Uniaxial compressive creep of a 90+% Al(,2)O(,3) and 50+% Al(,2)O(,3) refractory concrete was investigated. The creep behavior was evaluated as a function of aggregate packing, cement composition, and time at temperatures up to 2200(DEGREES)F and stresses up to 2500 psi;Refractory concrete crept more when heated the first time than on subsequent heating. On first heating, the hydrated calcium aluminate cement bond dehydrates and then sinters. The Arrhenius temperature dependence for creep is about 30 Kcal/mole on first heating and about 130 Kcal/mole thereafter for ISU-90 concretes. For ISU-50 concretes, the Arrhenius temperature dependence for creep varies from 20 to 45 Kcal/mole on initial heating and 100 to 180 Kcal/mole thereafter. The stress dependence of the creep strain for the ISU-90 is given by stress exponent n to about 0.6 on the first heating and to about 3 on subsequent heating. For ISU-50, the stress exponent n is about 0.8 on first heating and about 2.5 on subsequent heating. The creep strain is dominated by deformation of the cement phase, the aggregate being essentially inert at all temperatures. Replacing the high purity cement with intermediate or low purity cements caused increased strain at the same stress-temperature conditions, and failure occcurred at lower stress-temperature combinations than for the high purity cement;The steady-state creep behavior was analyzed by using the general creep equation: d(epsilon)/dt = (epsilon) = A f (s) (sigma)('n) exp (-(DELTA)H(,c)/RT);Here "(epsilon)" is the strain rate of the creep, "A" is a constant, "f(s)" is a structure term, "(sigma)" is the stress term, "n" is the stress exponent, "(DELTA)H(,c)" is the activation energy of the creep, "R" is the gas constant, and "T" is the absolute temperature;Microscopy, x-ray diffractometry, and porosity and density data were used to gain insight into creep behavior and possible mechanisms of this system.