[r-t] Hudson stuff

[Alexander Holroyd]

> I had then assumed that the intransitve group could be
> generated by changes if and only if the transitive groups
> making it up could be.  Whilst this is clearly necessary,
> thinking further about it, it doesn't seem obvious that it
> is, in general, sufficient.  Equally I can't think of an
> obvious counter example.  Can anyone else?

The group <13245768,21437856> (order 8) cannot be generated by changes,
even though the two copies of D_4 can: <1324,2143> and <5768,6587>.

Ander