The dipteran 'Giant fibre' pathway: neurons and signals

1. The uniquely identifiable pair of giant descending neurons of Musca domestica and Calliphora erythrocephala are compared and described in relation to homologues in Drosophila. 2. Cobalt coupling in the giant fibre system suggests that giant descending neurons receive inputs from antennal mechanosensory afferents and small-field visual interneurons. 3. In Musca and Calliphora, terminals of giant descending neurons invade several discrete areas of neuropil. This is in contrast to Drosophila where homologous neurons have a relatively simple terminal. 4. Despite differences in morphology, in all three species a homologous terminal region is cobalt-coupled to the largest of three motorneurons supplying the second leg's tergotrochanteral muscle. 5. Sensory connections revealed anatomically were confirmed electrophysiologically by intracellular recording and dye-filling. Mechanosensory inputs from the ipsilateral antenna, and small field motion stimuli, or light 'on' and 'off' stimuli to the ipsilateral compound eyes, produce subthreshold activation in the giant descending neuron. 6. As in Drosophila, light 'on' or 'off' stimuli presented to white-eyed mutants of Musca and Calliphora elicit a spiking response in the giant descending neuron. In red-eyed flies, spikes could only be routinely induced by electrical stimulation across the head between the compound eyes or by releasing the cell from hyperpolarization. 7. Double recordings from the giant (intracellular combined with lucifer or cobalt iontophoresis for cell identification) and from the tergotrochanteral motorneuron (extracellular in muscle) show that a spike in the giant neuron is usually sufficient to drive a spike in the tergotrochanteral muscle. Latencies are short, as reported in Drosophila. Releasing the giant from hyperpolarization can result in a motorneuron spike in the absence of a spike in the giant: this suggests the presence of an electrical synapse. 8. Spiking in the tergotrochanteral muscle and the giant can occur independently of each other. Thus, this interneuron alone is neither necessary nor invariably sufficient for initiating 'escape behaviour' in the species studied. This is also corroborated by the structure of tergotrochanteral motorneurons whose extensive dendrites receive a variety of other inputs.