Temperature Regulation and Respiration in the Ostrich

TEMPERATURE REGULATION AND RESPIRATION IN THE OSTRICH

KNUT SCHMIDT-NIELSEN, JOHN KANWISHER, lIepartment of Zoology Woods Hole Oceanographic Institute Duke University Woods Hole, Massachusetts 02453 Durham, North Carolina 27706

ROBERT C. LASIEWSKI, JEROME E. COHNl,

Department of Zoology University of Kentucky Medical Center Universitv of California Lexinnton.y , Kentucky 40506 Los Angeles, California 90024

AND

WILLIAM L. BRETZ

Department of Zoology Duke University Durham, North Carolina 27706

The Ostrich (S&u&o cumelus), the largest MATERIALS AND METHODS living bird, is an inhabitant of semi-arid and Birds desert areas of Africa and, until exterminated, the Near East and the Arabian Peninsula. Twenty semi-domesticated adult Ostriches weighing 63-104 kg (12 males and 8 females) were pur- When exposed to the heat stress of a hot chased from commercial sources near Oudtshoom in desert, it must use water for evaporation in South Africa and transported by rail to Onderstepoort order to avoid overheating. While its size Veterinary Research Institute near Pretoria. They prevents it from taking advantage of micro- were permitted to graze freely in a 30-acre enclosure climates to the extent that small desert birds and were fed daily with fresh alfalfa and dry corn (maize). Water was provided at all times, except and mammals can, its large size is an ad- when it was removed as part of the experimental vantage in its water economy, as has been procedure. Handling facilities permitted the separa- discussed previously ( Schmidt-Nielsen 1964). tion of smaller or larger numbers of birds as well as the Birds have no sweat glands, and under heat isolation and/or capture of a given individual. Ostriches are very powerful birds and difficult as stress they rely upon increased evaporation well as dangerous to handle. For all experimental from the respiratory system as a major ave- procedures that required a close approach, the de- nue for heat dissipation. We were interested sired bird was driven into a chute, provided with a in the role of the respiratory system in evapora- hood, and transferred to a horizontal V-shaped re- tion, and particularly in the sites of evapora- straining device, the design of which was based on the restraining procedures used for the commercial tion. Furthermore, while in mammals an plucking of Ostrich feathers from living birds. Once increased ventilation causes alkalosis, in birds the bird was completely restrained, the hood could the presence of large air sacs connected to the be removed. respiratory system may have radically different Most experimental procedures did not require the use of tranquilizers or general anesthesia. However, effects on the gas exchange in the lung. where catheterization of blood vessels or other cutting Finally, the fact that the ventilation of the procedures were involved, local anesthesia was respiratory system can be modified by heat achieved with xylocaine. During measurements of surface temperatures in the trachea and the air sacs, stress without change in the rate of oxygen the birds were tranquilized with M 99 (etorphine consumption may provide an avenue for in- hydrochloride, Reckitt & Sons, Ltd.). vestigation of the poorly understood air-sac Weighing. Body weights of individual birds were system of birds. determined with the animal in a portable restraining device placed on a platform scale (precision about The Ostrich, although a non-flying bird, has 0.1 kg). Accurate changes in body weight used for a well developed air-sac system, and its large the determination of evaporative water loss were de- size and slow breathing rate provide an o,p- termined by suspending the bird in its restraining portunity to undertake experimental proce- device from the arm of a beam balance (designed according to Krogh and Trolle 1936, and modified dures which in smaller birds result in great with magnetic damping). This weighing did not pro- technical difficulties or seem impossible. vide the absolute weight of the animal, but gave weight changes with a precision better than 5 g. 1Deceased 7 April 1967. Since the evaporation might reach 10 g/min, the

