,.^.6 



NA TURE 



[August i, 1901 



and eight minutes in Right Ascension. Photogravures for the 

 Lunar Atlas have been prepared for publication ; they refer 

 to the first and last quarter. On the occasion of the total eclipse 

 of May 28, MM. Hamy and Bigourdan were sent to Spain and 

 were favoured with splendid weather for their observations. 

 M. Bigourdan continues his observations of nebuK-e, and the 

 work of his great catalogue is progressing favourably. M. 

 Callandreau observed with the great equatorial of the western 

 tower, and used a wire illuminated only by points in the field 

 of his refractor ; he appears to be satisfied with this method, 

 which prevents the eye from being disturbed by too great a 

 quantity of light when observing feeble stars. 



M. Gaillot has worked at the theory of Saturn, using 

 Le Verrier's formulx, and has succeeded in showing that the 

 discrepancies between the results of computation and observation 

 should be attributed to the fact that a sufficient number of terms 

 had not been taken into consideration. 



A staff of six women observers, directed by Miss Dorothea 

 Klumpke, has determined the position of 29.627 stars for the 

 International Catalogue. This is the only department of the 

 Observatory where ladies have been admitted. To the meteoro- 

 logical department a new registering barometer has been 

 added : it is a mercury one, and the end of the index runs 

 through 3 mm. for a variation of i mm. in mercury. The 

 publication of the old observations from 1837 up to 18S6 will be 

 completed this year, and from 1886 on the observations will be 

 published regularly each year. The observations of 1S98 were 

 published in igoo, and those of 1S99 will appear shortly. 



PHOTOGR.-ii'HV BY THE LlcHT OF Venus — In the autumn 

 of last year several meagre accounts appeared in various journals 

 announcing that Dr. W. R. Brooks had succeeded in obtaining 

 good photographic records solely by means of light from the planet 

 Venus. In the Ce-K/«77,l/ajac!'w for August (1901), Dr. Brookshas 

 an article describing his experiences, illustrated by reproductions 

 of the photographs obtained at the Smith Observatory. These are 

 chiefly positives taken by placing a landscape or other negative 

 in a printing frame with a sensitive plate and exposing to the 

 light from the planet, care being taken to shield the frame from 

 all extraneous light. The results described were obtained when 

 the planet was a morning star, shortly after September 17, 1900. 

 Gelatine dry plates (speed not stated) were used, the exposures 

 given varying from thirty to forty-five minutes. A print on 

 bromide paper was obtained by exposure on five consecutive 

 clear mornings. The positives are all apparently well exposed, 

 and a portrait is also shown as being produced by the planet's 

 light, but by what procedure is not indicated. 



NkwNeeul.^. — In the Cow/to ««(/;« (cxxxiii. pp. 206-208), 

 M. Bigourdan continues his catalogue of new nebulre discovered 

 with the west equatorial of the Paris Observatory. Particulars 

 as to position, notes of special interest and comparisons with 

 other catalogues are given for twenty-three objects observed 

 between 1S84 and 1S98. 



THE CRYSTALLISATION OF SALT 

 SOLUTIONS. 



ALTHOUGH the processes of crystallisation have been known 

 to, and made use of by, chemists for ages, yet it is only 

 within the last few years that the phenomenon of crystallisation 

 from solution has been the subject of systematic investigation. 

 The pioneer work in connection with this systematic study on 

 the basis of modern principles has been done, for the most part, 

 by Dutch chemists. The researches of Roozeboom on the 

 equilibrium of systems in contact with water have shown 

 clearly the importance of the phase rule of WiUard Gibbs as a 

 guide in the study of the complex phenomena of heterogeneous 

 equilibrium. The study in van 't Hoffs laboratory of the con- 

 ditions of existence of crystallohydrates and of the phenomena 

 associated with the formation and decomposition of double salts 

 in contact with water has given us invaluable material for a 



the solution shall be saturated with regard to that substance. 

 In what follows it is presumed that the crystallisation takes place 

 so slowly that supersaturation phenomena can be neglected, and 

 the complications resulting from crystallisation of isomorphous 

 mixtures are also left out of account. Furthermore, we suppose 

 that the temperature of the solution remains constant during the 

 crystallisation. 



