644 SCIENCE PROGRESS 



occupies about 20,700 square miles, but the great majority of the 

 shocks originated within a more limited region of about 1,345 

 square miles, or 13 per cent, of the above. This may, for con- 

 venience, be termed the "earthquake zone." 



During the whole of the eight years the earthquakes of the 

 zone were, area for area, more frequent than in the region out- 

 side. But the relative frequency was far from constant. In 

 1885 earthquakes were 5^ times as frequent in the zone as in 

 equal areas outside; in 1886,4 times; in 1887, 2\ times. Pos- 

 sibly this decline in relative frequency during these three years 

 represents merely the fading activity of the after-shocks of the 

 last great earthquake in the same district, which occurred in 

 1859. At any rate, in 1887, it reached its lowest figure. In the 

 following year the relative frequency rose to 5§, in 1889 to 7, 

 and in 1890 and in 1891 up to October 27 to io|. On the next 

 day the great earthquake occurred, after which, during the re- 

 mainder of the year, the relative frequency was 139, and during 

 1892, 156, the latter higher figure being probably due to the 

 suppression of sympathetic earthquakes in the surrounding 

 district. Thus, the first and most obvious symptom of the 

 coming earthquake was a rapid increase in the frequency of 

 shocks in the earthquake zone with respect to that of the shocks 

 in the area immediately outside it. 



Another significant feature of the fore-shocks of this earth- 

 quake is their distribution along the fault-s}^stem. During the 

 five years 1885-89 they shunned as far as possible those areas 

 which, towards the end of 1891, became most prolific in after- 

 shocks ; their distribution in space was apparently without law. 

 But with the beginning of 1890 a remarkable change took place. 

 There were still one or two districts in which they were more 

 numerous than elsewhere; but, on the whole, the centres of the 

 fore-shocks cling to and mark out the fault-system that came into 

 action in 1891. Except for one portion of the whole area, and 

 that is occupied by mountains, the distribution of the centres 

 follows with remarkable uniformity the outline of the fault- 

 system — not only the actual course of the fault-scarp, but its 

 continuation to the south-east as well as the course of the deep- 

 seated fault of which there appeared no actual trace at the 

 surface. Then came the great earthquake, and immediately the 

 whole aspect of the distribution was changed. Thus the second 

 and no less significant feature of the fore-shocks is that, within 



