The anatomy of the developing zebrafish spinal cord is relatively simple but, despite this simplicity, it generates a sequence of three patterns of locomotive behaviors. The first behavior exhibited is spontaneous movement, then touch-evoked coiling, and finally swimming. Previous studies in zebrafish have suggested that spontaneous movements occur independent of supraspinal input and do not require chemical neurotransmission, while touch-evoked coiling and swimming depend on glycinergic neurotransmission as well as supraspinal input. In contrast, studies in other vertebrate preparations have shown that spontaneous movement requires glycine and other neurotransmitters and that later behaviors do not require supraspinal input. Here, we use lesion analysis combined with high-speed kinematic analysis to re-examine the role of glycine and supraspinal input in each of the three behaviors. We find that, similar to other vertebrate preparations, supraspinal input is not essential for spontaneous movement, touch-evoked coiling, or swimming behavior. Moreover, we find that blockade of glycinergic neurotransmission decreases the rate of spontaneous movement and impairs touch-evoked coiling and swimming, suggesting that glycinergic neurotransmission plays critical yet distinct roles for individual patterns of locomotive behaviors.