Control of Sound Production in the Syrinx of the Fowl Gallus Gallus

Electromyographic studies by Youngren, Peek & Phillips (1974) and Gaunt & Gaunt (1977) have shown that the extrinsic muscles of the fowl syrinx, the tracheo-lateralis (TL) and sternotrachealis (ST), are simultaneously active during a large part of the crowing cycle. This finding casts some doubt on the theory originally devised by Miskimen (1951), and since advocated by many authors, that it is the contraction of the ST that is responsible for bringing the syrinx into the sound-producing configuration and that this action is opposed by the TL. The theory predicts that, once the ST has brought about the initial displacement of the external tympaniform membranes into the syringeal lumen, further inward movement of the membranes will occur as a result of the passive suction forces generated by the accelerating airstream until a point is finally reached at which membrane vibration begins.

The observation that airway resistance increases markedly during vocalization has been taken as evidence of the syringeal constriction that would be expected of the original theory (Brackenbury, 1977, 1978 a, b) but the relative importance of active and passive factors in this process is not known. On the other hand several workers have demonstrated that spontaneous vocalization continues after de-activation of the extrinsic syringeal muscles. The present experiments were designed to shed further light on the interaction between muscular and aerodynamic forces during sound production. The method employed was to monitor the changes in airway resistance that result from stimulation of the muscles before and during artifical sound production and to relate these findings to alterations in the configuration of the syrinx.

Studies were carried out on a total of 14 male and female chickens aged 12–17 weeks. Each was anaesthetized with a mixture of sodium pentobarbitone (30 mg/ml) and ethyl carbamate (150 mg/ml) infused slowly into the wing vein, and prepared for unidirectional ventilation as follows. The skin was reflected from the interclavi-cular area, the crop displaced to one side, the interclavicular air sac (ICS) cannulated and a stream of warmed and humidified air led into the ICS and out of the beak. Airflow rate was measured before entering the ICS by a Fleisch pneumotachograph connected to a Grass PT5 manometer. A second manometer was used to monitor ICS pressure. Both signals were displayed on a Grass Model 7 recorder.

Spontaneous respiratory movements were abolished by raising the airstream to a level above the apnoeic threshold (approx, 1·2 1/min) and the skin incision was carried forward along a mid-ventral line from the crop to the larynx. The descending cervical branch of the left hypoglossal nerve was dissected free of the underlying TL muscle for a length of 2–3 cm in the mid or caudal cervical level, severed and its peripheral stump mounted on bipolar silver wire electrodes connected to a squarewave generator. Pulses of 5–7 V amplitude, and 0·3–0·5 ms duration, were delivered at 2, 5, 10, 20, 50 and 100 Hz. The nerve-muscle preparation was kept moist with Ringer at all times.

The ST was stimulated by paired, sharpened platinum-alloy electrodes inserted directly into the muscles (one pair to each muscle) after the ICS had been completely opened. In this case artificial ventilation was introduced via a mid-cervical tracheostomy, the direction of airflow was from the trachea to the lung air-sac system and pressure changes were monitored on the tracheal side of the syrinx. This is the reverse of the normal direction of airflow during vocalization and it is assumed that changes of syringeal configuration due to muscle stimulation would exert qualitatively similar effects on the measured flow resistance regardless of flow direction, i.e. a constriction of the syrinx would increase flow resistance and vice versa.

Some experiments involved raising the airflow rate to 3–5 1/min for brief periods in order to elicit passive sound production. In these cases, in order to alleviate the effects of excess CO₂ blow-off from the lung, artificial ventilation was discontinued between procedures thus permitting the resumption of normal breathing and the restoration of blood CO₂ levels.

Direct and indirect stimulation of the caudal 8–10 cm of the TL resulted in a caudo-cranial movement of the tympanum and a decrease in airway resistance (Fig. 1). The latter was presumably caused by the tensing and drawing apart of the ETMs and the consequent dilation of the syringeal lumen. No qualitative difference in response was observed when the muscle was stimulated indirectly at points 10 cm and 3 cm cranial to the ICS membrane. Since the descending cervical branch of the hypoglossal innervates both the TL and ST muscles (Youngren et al. 1974), the more caudal parts of this nerve might be expected to contain relatively greater numbers of motor fibres to the latter muscle. However, the present result suggests that if the ST were indeed being activated at the same time as the TL, its strength of contraction was never sufficient to reverse the action of the TL.

