@jwaller, many thanks for the links! I see I had the wrong impression about basisOfRecord in the early years of GBIF. I thought most occurrence records were from collections, whereas the analytics show that observations dominated in 2007:
It’s still true, though, that today’s GBIF is overwhelmingly a repository for contemporary observations. That suits today’s conservation planners, but collections managers might be feeling that the need to share their CMS records with GBIF isn’t as urgent as it once was.