TOL. XLIV.] PHILOSOPHICAL TRANSACTIONS. 403 



it will also stand at the same height when a hoar frost appears on the ground; 

 which he considers as a certain sign of the beginning of freezing. 



Having thus given an account of the several sorts of thermometers hitherto 

 used, and what degrees of heat they are proper to measure, we find none of 

 them capable of measuring the greater degrees of heat, which are the most com- 

 monly made use of by the chemists in many of their operations. Besides, all the 

 above instruments, being made of glass, are easily broken by accidents, and as 

 liable to crack of themselves, by being taken out of a great heat, and too suddenly 

 exposed to cold. Dr. M. therefore considered whether the above mentioned 

 property of solids, and especially of metals contracting with cold, and expanding 

 with heat, might not be applied to the construction of an instrument capable of 

 measuring all degrees even of the greatest cold, as well as the greatest heat, to 

 the melting copper or iron, which require more heat than any other metals to 

 melt them. Though the alterations in metals are but small, in respect of those 

 in spirits, or even mercury, yet it being found that iron, e. g. becomes -^ longer* 

 when red-hot, than when of its natural temperature; and Dr. Derham, in his 

 last paper read before the r. s. concerning the vibration of pendulums, says, 

 that a rod SQ-JJjSj- inches long, becomes -^V inch longer than its natural dimen- 

 sions in temperate air, by being exposed to heat equal to that of the human body; 

 "^-j-g- inch longer in hot sunshine; that it was 1*5- or -f inch longer than its natural 

 state, by being heated in a flaming heat; that it became -^ shorter than its 

 natural length by being quenched in cold water; and still -pf^ shorter by being 

 put into a mixture of salt and snow. From which experiments we may conclude, 

 that from Fahrenheit's cold of 40 below O, to the greatest heat iron can bear 

 without melting, a rod of three feet long will have about 4- inch increase; which 

 increase of length will be range enough to make all the intermediate degrees ob- 

 servable on an instrument. 



Suppose, in fig. 1, pi. Q, ab a rod of iron at its natural length by the heat of 

 the atmosphere, placed upright on one end; on the point of which rests a bar 

 CD moveable on an axis at a; and that, by making a fire about the end b of the 

 rod, till it is just ready to melt, the rod will increase in length Ab, and conse- 

 quently push the bar into the situation c d. Now it is obvious that though the 

 elongation of the rod Ab be even scarcely perceptible to the eye, yet if on the bar 

 CD the distance a a from the axis to the place where the rod ba pushes against it 

 be very small, and the other part of the bar aD very long, the arch od may be 

 increased at pleasure, so as to bear to be divided into any number of divisions that 

 shall be found necessary : for the arch Dd will always be to the arch cc in the same 

 proportion as the distance na is to ac ; and likewise the chords of these arches od and 

 Ab will be in the same proportion ; 7 <?, is the situation of the lever on the level ; and 



• Vide Sturm. Coll.— Orig. 

 3 f2 



