OP ARTS AND SCIENCES. 79 



change of seasons to the maximum density, when displacement occurs 

 to .the depth where density, due to contamination, equals density arising 

 from cooling. 



A moment's consideration of the law of diffusion in liquids at rest 

 will convince any one of the great importance which must be given to 

 the act of flowing, through which it is possible to have a comparatively 

 pure water as a supply, over a body of quiescent water in repose of 

 other composition. 



While the subject of source of supply was under consideration, an 

 opportunity existed for exact observations on the volume and position 

 of this mass of contaminated water, more sure than the results of 

 analyses of samples from different depths. 



Recurring to my observations on sea-water, I suggested to Mr. C. L. 

 Stevenson, acting as engineer to the Board of Commissioners, that a 

 copper wire, or a strip of silvered copper, weighted so as to descend 

 into the silt, and buoyed at the surface in a vertical position, would 

 take a polarized state. The lower portion would be in the condition 

 of zinc in a compound arc of zinc and copper in acid, being acted on 

 by the sulphur eliminated from a sulphur compound present. In this 

 way the wire, or strip, would write a true statement on itself of the 

 depth of the contaminated water at any point, and define exactly where 

 contamination ceased, by a bright continuous surface above this line, 

 after some days of immersion. 



Mr. Stevenson entered with zeal and care on the apjilication of this 

 then novel test, not only demonstrating the condition of gradations of 

 impurity below, but defining by the definite line on the wire, or sti'ip, 

 the exact depth of the purer water above. Many observations were 

 made, resulting in proving the adequacy and value of this mode of 

 testing a mass of water for variation of composition in its parts. 



The following year Mr. Stevenson and his assistants extended, 

 agreeably to his promise, the application of the method of observing 

 to other bodies of water, and gave me his original notes of results, 

 both positive and negative : positive in relation to the detection of the 

 smallest portion of contaminated water, and negative when pure natural 

 water alone existed. 



Many modifications of the original applications were made, with 

 instructive results observed. Among these results was the consider- 

 able conducting power of a hemp or long flax string, which enabled it 

 to replace in part of the length the copper wire or copper-silvered 

 band. He fotmd that the chemical action, when the testing line was 

 used with the whole copper wire, or parts of it, the ferric cyanide of 



