Sex-determining mechanism inBuergeria buergeri (Anura Rhacophoridae). III. Does the ZZW triploid frog become female or maleкод для вставкиСкачать
270 K. JACOBS JOURNAL ETOF AL. EXPERIMENTAL ZOOLOGY 283:270–285 (1999) Tympanal Receptor Cells of Schistocerca gregaria: Correlation of Soma Positions and Dendrite Attachment Sites, Central Projections and Physiologies KIRSTEN JACOBS, BEATRIX OTTE, AND REINHARD LAKES-HARLAN* Institut für Zoologie und Anthropologie, Georg-August-Universität Göttingen, 37073 Göttingen, Germany ABSTRACT By using neurobiotin as a marker in intracellular recordings, we were able to directly correlate soma positions and dendrite attachment sites as well as axonal morphologies and physiologies of single auditory receptor cells of Schistocerca gregaria. We could clearly discriminate three groups of receptor cells, differing in their orientation within the Müller’s organ, their central arborizations and their physiology: Group I comprises 20 receptor cells with their dendrites attached to the “folded body.” Their characteristic frequencies (CFs) lie at 400–700 Hz or at 1.5–2 kHz. Group II consists of 12–14 high frequency receptor cells (CFs 12–25 kHz) whose dendrites are attached to the “pyriform vesicle.” Group III receptor cells dendrites are attached to either the “elevated process” (EP) or to the “styliform body” (SB); their CFs lie at 3–4 kHz. There were no differences in physiology and central arborizations between those receptor cells of Group III whose dendrites are attached to the SB and those whose dendrites are attached to the EP. Our method renders it possible to combine previous classifications based on either exclusively morphological (a-, b-, c-, d-cells) or physiological (type 1–type 4 cells) findings. In contrast to the hitherto hypothetical indirect correlations, we correlate c-cells and type 1 cells (= group I; see above) and d-cells to type 4 cells (= group II). Furthermore, we demonstrate that a subdivision of a-cells and b-cells is not reflected in a subdivision of type 2 and type 3 cells. The latter have to be combined into one group (= group III). J. Exp. Zool. 283:270285, 1999. © 1999 Wiley-Liss, Inc. The grasshopper Schistocerca gregaria has one ear on each side of the first abdominal segment. An ear consists of a tympanal membrane, tracheal sacs and a so-called Müller’s organ, which contains the somata of the tympanal sensory cells (a-, b-, c-, and d-cells; Schwabe, ’06; Gray, ’60). The Müller’s organ is attached to several sclerotized structures on the inner surface of the tympanal membrane. According to Gray (’60) these structures are called “folded body” (FB), “elevated process” (EP), “styliform body” (SB) and “pyriform vesicle” (PV). Their functional role is apparently to transmit vibrations of the tympanum to the sensory cells leading to sensory transduction (Stephen and Bennet-Clark, ’82; Breckow and Sippel, ’85). Horridge (’61) and Popov (’65) showed first that locusts are able to discriminate not only the amplitude, but also the pitch of sounds. Later on, Michelsen (’71a) described different response spectra for cells situated in different parts of the Müller’s organ. He suggested (’71b) that the sclerites to which the sensory cells are attached might © 1999 WILEY-LISS, INC. act as “passive observers” of tympanal motion that he shows to be already frequency dependent. Therefore, the frequency-dependence is transmitted to the different groups of sensory cells. Michelsen (’71b) also presumed that the mass of the Müller’s organ affects the mechanical frequency response of the whole system. This idea has been confirmed by Stephen and Bennet-Clark (’82) and by Breckow and Sippel (’85). Using intracellular recordings of processes of tympanal receptor cells within the metathoracic ganglion and subsequent marking of the cells, four groups of receptor cells have been distinguished according to their physiological properties and central projections (type 1 – type 4 receptor cells; Römer, ’76; Römer, ’85; Halex et al., ’88). Due to Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: SFB 406, Teilprojekt A6. *Correspondence to: Reinhard Lakes-Harlan, Institut für Zoologie und Anthropologie, Georg-August-Universität Göttingen, Berliner Str. 28, 37073 Göttingen, Germany. E-mail: firstname.lastname@example.org Received 20 February 1998; Accepted 16 June 1998. TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA methodological difficulties it has so far been impossible to determine simultaneously physiological characteristics, central arborizations and the morphology within the Müller’s organ of single auditory receptor cells. Therefore, to date the correlation of these features has been made not with measurements at the same time, but only indirectly: Breckow and Sippel (’85) have measured resonant frequencies of the Müller’s organ and of the sclerotized regions of the tympanal membrane on different ages and species of locusts. Afterwards, they compared changes occurring in amplitude of oscillations with changes in sensitivity of physiologically distinguishable groups formerly described by other authors. Based upon this comparison, they presented a hypothetical correlation between the anatomical and physiological classification of the receptor cell groups (a-cells = type 1 cells; b-cells = type 3 cells; c-cells = type 2 cells; d-cells = type 4 cells). In the present paper we labeled single auditory receptor cells with neurobiotin during physiological recordings in order to obtain their anatomy within the Müller’s organ and their central projections together with their threshold curves. This method has been successfully used in bushcrickets to characterize the anatomy of tympanal receptor cells (Stumpner, ’96). With this method, it was possible to check how a classification of sensory cells within the Müller’s organ translates into a somatotopic or tonotopic organization within the auditory neuropil. Additionally, we investigated response characteristics of these cells to frequencies in a range between 100 Hz and 40 kHz, thus also concerning very low frequencies that have not been tested in all previous physiological studies on tympanal receptor cells. To a great extent, our results are not in accordance with the indirect correlation made by several authors. Therefore, we present and discuss an updated grouping of tympanal sensory cells. 