i88 Dr. Marcet on the specific gravity and temperature 
scription. The general conclusion drawn from four experi- 
ments, the results of which did not essentially differ from 
each other, was, that if a vessel filled with sea water of the 
specific gravity of about 1027, and of any temperature above 
the freezing point, be gradually and slowly cooled, the water 
contracts in bulk ; and that this contraction continues to pro- 
ceed, though in a diminishing ratio, till the temperature has 
reached 22 0 of Fahreneit’s scale. At this point the water 
appears* to expand a little, and continues to do so till its tem- 
perature is reduced to between 19 0 and 18 0 , at which point 
the fluid suddenly expands to a very considerable degree, 
shooting up with great rapidity, and forcing itself out at the 
open end of the tube. At the same moment the thermometer 
rises to 28°, and remains at that point. The liquid is now 
* I say appears, because the rise of the column, occasioned by the contraction of 
the glass, may in part account for this effect. It would have been extremely difficult 
to have estimated this circumstance with precision in the above experiment, because 
the tube belonging to my apparatus was not perfectly uniform in its bore. But by 
ascertaining the capacity of a given portion of the tube, as well as that of the bulb of 
the apparatus, and calculating the contraction produced in the glass by a reduction 
of four degrees of temperature, I have been able to satisfy myself that the effect 
arising from this contraction could only produce about one half of the rise of the 
column observed in this experiment. So that it can hardly be doubled but that some 
expansion, however small in its amount, takes place in sea water when cooled from 
22 0 to 18''. But I hope to be soon able to repeat the experiment in a more perfect 
manner, by a method similar to that employed, for an analogous object, by MM. 
Du long and Petit, and described in their excellent paper on the “ Mesures des 
Temperatures, &c. 1818.” 
It may also be objected to this experiment, that the bulb has not its interior cooled 
uniformly, since the surface must be acted upon by the application of cold before the 
central parts. This is true to a certain extent. But from the great slowness of the 
experiment (which lasted about three hours at each time), this source of error is in a 
great degree avoided ; and, that the greatest degree of cold actually reached the centre 
of the vessel, was proved by a nucleus of ice being formed in it, which closely in- 
vested the bulb of the thermometer. 