The contractile force of the TL, which could be measured by the total distance moved by the tympanum and the resultant fall in airway resistance, was a function of stimulation rate, rising from a negligible value at 5 Hz to a maximum at 100 Hz (Fig. 1). Vibratory movement of the muscle was visibly evident at the lower stimulation frequencies, transforming into a fused response at 100 Hz. The resistance traces, however, showed no signs of saw-tooth rise at subtetanic frequencies and summation was always smooth. It appeared that the motion of the trachea was highly damped by the viscoelastic drag of the extensible elements in the trachea and its confining connective tissue sheaths.

Direct tetanic stimulation of the cranial part of the TL produced a marked retraction of the larynx and a telescoping of the rostral parts of the trachea.

Direct bilateral stimulation of the paired ST muscles produced a graded rise in airway resistance according to stimulation rate (Fig. 2). The response time of the muscle was noticeably less than in the case of the TL and saw-tooth summation was evident in the resistance trace at subtetanic stimulus frequencies. A fused response appeared to occur at 50 Hz but overall contraction was greater at 100 Hz. The lower response time of the ST as compared to the TL may be due to differences in muscle loading rather than intrinsic differences in contractile properties. Since the ST acts upon only the relatively inextensible posterior parts of the trachea and the tympanum, its loading is relatively discrete and stiff, and its contraction markedly isometric. In contrast the trachea presents a comparatively extensible loading to the TL and the contraction of the latter is markedly isotonic. I am not aware of any histological differences between the TL and ST muscles but clearly information on this matter would be of value.

Direct stimulation of the cranial half of the TH produced a qualitatively similar effect to that of the TL, namely laryngeal retraction and tracheal shortening.

The syrinx can produce sound passively provided the instantaneous airflow rate and air pressure within the lung air-sac system exceed certain threshold values. The minimal values observed in these experiments were approximately 40 ml/s and 5 cm H₂O (5 × 10² N/m²) respectively, although in several animals larger values were necessary. Passive sound production could be abolished immediately by tetanic stimulation of the TL muscles (Fig. 3).

An alternative method for producing sound without increasing flow rate is to push the tympanum manually towards the pessulus, the cartilage lying at the opposite end of the syrinx (Fig. 6). This simulates the action of the ST muscle and produces an initial rise in airway resistance followed by a gradual rise in air sac pressure as the lung air sac system begins to accumulate air against the raised downstream pressure. At the same time the increased load on the supply stream may cause it to fall slightly but a point is eventually reached at which both ICS pressure and airflow rate are in excess of threshold values and the ETMs are triggered into vibration (Fig. 4a). The mechanism is sensitive to manipulation and over-displacement of the tympanum distorts the syrinx to such an extent that ICS pressure and airflow may rise and fall precipitously without achieving even momentarily a combination of values suitable for sound production (Fig. 4 b).

The latter is an example of gross impairment of the sound-production mechanism, but a situation can also be demonstrated in which marginal increases in pressure and decreases in airflow rate will lead to the abolition of existing sound. Fig. 5 shows the results of an experiment on a tracheostomised individual receiving unidirectional ventilation into the ICS. First, sound was induced by displacing the tympanum and fixing it in the effective position. Next, the resistance of the trachea was altered by tightening a screw-clip attached to the tracheostomy tube. Sound volume fell rapidly when a point was reached at which ICS pressure and airflow began to rise and fall almost imperceptibly; further tightening of the screw-clip produced measurable changes in the aerodynamic parameters and sound was abolished. Sound was restored by reopening the screw-clip and returning pressure and flow to their previous levels.

