the Vbe temperature coefficient for germanium is less than that of silicon, for germanium it's -1.8mV/degC, while for silicon it's -2.2mV/degC. it's not the change in Vbe that is destructive, but the changes in Icbo (C-B leakage) and the maximum operating temperature. germanium leakage current measured at room temperature is in the milliamp range, but in the nanoamp range for silicon. leakage current doubles for every 9degC increase in temperature. the maximum operating temperature of silicon is 150-200 degC, but for germanium it's 70-80 degC. so, while the change in Vbe is the thermal runaway mechanism for silicon transistors, for germanium, the hazard is from C-B leakage current. as far as parallel Ge transistors, and why it seemed to work, i'm still looking, but so far, it's conspicuous by it's absence in the texbooks and design guides. current hogging seems to be mentioned in reference to silicon transistors, but not germanium. if i were to make a guess why, it would be that germanium transistors also exhibited current hogging, but the leakage current problem was much larger in magnitude, and the current hogging problem went unnoticed until silicon came along and current hogging was more visible.