model averaging - parameter matrix not correct length

[Lyndon]

OK, I've worked it out. I first ran these models several years ago in an earlier version of Mark. Although I had opened the file in the latest version, I had not rerun the models. The estimates I had didn't make sense if I calculated the last pent by subtraction - more animals were estimated to have entered just prior to the last sample than were estimated to be there (n/p-hat) at the time! I've now re-run the models and the estimates make perfect sense.