13411 The Condor, 71:341-352, 1969 342 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ accuracy of the determinations of evaporation by of arterial and venous blood samples was also de- weighing was more than adequate. Correction for termined directly with the oxygen electrode. metabolic loss of carbon is unnecessary at such high The carbon dioxide tension of blood was determined rates of water loss and was not employed. by the Astrup method using Radiometer tonometer, Temperature-controlled room. All experiments which microelectrode, and PHM-4 pH meter. For unex- involved exposure to high temperatures were car- plained reasons the pH values of the Ostrich blood ried out in a room about 2.3 x 3.5 x 2.3 m high. did not consistently remain stable, although the same The temperature in this room was controlled to method has previously been used by us for blood of about f 1°C. The humidity in the room was not other birds without encountering this difficulty. How- under control, but high humidities were avoided by ever, by bracketing samples between suitable gas introducing outside air at a high rate. mixtures it was possible to obtain reliable determina- Temperature measwements. Cloaca1 temperatures tions of the CO, tension in the blood samples. were usually obtained by thermistor measurements Pressures in the airways. Air pressures were de- (Yellow Springs Instrument Co.) and occasionally termined by attaching a sensitive pressure transducer by mercury thermometers. During measurements of (Computer Instruments Corp., Type 6000, range tracheal temperatures the cloaca1 temperature was -0.3 to 0.3 psi) to a needle inserted into the ap- monitored with thermocouples made from 30-gauge propriate place, and the output was recorded on the copper-constantan wire. same Esterline-Angus Recorder used for thermocouples Temperatures in the respiratory system (surface as ( see above 1. Differential _nressures within the airwavs well as air) and skin temperatures were measured were recorded with the same transducer. with X-gauge copper-constantan thermocouples and Cutaneous water loss. Cutaneous water loss was recorded on an Esterline Angus Speed Servo AZAS determined by the weight increase of silica gel Recorder. The accuracy of these measurements was (Randall and McClure 1949). About 30 g silica gel about 0.5”C. Since the cloaca1 temperature was an was placed in a cylindrical aluminum container and important reference point in all these measurements, kept in place by a press-fitted piece of wire mesh. it was monitored simultaneously with the same re- The opening of the container covered a skin area corder. of 31.2 cm*. Cloaca1 temperature during exercise was measured Volumes of respiratory system. Tracheal volume with a temperature telemetering transmitter (range in was determined by filling excised tracheas with water excess of 2 km, accuracy about f. 0.2”C). and measuring the volume in a graduated cylinder. Respiration. Respiratory rates (frequency) were This method, because of expansion of the trachea determined either by counting and using a stopwatch, under pressure, will yield values somewhat too high, or by recording temperature oscillations at the nares but because of the exceptional rigidity of the Ostrich or in the trachea. trachea the error is not likely to be great. No at- Respiratory volumes were obtained by fitting a tempt was made to correct for this error. mask on the animal and collecting the expired air. Air sac volumes were determined either by weighing The mask was constructed from a plastic bottle of gelatin casts after filling the entire respiratory system suitable shape, fitted to the head, and provided with with gelatin and dissecting out the casts, or by a dilu- a sleeve of soft rubber dam which was kept closely tion technique which employed the quick injection assembled around the neck with a series of rubber into an air sac of 100 ml of oxygen (containing 2.2 bands. Pressure changes in the mask, and therefore per cent CO,). Continuous recording of the change leakage, were minimized by the use of a high flow, in oxygen concentration permitted an approximate six-element MacGregor valve. Expired air was col- determination of dilution volumes as well as the wash- lected in large meteorological balloons and the volume out time of the individual air sacs in normal respiration immediately measured with a dry gas meter ( ? 1 per at rest. During panting this approach could not be cent). When desired, air samples were withdrawn used because of the relatively slow response time from the respiratory system (air sacs and trachea) of the oxygen electrode (a few seconds). with 5-ml glass syringes provided with three-way metal valves and no. 16 needles of appropriate length. RESULTS For continuous recording of the oxygen tension in different air sacs, a no. 16 needle was inserted into BODY TEMPERATURE AND the appropriate air sac and the air drawn con- TOTAL EVAPORATION tinuously over a quick-responding oxygen electrode The Ostriches could maintain their body tem- by means of an aquarium pump. perature at a stable level during prolonged ex- Gas analysis. Air samples taken with glass syringes were analyzed on a Scholander 0.5 ml Gas Analyzer posure to a range of ambient temperatures of ( Scholander 1947 ). Continuous monitoring of oxygen 15 - 50°C (fig. 1). At the highest tempera- tension was done with an oxygen electrode according tures they became slightly hyperthermic, but to Kanwisher ( 1959), using a 12-p teflon membrane in order to obtain rapid response. The system of the many could still maintain their body tempera- oxygen electrode with the sampling needle attached ture at a stable level below 40°C for periods responded to changes in gas composition within 0.5 as long as 8 hr at 50°C ambient. In fact, if the set, while the recording of the full magnitude of a initial cloaca1 temperature were high (e.g., response required several seconds. Blood. The oxygen content of blood samples was due to struggling), some Ostriches were able measured with the oxygen electrode by recording the to lower their cloaca1 temperature at any change in oxygen tension when the oxygen was re- ambient temperature. These observations sug- leased from combination with hemoglobin with fer- gest that the body temperatures could have ricyanide ( Kanwisher and Carey, unpublished). The been maintained for longer periods if the ex- oxygen capacity was determined with the same method after equilibration with oxygen. The oxygen tension posure had been continued. The variability at TEMPERATURE REGULATION IN THE OSTRICH 343

. MEASURED OVTSIDE l IN tONST*NT TEMP ROOM

1 I ! \ , 1 t , 15 20 *5 30 35 40 45 50 AIR TEMPERATURE

FIGURE 1. Cloaca1 temperatures(“C) of 15 restrained Ostriches at various ambient temperatures. Valueswere selectedfrom individualsthat maintained stable cloaca1temperature for 0.5-R hr. a given ambient temperature, in addition to individual differences, undoubtedly is due to the fact that the birds were restrained and showed various degrees of nervousness or ex- citement. The maintenance of a stable body temperature below ambient temperature can only be FIGURE 2. Evaporative water loss of Ostriches at achieved by evaporation of water. The evap- various ambient temperatures determined by direct orative water loss in Ostriches increased with weighing. All values are from restrained birds and increasing heat load (fig. 2). There is con- represent rates sustained for 20 min or more. The siderable variation at any given temperature, dashed line indicates the mean evaporation from the particularly at higher temperatures. The low- respiratory tract of the Ostrich as previously reported by Crawford and Schmidt-Nielsen ( 1967). est values we recorded at any ambient temperature are similar to those previously recorded for Ostriches (Crawford and In contrast, the respiratory volumes during Schmidt-Nielsen 1967). The highest rate of heat load (see below) were of a magnitude evaporation we measured, 11 g/min, was sufficient for evaporation of the observed maintained for 40 min, and is about twice the water losses, provided the expired air is ex- maximum rate of 4.5 g/min obtained by Craw- haled saturated at body temperature. Further- ford and Schmidt-Nielsen ( 1967). Eleven more, the estimates of evaporation based on g/min from an 88-kg bird represents 0.75 per tidal volumes do not include exaporation from cent of the body weight per hour. In man the gular area, where temperature measure- evaporation rates frequently are twice as high ments revealed an important site of evapora- (1.5 per cent of body weight, or 1 literlhr). tion. Ambient temperature 40°C. Surface tem- SITES OF EVAPORATION peratures within the respiratory system give The skin plays a minor role in evaporative a great deal of information about the site of water loss. Determinations of water loss from evaporation. Figure 3 presents a diagrammatic bare skin areas of the upper leg showed representation of a typical situation when an evaporation of the magnitude of 0.05 g per Ostrich was exposed to an ambient tempera- m2/min at an ambient temperature of 40°C. ture equal to body temperature (40°C). In This loss from the skin is less than 2 per cent this situation no heat exchange can occur of the total water loss from the bird measured through conduction or radiation, and the entire at the same ambient temperature. Further- metabolic heat production must be dissipated more, skin surface temperatures during heat by evaporation of water. stress were no lower than cloaca1 tempera- The surface temperatures recorded within tures. When the skin is at the same tempera- the respiratory system did not fluctuate ap- ture as the body core, heat cannot flow from preciably during the respiratory cycle, al- core to skin, and consequently the skin cannot though in any particular experiment they serve in heat dissipation. might deviate appreciably from the figures in 344 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ

