94 



NATURE 



[November 2b, 1891 



volume) having been determined with considerable accuracy. 

 Some of the critical constants of the various substances examined 

 are given in the following table, the brackets indicating calculated 

 values — 



Fluorbenzene 

 Chlorobenzene ... 

 Bromobenzene 



lodobenzene 



Benzene 



Carbon tetrachloride 

 Stannic chloride ... 



Ether 



Methyl alcohol 

 Ethyl alcohol 

 Propyl alcohol 

 Acetic acid 



QH,F 



! QHjCi 



I C^HjBr 

 ' C«HJ 

 I CgHg 



I CC14 



SnCl^ 



I iCoHg)^ 

 I CH3OH 



I CHsOH 

 ! Cs'HyOH 

 CHo.COOH 



233 



(262) 

 (275) 

 (298) 



219 



147 



Other tables of experimental data — including boiling-points at 

 corresponding pressures, vapour pressures at corresponding 

 temperatures, molecular volumes of liquid 'and saturated 

 vapours at corresponding pressures and at corresponding tem- 

 peratures, and ratios calculated therefrom, accompany the 

 paper. From these the author infers : (i) that Van der Waals's 

 generalizations are nearly true for chloro-, bromo-, and lodo- 

 benzene when compared with fluorbenzene ; (2) that for 

 benzene, carbon tetrachloride, stannic chloride, and ether, the 

 generalizations may only be taken as rough approximations 

 to the truth ; and (3) that for the three alcohols and acetic 

 acid, they do not hold good at all. The tables further show 

 that more consistent results are obtained when the comparisons 

 are made at corresponding pressures rather than at corresponding 

 temperatures, particularly in the case of molecular volumes of 

 saturated vapours. The subject of saturated vapours is also 

 treated by another method. If Van der Waals's deductions 

 were strictly true, then the ratios of the actual densities of the 

 saturated vapours of different substances to their theoretical 

 densities should be equal at corresponding pressures. These 

 ratios have therefore been calculated, and show an approximate 

 agreement amongst benzene and its halogen derivatives, carbon 

 tetrachloride, stannic chloride, and ether. For the other 

 substances the agreement is less satisfactory. It is also noted 

 that the ratio of the actual critical density to the theoretical 

 density is for many substances about 4-4. The alcohols difi"ering 

 so widely from the other compounds, were compared amongst 

 themselves instead of with fluorbenzene, with the result that 

 somewhat closer agreement was found, but the deviations were 

 still far outside the limits of experimental error. Of the critical 

 constants the volumes are the most difficult to determine, because 

 at the critical point the curves connecting temperature and 

 volume, and pressure and volume, are parallel to the axes of 

 volume. Accordingly, the author, in some cases, has deduced 

 this quantity by plotting against temperature the numbers 

 representing the ratios of the molecular volumes both of liquid 

 and saturated vapour to those of fluorbenzene at corresponding 

 temperatures and also at corresponding pressures. Four curves 

 result, which should intersect at the critical temperature, and 

 the point of intersection gives the ratio of the molecular 

 critical volume of the substance to that of fluor-benzene. This 

 method leads to results in fair accord with direct determina- 

 tions. In the discussion which followed the reading of the 

 paper. Prof. Ramsay said the results proved that Van der 

 Waals's generalizations were only rough approximations, and he 

 suggested that some force had been neglected or a term omitted 

 from the equations. Perhaps the assumption that the molecules 

 are incompressible was not correct. He also strongly protested 

 against the tacit assumption of Van der Waals's laws, and 

 deductions made therefrom, which had recently become so 

 common, particularly in German text-books. Prof Perry 

 inquired whether the quantities a, b, and a, had been deter- 

 mined for different substances and found to be constant. Prof. 

 Ramsay said that for substances in states analogous to those of 

 perfect gases, the quantities were approximately constant, but 

 when the liquid state was approached this was no longer 



NO. II 5 2, VOL. 45] 



true. According to Prof. Tait, the two states were not con- 

 tinuous. Prof. Herschel remarked that Prof. Tait had estab- 

 lished his law on the assumption that the co-volume is four 

 times the volume occupied by the molecules. This law, he 

 said, had been amply verified by experiments on explosions. 

