January 30, 1902} 



NA TURE 



305 



OUR ASTRONOMICAL COLUMN. 

 Astronomical Occurrences in February. 

 Feb. I. igh. lotn. to 2oh. 29m. Moon occults j8- Scorpii 

 (mag. S"2). 

 I. igh. lom. to 2oh. 28m. Moon occults /3' Scorpii 



(mag. 3-0). 

 3. oh. Mercury at greatest elongation (18 17' E.). 

 6. I7h. Jupiter in conjunction with moon. Jupiter 

 5° 26' S. 



8. I7h. Venus in conjunction with moon. Venus 



3° 9' N. 



9. 9h. Mercury in conjunction with moon. Mercury 



2- 23' S. 

 12. 7h. 40m. to 8h. l8m. Moon occults e Piscium 



(mag. 4-5). 

 14. 5h. 53m. to 6h. 45m. Moon occults a Arietis 



(mag. 5-5). 

 14. iih. Venus in inferior conjunction with the sun. 



16. 6h. 24m. to 7h. i8m. Moon occults i Tauri 



(mag. 5-1). 



17. iih. 41m. Minimum of Algol ((8 Persei). 

 20. 8h. 30m. Minimum of Algol (S Persei). 



\'ARIABILnY OF THE SATELLITES OF SaTURN. — In the 

 Bulletin ilc la Soa'J/t' Aslronomiijite de /><z«i-f (January 1902), 

 M. L. Rudaux directs attention to the probable variation in the 

 brightness of the satellites Titan and Japetus, from evidence 

 furnished by observations since 1892. He concludes that in the 

 case of Titan the change is about half a magnitude, from 80 to 

 8"5, and, moreover, the variation appears to occur regularly at 

 the same parts of the satellite's orbit. Maxima occur near and 

 just after west elongation, and minima when the satellite is 

 between superior conjunction and east elongation. A suggested 

 explanation supposes the existence of fixed regions of different 

 brightness and the rotation period equal to the time of revolu- 

 tion, as in the case of our own moon. There is a rapid passage 

 from maximum to minimum. 



The satellite Japetus is also thought to have equal periods for 

 rotation and revolution, and the observation of Cassini is con- 

 firmed that the body almost becomes invisible in the eastern 

 portion of its orbit. The variation is from the ninth to twelfth 

 magnitude. 



Prof. T. J. J. See refers to the variable visibility of Japetus 

 in an article giving measures of the diameters of the satellites of 

 Jupiter and Saturn in Astronoinische Nachrichlen (Bd. 157, No. 

 3764). He says, "The disc of Titan is rather obscure, but that 

 of Japetus is even more so ; in fact, only one side gives sufficient 

 light to enable the observer to recognise a disc. This is visible 

 when the satellite precedes the planet." 



Magnetic Observations during Total Solar Eclipse, 

 May 18, 1901. — In a pamphlet reprinted from the Overgedrukt 

 Itit het Naluurlcundig Tijdschrift voor Ned- Indie (vol. Ixi., 

 part iii. pp. 173-193), Dr. W. van Bemmelen presents the 

 observations undertaken at Batavia and Karang Sago (Sumatra) 

 during the last total eclipse of the sun on May 18, 1901. The 

 determinations were made on behalf of the Batavia Observatory 

 in response to the appeal of Dr. L. A. Bauer for accurate 

 measures. The observations were made both visually and by 

 means of self-recording photographic apparatus, reproductions 

 from these latter being given. Although complicated by the 

 presence of various common disturbances, the curves show 

 decided irregularities about the time of eclipse which are thought 

 to be due to the occurrence of that phenomenon. The observa- 

 tions at Batavia were only rendered possible by the courtesy of 

 the Electric Car Company in stopping traffic over their lines 

 from 11.30 a.m. to 2.30 p.m. on the day of eclipse. 



Simultaneous Visibility of Sun and Total Lunar 

 Eclipse.— In the Transactions of the Vienna Academy of 

 Sciences (Section of Mathematics and Natural Sciences) 

 No. xxiv. pp. 263-271, Herr Dr. C. Hillebrand directs atten- 

 tion to the possible observation of both sun and total lunar 

 eclipse at such times when the phenomenon occurs as the moon 

 is rising or setting. As the refraction at the horizon is greater 

 than the diameter of the lunar or solar disc, the sun will be 

 visible after true sunset or before true sunrise ; the conditions 

 for the eclipsed moon to be thus seen may be fulfilled at 

 certain localities during the lunar eclipses of April 22 and 

 October 16 of the present year. 



