110 DR HUGH ROBERT MILL ON THE 



one exception at flood-tide) dip landward from Row I. to II.; in six of the cases (with 

 practically only one exception at ebb-tide) they dipped seaward from Row II. to I. ; and 

 one case was indeterminate. The indeterminate case was Section XVI., the observations 

 being made at low- water, when the whole mass of water was within o, 2 of 42°. 



The six instances of landward-dipping isotherms were observed in February 1888, 

 August 188(5, August 1887, September 1888, November 1886, and November 1887. 

 They are thus all cases in which the water, as a whole, was at a maximum or cooling. 

 The February instance is peculiar in showing a minimum temperature in the middle 

 depths inside the loch, and in having too few isotherms to justify discussion. It 

 may accordingly be passed over. The September Section, No. XIX., is also excep- 

 tional, as it shows a maximum temperature in the middle layers outside the bar, and 

 it is also the only one in which this form of dip occurs with an ebb-tide. It shows 

 the uprising and outflowing of water between 54° and 55° over the bar, but at the time 

 the wind was blowing strongly against the tide, and there are signs of this action driving 

 in rather cooler surface water, and compelling the ebb-tide to escape as an undercurrent. 

 Distinct traces of banking up by wind at the head of the loch are shown. The remaining 

 four cases may be taken as typical. In both the August cases the tide at Row was in 

 the first quarter of flood, in one case (No. IV.) with a wind blowing into the loch, in 

 the other (No. XII.) it was calm. In both, colder water is entering, and in both the 

 bar seems to stop the deeper and colder layers outside from getting in. In fact the 

 surface temperature at Row I. is found (judging by the isotherms) on the sill of the 

 bar; the horizontal stratum in line with the bar being 0°"5 in one case and 1°"5 in 

 the other colder at Row I. than on the bar. The two November sections show a 

 substantially similar state of matters, except that the water entering was then warmer 

 than that inside, but it entered in the same way, affecting most the surface about 

 Row II. 



If we may look on the isotherms as representing originally-horizontal planes between 

 layers of water of different temperature, and suppose that these layers do not change in 

 temperature, the inflection of the isotherms indicates that the water from outside pushes 

 against and bends up those layers, driving the bent-up portions back into the loch on 

 the surface, and so thickening them at the lines of inflection. In other words, it proves 

 that the tide is mainly a mass movement, bringing up the outer water — warmer or colder 

 according to the season — against the inner waters, pushing the latter forward bodily, not 

 slipping under and lifting them up, as is the case in a tidal river. The bar effectually 

 prevents mixture below its level, and also prevents the mere slipping to and fro of a 

 mass of water cut off horizontally above. It acts like the continental edge on oceanic 

 isotherms, causing a disturbance in the distribution of the water to a considerable 

 depth, and forcing up the deep layers to the surface, so that in these conditions of 

 flood-tide Row I. has the deep temperature raised much nearer the surface than Row II., 

 showing strong up welling. 



It is unfortunate that observations were not made on the shallowest point of 



