Chain semigroup fix

The truly marvellous BECKS automaton again did a good service by showing that there was a false assumption about finding the gaps in reduced inclusion relation.
Now, morphic relations work fine for the monster semigroups as well.