ON THE ACTION OF AIR AND WATER UPON IRON. 207 
A thermometer sheltered from radiation and 
on land does not rise, in any part of the 
Bloke above. bac ani oame) wie. ela) 400148 Fahr. 
One similarly situated on the open sea, 
no-where rises above . . +. . . . . + 87°8 
Lowest observed temperature on land . . — 58° 
Temperature of the sea in any latitude or sea- 
son, never rises above. . . . . . « + 86° 
Nor has been observed lower than . . . — 29°* 
At the mouths of tropical tidal rivers, or in lagoons, the tempe- 
rature of the water, however, may reach a much higher limit. 
157. I would here remark a cause of increased corrosive ac- 
tion, affecting castings, such as cast-iron piling, &c., at the 
mouths of tidal rivers, which has not, to my knowledge, struck 
previous observers. 
It is well known that the sea water, during the flowing of the 
tide, from its greater density, forces itself beneath the river water 
like a wedge, and slowly and imperfectly mixes with it, hence 
two strata, one of fresh or brackish water, the other of salt 
water belowit. Thus while engaged in a diving-bell survey of 
part of the bed of the river Bann, in the North of Ireland, last 
year, I found, during the flow of tide, the water strongly saline 
at the bottom of the river, and yet fresh enough to drink within 
three feet of the surface, the total depth of water being about 
25 feet; and in the Proceedings of the Royal Society of Edin- 
burgh (April, 1817) will be found a paper by Mr. Stevenson, 
C.E., in which he describes analogous phenomena as occurring 
at the mouth of the river Dee, at Aberdeen, in the rivers Forth 
and Tay, and at Loch Eil, where the Caledonian Canal joins 
the Western Sea. On taking water up at various depths at Fort 
William, he found the specific gravity 
‘ At the surface = 1008-2, 
At 9 fathoms = 1025°5, 
At 30 fathoms = 1027-2, 
or completely fresh at top, and salt as the sea itself beneath. 
‘Now Becquerel has proved that a homogeneous metallic sur- 
face (a rod or wire for instance) exposed to the action of a fluid 
“menstruum, will assume a state of electrical tension, provided 
‘that the fluid in which it is immersed be of different density in 
_ two strata, 7. e. of different corrosive power. In fact, the 
“metal and the two layers of fluid constitute a voltaic pile 
of one solid and two fiuid elements; hence as one end of the 
‘metallic rod will be in a positive state with respect to the 
‘other, it will be corroded faster than the other. Now this 
* Annales de Chimie, xxvii., 482. 
Q2 
