374 Mr. T. Andrews. Action of Tidal Streams on [June 18, 



The plate A standing alone and unconnected was immersed in the 

 vessel S containing a measured quantity of sea water. 



The plate D also stood unconnected in the vessel F, and the action 

 on this plate was similar to that on plate C, though somewhat 

 modified in extent. This plate, although standing isolated, afforded 

 another confirmatory indication of the extensive corrosion of metals 

 exposed in water less saline at the surface than near the bottom, not- 

 withstanding that the total salinity of the water in vessel F was 

 greatly below that of the sea water. 



The plate B was immersed in the sea water, but galvanically con- 

 nected with the plate C in the vessel F. 



In the case of the last two plates B and C, circuit was completed, 

 and slow bat continuous diffusion between the two solutions accom- 

 plished by means of a triple fold of soft moistened chamois -leather L 

 (1 inch wide), one end of which was immersed deeply in the sea 

 water, the other to an equal depth in the distilled water, see fig. 1. 

 The sea water was obtained regularly from Filey Bay. The vessel 

 was filled at the commencement with 25 fluid ounces of sea water, the 

 vessel F with the same volume of distilled water. Both were 

 emptied by siphon (without disturbing the plates) and each carefully 

 refilled with their respective waters periodically every fourteen days 

 for a total period of fifty -two weeks. Seven pairs of such diffusion 

 cells of equal size were employed in the investigation. At the termi- 

 nation the plates were taken out, washed, carefully cleaned, dried at' 

 212° F., and again weighed on the balance, and the corrosive effects 

 determined by their relative loss in weight — the results are sum- 

 marised in Table C. 



The arrangement, fig. 1, was made so that the plate C, from its 

 position, was exposed near the bottom at (a) to the action of water 

 containing rather more salt than the water adjacent to the top at (&), 

 whilst constant diffusion was at the same time proceeding, the salt 

 water from vessel S also coming over into the distilled water at an 

 increasing ratio, by means of the soft moist leather connexion ; the 

 water, however, about the position (b) nearer the surface, maintained 

 a less saline composition than at (a), and the estimation of this dif- 

 ference between top and bottom water during the periodic slow dif- 

 fusions, is shown in the following analysis, selected as an example 

 from a number of similar observations taken at intervals during the 

 progress of the investigation. 



It will be seen that the diffusion cell arrangements allowed equili- 

 brium of composition between bottom and surface to be restored in 

 about fourteen days. As there was easy communication (about 

 •J inch) around the edges, not only of the plate C, but of the others, 

 the fluid during diffusion in vessel F would probably arrange itself 

 in layers of fairly equal density therein. These diffusion results also 