The simplest conditions are then met with in the case of a 

 solution containing a single substance, say a salt, which is not 

 capable of combining with water of crystallisition. If an un- 

 saturated solution of such a salt is evaporated, the commence- 

 ment of crystallisation is conditioned solely at a given temper- 

 ature by the attainment of a definite concentration. As evapor- 

 ation proceeds the salt then separates out continuously, the 

 composition of the solution undergoing no change until the last 

 trace of water has been removed. 



If the dissolved salt forms crystallohydrates, i.e. salts with 

 water of crystallisation, then the products of isothermal evapor- 

 ation are dependent upon the temperature, a less hydrated form 

 separating as the temperature is higher. Thus solutions of 

 manganese chloride yield the tetrahydrate if the temperature 

 does not exceed 58^ C., whereas at higher temperatures the' 

 dihydrate crystallises out. It is well known that salts containing 

 water of crystallisation undergo at a definite temperature a 

 change in which the whole or part of the water of crystallisation 

 is split off. Glauber's salt loses its ten molecules of water at 

 32'4° C. ; ordinary zinc sulphate containing seven molecules of 

 water yields the hexahydrate at 39° C. , and this again, at a higher 

 temperature, yields a lower hydrate. These temperatures are 

 known as the transition temperatures of the salt hydrates, and 

 have a far-reaching analogy with the melting points of solid 

 substances. 



The limiting temperatures corresponding to the crystallisation 

 of a definite hydrate from the salt solution are determined by the 

 transition temperatures of the solid hydrates. 



If supersaturation phenomena intervene we may observe 

 the separation of hydrates from solution at temperatures below 

 the normal limiting temperature. It is, however, only under 

 this condition that crystallisation of such unstable hydrates 

 takes place, for at a given temperature the unstable hydrates are 

 more soluble than the stable hydrate. If the solution from 

 which an unstable hydrate has begun to crystallise out be im- 

 pregnated with the hydrate of smaller solubility, the unstable 

 hydrate will redissolve and crystallisation of the normal hydrate 

 ensues 



If a solution contains two dissolved salts having a common 

 ion, the phenomena of crystallisation are about as simple as in 

 the case of a solution containing a single salt. Let us suppose, 

 in the first instance, that these salts do not unite to form a 

 double salt, and that they do not form crystallohydrates. Such 

 a solution is one containing the chlorides of sodium and potass- 

 ium, and in this case a knowledge of the composition of the 

 three solutions, saturated respectively with regard to each single 

 salt and with regard to both simultaneously, enables us to pre- 

 dict what will take place on isothermal evaporation. A graphic 

 representation of the solubility data facilitates the tracing of the 

 crystallisation process very considerably, and the composition 

 of the various solutions is conveniently expressed by the number 

 of molecules of dissolved salt per 1000 molecules of water. 

 Fig. I contains the data for the system consisting of w.ater, 

 potassium chloride and sodium chloride at 25' C., A repre- 

 senting the .saturated solution of sodium chloride, B that of 

 potassium chloride, and c the solution saturated with regard 

 to both. 



Along the curve .\ c we have solutions saturated with regard 

 to sodium chloride in which the potassium chloride concentration 

 gradually increases. Similarly, the points along the curve li c 

 represent solutions containing increasing quantities of sodium 

 chloride, all of which are saturated with reference to potassium 

 chloride. 



All points within the figure oacb represent unsaturated 



correct understanding of the processes of crystallisation. Not j solutions, the quantities of the dissolved salts being given by 

 only is the systematic investigation of this phenomenon of ! the lengths of the projections on the axes. If a solution 



importance to the chemist, but the geologist is also dependent 

 on such knowledge for the final explanation of the conditions of 

 formation of the vast oceanic salt deposits. 



A knowledge of the composition of the solution in equilibrium 

 with a system of solid substances is obviously an all important 

 factor for the study of the processes of crystallisation, for the 

 separation of any solid substance from the solution requires that 



NO. 1657, VOL. 64] 



corresponding«to the point c is slowly evaporated, at 25° C., 

 the change in the composition of the solution will be repre- 

 sented by the continuation of the line o .• (o corresponding to 

 pure water). At the point d, where this line meets the curve 

 B c, the solution becomes saturated with potassium chloride 

 and the latter crystallises from solution. By the cotitinued 

 separation of potassium chloride the relative proportion of 