NO. 1683. VOL. 65] 



THE 



VALIDITY OF THE 

 THEOR Y} 



ION IS A TION 



Introduction. — The theory of electrolytic dissociation as 

 advanced by Arrhenius in 1887 is based primarily upon the 

 facts that the molecular conductivity of solutions increases with 

 the dilution, that substances which, when dissolved, conduct 

 electricity also have abnormally low molecular weights in such 

 solutions when tested by osmotic or freezing- or boiling-point 

 methods, and that the so-called degree of dissociation may be 

 calculated from the electrical conductivity or the results of the 

 molecular weight determinations. In his original article, 

 Arrhenius states that the phenomena of electrolysis, when 

 viewed from the standpoint of thermodynamics, require the 

 assumption of free ions, as was pointed out by Clausius, and 

 that the well-known additive properties of solutions support 

 this hypothesis. Arrhenius sought to save van 't Hoff's theory 

 of solutions from having but a limited application, and to extend 

 it. 



Van 't HofT found it necessary to introduce the factor 2 in the 

 case of electrolytic aqueous solutions in order to make them 

 conform to the gas equation. This factor Arrhenius calculated 

 from the electrical conductivity on the one hand and from 

 molecular weight determinations on the other, the resulting 

 figures showing an agreement to within S 'o '5 P^'' cent. The 

 agreement was not good, and the results were obtained 

 exclusively from aqueous solutions. The non-aqueous solutions 

 then known were practically non-conductors, and appeared to 

 be fairly normal as regards van 't Hoff's theory, so that non- 

 aqueous solutions in general came to be regarded as having 

 normal molecular weights and as being non-conductors. 



Beha~>ionr of Non-aqueous Electrolytic Solutions. — Previous 

 work has already shown that non-aqueous electrolytic 

 solutions are frequently abnormal in the light of the ionisation 

 theory. Thus in many cases the molecular conductivity 

 decreases with increased dilution, e.g. Nal and NaBr in benzo- 

 nitrile, AgNOo in piperidine, FeCf, in pyridine and in benzalde- 

 hyde, and Col., in POCf,. In other cases the molecular 

 conductivity at first increases and then decreases with dilution, 

 e.g. FeCf, in paraldehyde, CBr.,COOn in POCf,. Many 

 solutions which, according to molecular weight determinations, 

 are undissocialed, conduct well. Thus AgNO^, has a normal 

 molecular weight in pyridine and benzonitrile, yet it conducts 

 fairly well. According to Dutoit and Friderich, Cdl.,, LiCl, 

 Nal, HgCU and NH,CNS have normal molecular weights in 

 acetone, and yet these solutions are conductors. Walden has 

 found that KI, Nal, Rbl, NHjI and KCNS conduct well in liquid 

 SO.j, and yet have abnormally large molecular weights in this 

 solvent. Franklin and Kraus have found that while NH4NO3, 

 NaNO;, and KI dissolved in liquid ammonia are excellent 

 conductors, the boiling points of the solutions are not nearly so 

 high as they ought to be according to the ionisation theory. 

 Nicolo Castoro found by means of the freezing-point method 

 that AgNOo, CdCl.,, HgCl.j and ZnCl., have normal molecular 

 weights in urethane ; yet the author has found that the first 

 three of these solutions are conductors. Recently, Innes found 

 the molecular weights of succinic, salicylic and tartaric acids 

 to be normal in pyridine according to the boiling-point method ; 

 preliminary tests by the author have shown that all three of 

 these solutions are fairly good conductors. 



In the case of non-aqueous solutions the various methods of 

 observing ionisation do not always give the same indication 

 with increasing dilution ; it is sometimes in one direction and 

 sometimes in the other. Occasionally simple substances in 

 solution show abnormally low molecular weights, and yet are 

 non-conductors. The author has found this to be so in the case 

 of solutions of diphenylamine in methyl cyanide. 



The abnormal behaviour, according to the theory, of non- 

 aqueous solutions led the author to investigate aqueous solutions 

 somewhat further. 



Experimental Part. — The investigation consisted of four 

 parts : — (I) Determination of boiling pointsof aqueous solutions 

 of typical, common, chemical compounds from low to very high 

 concentrations, to see how the molecular weight changes with 

 the concentration. (2) Measurement of the conductivity of these 

 solutions at or near their boiling points. These two parts of the 

 work were carried out by Mr. A. A. Koch. (3) Measurement 



J Abstract of a paper by Mr. Louis Kahlenbergin the/tr, 

 Clicmistyy (vol. v. PP.J339-392, June, igoi). 



at 0/ Physical 



