's shadow is due to desert and vol- 

 canic dust. 



The observations do not yield any very definite conclusion as to 

 a difference in brightness of the eastern, or sunrise, side of the umbra, 



