692 



DR HUGH ROBERT MILL ON THE 



if this argument holds, a fall of more than 500 inches would be necessary to completely 

 freshen the area in two or three months. Granting that this result would follow, it would 

 necessitate a fall of 5 inches of rain to diminish the density by '00025, or 1 inch of 

 rain would diminish the density by "00005. This may be taken as the maximum 

 freshening power of rain, and applying the datum to interpret the relation of the differ- 

 ence between the smoothed rainfall curve (taking two-monthly means) and the normal 

 rainfall curve to the actual density curves for bottom water, we get the general statement 

 in Table XXIII., allowing that change of rainfall takes two months to produce its full 

 effect on salinity. 



Table XXIII. — Rainfall and Salinity (Supply 10). 





1886. 



1887. 





April. 



May. 



June. 



July. 



Aug. 



Sept 



Oct. 



Nov. 



Dec. 



Jan. 



Feb. 



Mar. 



April. 



May. 



June. 



July. 



Aug. 



Sept 



Actual Bottom Salinity,} 

 Landward Division, from > 



Normal deduced from Ac- \ 

 tual Rainfall Curve, . . ) 



Normal deduced from Rain- \ 

 fall Smoothed Curve, . / 



2455 



2441 

 2448 



2455 



2 453 

 2450 



2460 



2458 

 2460 



2470 



2472 

 2469 



2475 



2468 

 2470 



2470 



2466 

 2463 



2460 



2451 

 2455 



2445 



2450 

 2444 



2440 



2433 

 2437 



2430 



2430 

 2428 



2400 



2391 

 2393 



2420 



241X 

 2410 



2430 



2423 

 2423 



2450 



2443 

 2443 



2565 



2463 

 2462 



2465 



2458 

 2459 



2465 



2456 

 2455 



2465 



2464 

 2461 



The result for the landward area is unsatisfactory, inasmuch as it fails to give closely 

 accordant* values for the same months of the two years ; but in the necessary uncertainty 

 attendant on the inexact data at our disposal, it is still possible to draw a Curve freehand, 

 taking the mean value of months of the same name, and so arrive at the best idea 

 practicable of the normal distribution of salinity. When this is done, we get as a result 

 the following estimate for average normal density. 



Table XXIV. — Normal Density of Bottom, Water in Clyde Sea Area. 





Jan. 



Feb. 



Mar. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



Year. 



Landward. ) 

 Density, 1'0, J 



Seaward, . . . 



2425 

 2480 



2410 

 2465 



2416 



2470 



2434 

 2490 



2445 

 2500 



2458 

 2515 



2464 

 2520 



2464 

 2520 



2460 

 2515 



2452 

 2505 



2440 

 2495 



2430 



2485 



2442 

 2497 



The figures for the landward division were calculated as explained above, those for the 

 seaward division were arrived at by the assumption that the normal curves would run 

 practically parallel, and be at the same interval apart as the actual curves. The upwelling 

 and heading back of the water by wind make it impossible to give more than a rough guess 

 at the normal density of the surface water, which it would not be worth while expressing 

 in figures, but the dotted curves on Plates XL and XII. show the probable arrangement. 

 The Increase of Salinity with Depth. — The general tables giving the average salinity 

 for each station, show that from the Channel landward the difference of salinity between 



