The medullary raphe nuclei contain putative central respiratory chemoreceptor neurones that are highly sensitive to acidosis. To define the primary stimulus for chemosensitivity in these neurones, the response to hypercapnic acidosis was quantified and compared with the response to independent changes in PCO2 and extracellular pH (pHo). Neurones from the ventromedial medulla of neonatal rats (P0-P2) were dissociated and maintained in tissue culture for long enough to develop a mature response (up to 70 days). Perforated patch clamp recordings were used to record membrane potential and firing rate while changes were made in pHo, PCO2 and/or [NaHCO3]o from baseline values of 7.4, 5 % and 26 mm, respectively. Hypercapnic acidosis (PCO2 9 %; pHo 7.17) induced an increase in firing rate to 285 % of control in one subset of neurones (‘stimulated neurones’) and induced a decrease in firing rate to 21 % of control in a different subset of neurones (‘inhibited neurones’). Isocapnic acidosis (pHo 7.16; [NaHCO3]o 15 mm) induced an increase in firing rate of stimulated neurones to 309 % of control, and a decrease in firing rate of inhibited neurones to 38 % of control. In a different group of neurones, isohydric hypercapnia (9 % PCO2; [NaHCO3]o 40 mm) induced an increase in firing rate of stimulated neurones by the same amount (to 384 % of control) as in response to hypercapnic acidosis (to 327 % of control). Inhibited neurones also responded to isohydric hypercapnia in the same way as they did to hypercapnic acidosis. In Hepes-buffered solution, both types of neurone responded to changes in pHo in the same way as they responded to changes in pHo in bicarbonate-buffered Ringer solution. It has previously been shown that all acidosis-stimulated neurones in the medullary raphe are immunoreactive for tryptophan hydroxylase (TpOH-ir). Here it was found that TpOH-ir neurones in the medullary raphe were immunoreactive for carbonic anhydrase type II and type IV (CA II and CA IV). However, CA immunoreactivity was also common in neurones of the hypoglossal motor nucleus, inferior olive, hippocampus and cerebellum, indicating that its presence is not uniquely associated with chemosensitive neurones. In addition, under the conditions used here, acetazolamide (100 μm) did not have a significant effect on the response to hypercapnic acidosis. We conclude that chemosensitivity of raphe neurones can occur independently of changes in pHo, PCO2 or bicarbonate. The results suggest that a change in intracellular pH (pHi) may be the primary stimulus for chemosensitivity in these putative central respiratory chemoreceptor neurones.