Creep resistance is an important design criterion at high temperatures especially when continuous attempts are made to increase the efficiencies by increasing the operating temperatures. Alumina is an important high temperature material and in addition to that it is used in wide variety of applications such as substrates for electronic packaging, spark plugs, envelopes for sodium vapour lamps, cutting tools (when reinforced with silicon carbide) and in artificial joint prostheses. Studies on creep in alumina. have started as early as 1961. There are differing mechanisms proposed to explain the creep behaviour of alumina in the literature, but until now there is no any unanimous decision as to what the rate controlling mechanism is. Magnesia doped at ppm levels can produce significant changes in the microstructure of alumina, the most important consequence of that being the grain growth inhibition, which renders alumina superplastic. However, in a stoichiometric oxide like alumina, small impurities can create extrinsic defects which would change the diffusivities and creep rates. Therefore the background impurities in alumina should be kept to a minimum, if small dopant effects have to be studied. The present study was undertaken making use of high purity alumina powder and comparing the grain growth and creep properties of pure and magnesia doped alumina, especially since no such investigation was carried out in the recent past with high purity alumina. Pure alumina was processed by cold compaction followed by cold isostatic pressing (CIP) and pressureless sintering in air at 1773 K for 1 hour. Magnesia doped alumina was prepared by calcining a mixture of alumina and magnesium nitrate at 973 K for 2 hours followed by cold compaction, CIPing and pressureless sintering in air at 1773 K. Both pure and magnesia doped alumina were further annealed at 1873 K for various times to get grain sizes in the ranges of 1-5 μm. Grain growth kinetics of pure and magnesia doped alumina were studied at 1823 and 1873 K. The parameter Kg which quantifies the mobility of the grain boundary was got. It was found that Kg had decreased in the magnesia doped alumina (in comparison with pure alumina) by a factor of about 3 to 4 which was marginal and insignificant. The grain sizes followed a log normal distribution in both the cases, indicative of normal grain growth. Creep studies were conducted on pure and magnesia doped alumina in three modes, namely, constant stress, temperature jump and stress jump test. The temperature range used was 1673 to 1773 K and the stress range used was 10 to 100 MPa. The creep parameters were found to be n~1.6, p~3.7 and Q-545 kJ mol"1 for pure alumina and n~l .3, p~3.0 and Q~460 kJ mol-1 for magnesia doped alumina. The creep rates in the case of magnesia doped alumina were found to have increased by a factor of 2 to 3, in comparison with pure alumina. The increase in creep rates were found to be insignificant. The creep data were analyzed and the possibility of the dislocation and interface reaction controlled creep mechanisms were ruled out since they were inconsistent with the data. It was found, from creep parameters and the comparison of theoretical Coble and Nabarro-Herring creep rates with the experimental rates, that Coble creep might be rate controlling. The activation energy values suggested that aluminium ion diffusing along grain boundary might be the rate controlling species. However, when the theoretical creep rates considering various species were compared, the rate controlling species turned out to be oxygen ion diffusing along the grain boundary.