low temperatures give excellent conditions for transfer of heat from blood to surface, but as stated above, low temperatures may indicate either a relatively low blood flow at the moment of measurement or a high rate of evaporation. The temperature of the air stream may be quite different from that of the adjacent surface and fluctuate greatly with the respiratory cycle. The mean air velocity in the trachea can be calculated from the tidal volume, the duration of the respiratory cycle, and the cross- sectional area of the trachea. Such estimates show that the air velocity may reach values as high as 29 m/set. Temperature equilibration between tracheal wall and the core of the air stream cannot be expected under these conditions, and we had many measurements of air temperatures in the tracheal air stream far higher than the surface temperature. In the air sacs of panting Ostriches we recorded surface temperature ( post-thoracic and interclavicular sacs) as much as 2°C below body temperature. These low surface temperatures implicate the air sacs as possible sites of heat transfer and/or evaporation, but FIGURE 3. Representative surface temperatures in the respiratory tract of an Ostrich kept at an air their role relative to the trachea cannot be temperature (40°C) equal to the body temperature. evaluated in the absence of information on The tracheal locations marked 10 cm, 40 cm, and 60 blood flow. Although the surface area of the cm refer to the distance from the glottis. air sacs is large, they appear to be poorly vascularized. Ambient temperature 50°C. At an ambient the diagrammatic representation of figure 3. temperature of 50°C the tracheal wall might The surface temperatures describe the pos- still be several degrees belomwbody tempera- sibilities for heat exchange between circulat- ture (about 40-42”C) at all levels measured. ing blood and the cooling surface, and it can Under these hot conditions the core of the be seen that the blood can lose heat in the inspired air steam might be several degrees gular area, throughout most of the trachea, hotter than the tracheal surface and thus also and in the walls of the air sacs. above body temperature. The hot ambient Although temperatures lower than ambient air could thus penetrate to more than 60 cm and body indicate sites of evaporation, they do below the glottis (which was the lowest not indicate the amount of heat transfer in point in the trachea where it was practical any given place, for the amount of heat trans- for us to make measurements). The presence fer also depends on surface area, rate of blood of a hot core of air at this level suggests that flow through this area, and the temperature of additional evaporation could take place lower the blood. Surface temperatures within the air- down, presumably in the air sacs. ways were quite variable, particularly in the When Ostriches were panting at 50°C region of the trachea. In the Ostrich the tracheal wall is highly vascularized and the ambient temperature, the temperature in the blood flow changes visibly with heat exposure. post-thoracic sac oscillated as much as 5” with The surface temperature is lowered by in- each breath, and in the interclavicular sac creased evaporation and raised by increased as much as 1”. From this we conclude that blood flow; a low temperature is therefore not a hot core of air from the tracheal air stream a measure of the amount of cooling taking may penetrate directly all the way to the post- place. Under the conditions represented in thoracic sac, bypassing the lung surfaces. It figure 3, surface temperatures as low as 31.9”C is less certain whether the oscillations in the have been recorded, both in the upper part interclavicular sac are due to the same hot of the trachea (20 cm below glottis) as well air entering directly into this sac. as lower down (60 cm below glottis). These The surface temperature of the gular area TEMPERATURE REGULATION IN THE OSTRICH 345 is surprisingly low, and this area must also TABLE 1. Respiration rates of an Ostrich exposed be important in heat dissipation. We recorded for 3 hr to a series of air temperatures (Ta) at or above cloaca1 temperature (To, ). gular temperatures more than 3” below body’ temperature when ambient temperature was TA Tr1 Resp. rate “C “C cycledmin 54°C (gular surface 38”C, cloaca1 41.3%). The gular area is heavily vascularized, which 39.2 39.9 58 39.2 39.9 59 permits the cooling of large quantities of 39.8 39.8 57 blood. 45.0 39.6 59 The movements of the gular area are 49.6 39.7 58 synchronized with the respiratory cycle. At 56.0 40.2 58 the very end of each inspiration the gular membrane and associated structures suddenly interfered with the mechanics of breathing. drop, bringing into the oral cavity an addi- It also seems unlikely that tidal volumes during tional volume of air, over and above the exercise or heat stress should not increase tidal volume. At the beginning of expiration, more than two-fold. Nevertheless, if the ob- the entire gular structure is again raised. The served tidal volumes and respiratory rates quantitative importance of this mechanism are used to estimate the rates of evaporation (henceforth designated “gular pumping”) is (assuming that expired air is saturated at difficult to estimate, but the low surface tem- body temperature) the results correspond to peratures, even at 54°C ambient, clearly show the measured water losses. that the gular area must be a significant site AIR SAC AND LUNG FUNCTION of evaporation. Anatomy of air-sac system. The Ostrich RESPIRATORY RATES AND VOLUMES has a well developed air-sac system which The majority of respiratory rates observed in is diagrammatically presented in figure 4. The our Ostriches fell within two ranges, either a trachea branches into two primary bronchi, low rate of about 6-12 cycleslmin or a high one to each lung. These primary bronchi con- rate of 40-66 cycles/min. These rates were tinue directly through the lungs (where they very easily disturbed by experimental pro- are called mesobronchi) and connect to the cedures such as covering the bird with a hood most posterior air sacs (abdominal and post- or the use of the respiratory mask, and some- thoracic sacs). The connection from the main times even by activities of the experimenter mesobronchi to the more anterior sacs (the which did not directly involve the bird. Inter- interclavicular, lateral clavicular, and pre- mediate rates were observed quite frequently, thoracic sacs) is via secondary branches (at and many (but not all) of these were in con- times called ventrobronchi ) . nection with the use of the hood or the Volumes of the air sacs. The total volume respiratory mask. of the respiratory system in a IWkg Ostrich It is worth noting that the high respiratory is about 15 liters, as determined by the volume rates during heat load remained quite stable, of liquefied gelatin that would fill the system and for any given individual did not seem in dead birds. The largest sacs are the post- to change with increasing heat load. In one thoracic; the others, in order of decreasing size, particular case a bird exposed to an ambient are the interclavicular (which is unpaired), temperature rising from 39.2 to 56°C over abdominal, prethoracic, and lateral clavicular 3 hr kept its respiration quite constant at 58 sacs. The volumes of the major air sacs were cycles/min with a total variation of only * 1 determined by weighing gelatin casts made The tidal volume of cycle/min (table 1) . by filling the entire respiratory system with resting birds not exposed to heat load was liquefied gelatin. about 1.2 to 1.5 liters. During panting at moderate heat load the tidal volume was less, but it increased during more severe heat loads. The largest tidal volume we measured in a bird ,during heat stress was 2.6 liters, but it seems unlikely that this represents the maximal tidal volume for birds that weigh 70-160 kg and have a total volume of the respiratory system (lungs and air sacs) of some 15 liters. We had a strong subjective impression that FIGURE 4. Diagrammatic location of the air sacs the use of a mask for measuring tidal volumes of the Ostrich. 346 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ

TABLE 2. Approximate air-sac volumes of Ostriches, determined from gelatin casts, or as dilution volumes derived from oxygen injection into the sacs. - Approx.vol, liters’ Dilution vol, ml On.2 Both Gelatin WI, ml Bird A Bird B side sides Interclavicular unpaired 470 1200 870 1.0 1 Lateral clavicular paired 450 6.30 0.5 1 Prethoracic paired 530 1340 1960 0.5 1 Post-thoracic paired 2400 2660 2900 2.5 5 Abdominal paired 1020 1500 910 1.0 2 Lung 1.5 3 Total: major air sacs + lungs 5 ’ ’ These two columns represent compromise values given as rounded-off magnitudes to reflect fhe uncertainty of the estimates.

Air sac volumes were also determined as the can be considerable oscillation in the 02 con- dilution volume obtained from the injection of centration in a given air sac, a fact not re- 100 ml 02 into a sac, and the results yielded vealed in series of single samples which are volumes similar to those obtained from gelatin withdrawn and analyzed. These oscillations casts (table 2). These determinations are may be related to the location of the sampling less accurate than gelatin casts since they needle (whether or not it is near the incoming assume instantaneous mixing in a fixed volume, air stream) although the oscillations were not which is incorrect because respiration and air consistently related to the respiratory cycle. exchange continue during the measurements. Whatever the reasons for the oscillations in Any loss of oxygen from the sacs would lead Oa concentration, their existence points to two to an increase in the calculated dilution dangers: ( 1) the use of concentrations ob- volume, and we believe that the large dilution tained by spot sampling as representative of volume obtained for the prethoracic sac may the gas composition in a given sac, and (2) be due to an immediate oxygen loss through the use of observed differences between the the air sac connection. various sacs for the interpretation of function. The volumes given in table 2 do not repre- Arrival of inspired gas at the air sacs. The sent the total volume of the respiratory system, arrival times of oxygen at different air sacs, for they do not include the many small obtained when Ostriches inhaled a single diverticuli from the air sacs, the spaces in breath of oxygen, are given in figure 5. Some pneumatized bones, or the volumes of the of the inhaled oxygen went directly to the trachea and lungs. The volume of gelatin abdominal and post-thoracic sacs, and ap- (about 15 liter) used for the casts is probably peared by the end of the first inspiration. The larger than the actual volume of the living rise in oxygen concentration continued through system due to expansion caused by the weight the expiratory phase of this first cycle, but of the gelatin mass. there was no further increase during the Gas concentrations in the air sacs. Gas con- second cycle. This confirms that the inhaled centrations in the air sacs during quiet breath- air, which during the second cycle again is ing at rest tended to fall into groups. The atmospheric air, arrives directly at the poste- most anterior sacs (interclavicular, lateral rior air sacs. clavicular, and prethoracic) had lower 02 There was a delay of more than one full concentrations and higher CO2 concentrations respiratory cycle before the inhaled oxygen than the posterior sacs (post-thoracic and caused any increase in oxygen tension in the abdominal). Typical values are seen at the anterior sacs. Since only one single breath right side of figure 5. The magnitude of these of oxygen was inhaled, the oxygen which ap- values is substantiated by several hundred peared in the anterior sacs during the second measurements of the various air sacs on and third respiratory cycle must have been Ostriches at rest, but as explained below, we located elsewhere during the first cycle. Pre- consider reported means and standard de- sumably the air entering the anterior sacs is viations as arbitrary and misleading. derived from air which has passed through Information about 02 concentration alone posterior sacs and lungs. was also obtained with the oxygen electrode. These arrival times of the inhaled gas, taken On the whole this method confirmed the in- in conjunction with the normal concentra- formation obtained by analysis of withdrawn tion of oxygen and carbon dioxide in the air samples, but on occasion it showed that there sacs, suggest that the following sequence of TEMPERATURE REGULATION IN THE OSTRICH 347

USUAL GAS COMPOSITION

(Abdominal1

If.4 3.6

(LUNG) (?I 3.9(?)

Prethorocic 14.1 6,6

I ntercloviculor 14.9 7.0

Later01 Cloviculor I 13.2 6.6

TIME OF ARRIVAL OF INHALED OXYGEN RELATIVE TO BREATHING CYCLE (TOPOF GRAPH)

FIGURE 5. Summary of times of arrival in the various air sacs of inhaled 02 in relation to the breathing cycle. After one single inhalation of Oa the arrival of this gas in the air sac was recorded with an oxygen electrode. The left end of each horizontal bar indicates the earliest observed initial arrival of the 0% in that particular sac. The right end of the bar indicates the latest time at which the 0, concentration in the sac was observed to increase. The respiratory cycle is indicated at the top of the graph; the shaded portion of the first cycle denotes a single inspiration of 0,. Respiratory rate about 6 cycle/min, i.e., each cycle about 10 sec. air flow may occur: inhaled air arrives first and one-half respiratory cycles in all the other at the posterior sacs (although some may enter sacs. The post-thoracic sac is the largest of the lung directly ) ; then, during expiration, the air sacs, and the long washout time sug- much of the air from the posterior sacs enters gests that it is not ventilated in proportion to, the lung, an,d the anterior sacs receive air that its volume. has passed across the respiratory surface of the The relatively short washout time of the lung. anterior sacs provides evidence that these sacs All air sacs in the resting Ostrich are highly do not receive outside atmospheric air during ventilated, as shown by the time required normal respiration at rest. The high renewal for return to normal gas composition after rate of air in conjunction with the observed injection of 100 ml 02. The time required to, gas composition (relatively high CO2 and low reduce to one-half the originally achieved in- 02) can best be explained by postulating that crease in O2 concentration varied from less the air entering these sacs has passed across the than two to about four respiratory cycles. This gas exchange surfaces of the lung. rapid washout shows that none of the air sacs Pressures in the airways. The pressures contains an inert air mass. The longest time within the respiratory system were below for the washout was observed in the post- atmospheric during inspiration, above during thoracic sac where its duration was the equiva- expiration. All pressure differences ( n P) lent of about four respiratory cycles, as op- were relatively small; in quiet breathing, n P posed to between one and one-half and two between air sacs and outside reached about 348 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ

