Physical models of graptolites have been constructed for a range of morphologies, with emphasis on planar, multiramous forms. The models are life size and have the density of a living graptolite, based on the now-established collagenous nature of the periderm and unavoidable assumptions about the amount of extrathecal tissue present in the living colony. These models have been used to test the two main hypotheses of graptolite life habit developed by Bulman, Rickards, Kirk, and others. Testing of graptoloid models in water suggests that many rhabdosome shapes were designed for passive rotation within the water column. This is caused in the models by a variety of modifications, including changes in thecal and stipe orientation, alterations of colony shape and the addition of vanes and hooks. Rotation would only have been useful when the rhabdosome was in directional motion and the frequency of such modifications seems anomalous if no such movement occurred. Thus movement by some means is required, either passively, by changes in buoyancy, or by automobility. Spiralling action would increase the harvesting path of an individual living on a planar, multiramous colony, making this a theoretically advantageous mode of life for the morphology. It would prevent the individual zooids of scandent biserial and uniserial colonies from feeding from the same narrow band of water.