Appendices io Pourth Annual Meport 



servation of samples of mineral water, is superior, I believe, to anything 

 short of hermetical sealing. 



During the last few days I was forced to use bottles of a different descrip- 

 tion, my stock having run out. The samples from stations 80 and 84 

 were collected in black pint bottles, and closed in the above manner. The 

 samples from stations 69, 70, and 73 were collected in wide-mouthed 

 bottles having good glass stoppers. A layer of sealing-wax covered the 

 whole of the upper surface of the flat stoppers, which were, moreover, 

 covered and held down with a piece of oiled silk firmly tied round the 

 neck of the bottle. 



I if 



Method for tlie Determination of the Specific Gravities. 



In choosing a method by which to determine the specific gravities, I 

 was guided by the following considerations : — 



All methods involving corrections, based on observations of the expan- 

 sion of sea water by heat, are liable to serious objections. In the first 

 place, the observations as yet made, though numerous and elaborate, do 

 not agree sufficiently with each other. In the second place, all such 

 observations have been made either with sea water, or with sea water 

 diluted with pure water. 



Now the samples which I had collected were for the greater part 

 largely diluted with river water, the character of which is very imperfectly 

 known ; and thinking it a matter of great interest to ascertain w^hether the 

 brackish waters of the Moray Firth differ appreciably from sea water 

 diluted with pure water, it seemed to me improper to use any method 

 involving the assumption that the coefficient of expansion by heat is 

 identical in both (Jases. 



Now the only way to avoid the necessity for such corrections, and at 

 the same time to obtain results strictly comparable with each other, is to 

 exclude altogether the influence of difference of temperature by comparing 

 the different waters with each other, and with pure water at one and the 

 same temperature throughout. 



After some trials, I resolved to determine at 0° C. the specific gravities 

 of the samples I had collected, and to employ for this purpose the 

 elegant form of pyknometer devised by Sprengel. Although this apparatus 

 is well known, I may be permitted to give a short description of it, for the 

 purpose of making my mode of working intelligible. 



It consists simply of a U tube, the open ends of which terminate in two 

 tubes of narrow bore, bent at right angles in opposite directions. 



The U shape is adopted for the sake of presenting a large surface, and 

 thus rendering the instrument sensitive to changes of temperature. 



One of the narrow terminal tubes is drawn out at the end to a fine 

 capillary ; the other, and wider terminal tube, has a fine mark etched on 

 it, about an inch below the bend. 



For brevity I shall, in what follows, refer to the apparatus simply as 

 the tube, and to the narrow-bore terminal tubes as the cajnllary terminal 

 and ivider terminal respectively. 



In order to fill the tube the wider terminal is dipped into the liquid, 

 and suction applied to the capillary terminal. 



The level of the water in the wider terminal can be adjusted to the 

 mark by applying a piece of blotting paper to the capillary terminal. 



The great advantage of this form of apparatus is that, after being filled 

 with liquid at a given temperature, it may be allowed to assume the 

 temperature of the balance before weighing ; for, as the liquid and tube 



