SEICHES AND OTHER OSCILLATIONS 
89 
vibrations, but we have no evidence to produce for or against such 
a hypothesis. 
5. Towards the end of the survey of Loch Earn, some obser- 
vations were made with the statohmnograph (unfortiniately there 
was time to make only a few) which point to yet another explanation 
of some part, especially the more irregular part, of the embroidery 
on the limnograph. 
In fig. SS are placed together two statolimnograms, which were 
taken in close succession at two stations near to each other on the 
northern shore of Loch Earn, during a moderate westerly breeze 
[mean velocity 12 to 16 (mile/hour), extreme velocity occasionally 
24 (mile/hour)]. The upper one was taken in a sheltered bay to 
leeward of the delta of the Glentarken Burn, the lower about 100 
yards farther w^est, to windward of the delta. The bay was com- 
paratively calm, disturbed only by the swell propagated into it from 
the wind waves rolling; outside. The difference between the two 
limnograms is very striking. The maximum range of the embroidery 
to windward is much greater, and the pattern is much more irregular 
and complicated. What remains to leeward has much the same 
prominent periods as observed to windward, viz. '4™ to '5™, but it is 
obvious that the intervening promontory has screened off a great 
part of the vibrations. The part thus screened off* could only consist 
of surface waves of short length, and could not consist either of 
longitudinal or of transverse seiches. 
Again, I often watched the statolimnograph slowly inscribing 
indentations, such as those which are so marked in the lower 
limnogram in fig. 33, and noticed over and over again that it would 
set one down in an interval of total or comparative calm. On 
looking to windward when this happened, a black line would be seen 
on the water some distance off, indicating a coming wind-squall ; 
then presently would be heard the rustle of the wind in the trees 
overhead, and the increased prattle of the waves among the pebbles 
on the beach would show that the squall had reached the observer. 
In short, the lake-vibration had gone before, and the wind had 
followed after. The explanation seems to be that the squall exerts a 
horizontal traction on the water and causes a drift current. By and 
by this current becomes greater than the compensating return 
current underneath. Thus a hump (or a group of waves) is raised on 
the surface, which is propagated in the water with a speed usually 
exceeding the velocity of the wind in a moderate breeze. This is, in 
fact, in small a phenomenon with which sailors are familiar on a 
large scale, when they point to the long swell which records or 
presages a distant storm at sea. 
I obtained a striking confirmation of this view in the course of an 
