two rectangles marked 2 and 6 above it. Therefore in calculating the data for the three curves of Figs. 2 and 3 it counts one in every case. So too does 1868, although it was a year of greater severity as is indicated by the solid square below the curve and the rectangles marked 12 and 8 above it. The year 1869, however, having only a single rectangle with a value of two, counts only in the computation of the data for the solid line of Fig. 3. 1870, on the other hand, is reckoned as one in computing the curve of Fig. 2 and the dotted line of Fig. 3. With 1870, which was a maximum year, we cease to count the years as being after the preceding minimum. 1871 is reckoned not as four years after 1867, but as seven years before the minimum of 1878, and so forth. By adding the figures for all the sunspot waves, and plotting the results, we get the simple frequency curves of Figs. 2 and 3. Figs. 4 and 5 are derived in the same way, except for one thing. Instead of reckoning each year of the occurrence of earthquakes or eruptions as having a value of only one, each is reckoned according to the value given it by Sayles or Jensen, respectively, as shown by the character or size of the spots and rectangles of Fig. 1. An inspection of the four curves of Figs. 2 to 5 shows that they agree in essential points. Each of the six curves, two for Sayles, and four for Jensen, has a pronounced maximum at or within a year of the time of sun-spot minimum. That is, when sunspots are fewest, earthquakes and volcanic eruptions are most numerous and most severe.
The four curves of Figs. 6 to 9 on the right-hand side of page—were drawn in exactly the same way as the four which lie beside them (Figs. 2–5), except that the sun-spot maxima were used as the reference points instead of the minima. They are introduced by way of contrast. It is evident that telluric activity is weak at times of sun-spot maxima. All the curves of Figs. 2 to 9 show the lack of symmetry characteristic of sun-spot variations. The lapse of time from maximum to minimum is greater than from minimum to maximum.
Having seen that there is a coincidence of some sort between sun-spot minima and seismo-volcanic maxima, the next step is to compare the mean sun-spot curve from maximum to maximum with the mean seismo-volcanic curve for the same period. The mean sun-spot curve is, of course, easy to obtain. Figs. 10 to 13 show the first stages in the construction of the mean seismo-volcanic curve. The time from one sun-spot maximum to the next is divided into eight periods as follows:
1. The year of maximum spots.
2. The year succeeding that of maximum spots.
3. An intermediate period of decreasing number of spots,—average length about 3½ years.