1832.] Taken at Sea. 203 



departure the deviation of the chronometer is ascertained by both 

 morning and evening observations, and these be compared with alti- 

 tudes taken at the new station, corresponding as to the time of the 

 day, the results if not agreeing will differ but little. 



On one occasion I found a difference of the nature alluded to above 

 amounting to eight miles ; for I anchored off one of the Piscadare Islands, 

 near Formosa, and at about 8 h. a. m. took several altitudes for the chro- 

 nometers, after which we proceeded nearly west from the anchorage, appa- 

 rently 16 miles both by log and by the distance of the Island from us ; we 

 then anchored, and at 3. 30 p. m. the result of several observations taken 

 and compared with those taken at 8 A. m. placed the ship but 8 miles west 

 of the anchorage, when it was quite evident we were at least double that 

 distance. Thinking these observations might be faulty, I took several 

 others, tut at about 5 p. m. and the results of the latter corresponded so 

 nearly with those taken at 3h. 30m. p. m. as to satisfy me that the differ- 

 ence did not depend on error in the observations, and made me continue 

 at anchor until the next morning, when observations were again taken 

 and compared with those made the previous morning at the anchorage 

 near the Island ; they gave our difference of longitude 1 6 miles, agreeing 

 with the log and apparent distance of the Island from us ; here was an 

 instance in which had not the island been in sight, I should have laid down 

 a danger 8 miles short of what it should be, and should undoubtedly have 

 attributed the difference to the effect of a current setting against us. 



It is well known that in straits or confined navigation, owing to the 

 changes which take place in the horizon, correctness is scarcely attain- 

 able in observations taken on board ship : I have often found it so ? 

 and am of opinion, that in the vicinity of shallow banks, the result of 

 observations taken on ship board will be also generally found much at 

 variance. 



In the instance I have just stated, the bottom to the westward of the 

 ship was formed by several ridges of sand lying in a north and south 

 direction, having 7 or 8 fathoms water on them, with deeper channels 

 between. 



The part of the horizon above which the sun's altitude was measured 

 was remarkably bright, and I have thought that some unequal eva- 

 poration might have caused the horizon to appear higher in that quarter. 

 I have found that observations taken in the vicinity of Sagar were 

 not to be depended on, although there is apparently a clear horizon 

 before you. 



Latitudes taken afloat with much care will never agree with those 

 taken on shore with an artificial horizon, and when the sun is to the 



d d 2 



