The fact that about 80% of trilobites possessed genal spines indicates that these structures must have served at least one important function. The spines were not for predator deterrence because they extended nearly parallel to a trilobite’s body, and no other function for these structures seems previously to have been identified. Here I propose a common function for genal spines and enrollment by trilobites. (Of course, the primary role of enrollment was protection.) Being overturned on the seafloor would have been problematical for trilobites because their feet could not have righted them, but genal spines as well as enrollment made it possible for them to regain their normal position. Overturned trilobites could have used genal spines to stabilize themselves while they flexed their bodies and, in effect, undertook a backward somersault to stretch out in life position without turning sideways along the way because of their rounded pygidium. Instead, some trilobites could have used enrollment to do so. After enrolling, if fortunate, these forms could simply have allowed water movements to rotate them to a position in which unenrolling placed them in life position. Some trilobites that could enroll also possessed genal spines that were short enough to permit rolling, so they could have righted themselves by either method. (A few genera lacked roughly round cross-sections when enrolled and could not have rolled effectively on the seafloor.) The distribution of morphologic features among trilobites provides a test for these ideas. Species that could protect themselves by secure enrollment could do so only by virtue of their shapes. In fact, the outlines of trilobites that enroll are dictated by this behavior. One can predict that taxa with pygidia too narrow for protection by enrollment would have been required to possess genal spines to right themselves when overturned, whereas taxa that lacked genal spines would have been required to have pygidia wide enough for successful enrolling. Observations bear out these predictions. For trilobites that could not enroll for protection, partial enrollment in the presence of strong water movements would have provided the genal spines with a secondary function. Projecting out from the body and sticking in the sediment, they would have provided temporary stability.