Jan. 12, i888] 



NATURE 



249 



by Welsh in 1857-58 it was necessary to make use of a 

 special altazimuth instrument, or of a sextant and artificial 

 horizon, in order to determine the sun's altitude at the 

 time of observation, and a similar method was employed 

 by the Rev. P'ather Perry in the course of the magnetic 

 survey of France to which reference has already been 

 made. Thanks to the admirable arrangement of our 

 Post Office by which signals givini^ Greenwich mean 

 time are sent to all the postal telegraph stations in the 

 United Kingdom, it is possible for an observer engaged 

 in magnetic work in the British Isles to determine the 

 error and rate of his chronometer with an accuracy 

 sufficient to enable him to dispense with the labour and 

 trouble involved in the use of an altazimuth instrument. 

 But unfortunately Greenwich mean time is not yet flashed 

 all over the world, and a surveyor making use of the Kew 

 magnetometer in distant countries would be under the 

 necessity of making independent observations for solar 



altitude, and hence of adding to his iinpedimenta some 

 such arrangement as those used in former surveys. Nor 

 does this diminution in weight of the French instrument 

 materially influence the accuracy of the observations, at 

 all events so far as declination is concerned. It is hardly 

 possible with the English instrument, even under favour- 

 able conditions, to obtain a declination observation 

 which shall be accurate to within 2'. And yet, so far as 

 an analysis of the data given by M. Moureaux enables us 

 to judge, his instrument, of which the circle is only 

 o oS m. in diameter, gives results which are in at least as 

 close accordance with the truth. The method of observa- 

 tion which M. Moureaux adopts in determining the mag- 

 netic meridian allows him to read the position of both 

 ends of the magnet both when erect and inverted in its 

 stirrup. The magnets are solid and cylindrical in form, 

 65 cm. in length and 04 cm. in diameter, and weigh 

 about 7i grammes, and are suspended by a single thread 



Iclinometer. m m, reading microscopes ; i., lifting apparatus for needle ; i, dipping needle ; c, cover ; n, level. 



■of silk 01 1 m. in length. The ends of the magnets are 

 made slightly concave, and are p^lished so as to reflect 

 the cross-wire placed in each of the microscope^, through 

 which the readings for position are made. Each 

 determination of geographical meridian is the mean 

 result of from four to six independent observations, 

 which rarely differ among themselves by more than i' 

 of arc. 



The same magnet which serves for the observation of 

 declination is used as in the Kew instrument for the 

 determination of the horizontal component, which is 



done, as with us, by finding the relation - bv Gauss's 



M ' 



method of deflections, and the proluct HM by the 

 method of vibrations, whence H can be deduced. For 

 this particular determination it seems to us that the Kew 

 mode! is distinctly to be preferred. Indeed, in the 



observation of deflections the Kew instrument leaves very 

 little to be desired, provided that care is taken to avoid. 

 sudden alterations of temperature, say by exposure to 

 sunshine. The main error in the estimation of the period 

 of vibration of the magnet also arises from the uncertainty 

 of its temperature when observing in the field. But in 

 the French instrument no special pains are taken either 

 to ascertain or to correct for te nperature. M. Moureaux 

 indeed is of opinion that, under the conditions of observa- 

 tion, the error committed by neglecting the correction is 

 not greater than that which results from the difficulty of 

 knowing whether the temperature of the magnet is repre- 

 sented by that of the outside thermometer. This is no 

 doubt true of the instrument employed by the French 

 observer, but in the Kew pattern special attention is paid 

 to this point, and, although the arrangement leaves some- 

 thing to be desired, there is no doubt that with care the 

 temperature may be determined with a fairly close 



