Dear Ralf, I cannot find the mail anymore in which you explained why you want to have several labeltypes for multisort species. I think the argument was because of functorial composition, but I cannot remember well... Maybe you could repeat. Thanks, Martin