Data were obtained from 45 anesthetized (Dial), paralyzed, artificially ventilated, bilaterally vagotomized cats. Arrays of extracellular electrodes were used to monitor simultaneously the activities of lateral medullary respiratory neurons located in the rostral and caudal regions of the ventral respiratory group. The average discharge rate as a function of time in the respiratory cycle was determined for each neuron and concurrent phrenic nerve activity. Most cells were tested for axonal projections to the spinal cord or the ipsilateral vagus nerve using antidromic stimulation techniques. Seven hundred and sixty-one pairs of ipsilateral respiratory neurons that contained at least one neuron whose maximum discharge rate occurred during the inspiratory phase were analyzed by cross-correlation of the simultaneously recorded spike trains. Twenty-three percent of the 410 pairs of inspiratory (I) neurons showed short time scale correlations indicative of functional association due to paucisynaptic connections or shared inputs. Eight per cent of the 351 pairs composed of an I cell and and expiratory (E) neuron were correlated. We found evidence for excitation of both bulbospinal I neurons and I cells that were not antidromically activated by stimulation of the spinal cord and vagus nerve (NAA neurons) by NAA I cells. We also obtained data suggesting inhibitory actions of cells whose maximum discharge rate occurred in the first half of the I phase (I-DEC neurons). These actions included inhibition of other I-DEC neurons, inhibition of cells whose greatest firing rate occurred in the last half of the I phase (I-AUG neurons), inhibition of E-DEC neurons, and inhibition of E-AUG cells. Sixty-two percent (31/50) of the correlations that could be interpreted as evidence for an excitatory or inhibitory paucisynaptic connection were detected in pairs composed of a caudal and a rostral ventral respiratory group neuron. Eighty-eight percent (14/16) of proposed intergroup excitatory connections involved a projection from the rostral neuron of the pair to the caudal cell, whereas 73% (11/15) of proposed inhibitory connections involved a caudal-to-rostral projection. These results support and suggest several hypotheses for mechanisms that may help to control the development of augmenting activity in and the timing of each phase of the respiratory cycle.