FROM TONKIN TO INDIA 



utterly at variance with each other and with the probable result, that I was obliged 

 to reject them altogether. On every other occasion constant verifications admitted 

 of my placing entire reliance upon the readings of the compass. 



Being actually the first traveller in Viinnan to take precise declinations (Francis 

 (larnier had no theodolite), I had no data for a comparison with the annual increase 

 or decrease of magnetic intensity in that country. It is to be hoped that future 

 •observations made at the same points for several years may supply what is desirable. 



My instruments having been stolen at the end of July by the Lamasjen, my 

 astronomical observations ceased from that date. 



III. LONGITUDES 



The longitude of Fey-Long-Kiao was the last observed, owing, as above, to the 

 loss of my theodolite. 



Insti-umeiit and Method Employed. — The foregoing longitudes were obtained by 

 the use of the small Hurlimann theodolite graduated to 1'. For their determina- 

 tion I used the method of equal altitudes of moon and stars ably set forth by 

 M. Caspari, ingenieur hydrographe de la marine, in the second part of his 

 Coiirs d'astroitoiiite pratique (Paris, Gauthier-X'illars, p. 155). The principle of this 

 system is essentially the same as that of the method of lunar altitudes : but errors 

 of refraction, reading, and graduation are allowed for in the following manner : — 



Fix the telescope of the theodolite at a certain height, which it is not necessary 

 to know exactly. Choose a star, for purposes of comparison, as near as possible 

 to the moon's trajectory, and note the time at which the foremost of the two 

 bodies, in the direction of the diurnal motion, pa.sses beneath the horizontal thread. 

 Then displacing the telescope in azimuth, without altering its height, await the transit 

 of the second body, and again note the time. At each of such junctures ob.serve the 

 level, to check the stability or variations of inclination of the axis of the telescope. 



As shown by the formute of M. Caspari, the refractions, which owing to the 

 proximity of the two bodies may be considered as identical, do not appear in the 

 calculation ; the absolute altitudes being immaterial, the errors of graduation are 

 removed ; and it is enough to observe one side of the axis only, whether with the 

 telescope on the right or on the left. Immediately before or after, the local time 

 should be ascertained. 



In the case where the two bodies are near meridian passage, the observations 



366 



