526 



NA TURE 



{April 3. 1890 



I 



-Zoarces, where the brain is not flattened. Tn Zoarces, also, from 

 the same cause, the limbs of the fold are less closely applied to 

 -each other and much thicker. 



.The pineal body is roundish and solid in the early larval stage 

 in the herring. It is vertically flattened in the early post-larval 

 stage. In the i-inch sta^e it is much larger and contains a 

 lumen ; it shows signs of constriction into proxinnal and distal 

 elements, and the lumen contains a coagulable albuminous fluid, 

 as in P.ti'omyzon. In the ifV-inch stage the constriction is still 

 visible, and the walls are generally crenated. The tissues of the 

 pineal wall are now divided into three layers, and a/e of varying 

 thickness. The cartilage of the te<junien cranii overlies the body 

 at this stage. The constriction of the body appears to be an 

 exaggeration of the crenation of the pineal wall met with in 

 Sa'mo ; it has not, probably, the morphological value of the 

 constriction of the body in Pdroviyzon. 



March 27. — "On the Stability of a Rotating Spheroid of 

 Perfect Liquid." By G. H. liryan. Communicated by Prof. 

 G. H. Parwin. 



.The investigations of Riemann, Basset, and ■ others have 

 proved that Maclaurin's spheroid, when composed of ftictionless 

 liquid, ceases to be stable for an " ellipsoidal " type- of .disturb- 

 ance when its eccentricity attains the value o'9528867. The 

 ■object of the present paper is to discuss the cmditions of 

 stability with reference to disturbances of a general type 

 expressible in terms of spheroidal harmonics, with the view of 

 examining whether Riemann's condition is sufficient to ensure 

 stability for displacements other than ellipsoidal. 



Taking the criteria of stability determined in a previous 

 communication (Phil. Trans., A., 1889), the author shows 

 by numerical calculation that the form which is critical for 

 an ellipsoidal disturbance is stable for disturbances determined 

 by several of the lower harmonics. These results are then 

 extended by a perfectly general investigation to all other types 

 of displacement. 



The c )nclusion is that Riemann's and Basset's condition 

 of stability is sufficient to ensure the absolute stability of 

 Maclaurin's rotating spheroid for eviry possil)le displacement. 

 Also that, unless the liquid is subject to hypothetical constraints, 

 we cannot initially obtain any form other tha'i ellipsoidal from 

 the instability of the spheroidal form. In the case considered of 

 perfect liquid this ellipsoid does not rotate as if rigid, but its 

 principal axes rotate with half the an.jular velocity of the liquid. 



Physical Society, March 7. —Prof. W. E. Ayrton, F.R.S., 

 President, in the chair. — Dr. S. P. Thompson described Ber- 

 trand's refractometer, and exhibited the capabilities of the 

 instrument before the Society. Its action depends on total 

 reflection. The refractometer consists of a hemisphere of glass, 

 about 8 mm. diameter, set at the end of a tube, the plane face 

 being outwards and inclined at about 30° with the axis. One 

 side of the convex surface of the hemisphere is illuminated 

 through a piece of ground glass set about perpendicular to the 

 plane face. The hemisphere is viewed through an eye-piece 

 focussed on a scale divided to tenths of millimetres placed within 

 the tube. The instrument is particularly useful for minera- 

 logical specimens and liquids. The procedure in the latter case 

 is to smear a film of the liquid over the plane face of the hemi- 

 -sphere, and by looking through the eye-piece determine the scale 

 reading of the line which separates the light and darker portions 

 of the field. A reference to a calibration ta!)le gives the refrac- 

 tive index. In experimenting with solids a thm film of a very 

 ■dense liquid (supplied with the instrument) is placed between the 

 specimen and the glass, and the procedure is then as above. The 

 refractive index of opaque solids can be determined in this way. 

 In using the instrument for minerals great care must be taken 

 not to scratch the glass. The handiness of the refracto- 

 meter and its perfect portability (its dimensions being about 5 

 eentimetres long by 2^ cm. diameter) are great recommenda- 

 tions. Mr. Biakesley asked to what accuracy the scale could be 

 read, and whether the sensitiveness of the instrument was at 

 all comparable with that of other methods. Prof. Dunstan 

 inquired if it could be used with volatile liquids. In reply Dr. 

