A critical examination is made of possible ways in which a standard of light of high intensity could be developed using the positive crater of an electric arc between carbon electrodes in air. An account is given of much hitherto unpublished experimental work bearing on this problem.The paper describes a 3-electrode arc which was adopted in these researches as being the most suitable for the purpose. A physical photometer is used in which a galvanometer deflection automatically gives a measure of the light output of the arc. The accuracy exceeds that possible with visual photometry, and rapid variations of candle-power can be observed. Graphite electrodes and a wide range of uncored soot electrodes are used, and conditions for obtaining optimum results are described.For a standard, either the candle-power of the positive crater measured along the axis of the anode might be used, or the brightness of a small portion of the crater near the centre.Measurements are given of the effect on candle-power and brightness of many factors including diameter, resistivity and purity of the anode, arc current, negative electrodes, atmospheric pressure and absorption due to arc flame. Special attention is paid to the elimination of a phenomenon whose existence is generally not obvious, namely the rapid rotation of the arc stream, which is accompanied by a reduction of candle-power and brightness of uncertain amount.A critical comparison is made with other published work especially work in other countries, and it is concluded that a standard can best be developed utilizing brightness, an accuracy of 1% being so far attained. It is hoped to give in a subsequent paper the form that such a standard should take.