Error estimate for the valence approximation to lattice QCD

A Monte Carlo algorithm is used to calculate msub(N)/msub(rho), msub(δ)/msub(rho), msub(q)msub(rho)/msub(π)2, fsub(π)/msub(rho), fsub(rho)/msub(rho) and #betta#sub(mo) for lattice QCD with the number of flavors Nsub(f) of quarks contributing virtual loops given by zero, the valence approximation, and with Nsub(f) chosen, in effect to be -2. The difference between corresponding pairs of quantities with Nsub(f) of zero and Nsub(f) of -2 we take to be an estimate of the difference between the value for a realistic theory with Nsub(f) of +2 and for the approximate theory with Nsub(f) of zero. The changes in mass ratios and #betta#sub(mo) are consistent with zero and typically (20 +- 20)% while decay constant ratios change by as much as (88 +- 24)%. (orig.)