146 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



weight of chlorine g.as, if only the temperature be kept from 

 rising. If, now, a current of hydrogen gas, thoroughly dried, 

 be passed over the coke thus charged with chlorine, even in 

 absolute darkness, it will combine with the chlorine, and form 

 hydrochloric acid gas. Using an amount of the carbon weigh- 

 ing fifty grammes, the temperature lowered twenty degrees. 

 Here, then, was an actual combustion of hydrogen by chlo- 

 rine taking place in the dark, without heat, and actually pro- 

 ducing cold. This cold, however, is due to the absorption of 

 heat by the condensed chlorine, which resumes the gaseous 

 state. 6 i?, 1 872, 92. 



KECENT OBSERVATIONS ON SHIPS' MAGNETISM. 



In 1865, Professor William Harkness, of the United States 

 Naval Observatory, was ordered to the iron-clad 3Ionadnoc, 

 for the purpose of making observations on the action of her 

 compasses during the cruise that she was about to undertake 

 from Philadelphia to San Francisco, by the way of the Strait 

 of Magellan. The observations made by Professor Harkness 

 are of a novel character in the history of American naval 

 science, although it is well known that in England the subject 

 of ships' magnetism, and its effect on the compasses of the ves- 

 sels, has attracted a great deal of attention for the past 

 twenty years, and, in the hands of scientific men, especially of 

 Airy, has led to a fair understanding of the whole question. 

 We believe, however, that Professor Harkness' results have a 

 special interest as being the first, and, perhaps, the only ones^ 

 that have as yet been attained on vessels of the double-tur- 

 reted Monitor type. The conclusions to which he has arrived 

 may be briefly stated as follows: The deviations of seven 

 compasses were observed and compared with the theory at 

 ten different places, so situated as to afford very great changes 

 in the terrestrial magnetic elements. For all these compass- 

 es the coefficients depending upon the hard and soft iron 

 have been so far separated from each other as to render it 

 possible to predict the deviations in any part of the world, 

 and for the Admiralty standard compass, as well as for the 

 azimuth compass of the Monadnoc^ every one of the coeffi- 

 cients in Poisson's general equations has been determined 

 separately, with a considerable degree of accuracy. The 

 aoreement between the theoretical and observed variations 



