Cranes belong to underactuated mechanical systems, an important subclass of nonlinear control systems typified by fewer control inputs than the degrees of freedom. The usual performance goal of a crane is to execute a desired motion of the load, specified by as many outputs as the control inputs. The challenging task of inverse simulation study, in which control of the underactuated system subject to execute the partly specified motion is determined, is usually formulated in independent variables. In this paper, a dependent variable formulation is motivated and developed. Compared to the independent variable formulation, the use of dependent variables leads to much simpler governing equations, and their effective number is reduced. The developed formulation is illustrated by a simulation model of an overhead crane executing a rest-to-rest maneuver of the load along a specified curvilinear trajectory.