271 ings of nerve 6 of the metathoracic ganglion (tympanal nerve) with neurobiotin. The nerve was cut about 500 µm distal to the ganglion using a fine pair of scissors. The cut ends of the nerve were immersed in a neurobiotin solution (5% neurobiotin in 1 M potassium acetate [KAc]) for 12 hr at 4°C. Afterwards, the thoracic ganglia and the subesophageal ganglion as well as the ear were removed from the animal and fixed for 1 hr in 4% paraformaldehyde in PB (0.5 M Na2HPO4; 0.5 M NaH2PO4 × 2H2O). To permeabilize the tissue, the preparations were dehydrated, incubated in xylene for 5min, rehydrated and treated for 1 hr with collagenase/hyaluronidase (Sigma; each 1 mg/ml in PB). After this treatment, the preparations were incubated overnight in a solution of avidin with biotinylated horseradish peroxidase (Vector Elite Kit) at room temperature. The preparations were washed in PB several times and incubated in diaminobenzidine and nickel chloride (Vector Kit SK 4100) for 1 hr at room temperature before transferring the preparations to a solution of H2O2 in PB. When labeled neurons became visible, the reaction was stopped with PB. The preparations were again dehydrated, cleared in methyl salicylate and viewed using a Leica DMR microscope. Acoustic stimulation Recordings were performed in an unechoic chamber. Sound stimuli were presented either by a high-frequency loud speaker (Dynaudio DF21; 2–40 kHz flat frequency response) or by a low frequency speaker (W111; 100 Hz – 7 kHz flat frequency response). Pure sine-wave tones (100 Hz steps within 100 Hz to 1 kHz, 500 Hz steps within 1 to 4 kHz, 1 kHz steps within 4 to 12 kHz, and 5 kHz steps from 15 to 40 kHz) were generated by a personal computer (Lang et al., ’93). They were presented ipsilateral to the recording site at intensities of 35–90 dBSPL (5 or 10 dB increments). Duration of tone pulses was either 100 ms or 20 ms with a repetition rate of 3.3 Hz. MATERIALS AND METHODS Animals Recording of tympanal nerve activity All experiments were performed with adult male and female Schistocerca gregaria (L.), 90 animals in total. The animals were obtained from crowded cultures at the I. Zoological Institute, University of Göttingen. For extracellular recordings, animals were fixed ventral side up on a platform. A small window in the sternum was made by cutting the cuticle with a razor blade, and a hook electrode made of steel wire (diameter 30 µm) was placed underneath the tympanal nerve. Silicone paste (Roth) was applied to insulate the recording from the hemolymph and protect the nerve from drying out. The reference electrode was placed in the adjacent hemolymph. Neurobiotin backfills The anatomy of tympanal receptor cells was revealed by retrograde and orthograde axonal fill- 272 K. JACOBS ET AL. Fig. 1. A: Schematic drawing of housing the somata of auditory sensory cells in the Müller’s organ (MüO) and of their projections in the central nervous system. On the left side the metathoracic (TH3), mesothoracic (TH2), prothoracic (TH1) and subesophageal ganglion (SEG) are given in whole mount representation. Within the metathoracic ganglion, the auditory sensory fibers form two arborization areas, one within the first abdominal neuromere (caudal auditory neuropil; cNP) and one within the metathoracic neuromere (frontal auditory neuropil; fNP). In the other ganglia, only one arborization area can be found. B, C: Sagittal sections through the subesophageal (B) and the metathoracic (C) ganglion at the levels indicated in A (arrows). Fibers arborize in all ganglia within the ventral association centre (VAC) or within the median ventral association centre (mVAC). aVAC: anterior ventral association centre; c: caudal; d: dorsal; DIT: dorsal intermediate tract; DMT: dorsal median tract; f: frontal; Lb: labial neuromere; MVT: median ventral tract; Md: mandibular neuromere; Mx: maxillar neuromere; TyN: tympanal nerve; v: ventral; VIT: ventral intermediate tract; VMT: ventral median tract. Fig. 2. Photographs of Neurobiotin-marked auditory sensory neurones of Schistocerca gregaria. A, B: Zentrifugal backfill of the tympanal nerve (n) reveals positions of somata (s) and dendrites (d) of sensory neurones within the Müller’s organ. The dendrites are either projecting in direction of the folded body (FB), the elevated process (EP) or the styliform body (SB); dendrites projecting in direction of the pyriform vesicle are not shown. Cell bodies that belong to cells with dendrites that are attached to the EP or the SB cannot clearly be separated. Neurobiotin marks the whole dendrite up to its outer segment with the scolopale (arrowhead) and sometimes also the scolopale cell (c). Scale bars: 100 µm. C: Central projection of tympanal fibers within the subesophageal ganglion (arrowhead) in wholemount. The fibers arborize within the labial neuromere. In addition to tympanal sensory fibers, the cell body of one interneuron (in) is marked. TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA Scale bar: 100 µm. D: Sagittal section through the subesophageal ganglion shown in C. The arborization area of the marked fibers (arrowheads) dorsal to the ventral association centre (VAC) and the median ventral tract (MVT) might be homologous to the median VAC of the thoracic ganglia. Scale bar: 50 µm. E–J: Morphology within the Müller’s organ (F, G, I) and central projection within the metathoracic ganglion (E, H, J) of single tympanal sensory fibers. E: Arborization of group I sensory fibers within the caudal auditory neuropil (cn) and the frontal part of the frontal auditory neuropil (fn). Scale bar: 100 µm. H, J: Arborizations of group II sensory 273 fibers (J) and group III sensory fibers (H) within the frontal auditory neuropil (dotted line). Group II fiber arborizations are found caudal within the frontal auditory neuropil; group III fiber arborization lies between the arborization areas of the other two groups. Scale bars: 50 µm. F, G, I: The dendrite of a group I sensory fiber (F) projects to the folded body (FB), that of a group II sensory cell (G) to the pyriform vesicle (PV) and the dendrite of a group III sensory cell (I) either to the elevated process (EP) or the styliform body (SB). Scale bars: 100 µm. DMT: dorsal intermediate tract; VMT: ventral intermediate tract. 