I don't think polling the abstractors would yield any useful numbers.  Besides the fall throughs where no deed or mortgage is recorded, you would also need to account for alll of the out-of-state and offshore companies that work from their own title plants.  In some states, title agents are required to maintain title plants exclusive of the recorders/clerks office, or at least have use of the outside plants.

I think the only way to even get a ballpark figure is, similar to Robert's answer, to look at a month's sampling and average it out per day.  I would say for a purchase deal figure in a title search and two updates.  For a refinance, a title search and one or two updates.  A title search and several updates for each new foreclosure filing.  In some states, a title search for every condo assessment lien filed.  On the day a deed or mortgage is filed, chances are a bringdown search has been performed to check for intervening liens.  The title search was likely conducted weeks or months in advance of the filing.

I think it would be a lot of work, and the numbers would not hold steady from month to month, which would make the usefulness of the information fleeting.