666 DR HUGH ROBERT MILL ON THE 



bottle it is only necessary to rotate the cylinder until the locking springs fall into the 

 notches, when the cylinder can be raised. The base-plate is perforated with two 

 apertures; one closed by an ordinary stopcock is connected with a narrow brass tube 

 which opens close below the top-plate. Its function is to admit air when the bottle is 

 being emptied ; and, if left open when the apparatus is sent to a great depth, it acts as a 

 safety-valve, allowing the excess of contained water to be expelled as it expands on the 

 diminution of pressure. The other opening simply perforates the base-plate, and is used 

 for drawing off the water. The stopcock is turned by a key — the ordinary broad lever 

 being apt to be opened by pressure of the external water in hauling up — and a few 

 inches of indiarubber tubing serve to conduct the water into the glass bottle in which it 

 is preserved. Without some such precaution it is difficult to avoid spilling a good deal 

 of the water when the sea is rough and the vessel unsteady. 



The water-bottle is fixed by the sounding line being passed through the central tube, 

 the lower end of which is bell-shaped and rests on a knot or on a " toggle," a short bar of 

 wood spliced into the rope. The detaching tube is depressed by the fall of a heavy brass 

 messenger dropped along the line from the vessel or from a reversing thermometer. 



Preservation of Samples. — When a sample of water was obtained in the water-bottle, 

 it was immediately run off into a blue-glass stoppered bottle of about 1-| litres capacity, 

 similar to those employed on the " Challenger " expedition. A small portion of the 

 sample was first run in and the bottle thoroughly rinsed out with it, then the glass bottle 

 was completely filled, the ground-glass stopper labelled and dropped into its place, and 

 the bottle placed in a wooden box divided into fifteen cork-padded compartments and 

 supplied with a lid which protected the bottles completely from rain or sun. As a rule, 

 the determination of density was proceeded with at once, but when the weather was bad 

 it was necessary to keep the samples until in harbour for the night, or until another 

 opportunity presented itself. If any delay seemed probable, the stoppers were firmly 

 tied down when the sample was secured. 



After many experiments with adhesive labels of several kinds, and with parchment 

 strips tied round the necks of the bottles, the plan was adopted of simply writing the 

 necessary data in pencil on the grouDd-glass stopper. This was easily read through the 

 neck of the bottle, and was quite secure from risk of obliteration when in its place, and 

 as soon as the sample was done with, a rub with the finger or with a towel left the stopper 

 clean and ready for use again. The data recorded on the stopper were the date, the 

 index letter of the sample, and the position from which it came. Thus " 22.12.86, C, Bot." 

 indicated that it was the third sample collected on December 22, and that it came from 

 the bottom. In the record of the temperature sounding, which was invariably taken 

 when samples were collected, the same letters were entered, and reference could thus be 

 made at any time to the hour of collection, the position, and the temperature at the place. 

 When a sample of water was reserved for chemical analysis, it was a duplicate of that 

 taken for determining the density, and was at once securely tied down and the stopper 

 covered by a cap of vegetable parchment firmly tied on, so as to reduce the risk of acci- 



