This critical review of the specification and measurement of polarization errors in Adcock-type direction-finders was prepared in 1944. The respective merits of the “standard-wave error” introduced by Barfield, and the “pick-up ratio” for wanted and unwanted fields suggested by several other workers, are examined. It is concluded and recommended that the former is to be used for Adcock direction-finders erected not higher than about ¼λ above ground, while the latter is appropriate for more elevated systems. Comparison of direction-finders falling into different categories can only be carried out if a complete specification of polarization error performance over all relevant angles of incidence is available.The problems involved in using a relatively nearby test transmitter are discussed. For frequencies above about 3 Mc/s it is shown that a suitable local radiator consists of a rotatable closed loop used at moderate elevations, and at distances not less than a wavelength. The effects of loop size are discussed, and it is shown that adequately large loops are practicable; for example, for frequencies in the range 3–30 Mc/s, a loop up to about 5 ft in diameter may be used, providing the distance is not less than 100 m. For frequencies below about 2–3 Mc/s it is suggested that the “local-injection” method due to Barfield may be more practicable than a radiation method.It is pointed out that the performance of a direction-finder is in general intimately bound up with the electrical properties of the site on which it is erected.