350 



NATURE 



[August 13, 1908 



lucent diatoms at high magnifications. The necessary high 

 magnification requires the use of rapid plates, and with 

 these the image of each dot or pearl is seen to be 

 surrounded by a halo of fog, due to optical causes. To 

 remove this the author makes a copy of the negative 

 through a positive, using at one or both stages of the 

 process a plate of much lower speed than could be used 

 for the original negative. By this means the contrast is 

 increased, and the foggy appearance thus eliminated. 



In a pamphlet entitled " Ratio Coordinates and Carnot's 

 Theorem " (London : Whittal<er and Co., 1908, price \s. 

 net), "J. L. S. H.," of East London College, suggests 

 a method of dealing with certain theorems in analytical 

 projective geometry of quadric curves associated with 

 triangles. His " ratio coordinates " of a point practically 

 represent the ratios, with their signs changed, of the 

 triangular or areal coordinates, their product thus being 

 — I. Equations expressed in terms of these coordinates are 

 neither homogeneous nor symmetrical, the equations of a 

 straight line and a quadric being of the form of linear 

 and quadric equations in two variables, namely, one of the 

 coordinates and the reciprocal of another coordinate. 



We have received from the director of the Zi-ka-wei 

 Observatory (near Shanghai) a copy of a new and care- 

 fully prepared word-code for transmitting typhoon and gale 

 warnings, with particulars as to direction of motion, &c., 

 to such lighthouses on the Chinese coast as possess tele 

 graphic connection, and to some foreign places. A code 

 has been in use for many years, and has from time to 

 time been improved, but it became necessary to reduce, so 

 far as possible, the work of the telegraph companies, who 

 generously transmit the messages gratuitously. In our 

 note of May 23, 1907, it was explained that captains of 

 vessels leaving port were invited to repeat, under certain 

 restrictions as to time, &c., the warnings they had seen 

 in the harbours to lighthouses not yet electrically con- 

 nected. The plan appears to have met with success, and 

 orders to cooperate in passing on the signals in a modified 

 and simple form have been issued by several naval com- 

 manders fo their respective fleets. 



The results of an investigation of the connection between 

 band and line spectra of the same metallic elements were 

 described by Prof. W. N. Hartley. F.R.S., before the 

 Royal Dublin Society on June 16. The author concludes 

 that band spectra belong to the elementary atoms, and may 

 be considered as caused, not only by the motion of trans- 

 lation of the atoms themselves, but also of the vibratory 

 movements of the component parts within the atoms. The 

 banded flame spectra of lead, antimony, bismuth, tin, zinc, 

 cadmium, copper, silver, and gold are the spectra of the 

 atoms. The lines of these elements also are the spectra 

 of the atoms, and it is inferred therefrom that the com- 

 ponent parts of the atoms are in a state of dissociation. 

 It is concluded that the metallic elements with monatomic 

 molecules which exhibit two spectra — one of lines and the 

 other of bands — can exist in two different conditions, the 

 difference being in the larger amount of energy associated 

 with the atom which exhibits a line spectrum. 



The Journal of the Rontgen Society for July contains 

 a paper by Mr. W. Duddell, the president of the society, 

 on the measurement of the current through a Rontgen- 

 ray tube. The current was derived from a 12-inch induc- 

 tion coil, the primary of which was supplied from the 

 mains at 200 volts through a mercury turbine interrupter 

 giving about seventy-five interruptions per second. The 



NO. 2024, VOL. 78] 



secondary current was rectified by a point and cup spark 

 gap, and passed in series through the tube, an oscillograph, 

 a thermoammetcr, and a galvanometer, by means of which 

 the .shape of the current curve, the mean square current, 

 and the mean current were respectively determined. The 

 current curve was almost invariably a triangle with a very 

 short base and a height from 30 to 45 milliamperes, the 

 mean current varied between 0-5 and 1-2 milliamperes, and 

 the root mean square current between 2 and 5 milliamperes. 

 Mr. Duddell is disposed to think that there is a connec- 

 tion between the maximum value of the current and the 

 hardness " of a tube for a given mean current. 



The most recent experiments of Mr. H. N. Morse on 

 the osmotic pressures of sugar solutions, as described in 

 the American Chemical Journal for June and July, are 

 characterised by so marlced aji improvement on the excel- 

 lent results already obtained as to merit much greatei' 

 attention than the titles of the papers would suggest. 

 The earlier sources of error due to variations of tempera- 

 ture have been overcome by improved thermostatic methods, 

 those due to the change of volume resulting from the dis- 

 placement of the manometer connections have been re- 

 duced to insignificant proportions by improved mechanical 

 arrangements, and the last remaining disturbance, due to 

 the dilution of the contents of the cell during the periods 

 of closing and opening, has now been got rid of. As a 

 result, the errors of measurement have been reduced from 

 an atmosphere or more to a few hundredths of an atmo- 

 sphere only on pressures up to 24 atmospheres. In the 

 experiments on glucose, eight duplicate determinations 

 showed a deviation of o-oi atmosphere only, whilst the 

 two remaining pairs differed by 0-04 atmosphere. Observa- 

 tions of this degree of accuracy make it possible to follow, 

 not only the main course of the osmotic pressure laws, but 

 also the deviations from these laws, to which so much 

 attention is being paid at the present time. 



Prof. W. F. Osgood has given in the Annals of 

 Mathematics (2), ix., 3, a simple proof of the rule for 

 the differentiation of an integral when the independent 

 variable enters into the integrand and also into the limits. 

 The proof in question involves an application of Green's 

 theorem. It may be noticed that for teaching purposes a 

 geometrical proof is very useful, and probably sufl^cient. 



At the meeting of the Soci^t^ fran^aise de Physique on 

 July 3, an absolute torsion electrometer due to M. E. 

 Salmon was exhibited. The arrangement follows that of 

 the Kelvin absolute electrometer, the attracting plate, the 

 attracted disc, and the guard ring being, however, vertical 

 instead of horizontal. The attracted disc is supported by 

 a fibre in such a way that when no electrical force acts on 

 the disc its front surface is coplanar with that of the 

 guard ring. When the potential to be determined is 

 applied to the attracting plate, the disc moves outwards 

 from its normal position, and is brought back by twisting 

 the head of the suspending fibre through the requisite 

 angle. The torsional constant of the fibre is determined 

 by means of a thread attached to the centre of the disc, 

 which passes over a pulley and supports a weight. The 

 instrument has been used to measure potentials between 

 005 volt and 40,000 volts, and has an accuracy of about 

 I per cent. 



Messrs. T. C. and E. C. Jack will publish shortly a 

 popular work entitled "The Wild Beasts of the World," 

 edited by Mr. Frank Finn. The work will be illustrated 

 with 100 reproductions in full colours from drawings, and 



