earthquakes and volcanic eruptions are most frequent and most intense during the years shortly before and after sunspot minima.
In order to estimate the reliability of Jensen's conclusion, it is necessary, first, to eliminate the personal equation by comparing his data with another set compiled independently; and, second, to eliminate accidental or sporadic occurrences and faulty observations by an appeal to averages. It has been possible to accomplish the first result by means of data which Mr. Robert W. Sayles, of Harvard University, has kindly put at my disposal. In the pursuit of certain researches having no immediate connection with the problem in hand, he had prepared a table showing the years of occurrence of notable earthquakes and eruptions from 1755 to 1902. He had divided the years into three classes according to the severity and number of the phenomena of both sorts in each year. He has kindly prepared the accompanying diagram (Fig. 1), showing, on the one hand, the years of telluric activity by means of the row of dots at the bottom, and, on the other, the number of sunspots by means of the wavy line. The open circles indicate years when notable earthquakes or eruptions occurred, although not in large numbers, nor of exceptional severity. They are reckoned as unity. The solid round dots represent years of greater severity than the preceding, and are reckoned as having a value of two in computation. The solid squares indicate extreme severity, and are reckoned at three. To the diagram as prepared by Sayles, I have added Jensen's data, as appears in the small rectangles above the sunspot curve. Jensen, unlike Sayles, has separated earthquakes and volcanic eruptions. Earthquakes are shown above the heavy horizontal line, and eruptions below. The size of the rectangles indicates Jensen's estimate of severity and frequency combined. For convenience of reference I have added the appropriate numerals.
The data of Sayles and Jensen supplement each other admirably. Neither investigator lays claim to absolute completeness in his data; but, on the contrary, both express regret that they have not been able to obtain fuller information. Nevertheless neither appears to have omitted any phenomenon of first-class importance. The method of compilation was quite different in the two cases. Sayles lays special stress upon the severity of individual earthquakes or eruptions; while Jensen emphasizes the total number of occurrences in a given year. The two sets of data were prepared without reference to each other; and different sources of information were evidently used, as appears from the relatively large importance which Jensen, an Australian, naturally assigns to the phenomena, of Oceania. For instance, to illustrate the difference in the point of view, Jensen gives to 1835 the value of 6; while Sayles makes it one of the severest years, which on Jensen's
