of the Fishery Board for Scotland. 



193 



expand or contract with changes of temperature, the level of the liquid in 

 the wider terminal rises and falls, the capillary terminal renjaining full all 

 the while, so that none of the liquid is expelled unless thp expansion is so 

 great as to cause the liquid to more than fill the whole of the wider 

 terminal. 



My reasons for preferring this method were the following : — 



In the first place, the results were sure to be at least as accurate as 

 those obtained by any other method hitherto adopted for the purpose. 

 They would, moreover, be free from the sources of error attaching to the 

 determination of the coefficient of expansion of sea water, and from any 

 doubt as to the applicability of such determinations to the individual 

 samples. Finally, the results themselves, apart from any value they might 

 have as bearing upon the hydrography of the Moray Firth, would have 

 considerable interest, as furnishing a severe test of the accuracy and 

 general value of the method. 



The mode of working was as follows : — The tubes, which were of not less 

 than 20 c.c. capacity, were first thoroughly cleaned by repeated rinsing 

 with distilled water, and, when necessary, with a little dilute hydrochloric 

 acid. They were then carefully dried inside by placing them in hot dis- 

 tilled water, and then passing a current of pure dry air through them. 

 Some time after every sign of moisture had disappeared they were taken 

 out, carefully wiped with a fine old linen handkerchief, and allowed to 

 stand for at least an hour in the balance-room before weighing. 



They were next filled with pure water, by dipping the wider terminals 

 into the water and connecting the capillary terminals with a water suction- 

 pump, by means of a piece of black crowquill india-rubber tubing. The 

 white vulcanised rubber balls and tubing usually supplied for this pur- 

 pose are quite unsuitable, as it is almost impossible to prevent some of the 

 loose adherent matter, with which such rubber is weighted, from entering 

 the tubes. 



The tubes having been filled considerably over the finely etched marks 

 on the wider terminals, were placed four, six, and even eight at a time 

 in a semi-solid mixture of pounded ice and water, contained in a tub 

 provided with a false bottom of perforated lead, beneath which was a layer 

 of ice. Without this false bottom perfect uniformity of temperature 

 from top to bottom could not be attained whenever the surrounding 

 temperature was much above 0° C. Tbe temperature of the ice and water 

 surrounding the tubes, which latter were always placed so as to be at 

 least 4 inches from the side of the tub, never varied perceptibly during the 

 process of cooling down to 0° C, although the thermometer employed 

 was a Kew standard, divided into half degrees Centigrade of nearly 

 2-4 mm. in length. 



The tubes while in the mixture of ice and water were supported in 

 cylindrical cages of widely perforated zinc. Each of these cages was 

 capable of supporting two or three tubes, and as they could be moved 

 freely through the ice and water, local differences of temperature were 

 easily avoided. 



The level of the ice and water was kept about an inch above the etched 

 marks on the wider terminals. As soon as the level of the water in 

 the wider terminal of a tube became constant, a piece of blotting paper 

 was applied to the capillary terminal, and in this manner water was drawn 

 off until the level in the wider terminal was only slightly above the etched 

 mark. 



After a few minutes the tube was slightly raised by means of the 

 attached platinum wire, so as to bring the etched mark just above the 

 level of the ice and water. 



N 



