December 30, 1915] 



NATURE 



49 



will thus be evident that the word "efficiency" in 

 connection with heating apparatus can have only a 

 relative and not an absolute significance. At the best 

 we can only compare the efficiency of different methods 

 of heating relatively. But even in this the difficulties 

 are at least equally great. 



The only possible experimental means would be to 

 maintain the same room in the same conditions in the 

 same circumstances by each of the two methods, and 

 to compare the amounts of energy expended in the two 

 cases. But different methods of heating produce in 

 the room widely different and varying thermal results 

 which are very difficult either to compare, control, 

 detect, or even to define. In a practical experiment 

 on an existing room, it is impossible to exercise any 

 control over the out-of-door conditions, such as the 

 temperature, the humidity, and the air movement, 

 otherwise than by enclosing the test building in 

 another outer shell. This at once makes the experi- 

 ment unreal and artificial, and could only be applied 

 to a very few specially constructed rooms. 



As regards the indoor conditions, we have simul- 

 taneously and independently to control and to measure 

 the interchange, the humidity, and the movement, as 

 well as certain obscure electrical conditions of the air, 

 the air temperature, and what may be called the 

 radiant temperature, which is an entirely independent 

 function. All these factors are extremely difficult, 

 either to control with accuracy or to measure. They 

 all vary in different parts of the same room ; the 

 variation of any one of them would be sufficient to 

 vitiate the scientific accuracy of any exact experiment. 

 The loss of heat from a room depends to a notable 

 degree on variations in all such factors, especially on 

 the uncontrollable exterior conditions. Even so appar- 

 ently simple a function as the internal temperature 

 cannot be obtained even approximately by that of a 

 simple thermometer. This Tatter reading depends on 

 the shape, the nature of the surface, the mass, and 

 material, of the thermometer itself. The radiant tem- 

 perature esjDecially varies from point to point over the 

 whole area of a room heated by a hot body within it. 

 A scientific comparison would therefore be of so un- 

 wieldv a. nature that it would be almost useless as a 

 practical gfuide. 



A practical comparison can only be made with the 

 reservation that such a comparison would not bear 

 the brunt of scientific criticism. It would be impos- 

 sible without some pedantry to leave out of count the 

 practical object of all systems of heating, namely, to 

 render the room heated comfortable to inhabit. A 

 system which did not produce this result would be 

 valueless, however theoretically "efficient" it mifrht 

 be. This introduces variations of individual idio- 

 syncrasy. What is comfortable for one person is not 

 so regarded by another. The physiologist and 

 hv£jienist have also to be heard on the question. Con- 

 ditions which are often regarded as comfortable may 

 be notoriouslv unhealthy. Amid this diversity of con- 

 siderations, it is difficult to establish a basis even of 

 relative heating efficiency without laving down arbi- 

 trary conditions which may have little or no relation 

 to the real problems involved. It is evident that the 

 conditions to be produced must be specified in terms 

 of physics and chemistry, so as to take the matter out 

 of the subjective region. 



The temperature condition of a room is very un- 

 certain and difficult to ascertain, as the naked ther- 

 mometer is a very untrustworthy criterion of the tem- 

 perature of the air; it is equally so of the feeling of 

 warmth in a room. Indeed its reading in a heated 

 room indicates nothing but its own temperature. It 

 is possible to obtain from different kinds of correct 



NO. 2409, VOL. 96] 



thermometers at the same point in a heated room indi- 

 cations which vary by as much as io° or 12°. The 

 true ^ir temperature at the same point may be far 

 different from either of the two. It is impossible to 

 say which of these is the "correct temperature" with- 

 out laying down arbitrary conditions. Thus it is 

 absurd to take the identity of the reading of ordinary 

 thermometers on two separate occasions as an indica- 

 tion that the room is in the same temperature condi- 

 tion on the two occasions. If it could be proved that 

 I the identity of the thermometer reading produced a 

 i similar sensation of warmth the practical significance 

 i of this objection might vanish, but this is notoriously 

 untrue. The feeling of warmth is an exceedingly 

 complex matter, and can only be measured by a 

 different instrument altogether — a kind of electrical 

 calorimeter, which has to be proved by physiological 

 experiments of the most difficult character to give a 

 true criterion of the feeling of warmth. 



It would be futile even to make any comparison 

 without taking account of the interchange of air. To 

 estimate or control this is as difficult an experimental 

 problem as any previously alluded to. It can only be 

 done by introducing- into and thoroughly mixing with 

 the air of the room known quantities of an easily 

 recognised gas, and afterwards making periodic 

 analyses of the air at intervals. Mathematical cal- 

 culations based on the results of these observations 

 will then give the actual interchange. This com- 

 plicated determination must be made, otherwise any 

 possible kind of test would be altogether untrust- 

 worthy. The interchange of air generally accounts 

 for half the total loss of heat. 



As a rough practical basis of comparison between 

 different systems of heating, we may take the relative 

 amounts of energy necessary to be employed in a 

 room in order to produce the same feeling of warmth 

 as measured by a suitably calibrated instrument (not 

 a thermometer), while maintaining approximately the 

 same interchange of air. 



The subjoined table Is based on direct experiment 

 in this sense, and refers to the cost of continuous 

 uniform heating in certain natural or normal condi- 

 tions, the full description of which would occupy much 

 space. An alteration of the conditions would un- 

 doubtedly alter the percentages given. These values 

 cannot, however, in any event be taken as the practical 

 relative costs, because of the manner in which the 

 respective methods of heating are commonly employed 

 in practice. A hot-water heater or anthracite stove is 

 generally burning all the day and frequently at night. 

 A gas fire or electrical stove is usually turned off when 

 the occupant leaves the room. When these latter 

 agents are so employed, the function of the heater 

 is not to maintain a continuous constant temperature 

 throughout the room, but to provide rapidly and tern- 

 porarily such radiant conditions, that the region near 

 j the heater is in a thermal condition (including the 

 radiant condition), tolerable for persons to sit there. 

 I There is a wide difference between this and the main- 

 tenance of a continuous temperature throughout the 

 room, and a set comparison between the "cost 

 efficiency " is therefore misleading. Indeed, a com- 

 parison of cost cannot be reasonably made without a 

 good deal of definition of requirements. It is prob- 

 able, for instance, that in certain circumstances a 

 good gas stove is a cheaper method of heating an 

 occasionally used room than is an anthracite stove, 

 although the cost efficiency of the latter is far below 

 (i.e. more economical than) the former. 



In order to enable the running cost to be calculated 

 at different prices of the fuel (including in that term 

 electrical power), it will be desirable to quote the rela- 



