Message written at 19 Jan 1999 07:38:19 +0000

In the k-rate output version I had a bug which meant that I read the
wrong part of the table.  Could that be a cause for the oddities?