Examination of the estimated standard errors for both the ordinary

least squares and two stage Aitken's estimators demonstrates that a

considerable gain in efficiency was realized by incorporating the addi-

tional information of contemporaneous correlation across states in the

estimation procedure. In all cases, the estimated standard errors

obtained via the two stage estimator are less than or equal to the cor-

responding standard error estimates obtained using ordinary least

squares. These gains in efficiency are especially significant given the

small sample size employed in the estimation procedure.

 The above results give considerable support to the specification of

the system of price equations and the choice of estimator when taken as

a whole. The incorporation of contemporaneous correlation across equa-

tions not only yield the expected signs on all but one parameter esti-

mate but also results in considerably more efficient parameter estimates.

 The small magnitude of the parameter estimates in Table 7 gives the

appearance that changes in catch have very little effect on price both

within and across states. The size of the coefficients, however, are

the result of the unit of measure used for the catch quantities. Thus,

a discussion of the estimated within and across state price effects is

best couched in terms of direct and across state price flexibilities.

Estimated price flexibilities evaluated at the 1957 to 1975 mean levels

of price and catch are shown in Table 8.

 Florida has by far the greatest within region price flexibility.

A 10 percent increase in Florida reef fish catch is estimated to

decrease the price of reef fish in the state by 8.8 percent. For

Mississippi and Texas, similar increases in reef fish catch are esti-

mated to decrease price by 3.8 and 1.3 percent, respectively. Changes in