CLYDE SEA AREA. . 669 



securing concordant results. The quantities to be measured are in most cases extremely 

 small, and the conditions affecting them are numerous and variable. 



Dittmar's estimate of the probable error of the method, founded on an examination 

 of the hydrometer used on the Challenger, and on the salinity of some samples of water 

 whose densities had been measured by it, is ± 0*07, where distilled water is taken as 

 1000*00 ; but as his investigation was not carried out with the view of improving the 

 hydrometer as an instrument of precision, he did not study all the causes which lead to 

 the observed uncertainty. My experiments give practically the same conclusion, as I 

 have found the uncertainty to be greater than ± 0*03, but less than i 0"09. 



To ascertain the density I followed Buchanan's method exactly. A tall jar,- of 1 litre 

 capacity, is rinsed out with the sample to be examined, then filled up to about 800 c.c, 

 and the temperature of the water taken by a thermometer the correction of which is 

 known. The hydrometer is then placed in the liquid, and loaded so as to immerse about 

 one-third of the stem. When it comes to a position of equilibrium it is read to half a 

 millimetre by looking along the under surface of the water. Another weight is then 

 added to immerse fully two-thirds of the stem, and a second reading is made. The 

 hydrometer is then removed, dried, replaced in its box, and the temperature of the 

 water again observed. The whole operation occupies about ten minutes to perform. 

 The bottles of water, having been left in the laboratory from a day to a week (in the 

 case of the work done on the Firth of Forth), had very nearly the temperature of 

 the air, and never changed more than half a degree during the determination. In the 

 subsequent work on the Clyde the observations were almost invariably made when at sea, 

 shortly after collection ; equilibrium of temperature was not so nearly attained, and the 

 change in course of the observations was somewhat greater. 



Density is calculated from the observed readings by the formula 



where fit is the density (or specific gravity compared with water at 4°) of the sample at 

 the observed temperature t° (mean of two readings), W + w the mass of the loaded hydro- 

 meter, Yt the volume of the body of the instrument at t°, and v the volume of the 

 portion of the stem immersed (100— r, where r is the reading, because the scale in this 

 instrument is graduated from above downward). For purposes of comparison the 

 densities are reduced by Dittmar's tables to their value at 15°*56 (60° F.), so that the 

 recorded densities are 4 S 15 .5 6 . 



The following example of an actual case will illustrate the manner in which all the 

 determinations of density recorded in this memoir have been obtained. It has been 

 found more rapid to work by contracted division than by logarithms with a table of 

 differences for the main calculation, but logarithms are used for the temperature 

 reduction. The "Specific Gravity Book" is ruled in columns, each corresponding to a 

 single determination : — 



