>Assuming you mean B-tree, Fox indexes are B-trees (actually B+trees?)
>
>The perfomance difference comes from the variable record length of the Access files.
>
>Incidentally, Btrieve files are ISAM.

So it *still* looks to me that there are only marginal differences between SQL Server's storage mechanism and ISAM in general.

And I would question to a minor degree that variable length records slow Access (or SQL Server, for that matter) down. Given the speed of today's CPUs versus HD, I can estimate a distinct advantage for variable length records, *especially* when the average content length for any field is well below the maximum length of a field.

Cheers,

JimN