Given that H_{0} was rejected for one of the 20 formulations, is it plausible that this formu-lation actually has no effect on wear?
Yes. It is plausible that none of the formulations, including the one for which H_{0} was rejected, have any effect on wear. Twenty hypothesis tests were made. For each test there was a 5% chance (i.e., 1 chance in 20) of a type I error. We therefore expect on the average that out of every 20 true null hypotheses, one will be rejected. So rejecting H_{0} in one out of the 20 tests is exactly what one would expect in the case that none of the formulations made any difference.