 Thompson said that with non-homogeneous light the scale could 

 be read to i division, but with a sodium flame one-tenth of a 

 <iivi ion could be estimated. For volatile liquids, a drop may 

 be used instead of a film, or the evaporation of a thick film may 

 be retarded by a cover-glass. — Mr. H, Tomlinson's paper, on 

 the Villari critical point in nickel, was postponed. — Prof, 

 Dunstan described an apparatus for distilling mercury in a 



vacuum, devised by himself and W. Dymond, and showed the 

 worl-ing of the arrangement. It consists of a 3 mm. soft glass 

 tube rather more than a metre long, having an oblate spheroidal 

 bulb blown at the upper end. The bulb is placed over a ring 

 burner. At the top of the bulb, a tube of l'5 mm. diameter is ■ 

 attached, and this passes outside the bulb, and descends close to 

 the larger tuVie. The part of the smaller or fall tube just below 

 the bulb is enlarged so as to firm a condensation chamber, and 

 the lower part serves as a Sprengel tube. A conical reservoir 

 containing the mercury to be distilled is in flexible connection 

 with the lower end of the large tube as in Clark's well-known 

 apparatus. The advantages claimed for the new apparatus are, 

 its relative shortness and portability, the small quantity remain- 

 ing, undistilled, and its non-liability to damage or derangement if 

 left unsupplied with mercury. To ensure satisfactory working a 

 constant pressure of gas is necessary, and this is obtained by 

 inserting a Sugg's dry governor in the sujiply pipe. During 

 distillation, peculiar green flashes are seen within the condensa- 

 tion chamber, and these are intensified by bringing it near an 

 electric machine in action. The apparatus also serves well to 

 show the character of an electric discharge through mercury 

 vapour, for the mercury in the two tubes may be used as elec- 

 trodes. Prof. Thompson said he devised a simple form of 

 distilling apparatus some time ago which answered fairly well, 

 and could be made by any amateur glass-worker. It cons'sted 

 of a double barometer, one leg of which was of small bore, so 

 as to act as a Sprengel tube. The rising part of the bend at the 

 top. of the larger tube was expanded and served as the evaporat- 

 ing chamber, below which a burner was placed. The President 

 asked why Clark's apparatus is made so lengthy. In reply to 

 this question Mr, Boys said that as the fall tube goes down 

 within the rising one, the mercury near the top of the latter is 

 heated by the condensing mercury (thus 'economising gas) and 

 hence condensation dues not take place until the vapour has 

 passed a considerable distance down the fall tube. — Prof. S. U. 

 I'ickering read a paper on the theory of osmotic pressure 

 and its bearing on the nature of solution. The author said 

 that considerable doubt exists as to the accuracy of the premises 

 on which the theory is based, and if the theory is to be regarded 

 as true and not merely a rough working hypothesis, the lollow- 

 ing conditions must be fulfilled by weak solutions — (i) The 

 molecular depression of the freezing-point must be independent 

 of the nature of the dissolved substance. (2) Any deviations 

 from (i) must be in the direction indicated by the theory. (3) 

 the depression must be independent of the nature of solvent. 

 (4) The depression must be independent of the amount of 

 solvent (all solutions being weak). (5) The deviations with 

 strong solutions should be in the theoretical direction. (6) 

 They should be regular. Prof. Pickering proceeded to show 

 that experiment, instead of confirming the above statements, 

 disproves them all. As regards (i), without counting abnormally 

 low (half) values, Raoult's results show variations of 60, 40, 30, 

 &c., per cent, in different cases, and the author quoted other 

 values where the variations were 500, 260, 230, &c., per cent. 

 These variations, he considered, were too great to be explained 

 by the fact of the solutions used being 3 or 4 times too strong. 

 Referring to (2), he said that low values are reasonably ex- 

 plained by the polymerization of the dissolved molecules, high 

 values by their dissociation into ions. He then argued that 

 there are no abnormally high values, for the view that such 

 exist, and that they are explainable by dissociation involves the 

 following conclusions : {a) that the more stable a subistance is, 

 the more easily is it dissociated ; {b) that solution dissociates 

 molecules which we know can exist undissociated as gases ; {c) 

 that water must consist of igHjO, and the atomic theory is 

 wrong ; {d) that energy can be created, and therefore the theory 

 of its conservation is untenable. With respect to (3), it was 

 pointed out that in many instances the same dissolved substance 

 gives the full depression with one solvent and half depression 

 with another. Cases were quoted where the depression pro- 

 duced by the same dissolved body in different solvents showed 

 variations of 36,000, 21,000, and 28,000 per cent. In discussing 

 (4), the author >aid that even with solutions weaker than that 

 corresponding to a gas, the law is not fulfilled. Taking the 

 case of sulphuric acid (the only one at present fully investi- 

 gated), the variations amount to 40 per cent., or about 28 

 times the experimental error. With reference to (5), it was stated 

 that with strong solutions the molecular depression should 

 become smaller, but in every known case (9 were qu )ted) 

 it becomes larger, the increase in one instance being 3,200 per 