274 K. JACOBS ET AL. Single cell recording and marking A combined recording and staining technique was used to generate information about the central projections, the soma position within the Müller’s organ and the physiology of single receptor cells. The recordings were performed using thick-walled glass microelectrodes (borosilicate) filled with 5% neurobiotin in 1 M KAc (resistance: 50–200 MΩ) positioned in the tympanal nerve about 500 µm distal of the metathoracic ganglion. For stabilization of the recording, a spoon made of steel was placed beneath the nerve and slightly raised. The spoon also served as reference electrode. The responses of the single receptor fibers to tone pulses were tracked on an oscilloscope either and stored on magnetic tape (Racal Store 4DS) for subsequent analyses. Threshold was defined as an average of one spike per stimulus above spontaneous activity. After physiological characterisation, the receptor cells were stained by switching to positive current of >1 nA for at least 5 min. The preparation was then transferred to a moist chamber for 24– 48 hr to allow the neurobiotin to spread out within the whole cell. Afterwards, the ganglia and ears were processed as described above. For histology, ganglia were embedded in polyester wax (for a detailed description see Jacobs and Lakes-Harlan, ’97) and cut into parasagittal sections of 14 µm. In addition to these findings, which are in agreement with results of Halex et al. (’88), we were able to stain fiber arborizations within the subesophageal ganglion by using neurobiotin: In 10 out of 15 preparations one or two ascending fibers were marked rostrally to the prothoracic ganglion. In most of these preparations the staining of fibers ended within the connective between the prothoracic and the subesophageal ganglion. Nevertheless, in one preparation three fibers reaching the subesophageal ganglion were stained (Fig. 1; Fig. 2C). However, at least one of the stained arborizations in this preparation might belong to an interneuron, which has its soma located within the subesophageal ganglion. This interneuron was probably stained transsynaptically in the mVAC of the prothoracic or mesothoracic ganglion. Such a transsynaptic staining of interneurons occurrs only in very rare cases. The remaining two fibers showed some short arborizations within a neuropil region situated within the labial neuromere (Fig. 1; Fig. 2D). This neuropil region is located ventral to the VIT, lateral to the VMT and dorsal to the MVT, and it might be homologous to the mVAC of the thoracic ganglia. One of these two marked fibers ascended within the VIT up to the maxillar neuromere, where it ended without further arborizations. We found no marking within the mandibular neuromere. RESULTS Threshold curves of the whole tympanal nerve Central projection of tympanal sensory cells The axons of all tympanal sensory cells enter the metathoracic ganglion via the tympanal nerve, and once within the metathoracic ganglion the axons join the ventral intermediate tract (VIT; Halex et al., ’88; Pflüger et al., ’88). Here, they form two arborization areas, the caudal and frontal auditory neuropil (Fig. 1). The caudal auditory neuropil is build up by collaterals projecting into the ventral association centre (VAC) of the first abdominal neuromere. The frontal auditory neuropil is restricted to a neuropil region called the median VAC (mVAC), which is located within the metathoracic neuromere. The mVAC is situated between the median ventral tract (MVT), the ventral median tract (VMT) and the VIT. Some axons of tympanal sensory cells ascend to form further arborization areas within the meso- and the prothoracic ganglion. These arborizations are situated in neuropil areas which are serially homologous to the mVAC of the metathoracic ganglion. In order to extend threshold curves into the low frequency range, we stimulated tympanal sensory fibers with sound frequencies from 200 Hz to 7 kHz. Extracellular recordings from the tympanal nerve of 12 different animals revealed the highest sensitivity of the hearing-threshold-curve between 2 and 5kHz with a threshold between 37–40 dB SPL (Fig. 3). Receptor cells reacted to frequencies as low as 200 Hz with a threshold between 60 and 70 dB SPL. Positions of tympanal sensory cell somata and dendrites within the Müller’s organ Since there are four sites of attachment of dendrites to the tympanal membrane, it has been postulated that there are four different groups of sensory cells within the Müller’s organ. However, zentrifugal filling of the tympanal nerve with neurobiotin suggests a division of the sensory receptor cells into three instead of four groups (Fig. 2A, B; Fig. 4): Group I: This very homogenous group comprises TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA 275 about 20 sensory cells whose cell bodies lie in the ventral and laterocaudal direction towards the exit of the tympanal nerve (Fig. 4B). All cells from this group have dendrites that are attached to the “folded body” (FB; Fig. 2A). Group II: This second very homogenous group of tympanal sensory cells consists of 12–14 sensory cells with a laterocaudal and dorsal soma position within the Müller’s organ (Fig. 4C). The dendrites of these cells run exclusively in a small protrusion of the Müller’s organ named the “fusiform body” (FuB; Gray, ’60) which is connected to the “pyriform vesicle” (PV). Group III: The remaining cells belong to the third and last receptor cell group (Fig. 4D). The dendrites of these cells run perpendicular to those of the other two groups of auditory receptor cells and project either in the direction of the “elevated process” (EP) or in the direction of the “styliform body” (SB). The dendrites can be seen as a unit with respect to their spatial arrangement, as they are both arranged perpendicular to the dendrites of the other two cell groups. A subdivision of the somata of these tympanal sensory cells into two Fig. 4. Schematic drawing of tympanal sensory cells within the Müller’s organ as revealed by zentrifugal backfilling the tympanal nerve (tyn). The Müller’s organ is connected to four sclerotized structures of the tympanal membrane, the “elevated process” (EP), “folded body” (FB), “pyriform