These experiments illustrate the necessity of achieving adequate simultaneous levels of ICS pressure and airflow in order to excite the ETMs. If the increase in ICS pressure is achieved only at the expense of an excessive decline in air flow, either by over-compression of the longitudinal axis of the syrinx or by an increase in downstream resistance as might result from active glottal constriction, the aero-mechanical coupling between the airstream and the ETMs fails to occur.

Workers interested in vocal mechanisms have classified the tracheal muscles as extrinsic muscles of the syrinx; others (McLelland, 1965; White, 1968: White & Chubb, 1968; King, 1975) have referred to them as the caudal extrinsic muscles of the larynx. The latter workers, moreover, regard the cranial attachment of the ST as being the larynx, not the trachea, and hence they call it the sternolaryngeus. They regard the ST and TL as forming a single muscle over the cranial length of the trachea, the sternotracheolaryngeus. However, since Youngren et al. (1974) and Gaunt and Gaunt (1977) have demonstrated the separate functional identity of these muscles, their terminology will be retained in the present paper in order to avoid confusion. It remains true, nevertheless, that the TL should properly be referred to as the tracheolaryngeus since it inserts cranially on to the larynx whilst the TH should be referred to as the sternolaryngeus since it attaches caudally on the sternum and cranially on the larynx.

Present results confirm that contraction of the ST leads to a shortening of the distance between the tympanum and the pessulus, a slackening of the ETMs and partial occlusion of the syringeal lumen (Fig. 2). They thus lend support to the view expressed by many workers that one of the important functions of the ST is to bring about the initial yielding of the ETMs that will ultimately lead to their becoming coupled in an exchange of energy with the airstream (Miskimen, 1951; Gross, 1964a; Chamberlain, Gross, Cornwell & Mosby, 1968; Greenewalt, 1968 ; Gaunt, Gaunt & Hector, 1976; Gaunt & Gaunt, 1977; Youngren et al. 1974; Brackenbury, 1978a, b). The ST could achieve these effects directly at its point of insertion by pulling the tympanum caudad, or indirectly by tensing the syringeal ligament which then pulls the pessulus forward, as described by Youngren et al and Gaunt & Gaunt. Whatever the precise mechanism, the coupling between the airstream and the ETMs can occur only if ICS pressures and airflow rates of sufficient magnitude exist at the time of muscular contraction. At subthreshold flow rates, such as those employed in Fig. 2, partial invasion by the ETMs takes place, as reflected by a rise in airway resistance, but the positive feedback of energy from the air to the elastic membranes is insufficient to trigger their oscillation.

The TL cannot be assigned a single role since both its extreme points of attachment, the caudal end of the trachea and the larynx, are capable of acting as either origin or insertion depending on their degree of stabilization by other muscles. Thus, if its posterior point were anchored by the ST, and if the rostral extrinsic laryngeal muscles were simultaneously relaxed, contraction of the TL would produce a retraction of the larynx. However, it is generally accepted that contraction of the more caudal part of the TL, by exerting a caudorostral force on the tympanum, tends to remove the syrinx from the vocal configuration. This is borne out by the measured decrease in syringeal resistance shown in Fig. 1 and the abolition of pre-existing sound shown in Fig. 3. The apparent contradiction in the finding by Youngren et al. (1974) and Gaunt and Gaunt (1977) that both the TL and ST may be active during vocalization can be resolved by assuming, along with these authors, that the muscles are able, by finely graded opposition of action, to exercise a greater degree of control over the axial length of the syrinx and thus over the effectiveness of the aeromechanical coupling process in any given conditions.

Gaunt & Gaunt (1977) have analysed the way in which such co-ordinated muscular activity may explain the production of certain types of high-pitched wailing sounds which occur in chickens. Simultaneous recordings of tracheal pressure and muscle activity suggest that airflow is relatively unimpeded and that the ETMs are being held taut by strong contraction of the TL. Gaunt and Gaunt propose that contraction of the TL, by stabilizing its joint insertion with the ST also permits the latter to exert sufficient tension on the syringeal ligament to draw the pessulus forward and relax the posterior margin of the membranes. Only a very slight movement of the pessulus appears to be necessary for sound production to occur in these circumstances.

