I do not know how we eventually got to the situation where we are only doing samples prooftest on recoated splices – I was not here when it was implemented - but having an average of ~50 kspi on prooftest shows some very serious weakness in the recoated splice strength…