1898 - 99 .] Prof. Knott on Earthquake Vibrations. 
579 
From a study of the seismograms obtained in liis Isle of Wight 
observatory, Milne has been led to a very curious result, which 
seems to point to a reflection of trains of waves at some well-marked 
boundary or barrier. Certain seismograms show a 1 repeating ’ 
character — the group of ‘large waves’ being followed at. definite 
intervals by one or more similar groups of much smaller wavelets. 
The manner in 'which the 1 grouping ’ is reproduced is, to say the 
least, very striking. A particularly instructive case is the Sliide 
seismogram of June 29, 1898, here reproduced, with Professor 
Milne’s kind permission. 
The first figure shows, on reduced scale, the complete seismogram. 
The motion recorded lasted for 3 hours. First we have the Pre- 
liminary Tremors, increasing in intensity until after a lapse of 
20 minutes the Large Waves enter abruptly upon the scene. 
These then die away in gradually diminishing tremors. Kow a 
careful inspection will show that the following tremors may be 
Fig. 1. 
divided broadly into two groups, and that each group is not unlike 
the group of serrations that constitute the large waves. The sug- 
gestion is that they are either direct consequences of these large 
waves, or at any rate referable in their origin to the same original 
disturbance. To facilitate comparison the corresponding parts of 
each group are numbered alike — 1, F, 1"; 2, 2', 2"; etc. 
It is, however, in the two groups succeeding the Large Waves: 
that the most striking correspondence is observed. This is clearly 
brought out in fig. 2, which contains enlarged representations 
of the tail groups of the complete seismogram. Here also, to 
facilitate comparison, the corresponding parts are numbered 
similarly, but, as will be readily seen, the numbering in these does 