0.3 X W3 atm, and during maximal ventilation the lung is not increased. This indicates the (heat stress) n P rose to 5 x 1O-3atm. The presence of a functional shunting system which pressure differences between anterior sacs by-passes the lungs of the Ostrich. The path- (interclavicular ) and posterior sacs (post- way of the shunt is obviously the meso- thoracic) were very small, in the magnitude of’ bronchus, but the anatomical basis for direct- 0.1 of the inhalation or exhalation pressures. ing the air flow one way or the other is The small pressure differences within the obscure. respiratory system during quiet breathing do Metabolic rate and exercise. Our determi- not suggest the movement of air from one sac nations of oxygen consumption in Ostriches system to another within any given moment cannot be considered as standard resting of the respiratory cycle, but since resistance values. The birds were semi-domesticated but is unknown, the possibility cannot be elimi- remained somewhat apprehensive under all nated. The Ostrich respiratory’ system is a experimental procedures. They were re- high velocity-low pressure system. Aerody- strained, remained standing,, and were not namic flow patterns therefore become more postabsorptive. The lowest values we ob- important than changes in resistance (valving) served under these conditions was 415 ml in directing flow. Anatomically, valves are OJmin in a bird weighing 88 kg. not evident; functionally, they seem unlikely, Although not measured during exercise, and in fact unnecessary. oxygen consumption was determined immedi- Ventilation during panting. The increased ately after the termination of running 11 km ventilation volume during panting resulted in at moderate speed ( 17km/hr). Within 2 min decreased COa concentrations and increased of the termination of the run, the oxygen con- O2 concentrations in all air sacs. After several sumption was 3,266 ml OP/min in a bird that hours of heat exposure and panting the CO2 weighed 104 kg. concentration in the post-thoracic sac reached After exercise the CO2 concentrations were values between 0.5 and 1 per cent, with cor- reduced in the air sacs and the O2 concentra- respondingly high O2 concentrations (26-20.4 tions were increased, corresponding to an in- per cent). In the interclavicular sac the CO2 creased ventilation of the air sacs. The COa concentration was also low, up to 1.5 per cent concentration in the arterial blood did not after long exposures, and far below the con- decrease after exercise, which again suggests centration at rest ( 6-7 per cent ) . an accurate control of pulmonary ventilation These results show that air sac ventilation independent of any increased air sac ventila- is greatly increased in panting, and that the tion. interclavicular sac participates in the increased Respiratory rates immediately after exer- air renewal. However, the somewhat higher cise varied between 38 and 66 cycles/min. CO, concentration in the interclavicular sac These rates are similar to those obtained by than in the post-thoracic suggests that the heat stress alone. The maximum heart rate former still may receive some air from the observed immediately after exercise was 176 lung. beats/mm representing a five-fold increase Apparently the ventilation of the lung itself over resting heart rates (3640 beats/min were is not increased during heat stress in spite common in quiescent birds). of the great increase in total ventilation. We DISCUSSION observed no decrease in arterial Pco, during Ostriches are quite tolerant of high tempera- heat stress, even when Ostriches panted tures, both under natural conditions and in heavily for periods up to 8 hr. If the gas the laboratory. When exposed to high ambient exchange surfaces of the lung were ventilated temperatures they increase evaporation by in- in proportion to the increased respiratory creasing the rate and amplitude of respiration. minute volumes, the Pco, in the blomodshould The flightless Ostrich has a well-developed decrease drastically and alkalosis should occur. and extensive air-sac system, and its size and The absence of alkalosis after long periods of low respiratory rate make it particularly well heavy panting in the Ostrich contrasts sharply suited for studies of avian respiratory physiol- with the situation observed in other birds ogy. (Calder and Schmidt-Nielsen 1966, 1968) The body temperature of Ostriches can be which showed extreme alkalosis after several maintained below 40°C even when the birds hours of heat stress. We conclude that, al- are exposed to ambient temperatures as high though large air volumes flow through the as 56°C for more than 6 hr. Dehydrated respiratory system of the heat-stressed Ostrich, Ostriches expend less water for evaporation the air flow over the gas exchange surfaces of than do hydrated ones, and are more likely TEMPERATURE REGULATION IN THE OSTRICH 349 to permit their body temperature to rise to ml/min (Eichna et al. 1945), a rate which hyperthermic levels (Crawford and Schmidt- almost certainly exceeds that of which an Nielsen 1967). In this respect Ostriches are Ostrich is capable. similar to camels. A temperature rise of 4°C The rate of evaporation in the Ostrich can over “normal” levels in a l(N)-kg ostrich con- be accounted for totally on the basis of mea- stitutes a storage of 320 kcal of heat (sp heat = sured respiratory volumes and frequencies, 0.8), which represents a saving of 550 ml of provided air is exhaled saturated at body tem- water. Furthermore, an elevated body temperature. Modulation of the amount of water perature diminishes the gradients between the evaporated during varying heat stress must body and a hot environment, thereby reducing be achieved by varying tidal volume, for pant- heat gain from the environment. It is probable ing frequencies do not change with changes that the reduced gradients are quantitatively in heat stress. Panting rates are relatively far more important for water economy than constant (40-60 cycles/min) and several times is heat storage. In camels this was found to higher than resting rates (&12 cycleslmin). be the case ( Schmidt-Nielsen 1964). Intermediate rates were observed, but these The lowest oxygen consumption we mea- appeared mostly during use of a respiratory sured (415 ml OZ/min) in an 88-kg bird falls mask or a “hood.”’ Whether Ostriches use within the values obtained by Crawford and intermediate respiratory frequencies during Schmidt-Nielsen (1967) on a tame loo-kg moderate exercise is not known. Ostrich. The relationship between avian The two ’ distinct levels of respiratory fre- standard metabolic rates and body weight has quencies in the Ostrich are reminiscent of the recently been reviewed by Lasiewski and situation in dogs, which pant at frequencies Dawson ( 1987). They presented an equation determined by the resonant characteristics for the metabolic rate of birds in relation to of their thoracic cavity (Crawford 1962). Sub- body weight, based on data from 72 non- jectively, Ostriches do not appear to use the passerine birds ranging from hummingbirds elastic components of the respiratory system to the Ostrich: M = 78.3 W0.rz3, where M is to the same degree as dogs, and it appears standard metabolism in kcal/day, and W is as if both inspiration and expiration are body weight in kg. This equation is statisti- achieved by the muscles driving the thoracic cally indistinguishable from similar equations cage. There is no support for the suggestion for mammals (Brody 1945; Kleiber 1947, that Ostriches pant at the natural resonant 1961). frequency of their thoracic cavity. The equation predicts an oxygen consump- Evaporation in the panting Ostrich occurs tion of 290 ml Oz/min for an 88-kg bird, and mainly in the trachea, gular region, and prob- the value we obtained in the present study ably in the air sacs. The amount of water lost is 43 per cent higher. The discrepancy be- through the skin constitutes less than 2 per tween predicted and measured rates can be cent of water evaporated from the respiratory ascribed to the fact that the birds were stand- surfaces. Temperature measurements indicate ing up, apprehensive and restless, and that that the trachea is a major site of evaporative they were not postabsorptive. water loss. The oxygen consumption measured at the At the very end of inspiration in the panting end of running (3260 ml OJmin) was ap- Ostrich, the gular membrane and associated proximately 10 times the standard metabolism structures suddenly drop, to be raised again predicted for a bird of this size ( 104 kg). It at the beginning of expiration. This “gular should be noted that this rate was not obtained pumping” is timed so that additional ambient, during activity but during the initial recovery non-saturated air is brought into contact with period. Nevertheless, the level is consistent the moist oral surfaces at the end of inspira- with measurements of metabolism during tion. This should result in higher rates of flight in small birds, which are 8-U times evaporation than would otherwise be obtained. higher than standard metabolism (Lasiewski The gular pumping is similar to, gular flutter- 1963; LeFebvre 1964; Tucker 1968). ing in many smaller birds (Dawson and The amount of water evaporated by Schmidt-Nielsen 1964; Lasiewski and Bartholo- Ostriches at high temperature is moderate mew 1966). The value of evaporation from when compared to man, who is similar in the oral surfaces can be inferred from the ob- weight. The highest rate we observed, 11 ml servation that air could be cooled by lO-15°C HBO/min at 50.5”C, is similar to that in non- by the time it reached the glottis during in- acclimatized man ( Lade11 et al. 1944a, 1944b ) , spiration, and from the low temperatures of Acclimatized man may sweat as much as 4050 the gular surface. Lasiewski and Bartholomew 350 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ

