I asked an ornithologist about testing an item like a feather, an eggshell, or a dropping. He tells me that DNA testing could most likely be done that would conclusively prove whether the item came from an Ivory-bill.
9/23/05 update: Here's a lab specifically advertising this type of testing for $23 with a turnaround time of five business days.
I think it's significant that Cornell has failed to gather any of this kind of evidence. After 20,000 hours in the field, it seems likely to me that you'd find something if the birds were there (especially if you spent a lot of time around likely feeding and roosting areas).
This is just one more issue in a long list of troubling issues from Arkansas (ie, the bird seems impossibly wary, it's incredibly non-vocal, it seems to fly without the characteristic loud wing noise, it has chosen poor-to-marginal habitat, it seems to look a lot like a normal or abnormal Pileated, it double-raps oddly and rarely, etc).