1897.] NEW YORK ACADEMY OF SCIENCES. 379 



water slack tide, with the sky overcast, breeze slight and prac- 

 tically no sea. The distance of 112 fathoms was divided into 

 four equal parts, so that three intermediate samples were taken 

 at regular intervals between the surface and bottom collections. 



The apparatus used in obtaining the samples of water from 

 the several depths was constructed from an earthenware jug 

 containing two litres, closed with a hard rubber cork of 

 the largest size that could be urged into the opening. This 

 cork is bored and fitted with a piece of glass tubing bent 

 just above the cork and hermetically sealed at the outer end. A 

 thermometer should also liave been placed in the jar through a 

 second perforation of the cork, but we lacked this important ac- 

 cessory. The jar is then heavily weighted with bars of pig-iron, 

 and swung with fine steel wire suitably measured off into fathoms, 

 and a light guide line also attached to the bent glass tube. Such 

 a device is represented in the accompanying cut, p. 380, showing 

 the jar in readiness for lowering. W is the steel wire by which 

 the stone jar is suspended, gl being the guide line attached to 

 the bent glass tube. The iron weights are so attached as to keep 

 the jar in its perpendicular position. The apparatus may then 

 be rapidly lowered to any given depth from the deck of the boat 

 by means of a windlass, while the guide line is simultaneously 

 paid out in excess. In our collections, the water being smooth, 

 the guide line was issued from a skiff drifting a few yards away. 

 As the required depth is reached, the guide line is carefully 

 hauled taut and then given a sharp pull, which breaks the bent 

 glass tube at the angle and so allows the water to rush rapidly 

 in, because of the surrounding pressure, and fill the jug, which is 

 then drawn to the surface. We believe that there is practically 

 no intermixture of water from other levels as the jar ascends, 

 even in the short glass tube that remains in the cork ; there is, 

 indeed, an escape of fine bubbles from the contents of the jar 

 because of the diminishing pressure in its ascent, and we believe 

 that the displacement from a full jar through so small a column 

 of water as fills the tube is corrected by simply emptying the 

 tube and neck of the jar as soon as it is received, and that there 

 is then a representative sample secured by this means. The ap- 

 paratus was tested each time by lowering it to the bottom and 

 returning it without breaking the tube, and it was found that its 

 strength was sufficient. Great care is necessary in the adjust- 

 ment of the cork before lowering, as the pressure at considerable 

 depths is very great. 



The method of obtaining samples of water by means of a pump 

 through hose let down to the required depth was used by us in 

 1894 in Buzzard's Bay to a depth of twenty fathoms, and the 



