NATUEAL PHILOSOPHY. 189 



the evaporation of water by combustion under a kettle, its effect would be 

 diminished by any substance intervening between the flame and the kettle, 

 and the flame ought to be made to strike directly on the kettle with consider- 

 able force. But a very different fire was required to warm a room. In that 

 case, radiating substances might be employed to advantage. 



OX THE DEGREES AXD ZERO OF THE THERMOMETER. 



tf 



The following communication was presented to the American Association 

 by Professor J. F. Holton : 



The thermometer is not, as its name would indicate, a measure of heat. It 

 merely shows that one temperature is the same as another, or higher, or 

 lower, but it does not measure the difference between them. The thermom- 

 eter shows, for instance, the temperature that we name 40, 50, 60, but it 

 does not follow that the difference between 40 and 50 is the same as that 

 between 50 and 60. Indeed, it is not proved that these distances are com- 

 mensurable quantities at all, or that the one is either equal to the other, 

 greater, or less. It may be as impossible to assert such equality or inequality 

 as to say that a pound is equal to an hour, or greater, or less. To make this 

 clear let us propose to divide the difference between the boiling and freezing 

 point into two equal parts. "We may attempt to fix the middle point by mix- 

 ing a pound of water at 212 with a pound at 32 ; or we may take a pint 

 of each with a different result. If we substitute some other substance for 

 water we have every time a new result. If we abandon mixtures and resort 

 to expansions we succeed no better. Take for the most satisfactory of these 

 the expansion of the perfect gases. The rule is commonly stated that 448 

 volumes of gas at Fahrenheit increase one volume for each degree. It 

 would foUow that 480 volumes at 32 becomes 570 at 122, and 660 at 212. 

 In other words, 100 volumes at 32 become, by an increase of 90, 118*75, 

 while 90 from 122 to 212 increase 100 volumes only to 115 - 79. Thus causes 

 which we assume to be equal produce unequal effects. But the assumed law 

 of expansion of gases may itself be called in question. For if 480 volumes at 

 32 lose one volume with each degree of temperature, at 447 they will be re- 

 duced to a single volume, and at 448 annihilated. But if we admit the law 

 for temperature above 32, and would so modify it that each increase of 1 of 

 temperature shall convert 100,000 volumes into 100,208, then must the divi- 

 sions on the scales, both of the air thermometer and also of the mercurial, be 

 made logarithmic and progressively larger as the temperature increases. This 

 would add to the difficulty of graduation without any compensating advan- 

 tage. If equality can be predicated of increments of heat it must presuppose 

 an ascertainable absolute zero, just as equal gradations on a hygrometer, were 

 there such an instrument, would be based on an absolute dryness of a gas. 

 I will not speculate on the possibility of ever measuring the absolute heat hi 

 any temperature. I have designed only to show that all the points on ah 1 our 

 thermometric scales are arbitrary, like those of "Wedgewood's pyrometer, and 

 all that we can do is to select such as can be best ascertained and repeated in 

 successive instruments. The snffvtf nnd mo^t ready of these is by using the 



