Finally the day came, after numerous rehearsals, when the long steel tube, or terminal, on the end of the movable cable could be moved out by the trolley to a middle point over the pit, where it would make contact with bubbling liquid lava when lowered. This was an extremely ticklish procedure, for the lava was a heavy mat of self-crusting liquid rock with the crust forming hard slabs; few places kept up an appearance of bubbling porridge. No one had ever made contact before with the liquid of a fountain like “Old Faithful.” It was fortunate that the apparatus, which was expensive, consisting of platinum wires imbedded in silica glass, was made in duplicate so that we had two of everything. The splashing liquid of “Old Faithful” looked as harmless as a kettle of boiling soup, but Perret and Shepherd were in for a surprise. When Shepherd lowered the terminal directly into the liquid, “Old Faithful” exploded, for the molten slag proved to be a suction whirlpool which threw tentacles of lava over the steel pipe. The apparatus went down to destruction “like a bass under a log,” and the cable was bitten off like a piece of line. The entire terminal vanished into the vortex, leaving only a corroded wire. To shorten a long story, the second terminal was lowered into a seemingly safer liquid place. A wave of the melt slapped and strained the pipe, and though it was recovered, no electric resistance reading was obtained at any time with the box at the rim of the pit. Close to $1,000 in equipment was lost. The resistance pyrometer is a sensitive tool in the laboratory, for giving precise degrees of temperature in the region of 1200° Centigrade, supposedly the melting point of basalt. But it was unsuited for the rugged bubbling of basalt slag, where flaming gases and chilling air play more important parts than mere melting. Fortunately Shepherd and Perret were not at the end of their resources. There still remained the thermocouple, a simpler pair of wires of platinum and iridium encased in a steel tube. The connectors from these go to a simple galvanometer in the hands of the operator. The trolley could still be used, and the thermocouple pipe had no glass inside it to be shattered. A temperature of 1000° Centigrade was recorded in a bubbling area, and this was considered good enough for an approximation. Another experiment was to lower an iron bucket into the liquid, and pull it up full and dripping with black lava glass. This was sent off to Washington for analysis. Afterwards the lava lake went down, no more experiments that year were possible, and Perret began the plotting of a curve of high and low in the rise and fall at the bottom of the pit.