0G4 DR HUGH ROBERT MILL ON THE 



Part II. — Salinity and Chemical Composition of the Water. 



1886-1887. 

 Collection of Samples. 



Surface samjiles can be most readily obtained by using an ordinary draw-bucket, 

 either when the steamer is in motion or at rest. In this way a fair mixture of the upper 

 six or eight inches of water may be procured. Throughout this memoir the term 

 " surface-water " is used to denote specimens collected in such a manner. In some instances 

 a film of almost pure rain or river-water, perhaps approaching an inch in thickness, has 

 been found overspreading a much Salter layer with which it had not had time to mix. 

 Consequently in taking samples for analysis, when the water-bottle about to be described 

 was employed, care was taken to enclose the contents at the depth of about 1 foot. 



Where it was desirable to take a sample of a definite layer, the plan adopted was to 

 attach a narrow -necked stoppered bottle to a sounding-line weighted with a 7-lb. lead (fig. 4, 

 Plate X.) and lower it to the exact depth required. The stopper was withdrawn by a thin 

 line, and the bottle quickly filled itself from the layer in which the mouth was situated. 

 This method only serves for depths under 10 fathoms and for calm water, as when sunk 

 to a great depth the two lines are apt to become twisted, and, if the stopper fits accurately, 

 the pressure of the superincumbent water often prevents it from being withdrawn. 



Samples beneath the surface were in almost all cases obtained by the use of the form 

 of slip water-bottle described to this Society in 1886 (Proc. Roy. Soc. Edin., vol. xiii. 

 pp. 539-546). It was adapted for use on the same line as the deep-sea reversing 

 thermometers, and consisted essentially of a brass cylinder which, when set free by the 

 fall of a weight, struck and was automatically clamped to a base-plate. The external 

 springs used in the clamping arrangement were apt to be damaged by striking against 

 the vessel's side ; and a reserve supply of springs had to be carried in case of accident. 

 When asked to assist the Fishery Board for Scotland in preparing new apparatus in 1886, 

 I devised a more satisfactory form, which is known as Mill's Self -Locking Water Bottle. 

 The instrument has been adopted by many investigators, in this country and abroad, 

 with complete satisfaction, and a detailed description may be suitably given (fig. 1, 

 Plate X.). 



The back-bone of the instrument is a strong brass tube, about 2 feet 6 inches long and 

 of f -inch bore, through the axis of which the sounding-line passes and on which the whole 

 apparatus is built up (fig. 2, Plate X.). About 4 inches from the lower end of the tube a 

 brass base-plate about 5 inches in diameter and £ of an inch thick is fixed, a wide ring of 

 soft vulcanised indiarubber of the same diameter being cemented in a recess on the upper side. 

 A brass top-plate of somewhat thinner metal is similarly fixed about 1 foot higher on the 

 tube, and is connected with the base-plate by three thin brass plates set radially to act as 



