CLYDE SEA AREA. 35 



mediate between those fixed by observation ; but the fixed lines prevent the error from 

 being great at any point. Since the isotherms are vertical, their values show not only 

 the temperatures of surface and bottom water, but the mean temperature of the whole 

 mass of water at all times from April 1886 to May 1888, a period comprising two 

 maxima and three minima. From this diagram of continuous range, checked by constant 

 references to the actual means given in the table, a curve (fig. 6, Plate XXIII.) was 

 drawn, showing the variation of the temperature of the mass of water, the position of 

 each degree being fixed for time by the diagram. The curve, slightly smoothed, may be 

 taken as fairly representative of the actual order and amount of temperature changes. 



From a minimum of 42° on April 16th, 1886, the temperature of the water mounted 

 steadily to a maximum of 55° on September 10th, a rise of 13° in 147 days, or at the 

 rate of almost 0°"090 per day. The rate of fall, at first comparatively rapid, fell off after 

 the end of September, but became quicker again after January 20th, and the temperature 

 reached a minimum of 43°*8 (or possibly lower) on February 28th. This was a loss of 

 11°'2 in 171 days, at the average rate of 0°'065 per day. The temperature continued 

 to rise in the spring of 1887, at first slowly, and then more rapidly, reaching a maximum 

 of 5 6° "2 on September 5th. Throughout the rise the water was on the average two 

 degrees warmer than at the same period of the previous year. The period from Spring 

 minimum to Autumn maximum was 189 days, and the range of temperature 12°*4, the 

 rate of warming thus averaging only 0°*065 per day, the same as the rate of cooling 

 during the previous winter. The fall of temperature in the winter of 1887 was much 

 more uniform than in 1886. Up till December 10th, 1887 remained warmer, but after 

 that date it became colder than 1886, reaching a minimum of 42°*6 on April 10th, 1888. 

 The interval from Autumn maximum to Spring minimum was on this occasion as 

 much as 217 days, and the total cooling 13°"6, an average loss of 0°*062 per day. In 

 1886, the period of cooling was 16 per cent., and in 1887, 15 per cent, longer than the 

 period of heating. 



The mean temperature of the water for 1886 (interpolating probable values for the 

 first three months) must have been about 48° '2, and the mean temperature of the air 

 at the Mull of Cantyre lighthouse for the same period was 46° '5. The mean water- 

 temperature for 1887 was 49° # 3, and the air-temperature 47°'6 ; the water thus appearing 

 to be on the whole about 1°*7 warmer than the air. This shows that the water of the 

 Channel is on the whole a warming agent during the whole year, and confirms the Gulf 

 Stream drift theory of its origin. 



The comparison of air and water temperature in detail brings out several points of 

 interest. The range of air-temperature is greater and its phase earlier than that of water- 

 temperature. Both years showed similar relations. The air curve rose much more steeply 

 than that of the water, which it crossed upward in April close to the water minimum, 

 and came to a maximum in July, from five to six weeks before^ the water maximum. 

 Descending, the air curve cut the water downward at its maximum, cooled much more 

 rapidly, and came to a double minimum in December and March, the former being the 



