252 
Proceedings of Royal Society of Edinburgh. [sess. 
On a New Water-Bottle for Collecting Samples of Sea- 
Water from Moderate Depths. By H. N. Dickson, 
B.K.G.S. 
(Read May 21, 1894.) 
While making physical observations during some recent cruises 
on board H.M.S. “Jackal,” I was on several occasions made to 
feel the want of a water-bottle which could be relied on to act with 
precision in rough weather. Mill’s slip water-bottle, which leaves 
nothing to be desired in its working in smooth water, at least up 
to depths of 400 or 500 fathoms, is apt to be closed prematurely 
by the jerk on the sounding line when the ship is rolling or 
pitching in a sea-way ; and considerable doubt arises as to the 
point from which samples are really obtained. 
It seemed that the object might be attained, where moderate 
depths only were concerned, by sending a vessel down closed, and 
opening it at the required depth by means of some modification of 
the reversing apparatus of the “Scottish” thermometer frame. 
Until the external pressure becomes sufficient to make the necessary 
stopcocks leak, there can be no doubt as to the purity of the sample 
admitted to the bottle, the small volume of compressed air can 
easily be got rid of, and if all the orifices opening into the bottle 
point in one direction there is no tendency for the sample to become 
mixed with other waters in hauling up, even although the stop- 
cocks remain open. 
Figs. 1 and 2 represent a small model I have had con- 
structed, which has been found to work satisfactorily at depths 
up to 40 fathoms, even in a heavy sea. The main axis of the 
instrument consists of a brass tube XT', through which the sound- 
ing line is rove. A is the cylindrical bottle, closed by taps BB' 
at its upper and lower ends. From B' an indiarubber tube B'C 
is led upwards and lashed alongside B, the extremities of the two 
tubes being in the same plane. Two arms attached to the plugs of 
the taps are connected by a rod DD'; and at D' is attached a spiral 
spring G, whose other end is hooked to a “ lug ” at T'. The releas- 
