no HYDROGRAPHICAL SURVEYING [chap. iv. 



causes contributing to this discrepancy, and assign to each 

 its due weight. 



The difference between the results by triangulation and by 

 astronomical observations is attributable on the one hand to 

 errors in triangulation, viz. : 



( 1 ) Errors in the base as measured ; 



(2) Accumulated effect of errors in the triangles involved ; 



(3) Errors in the true bearings ; 



and on the other hand to errors in the astronomical observa- 

 tions. 



If proper care has been exercised throughout all the opera- 

 tions, each should bear its fair share in the final adjustment 

 within the limits of its probable error. 



This is more particularly the case when it becomes a question 

 of altering the true bearings to any material extent, especially 

 if the bearings obtained at one end of the survey, on being 

 transferred to the other end, throughout the whole length of 

 the triangulation, are found to agree fairly closely with the 

 bearings observed at that end. 



We should also be reluctant to throw over entirely the 

 results by triangulation if a check base has been measured and 

 found to corroborate that part of the work. 



These considerations have greater weight in the case of a 

 survey where the Hmits are not very extended. An error of 

 a fraction of a second of arc in the astronomical position of 

 one or both of the terminal points of a survey of small extent 

 might produce a somewhat starthng discrepancy in bearing. 



Although the observed latitudes will not probably be much 

 in error, the sextant being capable of very great accuracy, yet 

 an error of 1" or even 2" may easily occur in either latitude, 

 involving a possible error of 2" or 4" in the resulting diff. lat. 



In the case of a chronometric meridian distance, the possible 

 error in diff. long, may be even greater, and, as stated above, 

 astronomical positions are liable to be seriously affected by 

 errors due to the deflection of the plumb-hne. 



We must therefore endeavour to obtain final results without 

 placing undue reliance on any particular element of the 

 problem to the exclusion of others when circumstances permit. 



