In addition to the wellknown fictive locomotion, a fictive respiration can also be obtained in decorticate, unanaesthetized rabbit preparations after curarization and vagotomy. Both patterns were abolished after high spinal (C2 or C3) transection. Spinal rhythmic capabilities could be disclosed after administration of nialamide and DOPA: together with the earlier demonstrated locomotor-like bursting in hindlimb and forelimb muscle nerves, two different types of phrenic bursting patterns could be observed, depending on endtidal CO2 levels: (1) short lasting phrenic bursts (SLPBs), coordinated with locomotor bursts, result of a locomotor driving process; (2) when end-tidal CO2 was slightly increased (4.5 instead of 4.0%), long lasting phrenic bursts (LLPBs) developed: they have no causal link with the locomotor bursts. Intraspinal interactions were shown to operate between these rhythmic patterns: (1) the already mentioned caudo-rostral driving from hindlimb or posterior locomotion generators (pLGs) onto forelimb bursting and onto phrenic activity too (providing SLPBs in the latter case); (2) the rostro-caudal inhibition of fore- and hindlimb locomotor activity throughout each LLPB. Since forelimb locomotor-like bursting and LLPBs could still be obtained after functional isolation of the cervico-thoracic cord (through C2 and Th12 spinal transections) with comparable interactions as before Th12 transection, it is concluded that: two categories of generators, forelimb or anterior locomotion generators (aLGs), and chemosensitive respiration generators (RGs) are both present in this part of the cord, on the one hand; interactions between RGs and pLGs are likely to be achieved via aLGs on the other.