Hmmm... I'll think about it. It just seems like a lot of work to create a seperate place to store the data and then link it into the treeview just to avoid something that may never happen. :)

Thanks,

Michelle


>OK. Well, I really don't know the answer to your original question.
>
>Populating the tree "on-demand", however, seems to me the best solution. If the user expands a node, populate it. If the user searches and locates a specific record in another way, you might make sure that that specific record is in the treeview, and expand it.
>
>This would both make the treeview faster, and, hopefully, overcome the array limitation.
>
>Hilmar.