90 



consideration, is that the observed movements are not necessarily due to tilting, but are due to 

 electrical attractions or repulsions. Factors to be taken into account when discussing this possibility are 

 to be found in ' Proceedings of the Royal Society,' vol. A 76, 1905, p. 286. 



(J) TREMORS AND PULSATIONS. 



As shown in the films brought home by the " Discovery," tremors usually commence as intermittent 

 slight thickenings. The thickenings recur at shorter and shorter intervals until there is a thickened 

 line. This may have a width of 0'2 millim. The period of the movements they represent is probably 

 near to that of the pendulum, or 15 seconds. The duration of a storm usually lies between 6 and 20 hours. 

 These thickenings may develop into serrations when we see that the period has been that of the pendulum. 

 Regular movements with amplitudes of about 0'5 millim., and periods of 60 or 120 seconds, are evidently 

 forced vibrations, and are referred to as pulsations. These various movements have been tabulated as a 

 register, and also entered on squared paper, with the curves showing changes in the vertical. They have 

 been placed in the charge of the Royal Society. 



(c) EARTHQUAKES. 



Between March 14, 1902, and December 31, 1903, although there were many days when the instrument 

 was not working, 136 earthquakes were recorded. As none of these were felt by the staff of the 

 " Discovery," it may be assumed that none of them originated within 50 miles of the station on Ross 

 Island. A certain number were recorded all over the world, whilst* many were noted at very distant 

 observatories. These latter must have originated at distances greater than 500 miles. The measurements 

 of the various seismograms are contained in the accompanying register, which, as far as possible, also 

 contains corresponding information from 43 other stations, 38 of which have seismographs similar to that 

 used by the " Discovery." 



The results of analyses point to the following conclusions : 



1. Distributions of Origins. (See Plate 2.) 



Out of the 136 records, 73 refer to disturbances which originated in a sub-oceanic region lying between 

 New Zealand and the " Discovery." A certain number of these were only recorded by the " Discovery," 

 and the exact location of their origin is very doubtful ; others were recorded at Christchurch and 

 Wellington, others again reached Perth, while some travelled as far as their antipodes. 



On the maps published annually by the British Association to indicate the positions of origin of large 

 earthquakes, 12 districts are shown. These are named by the letters of the alphabet from A to L. 

 Districts J, I, L are not of great importance. The extremely active locality, the existence of which has 

 been made known by the work of the " Discovery," I propose to call District M. The high frequency in 

 the relief of seismic strain in the latter region indicates pronounced brady-seismical movement, an 

 inference which is quite consistent with the existence of the active Erebus and many other recent volcanic 

 peaks. It also suggests that New Zealand may be continued towards the south-west as a sub-oceanic 

 ridge, accelerations in the changes of which are announced by sudden yieldings along its base. The 

 islands of Auckland, Macquarie, and others, may indicate the existence of such a ridge, but I am not 

 aware that there are any soundings to confirm the suggestion. 



Sixteen records refer to shocks which originated near Japan the Philippines and the Celebes. Five 

 had their centres in the Himalayan region, and six off the West Coast of South America. (See Map, 

 Plate 2.) 



2. Seasonal Frequency of Antarctic Eartliquakes. 



The relative frequency of disturbances with an Antarctic origin in different seasons and months for the 

 years 1902 and 1903 is shown in the following table. The numerals in the body of the table are the 

 index numbers of earthquakes in the " Discovery " register : 



