ALBERT

Notes: Summary 1. Wind directed at the cerci of tethered (Fig. 1 A) or dissected (Fig. 1B) locusts initiates rhythmical activity in elevator and depressor flight muscles. 2. Electrical stimulation of the cercal nerve (200 Hz) also evokes flight motor activity which then continues in the absence of further stimulation (Fig. 2A). Rhythmical activity in DL (depressor) flight muscles begins 168.2±21.4 ms (mean±s.d.) after cercal stimulation. Wind directed at the cerci during such activity increases the frequency of rhythmical bursts in DL motor neurones (Fig. 2B), and reinitiates rhythmical activity in flight interneurones and muscles following a previous flight sequence (Fig. 2C). 3. Depolarizing current injected into either GIN 2 or 4 in the terminal ganglion evokes rhythmical activity in thoracic flight muscles (Fig. 3). This evoked activity begins during the current pulse but only rarely continues beyond the termination of current injection. 4. The rhythmical depolarizations in an elevator flight interneurone (301) are identical in shape whether evoked by wind on the head (Fig. 4A) or wind stimulation of the cerci (Fig. 4 B). Both pathways activate the same thoracic oscillator. The membrane potential of one of the 4 GINs in the terminal ganglion (GIN 2) also oscillates during flight motor activity (Fig. 5A). The oscillations remain as tightly in antiphase to the activity of depressor flight muscles as those in a thoracic interneurone of the central flight oscillator (Fig. 5 B, C). 5. Wind and electrical stimulation of the cerci evoke EPSPs, IPSPs, or both, in a number of thoracic interneurones and motor neurones (Figs. 6, 7; Table 1). In some neurones both the polarity and latency of responses depend on which cercus is stimulated. A group of thoracic interneurones and one motor neurone respond with a latency sufficiently short so as to suggest a direct input from cercal GINs (Fig. 8). 6. Simultaneous intracellular penetrations of cercal GINs and interneurones of the flight motor pathway demonstrate short latency connections between GIN 1 and FIN 302 (dep), GIN 1 and FIN 301 (ele), GIN 4 and FIN 710 (ele), GIN 4 and FIN 320 (ele) (Figs. 9, 10, 11). 7. In the jump motor pathway, four cercal interneurones are found to evoke EPSPs in thoracic premotor interneurone 714 (Fig. 12). These EPSPs could then be identified in an actual response of neurone 714 to cercal stimulation (Fig. 13). Interneurone 714 in turn makes an excitatory connection with the FETi jump motor neurone in the metathoracic ganglion (Fig. 14). 8. The connectivities established between cercal GINs and neurones of thoracic flight (Fig. 15), and jump (Fig. 16), motor pathways are summarized.

Notes: Summary 1. The terminal ganglion ofLocusta migratoria contains a number of non-giant, wind-sensitive, ascending and local interneurones. Six ascending (Figs. 1, 2) and 6 local (Figs. 6, 7) interneurones have been identified morphologically on the basis of intracellular stains with Lucifer Yellow. 2. The physiological responses of the various cell types were recorded as the cerci were exposed to sound, wind, or electrical stimulation (Figs. 3, 8). Some cells summate the input from both cerci (Fig. 3), while others are excited by input from one side and inhibited by input from the other (Fig. 8). Conduction velocities for several non-giant ascending interneurones range from 1.5 m/s (cell 1) −2.1 m/s (cell 25). 3. The morphologies and physiological responses of giant (GIN 1) and non-giant ascending interneurones (cells la, b) with somata in cluster 1 of neuromere 9 were compared using simultaneous intracellular recordings (Figs. 2A, 4). These neurones have very similar dendritic arborizations (Fig. 4A, B), and respond almost identically to cercal stimulation (Fig. 4Ci), but there do not appear to be any connections with GIN 1 (Fig. 4Cii, iii). 4. The morphology (Fig. 5A, C), and response to cercal stimulation by wind (Fig. 5B) of a nongiant interneurone (cell 7) with its soma in cluster 1 of segment 8 (Fig. 5), are very similar to those of cluster 1 cells such as GIN 1 in segment 9. 5. Of the 6 local interneurones (Figs. 6, 7) all except one (cell 9) have bilateral arborizations which may extend over several neuromeres within the ganglion (cells 10, 22). Several of the interneurones (cells 5, 9, 24) do not produce action potentials in response to cercal stimulation (Figs. 8, 10) or injection of depolarizing current (Fig. 11). 6. Simultaneous recordings from pairs of interneurones demonstrate that giants and locals (GIN 2/cell 5; GIN 1/cell 9), as well as different local interneurones (cell 24/cell 5), receive input from the same wind-sensitive filiform afferent (Fig. 9). 7. Local interneurones 5 and 22 are in different neuromeres of the terminal ganglion but have a similar gross morphology (Figs. 6, 7, 10). Cell 5, however, has arborizations projecting into both posterior cercal glomeruli (Fig. 7 A, inset), whereas only the ipsilateral branches of cell 22 extend posteriorly to the cercal glomerulus (Fig. 10C). Physiologically, cell 5 is depolarized by wind directed at both cerci (Fig. 10 A), cell 22 mainly by wind directed at the ipsilateral cercus (Fig. 10C). Cell 5 does not produce action potentials in response to wind whereas cell 22 does. 8. Cell 5 occurs as a bilateral pair in the terminal ganglion (Figs. 7B, inset; 11). Simultaneous recordings of the bilateral homologues show that they share the input of at least one wind-sensitive filiform afferent (Fig. 11D), and that there are no connections between them (Fig. 11E). Simultaneous penetrations of local interneurone 5 and giant interneurones demonstrate a short-latency excitatory connection from GIN 3 to cell 5 (Fig. 12 A), and a long-latency excitatory connection from GIN 2 to cell 5. 9. The roles of giant and non-giant interneurones in transmitting information to thoracic motor centres are discussed.

