54 DR HUGH ROBERT MILL ON THE 



negative slope at the period of maximum. The curve A has mean temperature 51°*7, and 

 its form approaches the oceanic type, already showing equalisation of temperature in the 

 upper layer. Curve B, sixteen days later, shows that the surface layer of 5 fathoms has 

 cooled l° - 8 by radiation upward, and conduction downward, probably the latter predomi- 

 nating. At 22^ fathoms the temperature is the same in both, but below that very great heat- 

 ing has gone on, the rate of increase of temperature increasing with the depth, so that in 

 the bottom 5 fathoms there was a rise of 3° '5 — twice as much as the loss at the surface. 

 Curve B has mean temperature 5 3° "3, the highest temperature observed at this station 

 with a negative slope. Unfortunately, the density of the water was not determined on 

 either occasion. It might be supposed that, as the surface water cooled to a certain 

 temperature between 55° and 53°, its density became greater than that beneath, and 

 mixture resulting in equalisation of temperature immediately ensued. It is unlikely, 

 however, that so decided an increase of density would result, but a decrease of density 

 gradient would greatly facilitate the action of tide and wind in producing mixture. Figure 

 14, Plate XXIV., is an interesting case of the steady fall of temperature in winter, 

 the curves retaining a distinct negative slope. In A the surface cooling seems to have 

 been prevented from producing its full effect more than half-way down ; but in B and 

 C the curves show very regular cooling, the rate increasing towards the bottom. Curve 

 D shows the remarkably superficial effect of very cold weather when the surface water 

 is much fresher than that beneath, and curve E shows how the surface chilling was 

 annulled a few days later, probably by warmer water brought from the seaward part of 

 the Basin by the tide, while cooling went on steadily below. The mean temperature of 

 curves D and E is, however, practically unchanged : the effect might be due to mixture 

 only. 



Fig. 15, Plate XXV., illustrates with more detail than fig. 13 the gradual process by 

 which the great slope of the curve of heating diminishes as the maximum temperature 

 approaches, and how the fall of surface temperature goes on while the temperature of the 

 mass as a whole, and of the bottom temperature especially, continues to rise. Al- 

 though the volume of the water near the surface is much greater than that at greater 

 depths, it is not likely that the cooling of the upper 17 fathoms by 2° can give out 

 heat enough to warm the lower 40 fathoms by nearly 5°, and we must look elsewhere 

 than to the surface for the main source of heat. This in the present case was probably 

 the warm water filling the Dunoon Basin, which appeared to catch and store up the 

 comparatively warm upper layers of the Arran Basin, and by a return underflow affected 

 the Garroch Head depression. 



Fig. 16, Plate XXV., serves to show how gradually the homothermic curve resumes a 

 positive slope as surface heating recommences in Spring. The lower 45 fathoms 

 remain practically unchanged, while the surface layer has been warmed up by two 

 degrees, and the slow penetration of the heat downward is shown by the typical 

 paraboloid form of the upper part of the curve. 



Combining all the results shown above in figures and curves into a section illustrating 



