THE SOLAR HYPOTHESIS. 251 



SO far as our preseut knowledge is concerned, while in the second place he strongly empha- 

 sizes the importance of volcanic dust, a factor whose importance has not hitherto been 

 appreciated. Humphreys's main conclusion may be summed up in his own words: 



"Variations in the average temperature of the atmosphere depend jointly upon volcanic erup- 

 tions through the action of dust on radiation, * * * and upon sun-spot numbers, through, pre- 

 sumably, some intermediate action they have upon the atmosphere (Bulletin Mount Weather 

 Observatory, vol. 6, p. 25). * * * It appears, from various considerations, that, with a constant 

 or nearly constant output of solar energy, the earth itself possesses the inherent ability of profoundly 

 modifying its own climates, whether only local or world-wide. Thus, a,s the laws of radiation 

 indicate must be true, and as observations, at least back to 1750, the date of the earliest reliable 

 records, show, the temperature of the lower atmosphere is distinctly influenced by the amount of 

 volcanic dust in the upper atmosphere, in the sense that when this amount is great the average 

 temperature at the surface of the earth is abnormally low, and when the dust is absent this tem- 

 perature is comparatively high. Hence, as there appear to have been several periods of great 

 volcanic activity in the past with intervening periods of quiescence, it is inferred that volcanic dust 

 in the upper atmosphere was at least an important factor in some, if not all, of the great and uni- 

 versal climatic changes that have left their records in abandoned beaches and forsaken moraines." 



The work of Abbott, Fowle, and Humphreys seems so convmcing that we can scarcely 

 doubt that the presence of volcanic dust, temporarily at least, is an important factor in 

 determining clmiatic conditions. The degree of importance, however, is open to question. 

 This can be tested by two methods, first by seeing how far present conditions of terrestrial 

 temperature and climate actually vary in harmony with the amount of volcanic dust, 

 and second by ascertaining to what extent volcanic activity and glacial periods have been 

 coincident during geological times. The test according to the first method is easily 

 made by studying figure 85, which is a reproduction of the last part of Humphreys's main 

 diagram and is to all intents the same as the diagram of Abbott and Fowle. According 

 to these diagrams, volcanic dust does not appear to be the main factor in determining 

 climatic variations, although it seems to be an important contributing factor. In figure 85, 

 the upper curve P represents variations in the intensity of solar radiation as measured 

 by the pyroheliometer. The curve dips suddenly in 1884 just after the eruption of 

 Krakatoa, in 1902-3 when Pele, Santa Maria, and Colima were in eruption, and in 1912 

 when Katmai in Alaska belched out dust. Another dip occurs in 1890-91 and may perhaps 

 be due to Bandai-Sau in Japan and Bogoslof in Alaska, but this is by no means clear. 

 The second curve (S) is that of sun-spots, reversed in order to bring the maxima at low 

 levels and the minima at high. The third curve represents a combination of P and S. 

 The lowest curve is the average departure from mean temperature at 17 American stations 

 and 13 in other parts of the world. It seems to be representative of the world as a whole. 

 Manifestly the temperature curve is closely similar to the curve formed by combining the 

 pyroheliometer and sun-spot curves and its relationship to that can scarcely be doubted. 

 When the temperature curve is compared with P and S individually, however, one sees at 

 once that it bears a somewhat pronounced resemblance to S but very little to P. The 

 logical conclusion would therefore seem to be that variations in the sun are the main 

 factor in modifying terrestrial temperature, but their effect may be much modified by 

 the presence of volcanic dust in the atmosphere. 



My own investigations seem to confirm this conclusion. Before the appearance of 

 the articles by Abbott, Fowle, and Humphreys, I had tested the relation of tree growth and 

 volcanic eruptions according to the method employed with sun-spots and the growth of 

 trees as explained on pages 238 and 239. That is, taking all the known volcanic eruptions 

 since 1755 a. d., I gave each one a weight of 1, 2, or 3, according to its severity, and then 

 computed the intensity of volcanic activity at different portions of the sun-spot cycle. 

 The results appear as curve / in figure 76 on page 240, but all mention of the matter 



