Formalin injected subcutaneously into the paw is a frequently used pain assay; it evokes an initial period of licking and flinching followed by a period of quiescence and last by a second period of intense and protracted licking and flinching. The prominent second phase is believed to reflect the development of a central (spinal cord) facilitation. This conclusion is based on the assumption that formalin evokes an initial burst of activity in fine afferent fibers, followed by prolonged low levels of activity in C fibers. Detailed reports substantiating this essential assumption have not been published. Thus, we recorded in situ from single sural nerve fibers, identified by their conduction velocity and modality, in the barbiturate anesthetized rat. Following formalin (2.5%, 50 microliters) injection into their receptive fields, phase-1 activity was prominent in A beta and A delta fibers as well as in high-threshold C nociceptive afferent fibers. Phase-2 activity was observed in A delta fibers with receptive fields in hairy skin and in all mechanically sensitive C fibers. Mean phase-2 activity in these fibers was 1/2-2/3 of the magnitude achieved in phase 1. Mechanically insensitive fibers and A delta and C fibers with receptive field centers outside of the injection zone began firing 15 min or more post-injection and would contribute to phase-2, but not phase-1, behavioral activity. Intravenous infusion of low doses of lidocaine yielding plasma levels of 3.6-7.9 micrograms/ml administered during phase 2 blocked formalin-evoked activity in primary afferent fibers in a dose-related fashion without blocking either electrically or mechanically evoked activity. Effective plasma doses were comparable to those found to relieve neuropathic pain. These data indicate that phase 2 in the formalin test is more closely related to ongoing afferent input than had previously been thought.