AbstractIn ordinary survey work of a lower order, i.e., work which does not attempt to attain geodetic accuracy, the most popular methods of finding latitude astronomically involve observations of the altitudes of celestial bodies, and it has long been recognized that observed altitudes are liable to serious errors from two main sources, namely, altitude bubble error and errors in the assumed values of astronomical refraction. The former can arise from the bubble not indicating the true horizontal (instrumental error) and/or imperfect levelling up of the vertical circle alidade bubble (personal error), which is rather a delicate operation even when using a prism reader or other aids. The instrumental error may, of course, be eliminated by taking readings with both faces of the instrunlent, though there are instances when this may not be possible. Comprehensive refraction tables are published which give correcting factors for varying states of temperature and pressure. But these are not the only two condit...