As far as recovery goes, he
pointed out that the bulk of the US 1960 census data was eventually
recovered even though it was initially thought to be lost.
The data was always there and available. My Aunt (an urban planner)
wrote two books using that data way back then, and often still uses
the data today. She says that the getting data from the Census people
has never really been a problem (except for 1890, of course).
The problem was that the Census Bureau and NARA used different
formats. I would bet that there also was a certain amount of petty
politics involved with the "losing" and "recovery".
--
Will