56 
Proceedings of the Royal Society 
It will be noticed by the dotted curve (Plate V., H) that since 
July the density has been continuously diminishing, though the 
larger curve shows that the diminution is very far from regular. 
On the 20th and 21st the surface water was almost fresh, indeed 
ice formed over the whole quarry at low water, although the 
temperature by a continuously immersed minimum thermometer did 
not fall below 32° Fahr. The bulb of this thermometer was 3 inches 
below the surface ; and as there was a difference of over 3° Fahr. 
between the temperature at the depth of 6 inches, and when the 
bulb was just covered, at a time when the surface temperature was 
34°, it is probable that the extreme surface layer was considerably 
colder'than the minimum thermometer showed. The ice was about a 
centimetre thick, the upper third was hard, smooth, transparent, and 
quite free from air-bubbles, while the lower part was very spongy, 
seemingly composed of successive horizontal laminse with water spaces 
between. Part of the ice collected was washed with distilled water, 
allowed to drain, then melted, and the chlorine determined by 
titration : — 
I. II. Mean. 
Grammes per litre, 0-594 0'581 0-587 
It is unfortunate that, owing to arrears of work consequent on 
illness, I did not determine the density of the water collected on the 
20th and 21st until the 25th. The two bottles had in the interim 
been frozen solid during a cold night (22nd) ; they were found next 
morning with the cords securing the stopper burst, the stoppers lying 
on the table at a little distance, and the ice protruding above the 
neck of each. On the 25th the water in the quarry was as salt as 
usual, and of course it was too late to ascertain to what depth the 
freshening on the two days had extended. 
It was suggested that the infiltration of fresh water from springs 
might lower the density of the surface water in the quarry, and so 
account for the peculiar temperature phenomena which have been 
observed there,* but on the 20th and 21st, and for several days before 
and after, all the sides of the quarry were sheathed in ice, the tem- 
perature of the air was below the freezing point, and, of course, at 
the time of sampling the tide was full. At low tide in August the 
water in the quarry was almost always found a little salter than at 
Proc. Roy. Soc. Pd., xii. 927. 
