From P.K. Sharma: The first two sentences of the proof of Lemma 2.8
need to be reversed like this:
  "Since M - N generates M as a module,
  the sequence {f^m M} also contains only finitely many sets. But this sequence
  is decreasing, so f^m M = f^{m+1} M for some m."

From Mehmet Kiral: in the proof of Lemma 1.6, the inclusion 
    $M_d \subseteq cM_d \subseteq cM$
should read
    $M_d \subseteq cM$
In the last line of the proof, the statement "the submodule $M_d
\subseteq M$ maps isomorphically to $cM = M_d$" requires justification.
It holds because on one hand $c M$ is torsion-free, while on the other hand
$c M = ac M$ for $a=c$.