One additional note - getting back to Roland's original question
regarding the small shells left in bottles by Octopus - It seems to
me that the number of shells in inhabited vs. uninhabited bottles is
not as significant as the condition  of the shells.  If the shells,
once deposited by the octopus, begin to undergo dissolution, then the
shells in a given bottle, whether inhabited or not, should show
various stages of dissolution.  Do any of the shells appear to be
"half dissolved?  slightly dissolved?  almost dissolved?  If not,
then something other than dissolution must be the cause of the
smaller number of shells in abandoned bottles.
 
P.S.  Roland - what species of Octopus are you dealing with?  Just
curious.
 
Paul M.