66 DR HUGH ROBERT MILL ON THE 



and order of the changes going on. The observations were, however, so few that when 

 the great extent of the Arran Basin and Plateau are considered it appears impossible to 

 utilise the sections for any reasonably accurate measurement of the mean temperature of 

 the region, and accordingly I do not reproduce the calculations provisionally arrived 

 at, and published in 1887 # in a preliminary paper dealing with the first year's work. 

 An attempt was made to allow for some of the probable errors in an estimate of this 

 kind, but the result was unsatisfactory. In the case of the narrower divisions, however, 

 the necessary allowances could be made, and they are discussed in detail. The sections, 

 which are practically those of the Arran Basin and its neighbouring regions, are given 

 in Plates VIII. to XL The following is a brief summary of the main features : — 



1. April 1886. — The only isotherms shown on the section are those of 43° and 42° ; 

 beneath about 15 fathoms the temperature varied from 42° to 41 0, 5. The section presents 

 a perfect picture of the uniform minimum temperature just beginning to rise at the surface. 



2. June 1886. — Eegular warming is here shown from above downward, and still 

 more markedly from the ocean inward. The isotherms all dip strongly seaward, showing 

 the effect of the warm Channel water forcing its way into the Arran Basin. The 

 Central Arran Basin is much colder than the Western Branch at the same depth, and 

 there is an evident up-draught of cold water from below at Otter. 



3. August 1886. — The same general order of temperature distribution is clearly 

 shown. Warming is still taking place from above downward, and from the sea inward, 

 the deeper isotherms showing an obvious relation to the irregularities of the bottom. 

 The up-draught at Otter does not appear to affect more than the superficial 30 fathoms. 



4. September 1886. — The conditions here are similar to those in the August 

 section, perhaps a little more pronounced, on account of the higher temperature of 

 the surface water. 



5. November 1886. — Here rapid cooling is shown. The Channel water being homo- 

 thermic at 50°, is rapidly chilling the lower layers of the deep water of the Arran Basin, 

 which remains above 51°, but the Channel water is warming the surface of the Arran 

 Basin, which has fallen below 50°, and is colder toward Loch Fyne. The upwelling at 

 Otter is marked by the warm deep water coming nearest the surface at that point and 

 overflowing into the Gortans Basin of Loch Fyne, which, at the bottom, is warmer than 

 any other water at the same depth. 



6. December 1886. — Here rapid cooling has gone on in the Arran Basin from the 

 surface down, most rapidly in the centre. The Channel water not having cooled so fast, 

 is now warmer than that of any part of the Basin, and the isotherms at the mouth of 

 the Arran Basin dip landward. In November the Channel water was aiding seasonal 

 change, and by its high salinity and low temperature rapidly affected the whole Arran 

 Basin. But cooling by surface exchange, and downward convection has now outstripped 

 it, and the Channel water is left relatively warmer, feebly retarding the inevitable work 

 of advancing winter. At the head of the Basin the isotherms dip seaward, showing the 



* Proceedings, Glasgow Philosophical Society, xviii. (1887), 332-356. 



