for adjuflirig ’Thermometers . 827 
however, never amounted to more than a quarter of a 
degree, except in one thermometer, in which there were 
three obfervations out of eighteen which differed more 
than that; one of them differed fo much as 0.65° from 
fome of the reft. In the obfervations made with the ball 
immerfed a little way in the water, there was a greater dif- 
ference between the obfervations of different days, even 
negledfing thofe in which much of the fides of the pot 
were expofed to the fire. In two of the thermometers 
the different obfervations differed about -ff of a degree 
from each other; but in the other thermometer they 
varied Jbths. 
We do not at all know what this difference could be 
owing to, efpecially in the obfervations in fteam. It 
could not proceed intirely from fome unknown dif- 
ference in the water; for, if it did, the difference between 
the different thermometers fhould have been always the 
fame, which was not the cafe, though in general, on 
thofe days in which one thermometer flood high, the 
others did alfo, efpecially in the trials in fteam. More- 
over, as far as can be perceived from our experiments, 
there feems to be very little difference between different 
waters with refpecft to the heat which they acquire in boil- 
ing. We could not be fure that there was any difference be- 
tween rain or diftilled water and pump water, provided 
5 L 2 the 
