640 MR E. M. WEDDERBURN 
proximity of the uninode. In fig. 16 * the sixteen diagrams show cross sections of the 
loch, the depth scale being greatly exaggerated. The blackened portion shows the area 
between the isotherms for 9° C. and 11° C. (z.e. in the neighbourhood of the discon- 
tinuity), for successive hours, on 9th August from 1 am. to 4 p.m. The period of 
the oscillation is about 15 hours, and so the diagram covers a complete period. The 
gradual alteration in the tilt of the isotherms can be easily followed. It is difficult to 
get an accurate determination of the period of the oscillation, and, indeed, as the 
distribution of temperature is always varying slightly, the period cannot be absolutely 
uniform. The curves for Station V. are purer than at Station I., as westerly winds 
= 
& 0 if 4 BO A 16 & wo 2 A Gs 
Fic. 10.—Temperature-depth diagrams, August 24th-25th, 
prevailed. Station V. being the leeward end of the loch during west winds, the dis- 
continuity was usually well marked, which is probably the reason for the greater purity 
of the curves. At Station V. from noon on 6th August to 7 a.m. on the 10th (91 hours) : 
there were six complete oscillations, giving a period of 15°17 hours for each oscillation. 
Or we may take the Station I. curves, and we find that from 7 p.m. on the 6th to 8 a.m. 
on the 9th (61 hours) there were four oscillations giving a period of 15:25 hours, and 
we may take 15°2 hours as being approximately the period of the oscillations at this 
time. It is difticult to obtain a determination of the period later in the month, but 
at Station V. there are four fairly regular oscillations from 2 a.m. on the 14th to 
10 a.m. on the 16th, which gives a period of 14 hours—very appreciably less. As 
* See explanation of diagrams, p. 652. 