(1966) measured gular temperatures as much presentation, see King and Farner 1964. ) as 3” below body temperature in the Poor- Three main patterns for air movement in the will (Phalumwptilus nuttallii), demonstrating lung have been suggested: 1) back-and-forth the importance of this area as a site of heat loss tidal flow (e.g., Zeuthen 1942); 2) outside + in small birds that employ gular flutter. lung + air sacs+ outside (e.g., Shepard et al. The relative roles of oral surfaces, trachea, 1959) ; and 3) outside + air sacs+ lung + and air sacs in heat dissipation remain uncer- outside (e.g., Donoso and Cohn 1962; Cohn tain. The fact that air-sac temperatures are et al. 1963). low could be interpreted either as local evap The relatively small pressure differences oration, or as due to the arrival of cool air between air sacs in the Ostrich at rest and the from the trachea. Subjectively, the air-sac outside suggest that all air sacs may be filling wall looks poorly vascularized, while the and emptying at the same time. Synchronous tracheal wall is extremely well vascularized pressure changes in air sacs of other species and hyperemic during heat stress. On the have been found by Baer ( 1896)) Soum ( 1896, other hand, the total area of the air sacs is as quoted from Zeuthen I942), Francois- much greater than that of the trachea. Franck ( 1966), Victorow ( 1969), Sturkie The extent to which the inhaled air is cooled ( 1954)) Donoso and Cohn ( 1962), and Cohn during its passage through the trachea depends et al. (1963). on whether the flow is laminar or turbulent. The experiments on inhalation of pure In normal breathing with a tidal volume of 1.2- oxygen in the Ostrich suggest that inhaled 1.5 liters at a rate of 6-12 cycles/min, the flow air passes through the trachea and meso- rate in a tube of 1 cm radius gives Reynolds bronchus directly to the abdominal and post- numbers between 450 and 1160, and flow thoracic sacs (and perhaps to the lungs). The should still be laminar. If air velocity increases concentration of gas normally found in these considerably, as it does in panting, the sacs could be obtained by mixing approxi- Reynolds number will be some 10,606 and the mately equal volumes of dead-space air and flow must be turbulent. If the air flow in the outside air. No detectable amounts of in- trachea is laminar, heat transfer between the spired air seem to go directly to the anterior air stream and the wall is by conduction, a very sacs ( prethoracic, cervical, and interclavicu- slow process. For rapid heat transfer in the lar). From the abdominal and post-thoracic trachea the air flow must be turbulent, but it sacs the air supposedly passes across the must not be too turbulent or the work of respiratory surfaces of the lungs, probably breathing will increase too much. It seems via the dorsobronchi-parabronchi-ventrobron- that the dimensions of the trachea of the chi system proposed by Hazelhoff ( 1951), to Ostrich are such that they permit laminar the prethoracic, cervical, and interclavicular flow and low work of breathing at rest, but sacs, In spite of the high CO2 and low 02 in that the flow changes to turbulence during the anterior sacs, these must be well ventilated panting. (low wash-out time), which agrees with the It was mentioned above that modulation of observations of Graham ( 1939), Sturkie evaporation in the panting Ostrich must be ( 1954), Donoso and Cohn ( 1962), and Cohn achieved by changes in tidal volume, for et al. ( 1963)) but not with Zeuthen ( 1942). respiratory frequency remains constant. How- The complex pattern of air flow operates ever, such variations in tidal volume may also under high velocity and low pressure, and we have an indirect effect through the changes found no evidence of anatomical valves. The in turbulence which result from changes in direction of air flow seems to be controlled linear velocity. Thus, not only does the volume by aerodynamic forces and the spatial orienta- of tidal air change, but the heat exchange tion of the openings and passages. The pat- between the tracheal air and the wall is in- terns of flow which we have suggested in the fluenced as well. This suggests a modulation Ostrich are similar to those proposed by Dot- of heat transfer greater than that resulting terweich ( 1936) and Hazelhoff ( 1951), al- merely from the change in tidal volume. though contrary to the much-quoted theory of Our results on the Ostrich permit some Zeuthen ( 1942). conclusions about the patterns of air flow in The anterior sacs have a higher CO2 con- the air-sac system. Numerous previous studies centration (about 67 per cent) than the employing a variety of birds and techniques posterior sacs (about 4 per cent), a situation have led to several conflicting theories and previously noted by several investigators. The much controversy. Many extensive reviews arterial blood appears to be in equilibrium have been published. (For a brief and elegant with air similar to that in the posterior sacs TEMPERATURE REGULATION IN THE OSTRICH 351

