Hydrodynamic torque measurements on impellers of square cross section were conducted for 33 different impellers in 3 different cylindrical unbaffled tanks. The tanks differed in volume by two orders of magnitude, and the Reynolds number range of the data was 10(4)-10(5). A wide range of impeller sizes and aspect ratios was investigated, focusing on the decay, regime where the tank walls impacted the torque measurement, and the steady, state where the torque is constant on average. The number of revolutions required for spin-lip to steady state is a function of the tank radius and height and the impeller radius and height, and an expression is presented that describes the spin-up time in terms of these variables and is valid for all of the data. At steady state and for a Reynolds number of 10(5), the torque coefficient or power number depended primarily oil the height of the impeller and tank, bill not on the impeller radius. During the decay regime, the measured torque coefficient decreases with a power-law relation and is proportional to the number of revolutions raised to the k power. The k value did not depend on the height of the tank, bill only on the impeller radius, impeller height, arid tank radius. Expressions relating the steady-state torque coefficient and the decay exponent, k, to the geometric variables model all of the data quite well. With these relations, the effects of the varying impeller aspect ratios on the energy required for mixing of momentum, and on the time required for mixing to occur, cart be quantified.