vesicle” (PV) and “styliform body” (SB). A: Distribution of somata and dendrites of all tympanal sensory cells within the Müller’s organ (view from caudal). B: Tympanal sensory cells of group I with a dendrite projection to the FB are drawn in black. The inset shows the position of these cells in a transverse section of the Müller’s organ (as indicated by arrow). C: Soma position and dendrite projection of tympanal sensory cells of group II. The scolopidia of these cells form a structure called “fusiform body” (FuB). D: Somata of group III tympanal sensory cells are arranged in a ring, their dendrites are attached either to the EP or the SB. More frontal cells are drawn in black, more caudal ones in dark gray. c: caudal; d: dorsal; f: frontal; l: lateral; m: median; v: ventral. Fig. 3. Threshold curve for the response of the tympanal nerve to acoustic stimuli with frequencies from 200 Hz to 7 kHz. For 12 animals, the mean (±S.D.; shaded area) is shown. 276 K. JACOBS ET AL. groups due to the insertion sites of their dendrites is not practicable as they are arranged in an almost ring-like structure: this arrangement starts at the border to the cell bodies of the first group, ventral and laterocaudal within the Müller’s organ, proceeds from there dorsally and to the most median and frontal part of the Müller’s organ (Fig. 4D, black cells) and finally returns in lateral and caudal direction again, but further dorsally (Fig. 4D, dark grey cells). Somata of cells with dendrites attached to the EP as well as of those cells with dendrites attached to the SB can both be found at nearly every position in this arrangement. Single cell physiology and morphology Intracellular recordings from auditory receptor cells and simultaneous application of the marker neurobiotin render it possible to correlate three features of these cells: the position of the soma and dendrite within the Müller’s organ, the central projection and the physiological qualities. We performed recordings from each of the three groups of receptor cells just described in the Müller’s organ. Examples of four cells of each group are depicted in Figures 5–7. Group I: The central projections (n = 15; morphology within the Müller’s organ see above; Fig. 2F; Fig. 5) of these cells show strong arborizations in the caudal auditory neuropil, whereas arborizations in the frontal auditory neuropil are more sparse in comparison to that of tympanal receptor cells of group II and group III (see below; Fig. 5; cell 1–4; Fig. 2E). Only 1–3 branches grow from the primary axon into the frontal part of the mVAC. They show few varicosities. 60% of the investigated cells ascend to the mesothoracic ganglion where they terminate in the VAC with few branches. In one preparation, a cell of this group was marked up to the prothoracic ganglion. According to the projection of their dendrites and their pattern of central projections within the caudal auditory neuropil, cells of group I can be further divided into two subgroups: those cells with dendrites inserting in the caudal part of the FB have caudal arborizations which are limited to the VAC of the first abdominal neuromere (Fig. 5, cells 1 and 2; also Fig. 2E). Cells with dendrites projecting to the more frontal part of the FB have caudal arborizations in the first abdominal neuromere and additional single collaterals that run into the second (Fig. 5; cell 3) or even near to the third abdominal neuromere (Fig. 5; cell 4). On the basis of their physiological properties, these cells can be characterized as insensitive low frequency receptors, since their thresholds always lie above 60 dB SPL. The response of these cells is generally limited to a frequency range from 300 to 400 Hz up to 4 kHz, in few cases up to 7–10 kHz (Fig. 5, cell 4). In general, all cells of group I have two areas of high sensitivity: a characteristic frequency of 1.5 kHz and a second range of high sensitivity at 400–700 Hz. Cells of the first subgroup tend to be more sensitive at 1.5 kHz than at 400–700 Hz (Fig. 5, cells 1 and 2); cells of the second subgroup seem to react the other way round (Fig. 5, cells 3 and 4). Group II: Tympanal sensory cells of the second group (n = 8; morphology within the Müller’s organ as described above; Fig. 2G) have numerous and thin arborizations in the frontal neuropil (Fig. 6; Fig. 2H). These emerge from a primary axon at the caudal end of the mVAC and cover the complete caudal portion of the mVAC. The arborizations extend dorsally into the mVAC and terminate close to the midline of the metathoracic ganglion. Additional smaller and less dense arborizations in more frontal areas of the mVAC can also be found (Fig. 6, cells 1, 3, and 4). All group II cells recorded from, ascend towards the meso- or, in most cases, even the prothoracic ganglion. All group II cells respond to a very broad frequency range (4–7 kHz up to over 40 kHz). Their characteristic frequencies lie in the high frequency range between 12 and 25 kHz; in this range the thresholds are between 40 dB SPL (Fig. 6, cell 4) and 60 dB SPL (Fig. 6, cell 3). This variation in sensitivity is also visible in recordings from several high frequency cells within the same animal. Group III: Tympanal sensory cells of the third group (n = 26; morphology within the Müller’s organ as described above; Fig. 2I) have thick and numerous central arborizations in the frontal auditory neuropil. They extend to the midline of the ganglion and cover the intermediate space between cells of groups I and II in caudal-frontal direction. The arborization patterns, as well as the density of arborizations, are variable: cells 1 and 2 (Fig. 7) have at least two first order projections that branch off from the primary axon. They spread within the mVAC in dorso-ventral direction and send off neurites of higher order that arborize extensively. In contrast, in cells 3 and 4 (Fig. 7) a single first order branch extends from the primary axon within the mVAC and sends off a variable but high number of branches of higher order in the frontal direction. Only 31% of the cells of group III have arborizations in the caudal neuropil. Forty-six percent of the marked cells ascend TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA Fig. 5. Central projection, physiology and morphology within the Müller’s organ of four different tympanal sensory cells belonging to group I. Central projection is shown both in whole mount preparation of the metathoracic ganglion and in sagittal section through the frontal auditory neuropil (section level: arrow). On the figure’s right is shown the soma and dendrite of the cell within the Müller’s organ. Each cell is shown 277 in median (top of the two organs) as well as in caudal (bottom of the two organs) view of the Müller’s organ. For better comparison two preparations have been put into one metathoracic ganglion and the left and right cell of the metathoracic ganglion has been drawn as if belonging to a right Müller’s organ. c: caudal; d: dorsal; f: frontal; l: lateral; m: median; mVAC: median ventral association centre; v: ventral. 