April 25, 1895J 



NA rURE 



617 



A- 



the relative degrees of brightness at these times are i, \, and \. 

 and assuming that the eclipses are central, it is easily shown 

 that the observed magnitudes may be explained by supposing 

 that the two components are of equal si/,e, while one is twice as 

 bright as the other. The unequal duration of the minima 

 further indicates that the orbit is an ellipse with an eccentricity 

 of o 2475, ^f"! '' ''* calculated that the semi-axis major of the 

 orbit IS six times the diameter of the stars. The plane of the 

 orbit passes through the sun, and the line of apsides is inclined 

 at an angle of 4° to the line of sight. The stars revolve in this 

 orbit in a period of 3 days 23 hours 48 minutes 30 seconds. 



It seems probable that this variable may form a connecting 

 link between Algol, which consists of a bright and adaik body, 

 and Y Cygni, consisting of two stars of e^jual brightness. 



The Diameter of Nrpiune. — With the Lick telescope 

 and an eyepiece magnifying 1000 diameters. Prof. Barnard finds 

 the mean angular diameter of Nepiune, when reduced to the 

 mean distance from sun 300551, to be 2"'433. This corre- 

 sponds to an actual diameter ol 32 900 miles, which is from two 

 to four thousand miles less than that stated in most of our text- 

 books. — Astronomical yournal^ No. 342. 



INDUCED MAGNETISM IN VOLCANIC ROCKS. 



N interesting note by G. Fulgheraiter, on the magnetism 

 induced in volcanic rocks by the earth's magnetic field, 

 appears in ihe A/li delta A'eale Accatiemia dei Lincei (vol. iv. 

 part 5, March 3, 1S95). The author has performed ? number of 

 experiments on volcanic rocks, in order to determine the amount 

 of induced magnetism left when, after heating to such a tem- 

 perature that they entirely lo-e their permanent magnetism, 

 they are either allowed to cool slowly or are suddenly cooled, 

 in each case under the influence of the earth's field. From such 

 observations he hopes to be able to deduce some conclusions as 

 to the conditions under which the rocks experimented on, which 

 were originally permanently magnetised, became magnetised. 

 The rocks are cut into small pirallclopepedons weighing about 

 50 grams, and such that the length is about two or three times 

 the depth or breadth, care being always taken to cut the 

 rock so that the axes of these pieces were vertical when 

 the rock was in its place in the earth. The intensity of magneti- 

 sation was in every case measured by the method of deflection ; 

 a freely suspended magnetic needle being deflected by the 

 sample, which was placed with its length east and west. After 

 measuring the intensity of magnetisation of the sample , they 

 were heated to redness, and then either allowed to cool slowly, 

 or are rapidly quenched with their axes veri ical. Their magnetic 

 moment was determined, first immediately they were cool, and 

 then after they had stood under the influence of the earth's field 

 for three months. The specimens of basalt examined may be 

 divided into two groups : in the first may be placed those 

 specimens which were originally only slightly magnetised, and 

 in this case, after heating to reiness, the magnetisation is always 

 increased, but to a very different degree in the different speci- 

 mens. The second group includes those basalts which were 

 originally strongly magnetised, and in this case after heating 

 the magnetisation was considerably reduced. In both groups 

 the magnetisation underwent no change duiing thtee months, 

 and sudden cooling gave the same results as slow cooling. 

 Experiments have also been made on tuff and peperino. The 

 results obtained with the first of these rocks are similar to those 

 obtained with the first group of basalts. Peperino, however, 

 differs in that, before being heated, lis coercive force seems 

 ! Imost nil, the bar becoming only temporarily magnetised. 

 After healing, the character of the rock is altered, as 

 it can now become permanently magnetised and behaves just 

 like the tuff. From this the auihor concludes that peperino 

 has been formed at a low temperature, probably by the action 

 of water on cinders, Ac. 



THE FREEZING-POINT OF DILUTE 

 SOLUTIONS. 



(^ORRKCT determinations of the freezing-point of dilute 

 ^^ solutions are of fundamental importance in connection 

 with the general theory of the subject, and it is therefore any- 

 thing but satisfactory to find that, in spite of the closeness with 

 whicn the individual results of the same observer agree amongst 

 thennelves, the results of different observers are in many cases 



NO. 1330. VOL. 51] 



widely separated. For example, the following values have 

 been given as the molecular depression of the Ireezing-point 

 in ihe case of a i per cent, aqueous solution of sugar: — 

 202, .\irhenius ; 207, Raoult ; 201, Pickering; 218, W. C. 

