> Maybe. Or maybe you are in one of your ambiguous many:many conditions
> that, according to you, are AOK. Still, I'm not willing to go
> there. I say
> that a detail can have only one master and you deny it, so there's really
> no other place to go on this issue.
>
> Helen
>

if unknown cursor errors were the exclusive domain of situations with many
to many relationships - you'd have my ears pricked. But this error appears
when no such model exists. It's very infrequent and quite random in nature
and thus very difficult to track down.
Alan