The upstream gas temperature measured with the thermocouple shown in figure 2 was **f. The **f values are listed in Table 1 together with the measured surface temperatures and arc voltages. Simultaneously with the anode surface temperature and voltage measurements pyrometer readings were taken along the cylindrical surface of the carbon anode holder as indicated on figure 2. Some of these temperatures are plotted in figure 4. They showed no marked dependence on the flow rate within the accuracy of these measurements. Thus, the dotted line shown in figure 4 was taken as typical for the temperature distribution for all blowing rates.

The thermocouples in the aluminum disk shown in figure 2 indicated an equilibrium temperature of the surface of **f. This temperature was taken as environmental temperature to which the anode holder was exposed as far as radiation is concerned. It is sufficiently small compared with the surface temperature of the anode holder, to make the energy flux radiated from the environment toward the anode holder negligible within the accuracy of the present measurements. The reflection of radiation originating from the anode holder and reflected back to it by the surrounding metal surfaces should also be small because of the specular characteristic of the metal surfaces and of the specific geometry. The total heat loss through the anode holder included also the heat conducted through the base of the cylindrical piece into the adjacent metal parts. It was calculated from the temperature gradient **f at **f inch as **f. The total heat flux from the porous plug into the plug holder is thereby **f The temperature distribution of figure 4 gives **f for all blowing rates, assuming **f. The temperature dependent value of |e was taken from Ref. 7. The radiation loss from the anode surface was computed according to **f where **f is the mean of the fourth powers of the temperatures **f and **f calculated analogously to equation (1).