986 



SCIENCE. 



[N. S. Vol. XI. No. 286. 



does not generally hold. "We shall be in- 

 clined to consider the rule, that in gases with 

 polyatomic molecules the mean energy of an atom 

 is smaller than the translatory energy of amolecule, 

 to be a veritable law of nature, which, like 

 Boyle's and the other laws of gases, admits ex- 

 ceptions under certain circumstances" (p. 130). 

 " Since the bonds of the atoms by which they 

 are bound together in the molecule allow of 

 neither perfect freedom nor perfect fastness, it 

 does not seem admissible simply to count the 

 kinds of movability " (p. 144). 



Chaptee, VI. — Molecular Free Paths. 



There is little new except in § 71, where the 

 various hypotheses to account for the observed 

 relations between temperature and mean free 

 path are stated and that of Sutherland is dis- 

 cussed at some length. 



Chapter VII. — Viscosity. ■ 



We find (p. 187) "From these measures [by 

 Crookes] it resulted that Maxwell's law of the 

 constancy of the coefficient of viscosity in actual 

 gases holds down to pressures which are so small 

 that they can no longer be measured with accu- 

 racy. Only at a much higher rarefaction there 

 occurs a sudden drop in the value of the coeffi- 

 cient of viscosity." " Just as at very low pres- 

 sures, so also at very high pressures, Maxwell's 

 law of the constancy of viscosity-coefficient 

 loses its strict validity." A slightly smaller 

 value of the coefficient is now given than the 

 one stated in the first edition. 



"By the entrance of more atoms into the 

 molecule the section ^s'^ will in most cases in- 

 crease, while the molecular speed will diminish ; 

 the interval T between successive collisions may 

 therefore be constant. In this case the coeffi- 

 cient of viscosity must also be constant " * * * 

 "Hence all gaseous substances which contain 

 a large number of atoms in their molecule have 

 nearly equal coefficients of viscosity" (p. 200). 



§ 80, on viscosity of gaseous mixtures, de- 

 duces from theory a formula which is followed 

 by this passage : ' ' Puluj has made measure- 

 ments of the internal friction of mixtures of 

 carbonic acid and hydrogen, has compared his 

 results with the formula, and has found a really 

 good agreement. So, too, has Breitenbach. 

 By these observations a striking fact was con- 



firmed, which was first noticed by Graham as 

 he allowed mixtures of hydrogen with other 

 gases to flow through capillary tubes. Al- 

 though the viscosity of hydrogen is less than 

 that of carbonic acid, a slight admixture of 

 hydrogen has the eflTect of increasing and not 

 of decreasing the viscosity of carbonic acid ; 

 nor does a diminution begin until the mixture 

 contains a largish amount of hydrogen." 



In § 85, on the influence of temperature on 

 viscosity, the views of Sutherland in regard to 

 the effects of molecular attraction are again 

 made prominent. "The reasons for the as- 

 sumption that the sphere of action [of the 

 molecule] diminishes as the temperature rises 

 have been given." * * * " But in the question 

 of actual gases, those attempts at explanation 

 will certainly meet with most acceptance which 

 do not assume a real diminution of the mole- 

 cules or their spheres of action, but only an 

 apparent alteration. From this point of view 

 Sutherland's view deserves to be preferred 

 to all others. According to him we have 

 not to deal with the real sphere of action 

 [that is, the sphere of which the radius is the 

 smallest distance apart of the centres of the 

 particles at the moment of collision], but with 

 an apparent sphere which is [owing to inter- 

 molecular attraction] larger than the real one." 

 "The amount of this enlargement depends 

 on the speed with which the particles move, 

 and, therefore, on the temperature of the gas; 

 it is the greater the less the speed or the 

 lower the temperature." " la bringing forward 

 this hypothesis to explain the phenomena Suth- 

 erland had really the greatest success." 



g § 86-92, dealing with the viscosity of vapors 

 and with viscosity as affected by dissociation, 

 are made up of matter which did not appear 

 in the first edition. They contaiu, perhaps, 

 nothing more striking than the fact that con- 

 densation, occurring when a vapor, originally 

 saturated, flows with some expansion through 

 a small tube, makes the coefficient of viscosity, 

 as estimated from rate of delivery, smaller than 

 if the vapor were dry. 



Chapter VIII. — Diffusion. 



This chapter differs from the correspond- 

 ing part of the first edition mainly by the 