 Dr. Burton, referring to Prof. Ramsay's remarks on the com- 

 pressibility of molecules, said the law of force between attract- 

 ing molecules should be accurately known before any deductions 

 were made ; and he pointed out that, at constant volume, the 

 pressure should be proportional to the absolute temperature, if 

 allowance be made for the negative pressure of attraction. Mr. 

 Blakesley, in speaking of molecular forces, said he had observed 

 that, when water is allowed to evaporate from glass, a furrow is 

 formed in the glass, which marks out the original boundary of 

 the liquid. To all appearance, the particles of glass are torn 

 away by the molecular forces acting along the boundary. 



GeologicaljSociety, Nov. 1 1.— Sir Archibald Geikie, F. R. S. , 

 President, in the chair. — The following communications were 

 rt2id : — On Dacrytherium oviintin from the Isle of Wight and 

 Quercy, by R. Lydekker. The author described a cranium and 

 mandible of Dacrythcrium Cayhixi from the Quercy phos- 

 phorites, which proved the identity of this form with the Dicho- 

 bune ovina of Owen from the Oligocene of the Isle of Wight. 

 This species should thus be known as Dacrylheriuiii ovinum. It 

 was shown tha tthe mandible referred by Fiihol to D. Cayluxi 

 belongs to another animal. — A discussion followed, in which 

 Mr. Charlesworth and Mr. E. T. Newton took part. — Supple- 

 mentary remarks on Glen Roy, by T. F. Jamieson. The 

 author discusses the conditions that preceded the formation of 

 the Glen Roy Lake, and appeals to a rain-map of Scotland in 

 support of his contention that the main snowfall in glacial times 

 would be on the western mountains. He gives reasons for 

 supposing that, previously to the formation of the lake, the 

 valleys of the Lochaber lakes were occupied by ice, and that 

 the period of the formation of the lakes was that of the decay of 

 the last ice-sheet. He supports the correctness of the mapping 

 of the terraces by the officers of the Ordnance Survey, and 

 shows how the absence of the two upper terraces in Glen Spean 

 and of the highest terrace in Glen Glaster simplifies the explana- 

 tion of the formation of the lakes by ice-barriers. The alluvium 

 of Bohuntine is considered to be the gravel and mud that fell into 

 the lake from the front of the ice when it;stood at the mouth of 

 Glen Roy during the formation of the two upper lines. During 

 the last stage of the lake, the ice in the valley of the Caledonian 

 Canal is believed to have constituted the main barrier, whilst 

 the Corry N'Eoin glacier played only a subordinate part. The 

 author suggests the possibility of a debacle during the drop of 

 water from the level of the highest to that of the middle terrace, 

 and in support of this calls attention to the breaking down of 

 the moraines of the Treig glacier at the mouth of the Rough 

 Burn. He believes that when the water dropped to the level of 

 the lowest terrace, it drained away quietly, at any rate until it 

 receded from Upper Glen Roy. In discussing Nicol's objections, 

 he maintains that notches would not be cut at the level of the 

 cols, and observes that the discrepancy between the heights of 

 the terraces and those of the cols has probably been increased by 

 the growth of peat over most of the ground about the water- 

 sheds. The horizontality of the terraces is stated to be a fact, 

 and cases are given where waterworn pebbles are found in 

 connection with the "roads," these being especially noticeable 

 in places where the south-west winds would fully exert their 

 influence, and the structure of the terraces is considered to be 

 such as would be produced at the margins of ice-dammed lakes. 

 Further information is supplied concerning the distribution of 

 the boulders of Glen Spean syenite. These are found on the 

 north side of the Spean Valley, at the height of 2000 feet above 

 the sea and 1400 feet above the river, and fragments of the 

 syenite have been carried towards the north-east, north, and 

 north-west. In an appendix, the author discusses Prof. Prest- 

 wich's remarks on the deltas, and his theory of the formation of 

 the terraces. After some remarks from Prof. Bonney and Mr. 

 Marr, the President said he agreed that no explanation that had 

 yet been proposed for the parallel roads of Lochaber was free 

 from difficulties. Yet he had long felt that these were far fewer 

 and less formidable in the glacier theory than in any other. 

 Had the terraces been marine, there ought surely to be similar 

 terraces in some at least of the hundreds of sheltered glens in 

 the Scottish Highlands, where the conditions for their formation ' 

 and preservation were at least as favourable as in Glen Roy and 

 its adjacent valleys. And though the absence of marine shells 



