report for 1850, p. 66. It is referred to by Dr. :\Iallet as hav- 

 ing been compiled by L. F. Kontz and originally published in 

 the Allgemeine Encyklopadie der Wissenschaften und 

 Kiinste, von Ersch und Gruber, Thiel, 36. This list, compiled 

 by a German authority previous to 1851, comprehends a dif- 

 ferent series of earthquakes, and must be taken as indepen- 

 dent of the data employed by Mr. Taber. 



Fig-. 2. ]\[onthly Distribution of Earthquakes. German Statistics. 

 Total 882. 



Here we have a curve not so striking as that shown in 

 figure 1 but still showing the same characteristics, with the 

 maximum in the season of the autumnal equinox and minor 

 rises in the three other seasons. It should be borne in mind 

 that the old data employed, while gathered from a much 

 wider field and coinprising a larger number of earthquakes, 

 is probablv less reliable and complete. The curve formed 

 upon the Charleston data, gathered in a single restricted and 

 well defined seismic area, and doubtless more accurately com- 

 piled, ought to be the better testimony, furnishing the more 

 typical curve. 



For a third trial, we turn to another distinct and far-dis- 

 tant field, this time employing statistics compiled by Baron 

 Dairoku Kikuchi, and published by the Japanese Imperial 

 Earthquake Commission in its Report No. 19. 



Here again we have the same general characteristics in 

 a modified "form and with sufficient variation to make the 

 study interesting. Still the maximum point is in the autum- 

 nal equinox. See figure 3. 



Again we shift the field and take data of Pacific Coast 

 earthquakes from 1850 to 1887 inclusive, compiled by Profes- 



11 