278 K. JACOBS ET AL. Fig. 6. Central projection, physiology and morphology within the Müller’s organ of four different tympanal sensory cells belonging to group II. For details see Fig. 5. Soma posi- tion and dendrite projection of cell 3 confirms its belonging to group II, but it had not been drawn. towards the mesothoracic ganglion, 8% extend to the prothoracic ganglion. We could not find any correlation between the attachment of dendrites to either the EP or the SB and the caudal-frontal extension of the central arborizations within the frontal auditory neuropil. For example, the morphology of the cells 1 and 2 (Fig. 7) within the Müller’s organ are simi- TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA 279 Fig. 7. Central projection, physiology and morphology within the Müller’s organ of four different tympanal sensory cells belonging to group III. For details see Fig. 5. lar, but the central projection of cell 1 runs more caudally in the mVAC than that of cell 2. In contrast, cells 3 and 4 (Fig. 7) have different attachment sites of their dendrites (the dendrite of cell 3 projects towards the SB, that of cell 4 towards the EP), though they share the same area of arborization in the mVAC and have the same pattern of arborization. Only the density of their branching differs slightly. Figure 2J shows the central projections of two cells of this group that 280 K. JACOBS ET AL. have been recorded from in the same animal. The somata of these cells have clearly different positions within the Müller’s organ (not shown); the dendrite of one of these cells projects to the EP, the dendrite of the other to the SB. Within the mVAC though, the central projections of these two cells cover the same intermediate area. Physiologically the cells of group III can be described as low-frequency receptors with characteristic frequencies between 3 and 4 kHz. Within this group the threshold values as well as the covered frequency spectra vary considerably. Some cells are very insensitive and respond to a narrow range of frequencies between 2–7 and 10 kHz with a threshold of more than 60 dBSPL (Fig. 7; cells 1 and 3). With respect to their physiological properties these cells almost resemble group I cells. Other cells with a threshold of 40–50 dBSPL respond to frequencies up to 20–30 kHz (Fig. 7; cell 4). Most cells have been found in their sensitivity and covered range of frequencies to show values between these two extremes. There was no correlation between the dendrite projection sites and the sensitivities of these cells. Recordings and stainings from different cells of the same animal confirm the observation that cells of this group cover a wide range of frequencies and sensitivities. These results suggest that group III is very heterogeneous with respect to various parameters. No correlations between a given cell’s morphology in the Müller’s organ, its central projection pattern, and its characteristic frequency, threshold and frequency range were observed. Due to these results we can not corroborate with the often described subdivision of this group on the grounds of different insertion sites of the dendrites and different physiological properties of the cells. Fig. 8. Schematic summary of morphological and physiological characteristics of the three groups (I, II, III) of tympanal sensory cells. Group I comprises both the formerly described type-1 cells and c-cells, group II the formerly described type 4 cells and d-cells and group III the formerly described type-2 and -3 cells and a- and b-cells. A: Soma positions and dendrite attachment sites (EP: elevated process; FB: folded body; PV: pyriform vesicle; SB: styliform body) within the Müller’s organ. Their axons run within the tympanal nerve (TyN). B: Within the metathoracic ganglion, tympanal sensory cells form a caudal auditory neuropile (cNP) and a frontal auditory neuropile (fNP). They arborize tonotopic and somatotopic within the fNP. C: Schematic drawing of threshold curves. DISCUSSION Using neurobiotin as an intracellular marker we have determined central projections, morphologies within the Müller’s organ and physiologies of individual tympanal sensory cells. This correlation resulted in a classification of three groups of receptor cells primarily according to the positions of their somata and the attachment sites of their dendrites. This division is clearly reflected in the central projections and the physiological properties of the cells (Fig. 8). Since our results are not in accordance with TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA the common ideas about the morphology of the Müller’s organ, the arguments in favour of a division of the auditory receptor cells into three groups will be discussed later on. Classification of tympanal sensory cells The division of the tympanal receptor cells of the Müller’s organ into three groups differs from the former division into four groups that has been proposed by several authors. As early as 1960, Gray grouped the tympanal sensory cells of the Müller’s organ according to the insertion sites of their dendrites on the tympanal membrane (“styliform body,” “elevated process,” “folded body,” and “pyriform vesicle”). Gray named the four morphological groups a-, b-, c-, and d-cells (Table 1). Other authors divided the auditory sensory cells into four groups as well, by considering differing physiological characteristics and the arborization pattern within the frontal neuropil of the metathoracic ganglion (Table 2; type 1–type 4 cells; Römer, ’76; Petersen et al., ’82; Halex et al., ’88) as characteristics. Through a comparison of the physiological properties of cells with the resonance frequencies of the tympanal membrane at the attachment site of the different cell groups, (Michelsen, ’71b; Stephen and Bennet-Clark, ’82; Breckow and Sippel, ’85) to date Gray’s a-cells and the type 1 cells, his b-cells and the type 3 cells, his c-cells and the type 2 cells, and his d-cells and the type 4 cells have been correlated (Breckow and Sippel, ’85; Halex et al., ’88). These correlations were not completely confirmed by our experimental strategy of recording from the auditory sensory cells and marking them individually and completely with neurobiotin. Based on simultaneous examination of their soma postions, their central pro- 281 jections and their physiologies, we propose a subdivision into only three groups (Fig. 8). 