•GICEATEST DENSITY OF AVATER. ' J gQ 



The second experiment is by far the least satisfactory; in- The second ex- 

 -tleed it may be stievvn that the results are inconsistent with each P^'"i°J'^"'^ t^f^- 



/ . jected to o« 



Other; this arises from the difficulty, not to say irapossibility, vaiiousac- 

 of keeping up a steady ice-cold temperature in ajar contain- *^°""^^" 

 ing another with comparatively warm water in it. When we 

 •consider that the \yater in the jar, the air, and the table are 

 incessantly pouring heat into the ice-cold water, and that the 

 cwarmer water descends by the exterior side of the inner jar, 

 .-and by the rnterior side of the outer; also that the ice, which 

 is to regulate the teniperature, is every moment swimming on 

 the top of the v/ater when not .agitated, we must be in doubt 

 what was the true mean temperature of the water at the bot- 

 tom of the jar during the 80 minutes which this experiment 

 continued. From the circumstances, I judge that if the agi- 

 tation were discontinued 10 minutes, the temperature at bot- 

 tom would rise from 32 to 36° ; however this may ba\e been, 

 we cannot suppose that it could be kept uniformly at 32° by 

 ^' repeated cautious agitation ;" whilst the temperature at top 

 would be constantly at or near 32°. Hence the uncertainty 

 of any conclusion derived from this experiment. If we take 

 the experiment at the 46th minute, when the two thermome- 

 ters were at 40o, and exhibit the succeeding temperatures 

 with their differences, we shall immediately peiceiv€ an irre- 

 gularity unaccountable on the supposition of uniform exterior 

 toinperature of 32''. 



Minutes, 

 46 - 



Top Dif. 

 ■ 40 



Bot. Dif. 

 40 

 — 



52 36 — 40 



1+— — X 



58 35 — 39 



1- — 2 



65 



75 



103 



34 — 37 



34 



— — 1 



— 36 



— — 2 



— 34 



Vol. XIV.— JtJNE, 1806. 



I^ 



