An explanation was sought to explain the existence of the relatively large (∼2000 Å) fission gas bubbles found in the unrestructured region of an EBR-II-irradiated mixed-oxide pin following a TREAT transient in which peak temperatures stayed below melting. Using a code like GRASS, it was found difficult to explain their existence by employing the bubble mobility values fit to experimental measurements in the region of 1500°C. A rather good fit was obtained if the greater bubble mobility that theory gives for the surface-diffusion mechanism was assumed to be applicable at higher temperatures; e.g., above 1800°C. Sensitivity studies showed that swelling is very sensitive to peak temperatures and the duration of the transient and to hydrostatic pressures in the fuel. If the surface-diffusion mechanism is applicable, considerable fuel swelling can occur due to bubble growth and coalescence. In addition, bubble drift due to temperature gradient is found to equal or exceed the effects of Brownian motion.