We are running a hypothesis test to see if weight of African zebras differs significantly from that of Arabian horses. We assume weight in both populations is normally distributed and both populations...

We are running a hypothesis test to see if weight of African zebras differs significantly from that of Arabian horses.  We assume weight in both populations is normally distributed and both populations have the same population variance in weight which is unknown to us, so we use the sample variances to estimate the standard error for our calculations.  We have a sample of size 9 for the African zebras with mean 650 and a sample of size 16 for the Arabian horses with mean 658.  If the pooled variance is 81, then what is the standard error estimate for the estimator of the true mean difference in weight for the two populations?