AbstractThis paper attempts to analyse the usual accuracies which may be expected in the determination of distance by means of the subtense bar, and the improvement in accuracy consequent upon increasing the number of observations of the subtended angle, by examining sets of subtense observations taken by working surveyors under normal field conditions. Criteria for field rejections of outlying observations, and quick estimation of the expected accuracy by examining the range, are suggested. The tendency for outlying observations to occur early in a set, and the slightly high incidence of pairing of highest and lowest observations adjacent to each other are noted.Introduction.Theory.Sources of Samples.Values of Mean Square Errors: Before Rejections.The Rejection of Outlying Observations.Ranges.Recommended Criteria for Field Rejections.Estimate of Mean Square Error from Range.Incidence of Highest and Lowest Observations.Reliability of the Mean.Values of Mean Square Errors: After Rejections.Confidence Limits.Subtense Bar Traverses.Conclusion.