1. One group of tympanal sensory cells has their somata positioned laterocaudally and far ventrally within the Müller’s organ and their dendrites project onto the FB. Within the frontal auditory neuropil of the metathoracic ganglion they form but few arborizations more frontally. They respond preferentially to stimuli of very low frequencies and high intensity, their characteristic frequency lying between 400–700 Hz and at 1,500 Hz respectively. The positions of their somata and the fact that their dendrites project onto the FB indicate that they correspond to the tympanal sensory cells that have been described by Gray as c-cells (’60). Michelsen (’71a) recorded extracellularly from these cells in the isolated ear close to the soma. According to his results their frequency of highest sensitivity lies at 1.5 or 3.5 kHz; he describes these cells as the least sensitive of all tympanal receptors. On account of their central projections and physiological characteristics group I cells match the cells that have been described as type 1 cells by Römer (’85) in Locusta migratoria and by Halex et al. (’88) in Locusta migratoria and Schistocerca gregaria. Type 1 cells form sparse arborizations frontal within the mVAC and stronger caudal arborizations. They belong to the least sensitive cells with the lowest characteristic frequency of all tympanal sensory cells (Römer ’76). The characteristic frequencies and thresholds that have been described for this cell-type differ slightly between the different species that were investigated, but also from those that were found in our experiments (Table 2). These differences can partly be explained from TABLE 1. Correlation between morphology within the ear and physiology of single tympanal sensory cells as revealed by recording from the Müllers organ in different species of locusts* Source Gray (’60), (Schistocerca gregaria) Michelsen (’71a), (Schistocerca gregaria) Inglis and Oldfield (’88), (Valanga irregularis) Nomenclature CF (kHz) Threshold (dBSPL) Somaposition Dendrite attachment site a-cells b-cells c-cells d-cells a-cells b-cells c-cells d-cells a-cells c-cells d-cells — — — — 3.74 3.4 1.5–3.5 10–14 4–6 2–3.5 14–25 — — — — 46 51 >55 >40 >45 >70 >40 Median, dorsal Median, ventral Laterocaudal, ventral Laterocaudal, dorsal — — — — Median Laterocaudal, ventral Laterocaudal, dorsal EP SB FB PV — — — — EP FB PV *CF, characteristic frequency; FB, ‘‘folded body’’; EP, ‘‘elevated process’’; PV, ‘‘pyriform vesicle’’; SB, ‘‘styliform body.’’ 282 K. JACOBS ET AL. TABLE 2. Correlation between physiology and central arborization of single tympanal sensory cells as revealed by recording and marking* Source Römer (’76, ’85), (Locusta migratoria) Halex et al. (’88), (Schistocerca gregaria) Nomenclature CF (kHz) Threshold (dBSPL) Arborization pattern within the mVAC Type 1 Type 2 Type 3 Type 4 Type 1 Type 2 Type 3 Type 4 3.5–4 4 5.5–6 12–20 2–3.5 2–3.5 6 >10 50–60 >20 >40 >40 50–70 20–40 30–60 35–60 Frontal Caudal-intermediate Frontal-intermediate Caudal Frontal Frontal-intermediate Caudal-intermediate Caudal *CF: characteristics frequency; mVAC: median ventral association centre. the fact that in the other studies the frequencyrange below 1 kHz was not tested. Thus, the results of this study make it most likely, that group I cells correspond to type 1 cells and to c-cells. This directly contradicts the hypothesis of Breckow and Sippel (’85), who predicted that the cells of type 1 correspond to the a-cells that have been described by Gray (’60) and that the type 2 cells correspond to the c-cells. Their results were exclusively based on a determination of differences in resonance-frequency of the tympanal membrane and of different parts of the Müller’s organ in different ages and species of locusts. 2. Cells of group II are characterised in the Müller’s organ by having laterocaudally and dorsally positioned somata, and by dendrites projecting to the PV. Therefore they correspond to the d-cells as described by Gray (’60). Their central projection is very homogenous, forming extensive, but fine arborizations in the caudal part of the frontal auditory neuropil. Physiologically the cells of this group can be described as highfrequency receptors with a varying characteristic frequency between 12 and 25 kHz. The thresholds of the single cells belonging to group II show great variation in sensitivity. These results are in accordance with those of other authors. Group II cells correspond to the formerly characterized type 4 high-frequency receptor cells on the basis of physiological data and on the grounds of their central arborization pattern (Table 2; Römer, ’76; Petersen et al., ’82; Römer, ’85; Halex et al., ’88). Measurements of the tympanal membrane movement (Michelsen, ’71b; Stephen and Bennet-Clark, ’82; Breckow and Sippel, ’85) also led to the conclusion that the insertion point of the d-cells must be the insertion point of high-frequency receptors. 