[ r 53 ] 
that happen with any accuracy ; and yet the varia- 
tions in the night-time, which I have often found 
greater than in the day, have generally palled un- 
noticed; particularly, in one or two ftormy nights, 
I found the index point in the morning near the 
fame degree it did, when I placed the registers ; and 
yet it appeared, by the regifter it carried with it, 
that it had fallen feveral degrees during the ftorm. 
I Ihould imagine the metalline thermometer might 
be employed to fome ufeful purpofes, and at no very 
great expence. For inftance, a very plain inflrument 
of four fpelter bars, and three levers, might very ealily 
be contrived for hot-houfes, which, by a pin fixed in 
the fourth bar, at a proper place, adjufted by the 
botanical thermometer, might be made to raife a 
click, whenever the heat of the houfe raifed the bar 
to that point, fo as to let a ventilator operate by 
weights, until the air within the houfe became cool 
to the degree intended, by which the bars would be 
Contracted fo, as to draw back the click, and ftop the 
ventilation; by which means, the houfe might al- 
ways be kept within any two intended degrees of 
heat. The weight, which operates the ventilator, 
might be made to bear on a fpring, when it comes 
near the ground, to ring an alarm bell, to warn the 
attendant to wind up the weight, or awake him for 
the purpofe, if alleep. 
A like inflrument might probably be applied, with 
great benefit, to rooms where large affemblies are 
collected, and obliged to remain a long time. The 
unwholfomenefs of an over-heated air in fuch places, 
has been very fully proved, by the late moft worthy 
and ingenious Dr. Hales ; and yet the danger of 
Vol.LII, X fuddenly 
