~ 
254 Review of the Results of the U. S. Coast Survey. 
same class. By a consistent application of the method of least 
purpose and may affix to the observations a statement of facts 
affecting their quality, but here the influence of his judgment or 
and there are probably classes of errors which no num 
ations or variety of methods can entirely eliminate ; 
always be necessary to discriminate, and to apply smal] corree 
tions to the results in order to make them fulfil the theore 
relations existing between them. When this is done acco 2 
to fixed mathematical rules all uncertainty vanishes, and tru od 
must be the gainer, while on the other hand when it 1s allow t 
to be done according to personal judgment or bias, 3 
vary with different computers, and the door is opened to 1 
cation and fraud. ’ 
Se oe 
ES Saeed 
