'105 CONDUCTION OF HEAT 285 



were at that time obliged to yield to this view, because no 

 other possibility was seen of bringing the theoretical law 



f = KT]C 



into a complete agreement with the observations then 

 published. It had, in fact, been apparently established that 

 a smaller value must be assigned to the factor K for gases 

 whose atomic energy forms the greater part of the whole 

 energy than for the other gases in which the molecular 

 energy exceeds the atomic energy. These facts would really 

 be very simply explained on the hypothesis named, accord- 

 ing to which the gases with preponderating molecular 

 energy must have a better conductivity than the other 

 gases with greater atomic energy and smaller molecular 

 energy. 



Theoretical reasons could also be adduced for this hypo- 

 thesis which at that time seemed to be unavoidably necessary 

 from the results of the experimental measures that had been 

 made. If we consider the processes that occur during the 

 encounter of complicated molecular structures, we may 

 become inclined to the view that the motion of both centres 

 of gravity will chiefly be altered by the collision, while many 

 parts of each complex of atoms may scarcely be disturbed in 

 their own motions. The molecular energy would therefore 

 be carried over from place to place with greater speed or in 

 a higher degree, or, in other words, the conductivity of the 

 gases would be greater for the molecular energy than for the 

 atomic energy. 



In the first edition of this book, which was published in 

 1877, I explained these ideas as possible and admissible 

 because I knew no better explanation of the facts, although 

 even at that time I raised several weighty objections to 

 them. At the same time, however, I pointed to another 

 interpretation, in expressing my belief that all differences 

 between theory and observation might find their explanation 

 in the accumulation of the errors of experiment. 



The most important objection that can be urged against 

 this hypothesis, that gases have unequal conductivities for 

 their molecular and atomic energy, is that it contradicts an 