Youngren et al. (1974) concluded that, although it had a role during vocalization, the TL was primarily an accessory respiratory muscle since, unlike the ST, it was also active during normal breathing. They proposed that its chief function was to maintain the patency of the airway against adverse positive pressures in the ICS although of course this was also compatible with a steering role for the syrinx during vocalization. Gaunt & Gaunt (1977) expressed reservations on the normal respiratory function of the TL but agreed that it might prevent collapse of the ETMs during rapid, dyspnoeic inspirations. The present author can confirm the presence of strong inspiratory activity in the TL during stressed breathing in anaesthetized birds. Since resting respiratory pressures never exceed + 1–2 cm H₂O (1–2 × 10² N m^-2) there would seem little danger of ETM collapse in normal conditions. Moreover, owing to the existence of a low-resistance pathway from the ICS to the primary bronchi, via the third ventrobronchus of the lung (King, 1975), there is every chance that even large hydrostatic pressures in the ICS will be equalized across the ETMs, so long as concomitant airflow rates are small. When airflow rates are concomitantly large, local kinetic forces in the syrinx will of course tend to draw in the membranes during either expiration or inspiration.

White (1968) recognized the importance of the TL and TH (her ‘caudal extrinsic laryngeal muscles’) in bringing about a gross retraction of the larynx during crowing in cockerels. Powerful retraction of the larynx in response to direct stimulation of these muscles has been reported in the present study and also by Gaunt & Gaunt (1977). The function of laryngeal retraction during vocalization is not clear but it is possible that the resultant decrease in overall tracheal length and tracheal resistance might influence the efficiency of the syringeal mechanism. Fig. 5 demonstrates the sensitivity of the mechanism to alterations in downstream pressure and resistance; any means that led to a reduction in the potentially large tracheal pressures resulting from the dramatically increased airflows during vocalization (Brackenbury, 1977) would be advantageous to sound production.

White (1968) and Gaunt & Gaunt (1977) have suggested another possible function for laryngeal retraction: to influence sound quality by altering the shape of the pharyngeal cavity. Audiospectrograms of crowing in adult chickens show it to be very rich in overtones (Collias & Joos, 1953; Konishi, 1963; Wood-Gush, 1971) but these seem to be governed by tracheal length and ETM tension (Harris, Gross & Robeson, 1968; Abs, 1969; Gaunt & Wells, 1973; Lockner & Murrish, 1975). The characteristic tonal structure of the crow appears only after the voice has broken (Marler, Kreis & Willis, 1962) and further studies on voice-break in relation to the development of full activity in the tracheal muscles would help clarify this problem.

Several workers have shown that experimental de-activation of the vocal muscles will not silence birds (Miskimen, 1951 ; Youngren et al. 1974; Smith, 1977; Brackenbury, 1978 b) and Gross (19746) was led to the conclusion that it is almost impossible to abolish sound production by any means short of serious interference with the respiratory process itself. Simply by the expedient of raising their expiratory effort, operated animals appear to be able to compensate for the lack of muscular involvement. Normally, contraction of the ST facilitates the coupling between airflow and the ETMs; the extra effort in operated animals suffices to bring the relaxed membranes to the yield point and thus onto the force cascade leading to membrane oscillation.

Use of the vocal muscles economizes on effort since it allows triggering of the audiogenerator at relatively low air-sac pressures and airflow rates; indeed they may be indispensable for the production of low volume sounds that require minimal physical effort. In contrast, during crowing the normal muscular controls in favour of economy of effort appear to be sacrificed in order to obtain maximum sound output at maximum possible airflow rates. From their comparison of audio-spectrograms of normal crowing and of sounds elicited from excised syrinxes Harris et al. (1968) deduced that the ETMs must be held in a position of near maximal stretch throughout the crowing cycle. This would imply that the TL undergoes powerful contraction only weakly opposed by the much smaller ST muscle. The latter may serve partly as a ligamentar muscle which, by means of undergoing limited ‘active stretch’ by the TL, serves to provide a relatively stable anchorage from which the much larger muscle can exert maximal retractile force on the larynx. Moreover, the same anchorage may also allow the ST to stretch the syringeal ligament and relax the ETMs as described by Gaunt and Gaunt in their analysis of the wail in chickens.

