42 DR HUGH ROBERT MILL ON THE 



there was a gentle rise from 51 0, 4 at 20 h to 52°'5 at l h during ebb tide, and 

 then a gentle but fairly uniform fall with the rising tide to 51°*4 at 8 h 0. At 10 

 fathoms the changes were more pronounced than elsewhere. Starting at 49° - 2 the 

 water grew warmer somewhat irregularly until 2 h 0, when it was 51°'2. The next 

 sounding, at 3 h 0, when it was low tide, was 48°*2, a drop of 3°, and the subsequent 

 observations showed only a slight warming up to 48° "4 At 15 fathoms the temperature 

 fell from 49°'5 at 20 h to 48°'2 at 23 h ; then rose to 49°"2 at 4 h 0, and remained constant 

 at that temperature for the rest of the time. The bottom temperature was steady 

 throughout at 49° "2, the greatest variation being one-tenth of a degree. 



The average temperature of the whole section was 50°'4 at 20 h 0, gradually rose to 

 50 o, 7 at 3 h 0, and fell rather more rapidly to 49°*9 at 6 h 0, showing a tendency to increase 

 again later. The sudden dive of the 10-fathom curve happened while the mean curve was 

 stationary, thus showing that the cooling at 10 fathoms was not due to the intrusion of a 

 mass of cold water. (The actual observations were made a few minutes after each hour). 



The interpretation of the mean curve evidently is that as the tide was ebbing the 

 warmer surface water of the Arran Basin was tending to deepen on the Plateau, thus 

 slightly raising the temperature as a whole, but on the flood tide setting in the cooler 

 water of the Channel began to mix with the Plateau water. An attentive study of the 

 vertical curves (fig. 10, Plate XXIV.), explains how the singular fall of temperature at 10 

 fathoms occurred while the water as a whole remained unchanged in temperature. The 

 mean value of the curves is given in No. 19, Table IX., where the mean slope is seen to 

 be + 4°*8, the mean temperature of the surface 5 fathoms being 54 c, (extremes 55°*3 

 and 52°"9), while the mean of the bottom 5 fathoms was 49°'2 (extremes 49°'3 and 49°'l). 

 The first two soundings showed curves of a very simple type (a in fig. 10), where the 

 temperature fell in a paraboloid curve to 1 5 fathoms, and below that became apparently 

 homothermic. This simplicity entirely misled us, and, in spite of former experience, it 

 was, unfortunately, not thought necessary to fix more points in the lower 10 fathoms. 

 In all the subsequent soundings, however, this was done, and a remarkable distribution 

 came into sight. The minimum was not at the bottom at all, but in a sharp elbow of 

 the sickle-shaped curve, which at 23 h was found at 15 fathoms, with the value 48°'2 

 (b in fig. 10). This very abrupt inversion was a constant feature of the curve, and on 

 each successive sounding, the point at which it occurred was found to be higher up, and 

 the inflexion became more pronounced. It reached 10 fathoms at 3 h 0, and so produced 

 the sudden fall of temperature at that depth (c in fig. 10). There existed, in fact, 

 a thin layer of cold water between the warm heterothermic upper mass of water and the 

 cool homothermic lower layer. 



To show more clearly the changes in the distribution of temperature during the 

 period, a time-depth diagram was constructed. This is shown in Plate II. fig. 2. 

 In interpreting such a diagram it is necessary to remember that, when hours only are con- 

 sidered, water in any considerable masses cools or heats so slowly that little rearrangement 

 of the isotherms results from this cause. The isothermal sheets, traversino; a heterothermic 



