Page:Popular Science Monthly Volume 72.djvu/499

 scale would give it a value of 20. Again Jensen reckons 1855 at 17, while Sayles put it in his lowest class which would give it a value of only 7. The next year, on the other hand, is reckoned by Jensen at only 4, and by Sayles at 14. Other years, such as 1873–4–5–6 and 22 others are given values of from 1 to 16 by Jensen, but are not mentioned by Sayles; while the latter gives an open circle to 1803, 1805, 1853, 1861, 1863 and 1885, although Jensen does not assign them a value of even one. Other discrepancies might be mentioned. They are natural, indeed unavoidable, in a subject where there is so much opportunity for the personal equation, as well as for diverse authorities. A man's estimate of the severity of an earthquake or eruption is sure to depend largely upon the vividness of the account which he happens to read. Hence the great value of having two independent sets of data compiled for different purposes by men living at the antipodes, New England and New South Wales. The discrepancy in the two sets of data is an advantage because the one supplements the other, and because where the original sources of information are so diverse, the harmonious result derived from averages is highly remarkable. It can not be the result of chance.

Inspection of Fig. 1 shows that according to both Sayles and Jensen periods of minimum sunspots are times of maximum seismic and volcanic activity; whereas at periods of maximum sunspots, telluric activity almost ceases. There are certain glaring exceptions, such as 1883, or 1906 which does not appear in Fig. 1; but it should be noted that in 1883 the sunspot maximum was only about two thirds as high as the average. In order to estimate the true importance of such exceptions, I have plotted the curves shown in Figs. 2–5, showing the relative frequency and intensity of telluric activity in years of sunspot minima as compared with other years. Figs. 2 and 3 show the frequency of years in which one or more notable earthquakes or eruptions—combined in the case of Sayles, separate in that of Jensen—have occurred at the sun-spot minima and during the intervening years. To illustrate concretely, it appears in Fig. 1 that out of fourteen years of minima included in the period covered by the investigations of Sayles, 11, or 79 per cent., have, according to him, been characterized by notable earthquakes or eruptions. Out of the thirteen years immediately preceding a minimum only 5, or 38 per cent., have had noteworthy seismo-volcanic phenomena; out of those preceding a minimum by two years, 6, or 46 per cent., and so on. Of course, the curve soon comes to zero at either end, because, on an average, five or six years before or after a minimum we come to a maximum separating one wave of the sunspot curve from another. To illustrate again, 1867 was a minimum year, and has a solid circle below the sun-spot curve and