At the same time insurance against excessive tensile pull on the syrinx and ST muscles is procured by a novel means. For, as White (1968) reasoned, retraction of the larynx must be accompanied by relaxation of the rostral extrinsic laryngeal muscles and stretch of the aryteno-glossal ligaments, both of which connect the larynx to the hyoid. This means that the tension generated by the TL is relayed to elastic elements at the rostral end as well as at the caudal end (Fig. 6). Physical considerations dictate that the tension in the ST and syringeal membranes cannot exceed that in the rostral elastic elements. Both sets of elements also serve to limit the contractile force generated by the TL by allowing it to contract in an isotonic, as opposed to an isometric, manner. They are assisted in this task by the series of elastic elements within the trachea itself. The elastic loading of the TL by these various elements explains the rather sluggish contractile response of the muscles to electrical stimulation (Fig. 1).

Various circumstances accompanying crowing, such as gross laryngeal retraction, near maximal stretch of the ETMs and incomparably high expiratory flow rates, suggest an effort on the part of the bird to achieve the maximum possible unimpeded airflow. The advantages may be twofold: first, since the total fluid power available for driving the membranes is proportional to airflow² (Brackenbury, 1979a) it pays double dividends to minimise flow obstruction; second, it is possible that very high flow rates may permit the exploitation of a novel form of convective sound amplification which avoids the need to produce unduly large membrane vibration amplitudes and consequent overloading of the membranes (Brackenbury, 1979 b).

The observed increase in airway resistance during crowing has previously been ascribed mainly to the partial occlusion of the syringeal lumen by the ETMs but also partly to purely aerodynamic forces that cause the pressure/flow relationship to rise non-linearly at higher flow rates (Brackenbury, 1978 b). Amongst these forces, those due to turbulent airflow through the syrinx seem the strongest candidate. If, as a result of the contraction of the TL muscle, the syringeal lumen is not occluded to the extent previously envisaged, a greater attribution must be made to this turbulent source of resistance.

The ST, TL and TH are hypobranchial muscles derived from ventral extensions of the post-occipital somites of the embryo and thus share a common innervation by the hypoglossal nerve, with the intrinsic muscles of the hyoid and tongue and the rostral extrinsic muscles of the larynx (Weichert, 1970). Their primitive functional associations are therefore with acts involving adjustment to the shape of the anterior respiratory tract, such as deglutition, coughing and, in appropriate circumstances, respiration. The association between respiratory movements and tracheal muscle activity is probably very loose during normal respiration but it becomes more evident during dyspnea and thermal polypnea. The fluttering movements of the throat (gular flutter) of many birds during panting (Calder & Schmidt-Nielsen, 1968) involve the coordinated activity of the hyoid and extrinsic laryngeal muscles, including the TL and TH. It can be demonstrated that the cranial motor nuclei of the hypoglossal, as well as the vagus and glossopharyngeal nerves, receive dual control by the panting centre and the respiratory centre (Brackenbury, 1978c). The hypoglossal nuclei, and the muscles that they serve, are thus amenable to control by at least two other central areas in addition to the vocal centres of the mid-brain (Potash, 1970; Peek & Phillips, 1971 ; Phillips, Youngren & Peek, 1972). Only in a limited sense, therefore, can the extrinsic syringeal muscles be regarded as ‘vocal’ muscles. Truly vocal muscles exist in the syrinx of passerine birds but virtually nothing is known about their detailed innervation and control.

Natural physiological situations occur that involve the production of aerodynamic forces in excess of the threshold for passive sound production but for purposes entirely unrelated to vocalization, such as defaecation, preening and coughing. From the experimental situation illustrated in Fig. 5 it can be inferred that an effective device for reducing spurious sound production is active constriction of the glottis, for this elevates tracheal resistance and impairs the syringeal mechanism and this is precisely what appears to occur during defaecation and vigorous preening (Brackenbury 1978 b).

This work was supported by the Science Research Council.