Lots of older papers don't get read because they are not available online (a hypothesis that relies on the assumption that people read papers before citing them)

Seems like a partial reason. It took a bit of snooping to find a copy of this one online.