IN THIS PAPER, it is shown that the Cox, Ingersoll and Ross (CIR)[1977] and Vasicek[1977] closed-form solutions for equilibrium default-free bond prices are particular cases of a more general solution. This solution is characterized by separability in maturity and arguments of the bonds' semielasticity with respect to the state variables (or their instruments). The simple derivation of this result and several tests of the general solution are given in Section 2. It is shown, in Section 3, that particular solutions for bond prices which are expressed as functions of state variable instruments impose restrictions on the time-series processes for these instruments. Hence, the solutions may be tested by testing the implied restrictions. To further the application of this test, an extended model which includes two instrumental variables-real per capita consumption and the nominal short rate of interest-is formulated and solved here. In Section 4, a solution for the equilibrium bond price is obtained by numerical integration of the pricing function when stated in terms of per capita consumption. Two continuous-time processes for the consumption series-the lognormal and Ornstein-Uhlenbeck (O.U.)-are estimated by full information maximum likelihood methods. The results are used to illustrate deficiencies claimed here to exist in numerical methods. Finally, it is shown in Section 5 that a test of the expectations hypothesis follows immediately from the fundamental partial differential equation on which closed-form solutions for equilibrium bond prices are based (see, for example, CIR[1977]). Tentative evidence tends to reject the expectations hypothesis. All closed-form solutions for equilibrium bond pricing functions except the extended model referred to above and one given in CIR[1977] are single-state variable models. Whilst there appears little evidence to date that the dimensionality of the state variable vector cannot be treated as small, tests based on single-state variable processes preclude, at the least, a satisfactory approach to the evaluation of the relative importance of nominal versus real factors in the determination of the nominal term structure. An alternative methodology has