 Jones; i Si, Loomis. The results of Jones and Loomis, both 

 of whom claim increased accuracy for the methods they employ, 

 differ by some 18 per cent, fhe theoretical value of the 

 molecular depression, calculateil from the meliingpoint and 

 heat of fusion of ice, is I 86. The cause of these differences 

 has given rise to much discussion. Pickering has attempted to 

 show that Jones's results, wherein the temperature was read to 

 the ten-thousandth of a degree, were affecied tiy thermometer 

 errors. Jones has replied that his thermometer was tested. 

 Kohlrausch has drawn attention to probable sources of error in 

 Jones's method, but is compelled to admit that the differences 

 between the results of Jones and Loomis must, in ih; main, be 

 due to some unkrown source of error. 



A definite step in the direction of clearing up this point is 

 made in a recent number of the Zeit chrift fur phyiikalisclie 

 Chemic. Here Nernst and .\begg emphasise the lact that the 

 observed freezing-point must in general be different from the 

 true freezing-point, or the temperature at which solid and liquid 

 are in equilibiium. They point out that a partly-frozen liquid, 

 uninfluenced by the temperature of its surroundings, will strive 

 to reach the true freezing-point at a rate which, at any instant, 

 may be taken as proportional to the difference between its actual 

 temperature and the true freezing-point. Again, in practice, 

 on account of the limited amount ot substance employed, and 

 the effect of the temperature of the surroundings, &c. , unfrozen 

 liquid strives to reach a definite tempfrature, which may be 

 termed the "convergence temperature." On these assumptions 

 it is easy to show that the observed freezing-point, or the tem- 

 perature at which the thermometer becomes steady, will only be 

 the true Ireezing-point if the "convergence temperature " is 

 equal to the true freezing-point, or if R, the rate at which the 

 temperature of the partly frozen liquid approaches the freezing- 

 point, is infinitely great as compared with r, the rate at which 

 the temperature of the unfrozen liquid approaches the " con- 

 vergence temperature.' If one ol these conditions is not ful- 

 filled, corrections delerinmed experimentally have to be applied. 

 For dilute solutions of alcohol and common salt the corrections 

 were found to be inappreciable under the experimental con- 

 ditions described in the paper. Here the value obtained for 

 R, although, as is always ihe case, it was largely diminished 

 by the lag of the thermometer, still was sufficiently large 

 as compared with the value of r. in the case of sugar, 

 however, R was so small that by varying the experimental 

 conditions, a i per cent, solution gave molecular depressions 

 varying between 16 and 21 — limits which are even further 

 apart than those given by the results of previous observers. On 

 correcting the observed depressions in the manner described, 

 they all gave practically the theoretical value. 



X^ithout these corrections, observed freezing-points are thus 

 held to be functions of the size of the apparatus used, the 

 temperature of the cooling-bath, the rate of stirring which 

 largely affects the " temperature of convergence,'' &c. 



Kvidence is also given of the futility of expressing freezing- 

 points to the ten-thuusandth of a degree. It may readily happen 

 that the above correction is as high as 001, and as the mode of 

 deducing it is but approximate, in such a case 0°00I or o°oo2 

 would be a favourable estimate of the error of the end result, 

 even if satisfactory corrections could be applied for the alteration 

 in the concentration of the solution produced by freezing, and 

 the ordinary sources of error incidental to the method of ex- 

 periment. J. W. Rodger. 



THE EXAMINATION CURVE. 



T F the resulisof the examination of a mixed body of candidates 

 be plotted out on the graphic method, ihey will be found, 

 in accordance with a well-known law of statistics, to approxi- 

 mate to a curve having a more or less rapid gradient at either 

 end, and a mid-region of gentler ascent. Fig. I, for example, 

 shows the results of an examination of 27 students in physical 

 geography, the scale of marks running vertically fiom 10 to 90, 

 the examinees being arranged horizontally at equal distances 

 apart from the lowest to the highest. The larger the number 

 of candidates the more flattened ilocs the mid-region of the 

 curve tend to become, .\gain, in any series of examinations, 

 the mean results of which are plotted out, the more uniform 



