May 3, 1894] 



NATURE 



at Munich, in 1891, the wish was expressed for observations 

 ^especially telegraphic reports) from Bulgaria, and the publica- 

 tion of these results by Prof. Watsoff, director of the Central 

 Meteorological Station at Sophia, is a. good step in the right 

 direction. 



In a letter to the local press, Mr. H. C. Russell, F.R.S., the 

 Director of the Sydney Observatory, recently brought to- 

 gether some facts concerning icebergs and their relation 

 to weather. The letter was written with the view of show- 

 ing that the icebergs seen near Australia do not influence the 

 weather in the manner commonly supposed. All the reports of 

 icebergs seen within the last two years or so, were collected, and 

 it was found that they numbered eighty-four. In order to locate 

 these icebergs, the position of each ship when ice was seen, 

 was plotted on a map of the world. Of the eighty-four positions 

 thus marked on the map, sixty one, or 73 per cent., were under 

 the lee of Patagonia, as if they had been carried there by the 

 strong winds and the Cape Horn current, which sets northwards 

 after passing Cape Horn. Round the Cape of Good Hope was 

 another, though less numerous, group of positions, n"mbering 

 thirteen, and making 90 per cent, of the whole collected in the 

 two localities. About the Crozets two ships reported ice, and two 

 ships reported seeing ice on the same day and in the same longi- 

 tude at a point 500 miles south-west from Albany. Six other 

 locations were between New Zealand and Cape Horn. Of the 

 eighty-four ships reporting, only two saw ice anywhere near 

 Australia, and that was more than two years ago, on December 

 f4, 1891. As to the effect of icebergs in lowering the tempera- 

 ture, Mr. Russell points out that many observers have reported 

 that the thermometer is not a good indicator of the proximity 

 of ice, no matter whether it is placed in air or in the sea ; and 

 this is not surprising if it be remembered that fresh supplies of 

 ocean water and wind are always flowing past the icebergs, and 

 •distributing their cold over such wide areas that it become? in- 

 appreciable. Even under the lee of Patagonia, where, in a 

 sense, the ocean is full of icebergs for an area measured by 

 hundreds of miles each way, it does not appear that any very 

 remarkable effect on the temperature is produced, and near 

 Australia, where now and then a few scattered icebergs are seen 

 •in the distant offing, no appreciable effect can result from their 

 ■presence. 



At a recent meeting of the Academy of Sciences of St. 

 Petersburg, Dr. S. Wild read an interesting paper on some 

 improvements in the design and construction of magnetic in- 

 struments. The paper is divided into four parts, the first con- 

 taining the results of some experiments which show that 

 metallic wires are preferable to the silk threads usually employed 

 for the suspension of the magnets in magnetographs. The 

 second part deals with the question of the accuracy of the read- 

 ings given by the instruments employed to record the variations 

 •in the magnetic elements, when these instruments are contained 

 in buildings composed of a material (such .as ordinary red brick) 

 ■containing iron. The experiments on this subject, commenced 

 •in 1878, have been continued, and show very clearly that a 

 brick building has no influence on the measurements of the 

 strength of the field, while the influence on the dip is quite 

 negligible. The above result has also been deduced theoreti- 

 cally by M. Chwolson by means of the known value of the 

 susceptibility of the bricks. The third part of the paper con- 

 tains a description of a small magnetometer for determining the 

 value of the horizontal component of the earth's magnetism, 

 designed for the use of travellers in countries where it would be 

 ■difficult to carry the ordinary instruments. The instrument fits 

 on Hildebrandt's small universal theodolite, which instrument 

 is fitted with a compass. The combined instrument, although 

 of small weight, is capable of giving results of considerable 

 NO. 1279, VOL. 50] 



precision. Thus the dip can be determined to within half a 

 minute of arc, and the horizontal component to within one part 

 in four thousand. Lastly, the paper contains a description of 

 some changes which Dr. Wild has introduced into the portable 

 magnetometer. 



In the April number of the yournal de Physique, M. H. 

 Abraham gives an account of a method for measuring self and 

 mutual induction, which he says is capable of giving the quantity 

 which is being measured to within one part in a thousand. The 

 method employed consists essentially of balancing a steady 

 current passing through one pair of coils of a differential galvano- 

 meter against a succession of induced currents sent, by means of 

 a commutator rotating at a known speed, through the other coils. 

 The commutator then being stopped, the secondary circuit, con- 

 taining one pair of coils of the galvanometer, is connected with 

 the extremities of a resistance /' in the primary circuit. The 

 value of r is altered till the galvanometer is undeflected, then, 

 from a knowledge of the value of r and the speed of rotation 

 of the commutator, the value of the mutual induction can be 

 calculated. In many ways the method resembles that employed 

 by Profs. Ayrton and Perry in their secohm-meter. 



The line spectrum of oxygen has recently found another care- 

 ful investigator in Max Eisig, who gives an account of his work 

 in Wiedemann's Annalen. The wave-lengths photographed 

 and measured ranged from 24336 to 47104. The oxygen was 

 generated by the electrolysis of water acidulated with 

 pure phosphoric acid. The form of vacuum tube adopted 

 was that consisting of two vessels joined by a thin tube 

 at right angles to their length, which allowed the gas 

 to be seen through the length of the tube. The end was 

 closed by a plate of quartz to permit the passage of the ultra- 

 violet rays. The tube was cleaned by flushing it with oxygen 

 for several weeks. Of the four difterent oxygen spectra de- 

 scribed by Schuster, only the elementary line spectrum was 

 studied ; 93 lines were recorded and measured. A comparison 

 of these lines with a Rowland solar spectrum led to the con- 

 clusion that no identity between these oxygen lines and lines in 

 the solar spectrum can be established. 



Prof. Julujs Tiiomsen, universally known from his re- 

 searches in thermo-chemistry, has just finished a series of 

 observations on the ratio of the atomic weights of hydrogen and 

 oxygen (Zeit.ftirphys. C/icm. xiii. 398). In order to avoid the 

 errors incidental to the direct methods of estimating the ratio, 

 he employs an indirect method, which consists in estimating the 

 ratio of the molecular weights of hydrogen chloride and am- 

 monia. Pure hydrogen chloride is passed into a flask containing 

 water, and the increase in weight is determined. Pure dry 

 ammonia is then introduced into the solution until the hydro 

 chloric acid is neutralised, and the weight again determined. 

 After introducing all corrections, a large number of observa- 

 tions, in which the initial amount of hydrogen chloride varied 

 between 4 and 20 gr. indicate that the value for the ratio, 

 HCI/NH., is 2'I3934 ± 000009. This number leads to the 

 result that I : 16 represents the value of the ratio II : O with an 

 accuracy as great as is warranted by the numbers in use for the 

 atomic weights of chlorine and nitrogen, and which have of 

 course to be employed in the calculation. Dumas and Erdmann 

 and Marchand, from the synthesis of water from copper oxide 

 and hydrogen, found respectively i 0025 : 16 and I 0017 : 16 as 

 values of the ratio, and more recently Scott, from the volume 

 composition of water, gives the higher value 100S7 : 16. 



In a recent number of the Cenlraililall fiir Allgem. Pathologit 

 some exceedingly interesting experiments are recorded on 

 immunity from the attacks of cholera germs artificially induced 

 in the human subject. Drs. Sawtschenko and Sobolotny sue- 



