We studied the morphological and neurochemical properties, as well as the synaptic targets of glycinergic neurons in laminae I-IV of the mouse spinal dorsal horn by using transgenic technologies, immunohistochemistry, in situ hybridization, light and electron microscopy. We showed that although cell bodies of glycinergic neurons were rare in lamina II and only a slightly more numerous in lamina I, axon terminals were densely distributed in these superficial laminae, indicating that glycinergic neurons in the deeper layers of the dorsal horn send their axons into laminae I-II. The number of labelled neurons was higher in lamina III, but the highest density was observed in lamina IV. We demonstrated that there are at least 3 and 6 subtypes of glycinergic neurons showing distinct morphology in laminae I-II and III-IV, respectively. According to the classification scheme of Grudt and Perl (2002), all the labelled glycinergic neurons in laminae I-II and many of them in laminae III-IV were identified as islet, central and vertical neurons. Cell morphologies in laminae III-IV, however, were more diverse; some labelled neurons in laminae III-IV resembled inverted stalked cells of Gobel (1978) and pyramidal cells of Réthelyi and Szentágothai (1969), while the morphologies of others did not correspond to any of the previously identified cell types. We showed that distinct populations of glycinergic neurons in laminae I-IV express neuronal markers like PV, nNOS, RET and RORβ. First in the literature, we provided experimental evidence that there are glycinergic axon terminals in abundant numbers in laminae I-II that do not express GABA, and these glycineonly axon terminals may arise primarily, if not exclusively, from neurons in lamina IV. Thus, glycinergic neurons with cell bodies in lamina IV may contribute substantially to spinal pain processing. Our findings suggests that glycine-mediated postsynaptic inhibition must be significantly more dominant than glycinergic presynaptic inhibition in laminae I-III. Our results indicate that glycinergic postsynaptic inhibition, including glycinergic inhibition of inhibitory interneurons must be a common functional characteristic of neural circuits underlying spinal pain processing. On the other hand, it is possible to hypothesize that glycinergic presynaptic inhibition, even though can be weak at the cellular level, may be crucial for locally targeting functionally distinct subpopulations of primary afferent inputs.