104 Mr. Gr. F. Fitzgerald on certain Dimensional 



•e 



from those set forth in this paper have had supporters, and 

 this I believe to be erroneous. Against § 54 I must likewise 

 protest; for it is there asserted that with unlimited plates, " in 

 order to produce results similar to those produced with limited 

 plates, the gas between the two plates must maintain a greater 

 steady pressure on the plate H" (the hot one) "than that 

 which it exerts on the colder plate C." I wish Prof. Rey- 

 nolds had mentioned what result with any plates could possibly 

 be explained by such an absurd supposition. If he means 

 that such a supposition would be necessary to explain a repul- 

 sion between the plates and between them and the sides of the 

 containing vessel, he has not sufficiently considered the prob- 

 lem. I cannot believe that he is unable to see that if the gas 

 between the plates is in a state of stress, such as he himself 

 afterwards shows it might be, in which the pressure of the gas 

 on each plate is the same, but greater than the pressure on a 

 plane at right angles to the plates — if, I say, this were the 

 state of the gas, it is evident that the pressure on the cold 

 backs of the plates might be this less pressure, and so less than 

 that on their opposed surfaces, and they would tend to recede 

 from one another. Whether the gas between the plates would 

 actually be in this state is another question; but to suppose it 

 such is not absurd in the same manner as to suppose action 

 and reaction unequal though opposite. It is hardly worth 

 while pointing out that his whole illustration by means of 

 batteries is beside this question, and that the tendency of the 

 plates to separate is proportional to their area. 



In § 60 Prof. Reynolds acknowledges that stresses in the gas, 

 which turn out subsequently to be different pressures indifferent 

 directions, cause the phenomena he is investigating, although 

 in a letter to 'Nature' (Jan. 17, 1878, p. 220) he denied that 

 such a condition of stress can have any thing to do with the 

 cause of Mr. Crookes's results. It is this change in his views 

 that makes his theory practically the same as the one he was 

 then combating. With reference to the remaining paragraphs 

 of this section I have nothing but congratulations to offer, 

 the introduction of the " mean range " being in every way 

 admirable and a notable advance in the methods of treating 

 the subject. The assumption he makes as to the character of 

 the motions of the molecules is, like the analogous one of 

 Clausius, not only probable in itself, but justified in its results, 

 and, from these, seems sufficient as a first approximation. 



With the symbols and notation I have no fault to find; but 

 I must enter a protest against his elaborate and totally un- 

 necessary division of space into eight regions. He might 

 have perfectly well calculated equations (43) to (47) with- 