(fig. 5). Presumably the high CO2 concentra- be advantageous for the Ostrich to separate tion in the anterior sacs can be achieved only the need for increased ventilation due to heat by air reaching these sacs from the lung, with- stress from the need for increased oxygen up- out admixture of air from other sources. If take. In the flying bird, on the other hand, so, gas exchange between blood and air in the the demand for oxygen and for heat dissipation lung can hardly take place in the same way are frequently concurrent, and a disassociation as in the mammalian lung. On the other hand, of the two may be less imperative. a higher Poo2 in air coming from the lung SUMMARY than in arterial blood conld be achieved by a counter-current flow pattern. One condition The role of the avian respiratory system in for counter-current flow, a unidirectional temperature regulation and the air flow in the movement of the air, is possible because the air-sac system were studied in the Ostrich bird lung parenchyma does not have alveoli ( Stmthio cam&us). The Ostrich can maintain but rather air capillaries which permit a its body temperature below 40°C during 8 through-flow of air. The air passing out of the hr at ambient temperatures as high as 56°C. lung could thus be in equilibrium with the This is achieved by increased evaporation high Pcoz of incoming venous blood. Similar from the respiratory tract (panting). Evapora- counter-current exchange has been described tion increased with heat load to a maximum of for gas exchange in the placenta of certain about 0.75 per cent of the body weight per mammals and in the fish gill (Scholander hour at 50°C ambient temperature. In com- 1958). parison, a man under hot desert conditions During heat stress the air sacs were venti- may lose water (sweat) at twice this rate. lated more vigorously than at rest, and the Respiratory rates rose from 6-12 cycle/min gas concentrations of all sacs approached at rest to about 40-60 cycle/min during heat those of outside air. The CO2 concentrations load. The amount of’ water evaporated could remained slightly higher in the anterior than be accounted for on the basis of respiratory in the posterior sacs. A similar pattern has volumes, and skin loss was apparently insig- been found in heat-stressed pigeons ( Schamke nificant. 1938). The distribution of 02 and CO2 con- All air sacs are highly ventilated during both centrations in heat-stressed Ostriches sug- rest and panting. Experimental inhalation of gests that at least some of the air reaching pure O2 indicates that during inspiration out- the anterior sacs first passes through the lungs. side air passes directly to the posterior sacs, The complete absence of alkalosis in while the anteriolr sacs receive air that has Ostriches after panting for up to 8 hr suggests already passed over the respiratory surfaces of that during panting the lungs are bypassed the lung. Gas composition in the anterior sacs by much of the ventilation volume, and that suggests a counter-current flow of air and the air flow across the gas exchange surfaces blood in the lung. Sustained panting did not remains precisely adjusted to the metabolic cause alkalosis (decrease in arterial blond need for oxygen. In contrast, all other birds Pco,), although air sac Poe, fell markedly. examined (nine species) exhibit marked alka- This suggests a functional shunt system which losis after heat stress (Calder and Schmidt- permits a regulated by-pass of the lungs. Nielsen 1966, 1968). The regulation of air ACKNOWLEDGMENTS flow to the lung could be achieved by aerodynamic changes in flow patterns, or by This work was supported by NIH Research changes in resistance to air flow through the Grant HE-02228 to Schmidt-Nielsen, NSF Re- lungs as suggested by Zeuthen ( 1942). The search Grant GB 5347 to Lasiewski, NIH Re- parabronchial muscles discussed by Brandes search Career Award l-K6-GM-21,522 to (1924) could achieve the necessary resistance Schmidt-Nielsen, and an NSF Predoctoral changes. Fellowship to, Bretz. We acknowledge the The striking fact that the heat-stressed’ hospitality and valuable assistance extended Ostrich does not develop alkalosis, while all by the Onderstepoort Veterinary Research In- other birds examined do so, seems paradoxical, stitute through its Director, B. C. Jansen, as It is therefore tempting to speculate and sug- well as M. de Lange, K. van der Walt, H. P. A. gest that severe heat stress in the Ostrich de Boom, I. S. McFarlane, K. Riley, and many usually occurs when the bird is at rest, while others. We also received valuable help from in flying birds the greatest need for heat Cecil Luck and Koppel Furman, Department dissipation occurs during flight (i.e., severe of Physiology, University of Witwatersrand, exercise). Teleologically, it would therefore Johannesburg. 352 SCHMIDT-NIELSEN, KANWISHER, LASIEWSKI, COHN, AND BRETZ

