The basics of processing the results of observations obtained during tests of various optical-mechanical and optoelectronic devices are presented. Second edition (1st ed. - 1965) is distinguished by the expansion of the range of questions of the error theory itself and the least squares method - methods for assessing the parameters of the distribution of observation errors, questions of constructing confidence intervals, criteria for detecting systematic errors, etc. The chapter "Least Squares Alignment of Experimental Data" has been re-written.