of Edinburgh, Session 1884-85. 
41 
the value of an extreme case, reading over 80 and weight 7, is 
obvious ; as generally the reading is under 80, and when over it 
weight 6 is usually employed as an addition which immerses more 
of the stem. Had the scale been always used when dry greater 
accuracy could be expected ; but, on the other hand, it must not be 
supposed that the weight of the film of water was always equal to 
0-01814 grm., for often the reading was not taken until the greater 
part of the water carried up had run down again. 
The occurrence of the maximum deviation from the correct reading 
(1*5 mm.) in the same direction as the increase due to the water 
film in an extreme case would certainly affect the fourth place to the 
extent of 1, but the probability of such an event happening is very 
small, and the sum and difference of the effects of the error of read- 
ing and the mean error due to the water film are respectively 
0-00007 and 0*00001. 
The result of the discussion is that the discrepancies between the 
two observed densities of one sample of water are accounted for, 
perhaps not fully, but satisfactorily, in so much that it is beyond 
question that the hydrometer gives results which for the purposes 
of comparison are rigorously correct to the fourth decimal place. 
It also shows that by taking further precautions greater accuracy 
may be reasonably looked for. 
3. Determination of Alkalinity. 
Sea water has an alkaline reaction which is supposed to be due 
to the presence in solution of carbonates, and especially of carbonate 
of lime. The measurement of the alkalinity is then a measurement 
of the amount of dissolved carbonate. 
It is determined by Tornoe’s method, boiling with dilute hydro- 
chloric acid and titrating the excess with potash, using aurine as an 
indicator. 
Practically 250 c.c. of the water are measured in a flask, poured 
into a porcelain basin, a few drops of neutral aurine solution added, 
15 c.c. of standard hydrochloric acid run in from a burette, and the 
whole heated over a large spirit-lamp for from twenty to twenty-five 
minutes. It usually boils for from ten to fifteen minutes. The excess 
of free acid is then determined by means of potash, the red colour 
of alkaline aurine being dispelled by a few drops of acid, and 
