On the Contraction of IVater hj Heat. 167 



Bottom. Top. Air. 



Eleven o'clock 40° 40° 40" 



In 10 minutes 38+ 38 + 



— 20 38— 38 — 



— 30 37- 37- 



■ _ 40 36 36 



— 60 35-5 33-5 



— 80 33 33 



— 100 34-5 35 



_ 120 34— 34 



— 8 hours 34— 34 



A crust of ice be^an to form on the inside of the glass 

 Tthen the water in the axis of the bottom and of tlie lop 

 \v.as at 36". In the course of the experiment, it became at 

 least an inch thick. 



VV^e learn from this experiment, that cold applied to the 

 lowc-r part of a cylinder of water, nearly 18 inches long, 

 and having the temperature of 40°, is actually as speedily 

 perceived at the summit as in the axis of that part, on the 

 external surface of which it immediately acts. As fluids 

 conduct heat so very tardily, this can only arise from cur- 

 rents of cooled water ascending from the bottom, and these 

 cold currents could not move upwards, were not the water of 

 t-!)em spccificaliv lighter than that of the incumbent warmer 

 fiuid. 



The water, therefore, which at the bottom is cooled by 

 the contiguous frigorific mixture, must be expanded by the 

 loss of caloric. 



This experiment secures full force to the last, as it ob- 

 viates the objection already noticed, and also precludes 

 another. I have alreadv stated, that it may perhaps be al- 

 leged, that the fluid at the top, in experiment third, though 

 cooled to S'i'' did not descend, because below 40" the con- 

 traction is so trifling, that it does not occasion a diflerence 

 of specific gravity sufficiently great to cause the particles to 

 dt'Sfiend, when opposed by the inertia and tenacity of the 

 tiiud through which thev have to fall; or it may be con- 

 ceived, that the descent is so tardv, that lime is given to the 

 ambit nt air or subjacent fliud to furnish heat enough to 

 raise the temperature of the descending stream, and by that 

 arrest it in its downward course. 



But from the particulars above recorded, it is manifest, 

 that the chance of dcnsitv between the temperature of 32" 

 and 40" is quite sufficient to pat into niotion the particles, 

 and to enable them to overcome the obsiacle arising- from 

 inertia and tenacity, and to withstand the arresting effects 

 of atmospheric heat. 



I^ 4 Though 