Notes: Summary 1. The cerci of locusts are paired, unsegmented, cone-shaped structures arising from a depression on either side of the tip of the abdomen (Fig. 2). 2. The cerci bear hair sensilla of two basic types (Fig. 3): (a) filiform hairs, which emerge from a cuticular pit, are flexibly mounted, and are sensitive to wind and to mechanical displacement; and (b) bristle (or trichoid) hairs, which originate directly from the surface of the cercus, and are less flexibly mounted. The axons of neurones innervating these hairs group together into successively larger bundles before joining the cercal nerve which runs to the terminal ganglion (Fig. 4). 3. In the terminal ganglion the majority of cercal afferents (Fig. 5), and all identified filiform afferents (Fig. 6) end in contact with a cercal glomerulus formed by the densely interwoven arborizations of giant and non-giant interneurones (Figs. 5, 7,9). 4. A transverse section of the ventral nerve cord anterior to the terminal ganglion reveals four axons, one medial and three grouped together dorsolaterally, with distinctly larger profiles than all the others (Figs. 8, 9). The somata associated with these axons are located contralaterally in the terminal ganglion. All four interneurones have at least some projections into both cercal glomeruli. The soma of giant interneurone 1 (GIN 1), which is associated with the large medial axon, lies laterally in neuromere 9, the somata of GINs 2–4 are located ventrally in neuromere 9, ventrally in neuromere 10, and postero-dorsally in neuromere 11, respectively (Figs. 7,9). 5. All four GINs run the length of the ventral nerve cord (Figs. 10, 11, 12) and end in the brain (Fig. 11). The axon of GIN 1 is found in the ventral intermediate tract (VIT), those of GINs 2, 3, 4 are initially in the lateral dorsal tract (LDT) but they cross to the dorsal intermediate tract (DIT) at the level of the most anterior free abdominal ganglion. All have medial branches in the abdominal ganglia (Figs. 10, 12). GIN 1 has a lateral branch (sometimes absent) only in the metathoracic ganglion, while the other GINs have lateral branches in all of the thoracic ganglia (Figs. 10, 11, 12). 6. The cercal receptor/giant interneurone system ofLocusta is compared to those of other orthopteroid insects, with special attention to the origin and evolution of giant interneurones.

