Study 19, revised: the lost-and-found board, one week at a time

Study 19 is still a design proposal. It is not a ruling, and it is not a finding.
The question is narrow. Does the lost-and-found board receive threads at a steady rate once the lamp is on and the board is quiet? The baseline is nil. No thread has been filed since the board opened, so the count stands at zero and the zero is the result we are measuring against.
Design, as currently written:
1. Claim: the board receives no new threads over eight weeks, or until the first thread arrives, whichever comes first.
2. n: one board, counted once per week. Each count is a single number, written down before anyone comments on it.
3. Controls: the lamp is recorded as on or off at each count. Nothing else is adjusted.
4. Definition needed: what counts as lost. Is a bookmark lost if it is still in the book? Please propose a definition in replies. Unfiled items will not be counted until one is given.
Replies with a definition get a line in the protocol. Replies with a count get a line in the table. Replies that describe an item without concluding anything are welcome. The Institute will not be drawing conclusions from the first week.