LITERATURE CITED KROGH, A., AND C. TROLLE. 1936. A balance for the determination of insensible perspiration in BAER, M. 1896. Beitrlge zur Kenntnis der Anatomie man and its use. Skand. Arch. Physiol. 73:159- und Physiologie der Athemwerkzeuge bei den 162. Vogehr. Z. wiss. Zool. 61:420498. LADELL, W. S. S., J. C. WATERLOW, AND M. F. Hun- BRANDES, G. 1924. Beobachtungen und Reflexionen SON. 1944a. Desert climate. Physiological and iiber die Atmung der Vogel. Pfliigers’ Arch. ges. clinical observations. Lancet 2:491497. Physiol. 203: 492-51 I. LADELL, W. S. S.. 1. C. WATE~LOW. AND M. F. BRODY, S. 1945. Bioenergetics and growth. Rein- H&ON. 1944b: Desert climate. ’ Physiological hold, New York. 1023 p. and clinical observations. Heat exhaustion type II. CALDER, W. A., AND K. SCHM~T-NIELSEN. 1966. Lancet 2:527-531. Evaporative cooling and respiratory alkalosis in LASIXwSKI, R. C. 1963. Oxygen consumption of the pigeon. Proc. Natl. Acad. Sci. 55:750-756. torpid, resting, active, and flying hummingbirds. CALDE~, W. A., AND K. SCHMIDT-NIELSEN. 1968. Physiol. Zoiil. 36: 122,140. Panting and blood carbon dioxide in birds. Amer. LASIEUTSKI, R. C., AND G. A. BARTHOLOMEW. 1966. J. Physiol. 2X:477-482. Evaporative cooling in the Poor-will and the COHN, J. E., L. BURKE, AND H. MA~KESBERY. 1963. Tawnv Froemouth. Condor 68:25%262. Respiration in unanesthetized geese. Fed. Proc. LASIEWSKI: R. C, AND W. R. DAWSON. 1967. A re- 22:397. examination of the relation between standard CRAWFORD, E. C., JR. 1962. Mechanical aspects of metabolic rate and body weight in birds. Condor panting in dogs. J. Appl. Physiol. 17:249L251. 69: l&23. CRAWFORD, E. C., JR., AND K. SCHMIDT-NIELSEN. LEFEBVRE, E. A. 1964. The use of DzO” for mea- 1967. Temperature regulation and evaporative suring energy metabolism in Columba Zivia at rest cooling in the Ostrich. Amer. J. Physiol. 212: and in flight. Auk 81:403-416. 347-353. RANDALL, W. C., AND W. MCCLURE. 1949. Quanti- DAWSON, W. R., AND K. SCHMIDT-NIELSEN. 1964. tation of the output of individual sweat glands Terrestrial animals in dry heat: desert birds. p. and their response to stimulation. J. APP~. 481492. In D. B. Dill, E. F. Adolph, and C. G. Physiol. 2:72-80. Wilber [eds.] Handbook of physiology, Section SCHARNKE, H. 1938. Experimentelle Beitrlge zur 4: Adaptation to the environment. American Kenntnis der Vogelatmung. Z. vergl. Physiol. Physiological Society, Washington, D. C. 25: 548-583. DON~SO, H., AND J. E. COHN. 1962. Respiration in SCHOLANDER, P. F. 1947. Analyzer for accurate unanesthetized geese. Fed. Proc. 21:440. estimation of respiratory gases in one-half cubic DOTTERWEICH, H. 1936. Die Atmung der Vogel. centimeter samples. J. Biol. Chem. 167:235-250. Z. vergl. Physiol. 23:744-770. SCHOLANDER, P. F. 1958. Counter current ex- EICHNA, L. W., W. F. Asrm, W. B. BEAN, AND W. B. change. A principle in biology. Hvalradets SRELLY. 1945. The upper limits of environ- Skrifter 44: l-24. mental heat and humidity tolerated by ac- SCHM~T-NIELSEN, K. 1964. Desert animals. Phys- climatized men working in hot environments. J. iological problems of heat and water. Clarendon Indust. Hyg. Toxicol. 27:59-84. Press, Oxford, Errgland. 277 p. FRANCOIS-FRANCK, M. 1906. Etudes graphiques et SHEPARD, R. H., B. K. SLADEN, N. PEXE~SON, AND T. photographiques de mechanique respiratoire ENNS. 1959. Path taken by gases through the comparee. Compt. Rend. Sot. Biol. 61: 174-176. respiratory system of the chicken, J. Appl. GRAHAM, J. D. P. 1939. The air stream in the Physiol. 14:733-735. lung of the fowl. J. Physiol. 97:133-137. Soot, M. 1896. Recherches physiologiques sur HAZELHOFF, E. H. 1951. Structure and function of lappareil’ respiratoire des oiseaux. These Q la the lung of birds. Poultry Sci. 36 :3-10. Fat. Sci. Lyon, Paris. KANWISHER, J. 1959. Polarographic oxygen elec- STURKIE, P. D. 1954. Avian physiology. Comstock trode. Limnol. Oceanog. 4:210-217. Publishing Associates, Ithaca, N. Y. 423 p. KING, J. R., AND D. S. FARNER. 1964. Terrestrial TUCKER, V. A. 1968. Respiratory physiology of animals in humid heat: birds. p. 603-624. In House Sparrows in relation to high altitude flight. D. B. Dill, E. F. Adolph, and C. G. Wilber reds.1 J. Exp. Biol. 98:55-56. Handbook of physiology, Section 4: Adaptation VICTOROW, C. 1909. Die kiihlende Wirkung der to the environment. American Physiological Luftsacke bei den Vogeln. Pfliigers’ Arch. ges. Society, Washington, D. C. Physiol. 126:300-322. KLEIBER, M. 1947. Body size and metabolic rate. ZEUTHFN, E. 1942. The ventilation of the respira- Physiol. Rev. 27:511-541. tory tract in birds. K. Danske Videnskab. Selsk. KLEIBE~ M. 1961. The fire of life. An introduction Biol. Medd. 17:1-50. to animal energetics. John Wiley and Sons, New York. 454 p. Accepted for publication 20 May 1968.