Abstract According to Ornstein, Brinkman and Smit, the arc temperature can be determined with the aid of the band spectrum proceeding from the molecule CN. However, it can also be determined by means of the intensity ratio between two spectral lines of the same element with different excitation levels, whose relative transition probabilities are known. We employed the temperature-measurement with the aid of spectral lines of zinc (high ionizatioh potential!) whose transition probabilities have been determined by Schuttevaer and Smit. The measurements were carried out in order that we might ascertain the temperatures during the evaporation of various materials in the carbon arc, particularly as applied in the process used in our laboratory. A description is given of specially prepared electrodes permitting a steady supply of a small quantity of zinc during arcing. The result obtained for the centre of the arc, according to Addink's method, is a temperature of about 6100°K. Apart from alkali metals, the evaporating materials affect this temperature only slightly. In the case of the “Addink arc” the influence of sodium as main constituent is smaller, however, than in that of arcs used by others (e.g. Harvey). The temperature in the centre of an arc with a gap-width of 9 mm (analysis by means of X-values) has been found to be lower than the average temperature of an arc with a gap-width of 2 mm ( Q-values). Temperature-measurements and determination of the sensitivity in the arc with argon or with argon + oxygen as discharge gas show that for customary spectrochemical purposes the “argon arc” will probably afford no great improvement in comparison with the arc in air.