CLYDE SEA AREA. 9 



number of clearly-marked groups, which will be fully described in the sequel. In 

 drawing, the various points are connected by straight lines, not by a freehand curve. 

 The necessity of avoiding any theorising as to the form of the curve between fixed points 

 was demonstrated by repeated observations of extraordinarily rapid changes, and even 

 inversions of temperature gradient, producing sharp inflexions in the curve representing 

 them. In some of the enclosed loch basins the upper part of the temperature curve 

 was often sickle-shaped, and the lower perfectly straight — a state of matters which 

 demanded very close observations to clearly define. It was my custom latterly to plot 

 the temperatures roughly as the observations were made, and so see before leaving 

 the station where it was necessary to take closer observations, in order to lay down 

 the region of rapid change of curvature as completely as possible. 



From the curves the mean temperature of the vertical column of water from surface 

 to bottom was obtained by drawing a straight line cutting the curve so as to leave equal 

 areas between it and the curve above and below the intersection. The centre of this 

 line gave the mean temperature. The exact position of the line was found by repeated 

 trials, and the areas measured by counting the millimetre squares. These determinations 

 were checked in a number of cases selected at random, by taking the mean of each pair 

 of contiguous observations, and so by adding these means, multiplied by the distance 

 between respective pairs of readings, and dividing by the whole depth, getting the mean 

 temperature as accurately as possible. The two methods gave results corresponding to 

 one-tenth of a degree, the limit of accuracy of the component observations. After 

 the mean temperatures of about a thousand soundings had been calculated, it was found 

 that they could not be utilised in the manner originally intended, although in special 

 cases they furnish interesting conclusions. The mean temperature of the layers, five 

 fathoms deep, next the surface and the bottom, especially the former, yield results of 

 greater importance, particularly as concerns the relation of air and water temperature. 

 Seasonal curves showing the changes in mean temperature of the vertical columus, and 

 thus — assuming constant salinity and specific heat — the changes in total heat at the 

 station, were prepared from these data. 



When dealing with enclosed lochs, and sometimes with the more open basins, it was 

 convenient to draw isotherms, showing the distribution of temperature on a section of 

 the region. The degree of exaggeration necessary in order to give the diagram sufficient 

 breadth relative to its length varied in the different cases ; but in all it was made 

 sufficient to allow of isotherms for each degree Fahrenheit being clearly laid down. The 

 depth at which the temperature at each whole degree occurred was obtained from the 

 vertical curves. The chief difficulty in the case of a long section was the fact that there 

 was often a considerable interval of time between the various soundings. As the 

 distribution of temperature was in many cases profoundly altered by a single day of 

 strong wind, this fact occasionally led to great irregularities in the run of the isotherms. 

 On the whole, however, the sections give a fair idea of the total amount and general 

 distribution of heat in the water. In order to allow the distribution of temperature 



VOL. XXXVIII. PART I. (n t 0. 1). B 