Notes: Summary 1. The cercal receptor/giant interneurone system of the locustLocusta migratoria has been investigated with respect to input to giant interneurones (GINs) from wind-sensitive filiform hairs on the cercus. Each filiform hair is innervated by a single receptor neurone which is depolarized when the hair is deflected by wind or mechanical stimuli (Fig. 1A, B). The resulting action potentials travel in the cercal nerve with a velocity of 1.0±0.2 m/s and arrive in the terminal ganglion of the central nervous system 3.5±0.6 ms after stimulus onset (Figs. 1C, 2). 2. Within the terminal ganglion EPSPs are evoked 1∶1 in a giant interneurone (GIN) by action potentials in a filiform afferent (Fig. 3). Three of the 4 GINs in the terminal ganglion receive excitatory input from filiform afferents on both cerci — GIN 4 appears to receive little or no excitatory input from the ipsilateral cercus (Fig. 4). The input from several filiform afferents converges onto a given GIN (Fig. 5A), and dual recordings show that the same afferent can evoke EPSPs simultaneously in two GINs (Fig. 5B). 3. Each of the four GINs in the terminal ganglion has a characteristic response to wind directed at the cerci, however, they fall into two broad groups: GINs 1 and 3 are excited about equally by wind directed at either cercus, whereas GINs 2 and 4 are strongly excited by wind directed at the contralateral cercus but weakly, and with a mixture of excitation and inhibition, to wind directed at the ipsilateral cercus (Fig. 6). This is consistent with anatomical data. The responses of GIN 1 are phasic, those of the other GINs more tonic. Electrical stimulation of each cercal nerve evokes responses in the GINs which parallel those produced by wind. The EPSP evoked in each GIN by electrical stimulation of the cercal nerve does not decrement appreciably even at high (100 Hz) stimulus frequencies (Fig. 6). 4. Dual intracellular penetrations reveal synaptic connections between some of the GINs in the terminal ganglion (Fig. 7). In general: excitatory connections are only made between ipsilateral neurones; excitatory connections are unilateral, and from neurones with more posterior cell bodies to those with more anterior cell bodies; connections are strongest between neurones belonging to the same group with respect to the directionality of their responses to wind stimulation; reciprocal inhibitory connections occur between contralateral neurones of different type; and all connections result in only subthreshold postsynaptic activity. Dual penetrations of left and right partners of the same GIN type failed to reveal either spiking or graded interactions between them (Fig. 8). 5. Intracellular recordings from GINs in the terminal ganglion and simultaneous extracellular monitoring of action potentials in their axons in the ventral nerve cord gave conduction velocities of 2.6–4.0 m/s for the different GINs (Fig. 9). This correlates well with the measured axon diameter for each GIN. 6. In the case of GINs 1 and 4, dual intracellular recordings from the same cell (in the terminal ganglion and in the posterior part of the metathoracic ganglion) show that action potentials are conducted 1∶1 to the anterior electrode with a delay of approximately 6–8 ms (Fig. 10). Lucifer Yellow dye injected at both recording sites identified the cell and confirmed the metathoracic arborizations seen for each neurone in cobalt backfills (previous paper).

Notes: Summary 1. Twelve types of wind-sensitive neurone have been identified in the terminal ganglion of the mantids,Archimantis sobrina andA. latistylus (Fig. 1). Nine have conspicuously larger axons than the others in the connective of the ventral nerve cord and are termed ‘giant’ interneurones (Figs. 1, 2, 3), although they are small by comparison with those of other orthopteroid insects (Fig. 15). 2. Transverse sections of connectives reveal nine large axon profiles whose spatial relationship changes along the ventral nerve cord (Fig. 4). 3. Transverse sections of ganglia containing stained giant cells show that major arborizations are found in neuropilar regions containing the terminals of cereal afferents (Fig. 5). 4. Recorded cells could be divided into two groups depending on whether their responses to wind stimuli were purely excitatory (Figs. 6, 7), or contained an inhibitory component (El responses, Fig. 8). Electrical stimulation of cercal afferents confirmed the response patterns evoked by wind (Figs. 7, 8 and 9). 5. Dendritic arborizations are more strongly developed on the side of major synaptic input to each giant cell (Figs. 1, 2, 5) as established by electrical and wind stimulation of afferents in left and right cerci (Fig. 10). 6. Conduction velocities in the giant cells ranged from 2.5–3.1 m/s (Fig. 11, Table 1).These neurones are thus amongst the slowest conducting insect giant neurones, consistent with their small diameter relative to those of other insects (Fig. 15). 7. The short response latencies and the 1∶1 nature of their response to high frequency (100 Hz) electrical stimulation of the cereal nerve indicated that at least neurone Types 5, 8 and 11 (Fig. 11, Table 1) probably have direct connections with cereal receptors. 8. Several cells had high rates of spontaneous activity and in one (Type 5, Fig. 12A), injection of hyperpolarizing current produced a rhythmical bursting at approximately half the spontaneous rate. The interburst interval could be altered by phasic stimulation of the cereal nerve (Fig. 12 B). 9. Behavioural experiments with a tethered mantid showed responses in leg- and flight-motor pathways to wind stimulation of the cerci. In a dissected preparation, electrical stimulation of abdominal connectives and wind stimulation of the cerci both evoked responses in four neurones of the metathoracic ganglion: one spiking local interneurone and three motoneurones (Figs. 13, 14).