14^ 



NA TURE 



[July 28, 1923 



extended along the strike of the faults. Most of 

 the great earthquakes originating along the shores 

 of the islands have been accompanied by sea-waves, 

 each of which, so far as known, has been propagated 

 with the trough in advance of the crest. The wave 

 thus seems to indicate a sudden downward displace- 

 ment of the ocean -l)ed. Disastrous earthtjuakes 

 seldom recur in exactly the same place except after 

 long intervals. Thus, those parts of the zones of 

 active faulting near which severe earthquakes have 

 not occurred in historic times are to be regarded as 

 seismically the most dangerous. 

 VoLUMETKic Determination of Rainfall. — A 



Eaper on this subject by Mr. C. S. Salter was read 

 efore the Inland Navigation section of the thirteenth 

 International Congress of Navigation held recently in 

 London, and is published as a pamphlet. The sources 

 of error in rainfall records are three : design of rain- 

 gauge, exposure of rain-gauge, and interpretation of 

 records in terms of volume. Mr. Salter's paper deals 

 with the last consideration. Owing to the fact that 

 rainfall is extremely variable in its incidence in time 

 and its distribution in space, the reading of an indi- 

 vidual rain-gauge must be regarded as merely a 

 sample. The total rainfall of an individual month in 

 Britain may vary by 400 per cent, from the average 

 value, and that of an individual year by 70 per cent. 

 When the period of records is short no allowance for 

 the variation of time is possible, but a correction can 

 often be applied from adjacent long records. Gener- 

 ally speaking, in rainy districts, where thunder- 

 storms and sporadic rains do not bulk largely in the 

 total fall, from one to two years give a sufficiently 

 good basis for a factor of correction to be applied with 

 safety, provided that a long record is available at no 

 greater distance than five to ten miles. In districts 

 where the total fall is so small that a single thunder- 

 storm may introduce great local variations, from four 

 to five years' records are necessary. Variations in 

 space are relatively easily applied with the help of an 

 orographical map. A rainfall map is the best medium 

 for computing the volume of rainfall over a gathering- 

 ground as a whole, and the best and simplest method 

 is by planimeter measurement. 



The Winds of Hongkong. — A discussion under 

 the direction of Mr. T. F. Claxton " to ascertain 

 the difference in direction and velocity of the wind 

 at the Royal Observatory, Kowloon, and at Victoria 

 Peak, Hongkong, at different seasons of the year 

 and at different hours of the day," has been issued 

 by the Royal Observatory, Hongkong. ,The results 

 are based on the records of Beckley anemographs 

 for the period 1914-1918. The Royal Observatory 

 is situated on a hillock, 100 feet high, about 1000 

 yards from the harbour ; the surrounding country 

 is fiat, except to the north. Victoria Peak is 1840 

 feet above sea-level and is situated 3 miles to the 

 south-west of the Royal Observatory. To the north, 

 west, and south, the sides of the Peak are very steep, 

 and the easterly winds are affected by the Hongkong 

 hills. The anemograph records at the Observatory 

 are measured at the half hours, and the value set 

 against any hour is the run of the wind from 30 

 minutes before to 30 minutes after that hour. At 

 Victoria Peak the records are measured at the hour, 

 and the value set against any hour is the run of the 

 wind since the previous hour. This difference in 

 the method of registration seems likely to affect 

 the results for comparison. Detailed hourly ob- 

 servations of direction and velocity are given for 

 the two exposures for the years 1914-1918. The 

 different situations naturally give different results, 

 which are shown by numerous tables and diagrams. 

 For normal wind results, for comparison with other 



NO. 2804, VOL. I 12] 



where 0, 



world observatories, the results at Victoria Peak 

 should probably be preferred, althougli \nA\\ situa- 

 tions seem to leave much to be desired. 



Ionic Dissociation in Solution. — P. Debye and 

 E. Hueckel have investigated the electrostatic forces 

 between the ions of the solute, and the dipole action 

 of the molecules of the solvent {Phys. Zeits., May i). 

 They assume that the whole of the dissolved salt is 

 dissociated ; and for dilute solutions arrive at the 

 equation 



'"-'"'oDATN DAT''-"" (') 



the deviation from the classical theorv, 

 )At, At being the lowering of the frei / 

 point given by the classical theory and A, that acti: 1 ,y 

 observed ; w is a valency factor etjual to (2:»'<z,*/i;i'<)^. 

 The dissolved molecule is split up into vy -v^ - *, ions 

 of different kinds i -i -s, with valencies z, -z< - z„ 

 and w is calculated from these values ; '1=273 when 

 the solvent is water, 6 = 4-77x10-" e.s.u., n is 

 Loschmidt's number = 6-o6 x lo**, A = 1*346 x 10" ergs, 

 D, the dielectric constant, is 88-23 for water at 0° C, 

 2i'( = i/ is the number of ions into which a molecule 

 of the salt splits up. Using these values, 



d = o-z']ow ^Jv-f \ (2) 



and curves have been drawji^showing the experimental 

 relation between e and Jv-^ for a number of salts of 

 varying constitution, including magnesium sulphate, 

 lanthanum nitrate, potassium sulphate, and potassium 

 chloride. These curves follow the straight lines 

 obtained by giving w in (2) the proper values, for a 

 considerable distance from the origin. For higher 

 concentrations the deviations from the straight lines 

 depend on the individual properties of the ions ; and 

 particularly on their dimensions, which were neglected 

 in deriving (i) and (2). When the dimensions are 

 taken into account, theory is found to agree very 

 satisfactorily with experiment up to much higher 

 concentrations. For very high concentrations other 

 factors, previously neglected, have to be considered ; 

 there appears to be no doubt that, even in this case, 

 the molecules of the solute are split into their ions. 



Stereoscopic Projection. — Much attention has 

 been directed in recent years towards obtaining a 

 satisfactory method of stereoscopic projection. Many 

 of the devices proposed involve the use by the in- 

 dividual observer of spectacles or binoculars with 

 coloured glasses or interrupting shutters. The 

 Daponte Stereoscopic Projector or " Pulsograph," 

 which was exhibited by Mr. E. Sanger-Shepherd at 

 the Royal Society Conversazione on June 20, employs 

 an entirely different principle, whereby a " stereo- 

 scopic " effect can be readily observed by the unaided 

 eye of the spectator. Two photographs are taken 

 from two positions slightly separated, and projected 

 in register on an ordinary screen by two optical 

 systems. Between the source of light and the trans- 

 parency in each of the optical systems a rotating 

 shutter is placed, consisting of a glass disc with a 

 graduated grey film varying from black at zero to 

 clear at 180° and back to black at 360°. When one 

 shutter is passing maximum light the other is at 

 minimum transmission position, the rotating shutters 

 dissolving the right-hand picture into the left-hand 

 picture and vice versa. With the discs at the posi- 

 tion of equal transmission, that is at the 90 '^ position, 

 a double-image picture appears, since the two stereo- 

 scopic photographs are not exactly alike ; but on 

 the discs being rotated the " stereoscopic " relief 

 effect is immediately obtained. The " Pulsograph " 

 can be employed for the projection of lantern slides, 

 solid objects, or of kinematograph films. 



