Exposure to elevated environmental temperatures produces large increases in ventilation and respiratory evaporation in birds (Calder & Schmidt-Nielsen, 1967; Bouverot, Hildwein & LeGoff, 1974; Richards, 1976). Some workers have reported severe blood alkalosis in panting birds due to lung hyperventilation (Linsley & Burger, 1964; Calder & Schmidt-Nielsen, 1968; Frankel & Frascella, 1968), others only a small decrease in arterial (Marder, Arad & Gafni, 1974; Marder & Arad, 1975; Krausz, Bernstein & Marder, 1977). It has often been suggested, following Zeuthen (1942), that intrapulmonary valves may shunt air away from the parabronchi during panting thus alleviating the risk of hypocapnic alkalosis but such valves have not yet been demonstrated in vivo.
Recently Bech, Johansen & Maloiy (1979) showed that flamingoes, Phoenicopterus ruber, avoid hypocapnia during panting by restricting hyperventilation to the respiratory dead space. This is achieved by a 7-fold reduction of tidal volume, VT to a value which closely matches the tracheal dead space, VD. As a result of the diminished VT, carbon dioxide accumulated within the respiratory system and rapid, shallow panting was periodically interrupted by bursts of 1–3 deeper breaths which served to flush out the excess carbon dioxide.
We have observed a similar pattern of respiration in domestic fowl Gallus gallus exposed for 1 h periods to environmental temperatures Ta of 35–40 °C. After 15–30 min of panting the mean of the interclavicular air sac fell by 3·5 torr whilst that of the abdominal air sac rose by 10·4 torr (Fig. 1). During each burst of deep respiration the gas composition of the abdominal sac transiently returned towards its resting, non-panting value and the measured rates of oxygen consumption and carbon dioxide production transiently increased; at the same time end-tidal fell markedly (Fig. 1, inset). These observations show that during shallow panting stale air accumulates within the caudal air sacs. This gas is used to ventilate the palaeopulmo during expiration but it is clear from the slight freshening of interclavicular sac gas, which is similar in composition to end-parabronchial gas, that gas exchange in the parabronchi is not impeded. This is also borne out by the slight fall in arterial (Table 1).
Recordings such as that in Fig 1, inset, do not give an accurate reflection of the composition of gas leaving the lung during shallow panting since the VT is probably slightly larger than the tracheal volume and the stale gas front fails to register Wily on the gas analyser. However, the of end-expired gas during normal breathing and during sustained exercise is approximately 8 torr greater than that of interclavicular sac gas (J. Brackenbury, P. Avery & M. Gleeson, in preparation). Using interclavicular sac as a gauge of end-expired gas it is possible to calculate the effective lung ventilation from the measured as follows: ml BTPS min-1 where the figure 876 incorporates the factor for conversion from STPD to BTPS conditions. Dead space ventilation was calculated from the measured respiratory frequency f multiplied by the dead space volume VD.
Table 1 shows that despite a 2·6-fold increase in total ventilation increased by only 5 % as a result of a 4-fold drop in VT. The rise in the VD/VT ratio during panting automatically accounts for the observed increase in of the abdominal air sacs since, as Piiper (1978) has explained, the major source of CO2 in the caudal air sacs is dead space gas re-inhaled from the trachea and primary bronchus at the beginning of each inspiration.
ACKNOWLEDGEMENTS
This work was supported by the Science and Agricultural Research Councils.