3. The somata of the group III cells are arranged almost in a ring within the Müller’s or- gan. Their dendrites project in the direction of either the EP or the SB. Their central arborizations vary in form, but cover more or less the part of the frontal auditory neuropil between the other two groups. They can be classified as low frequency receptors with a characteristic frequency of 3 to 4 kHz and very divergent thresholds. Tympanal sensory cells of this group so far have been morphologically divided into a- and b-cells according to the description of Gray (’60). Other authors who have performed intracellular recordings in the tympanal nerve and the metathoracic ganglion found two cell groups with distinct physiological properties (type 2, type 3; Römer, ’76, ’85; Halex et al., ’88). For this reason it will be discussed in detail why in our opinion neither the morphology of the cells in the ear nor their physiology or central projection allow a further division and furthermore, why the cells of type 2 and 3 and the a- and b-cells can not be assigned. Morphology within the ear A morphological distinction between these cells has been introduced by Gray (’60), who described four separate groups of cell bodies of tympanal sensory cells. Their dendrites are attached to one of the four specialized structures on the tympanal membrane. However, our marking of all tympanal receptor cells within the Müller’s organ revealed that the cell bodies of those cells that project onto the SB and of those which connect in the area of the EP can not clearly be separated into two groups. Therefore only the different insertion points of the dendrites would justify segregation into two groups. When weighing this criterion it should be kept in mind that the EP and the SB and therefore all dendrites of this cellgroup have the same direction in space. Since this direction differs from that of the cells of group I and II, which are positioned perpendicular to each TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA other and to group III cells, it seems justified to view group III cells as a morphological unity. Physiology A clear distinction of the cells of group III, which comprises morphologically related a- and b-cells, on the grounds of their characteristic frequencies and their sensitivities is not possible. The characteristic frequencies of these cells all lie between 3 and 4 kHz with their thresholds spreading over a wide intensity-range. Neither in their characteristic frequency nor in their threshold a correlation with the insertion sites of the dendrites was found. This contradicts findings of Michelsen (’71a,c). He also found only slight differences in characteristic frequency of the a- and b-cells he recorded from (3.4 and 3.74 kHz), but he divided them into two physiologically distinct groups according to their different sensitivities. Since in his experiments the cells were not marked during recording, it seems possible that a clear distinction between the two types could not always be made, especially since they are arranged in a continuous ring and not in two separate groups. In addition, he performed his recordings of somata or dendrites of receptor cells directly in the ear, which might influence sensitivity of cells. Other publications report clear differences in characteristic frequencies and sensitivities of type 2 and type 3 cells (Römer, ’76; Petersen et al., ’82; Halex et al., ’88). We have never recorded from a cell with a characteristic frequency of 6 kHz (=type 3 in Schistocerca gregaria; Halex et al., ’88). It was possible though to show that cells with dendrites projecting onto the SB and onto the EP both have a characteristic frequency of 3 or 4 kHz, which corresponds to measurements of the resonant frequency at their attachment sites (Breckow and Sippel, ’85). Interestingly, Krahe (’91) in Locusta migratoria and Inglis and Oldfield (’88) in Valanga irregularis both identified only one further group of tympanal receptor cells in addition to the insensitive low-frequency receptor cells and the high frequency receptor cells (cells of group I and II). Central projection Römer (’85) and Halex et al. (’88) made a clear distinction between the cells of type 2 and 3, according to their central projections. The projections of both cell types were described to arborize in the metathoracic ganglion in the intermediate area of the frontal neuropil, one more fron- 283 tal, the other more caudal. The correlation of the frontal-intermediate and the caudal-intermediate arborizations with the corresponding cells characterized as type 2 or type 3 in the two publications is contradictory though (Table 2): Römer (’85) found type 2 cells arborizing more caudal than type 3 cells; Halex (’88) described type 2 cells to be located more frontal than type 3 cells. The cells that have been described as group III cells in this study arborize independently from the position of their somata and dendrites in overlapping areas in the intermediate part of the frontal arborization area. It can be concluded that the group I tympanal sensory cells correspond to the physiological characterized type 1 cells and the group II tympanal cells match the type 4 cells. Our results represent a first direct and unambiguous correlation of the c-cells to type 1 and that of the d-cells to type 4 cells. The a- and b-cells that have been described by Gray (’60) on the grounds of their morphology both belong to group III. With respect to their central arborization pattern, group III cells comprise both type 2 and type 3 cells. However, a-cells or b-cells do not correspond to type 2 or type 3 cells, or vice versa. Therefore, they can not be further subdivided neither according to the position of their somata nor to their physiology or their arborizations. Somatotopic order within the frontal auditory neuropil is tonotopic The somatotopic organization of tympanal receptor cells in the frontal auditory neuropil, which is described in this study, leads to a clear tonotopic organisation of the mVAC. Auditory sensory cells with a CF of 12–25 kHz arborize caudally, sensory cells with a CF of 3–4 kHz have an intermediate representation and insensitive receptor cells with a characteristic frequency of 400–1,500 Hz project frontally. In the most anterior part of the mVAC there are no arborizations of auditory receptor cells, but the axons of the vibration sensitive metathoracic myochordotonal organ and the axons of some abdominal chordotonal organs are shown to arborize in this part of the mVAC (Bräunig et al., ’81; Bickmeyer et al., ’92). Thus, the primary sensory information divides the mVAC into four parts, in which information of different frequency is processed. In 1985, Römer suggested a separation of high and low frequency information within the mVAC, which was confirmed by Halex et al. (’88). However, Römer (’85) concluded that the frontal au- 284 K. JACOBS ET AL. ditory neuropil is more importantly partitioned according to the threshold intensity of the receptor cells. Considering the new information presented in our study on the response of tympanal cells to frequencies below 2 kHz, we suggest that the auditory neuropil is indeed subdivided according to frequency rather than to intensity. Such a somatotopic order within a neuropil is realized in numerous sensory systems of insects, e.g. the cercal system of crickets (Bacon and Murphey, ’84; Jacobs and Theunissen, ’96) and the auditory system of bushcrickets (Römer, ’83; Oldfield, ’88; Stumpner, ’96). In bushcrickets, a tonotopic division of the auditory neuropil is also achieved by somatotopic arborizations of sensory cells which are aligned with this order (Römer, ’83; Oldfield, ’88). Auditory interneurons in the locust also show a frequency-specific tuning (Marquart, ’85; Römer et al., ’88). These interneurons partly achieve this tuning by having their dendrites restricted to certain parts of the mVAC. The best example for an overlapping of sensory cell projection and auditory interneurons with the same frequency tuning seems to be the SN5, whose threshold curve resembles that of the high-frequency receptors and whose postsynaptic structures are restricted to the caudal region of the mVAC (Römer et al., ’88). However, many interneurons achieve their specific tuning not by limiting their postsynaptic structures to a distinct neuropil areas, but also by a frequency dependent inhibition by other interneurons (Römer et al., ’81). ACKNOWLEDGMENTS The authors thank Drs. F. Lang, R. Heinrich and A. Stumpner for their critical comments on the manuscript, Dr. G. Ganter for improving the English language and Dr. N. Elsner for supporting the study. LITERATURE CITED Bacon JP, Murphey RK. 1984. Receptive fields of cricket giant interneurones are related to their dendritic structure. J Physiol Lond 352:601–623. Bickmeyer U, Kalmring K, Halex H, Mücke A. 1992. The bimodal auditory-vibratory system of the thoracic ventral nerve cord in Locusta migratoria (Acrididae, Locustinae, Oedipodini). J Exp Zool 264:381–394. Bräunig P, Hustert R, Pflüger H-J. 1981. Distribution and specific central projections of mechanoreceptors in the thorax and proximal leg joints of locusts. I. Morphology, location and innervation of internal proprioreceptors of proand metathorax and their central projections. Cell Tissue Res 216:57–77. Breckow J, Sippel M. 1985. Mechanics of the transduction of sound in the tympanal organ of adults and larvae of locusts. J Comp Physiol 157:619–629. Gray EG. 1960. The fine structure of the insect ear. Phil Trans R Soc (B) 243:75–94. Halex H, Kaiserm W, Kalmring K. 1988. Projection areas and branching patterns of the tympanal receptor cells in migratory locusts, Locusta migratoria and Schistocerca gregaria. Cell Tissue Res 253:517–528. Horridge GA. 1961. Pitch discrimination in locusts. Proc R Soc Lond B 155:218–231. Inglis M, Oldfield B. 1988. Tonotopic organisation of the auditory organ of the locust Valanga irregularis (Walker). J Comp Physiol A 164:49–53. Jacobs GA, Theunissen FE. 1996. Functional organization of a neural map in the cricket cercal sensory system. J Neurosci 16:769–784. Jacobs K, Lakes-Harlan R. 1997. Lectin histochemistry of the metathoracic ganglion of the locust Schistocerca gregaria before and after axotomy of the tympanal nerve. J Comp Neurol 387:255–265. Krahe R. 1991. Musterabbildungseigenschaften der tympanalen Rezeptoren von Locusta migratoria. Diploma thesis. Friedrich-Alexander-Universität NürnbergErlangen. Lang F, Brandt G, Glahe M. 1993. A versatile multichannel acoustic stimulator controlled by a personal computer. In: Elsner N, Heisenberg M, editors. Proceedings of the 25th Göttingen Neurobiology Conference. Stuttgart: Thieme Verlag. p 892. Marquart V. 1985. Auditorische interneurone im thorakalen nervensystem von heuschrecken: morphologie, physiologie und synaptische verbindungen. PhD thesis. Ruhr-Universität Bochum. Michelsen A. 1971a. The physiology of the locust ear. I. Frequency sensitivity of single cells in the isolated ear. Z Vergl Physiol 71:49–62. Michelsen A. 1971b. The physiology of the locust ear. II. Frequency discrimination based upon resonances in the tympanum. Z Vergl Physiol 71:63–101. Michelsen A. 1971c. The physiology of the locust ear. III. Acoustical properties of the intact ear. Z Vergl Physiol 71:102–128. Oldfield BP. 1988. Tonotopic organization of the insect auditory pathway. TINS 11:267–270. Petersen M, Kalmring K, Cokl A. 1982. The auditory system in larvae of the migratory locust. Physiol Entom 7:43–54. Pflüger H-J, Bräunig P, Hustert R. 1988. The organization of mechanosensory neuropils in locust thoracic ganglia. Phil Trans R Soc Lond B 321:1–26. Popov AV. 1965. Electrophysiological studies on peripheral auditory neurons in the locust. J Evol Biochem Physiol 1:239–250. Römer H. 1976. Die informationsverarbeitung tympanaler rezeptorelemente von Locusta migratoria (Acrididae, Orthoptera). J Comp Physiol 109:101–122. Römer H. 1983. Tonotopic organization of the auditory neuropile in the bushcricket Tettigonia viridissima. Nature 306:60–62. Römer H. 1985. Anatomical representation of frequency and intensity in the auditory system of Orthoptera. In: Kalmring K, Elsner N, editors. Acoustic and vibrational communication in insects. Berlin: Verlag Paul Parey. p 25–33. TYMPANAL RECEPTOR CELLS OF SCHISTOCERCA GREGARIA Römer H, Marquart V, Hardt M. 1988. Organization of a sensory neuropile in the auditory pathway of two groups of orthoptera. J Comp Neurol 275:201–215. Römer H, Rheinlaender J, Dronse R. 1981. Intracellular studies on auditory processing in the metathoracic ganglion of the locust. J Comp Physiol 144:305–312. Schwabe J. 1906. Beiträge zur morphologie und histologie 285 der tympanalen sinnesapparate der orthopteren. Zoologica 50:1–154. Stephen RO, Bennet-Clark HC. 1982. The anatomical and mechanical basis of stimulation and frequency analysis in the locust ear. J Exp Biol 99:279–314. Stumpner A. 1996. Tonotopic organization of the hearing organ in a bushcricket. Naturwissenschaften 83:81–84.