Power Input During Flight of the Fish Crow, Corvus Ossifragus

The energy cost of sustained bird flight has recently been determined for two species of birds, the budgerigar (Tucker, 1968) and the laughing gull (Tucker, 1972). We studied the fish crow because it has approximately the same body weight as the laughing gull, but has different wing shape, wingbeat frequency and flight pattern.

The fish crow (Corvus ossifragus) has shorter and broader wings than the laughing gull (Fig. 1). The gull has a long and narrow wing whose ratio of length to breadth, the aspect ratio, is 9·3 (Raspet, 1960); in contrast, the aspect ratio of the fish crow wing is 5·2 as determined from photographs of the birds in gliding flight. There is also a conspicuous difference in the arrangement of the primary feathers, which are more separated in the crow. The mean wingbeat frequency of the fish crow during level flight at speeds between 7·4 and 11·0m s^-1 was 5·2 s^-1, while Tucker (1972) reported 3·91 s^-1 for the laughing gull flying under similar conditions. Furthermore, the crow flaps its wings during most of its relatively short flights, while the gull spends more time soaring and covers longer distances.

Five adult fish crows (mean mass 275 g) were obtained commercially. They were kept in an outdoor aviary and fed on chopped beef kidney and dog chow. They appeared healthy and maintained their body weight for more than a year.

The crows were trained to fly in a wind-tunnel (see Tucker & Parrott, 1970), but only two would consistently fly for the required periods of 15−20 min. In initial training flights birds were released individually into the wind-tunnel’s working section, through which air streamed at about 8 m s^-1. Electric shocks were used as negative reinforcement for landing or for clinging to the in. mesh hardware cloth screen at the front of the working section of the tunnel. Flight times and range of speeds were gradually increased as training progressed. To keep the birds in condition it was necessary to fly them every day in a 1 h training session.

When the crows would fly continuously for 15−20 min, they were fitted with a mask and trailing air tube, similar to that described by Tucker (1972). The toes were taped in a closed position to discourage the birds from tearing at the mask with the claws. The crows accepted the mask after a few trials, and flew in an apparently normal manner in descending flight at angles of 4° and 6° from horizontal. At 2°only one crow would fly for long enough periods to obtain steady-state metabolic measurements. Neither crow would fly horizontally with the mask for sufficiently long periods to obtain reliable metabolic measurements.

Flight patterns varied from steady, flapping flight to alternate flapping and gliding. During the brief periods of gliding the crow drifted backwards, but caught up again by flapping to resume its position near the top and front of the working section of the wind-tunnel.

Oxygen consumption was measured in an open system in which room air was drawn in through the back of the mask, past the bird’s head, and into the tube attached to the front of the mask, at a precisely controlled rate of 255 ml s^-1. An aliquot stream of the mask air was drawn from the main stream at a rate of 25 ml s^-1 by a small, leak-free diaphragm pump, and directed at constant pressure through a desiccant (Drierite) into a Beckman G 2 paramagnetic oxygen analyser (full-scale deflexion for ⁠) For further details and calibration procedures, see Tucker (1972).

Oxygen concentration of excurrent mask air was determined from stable recorder tracings obtained after at least 5 min of flight. Oxygen concentration in ambient air was determined before and after a series of flights by drawing room air through the empty crow mask. Measurements preceding and following flights differed by less than 3 ×10−5 atm; the mean was taken as during the flights. Oxygen consumption was calculated according to Tucker’s equation (3) (Tucker, 1968), assuming an RQ of 0·80. If the RQ were actually 0·7, this assumption would entail an error in O₂ consumption of 2%. If RQ were an unlikely 1·0, the error would be 4%. All gas volumes were expressed in terms of standard temperature and pressure. Standard deviations of all oxygen consumption data, calculated separately at each air speed and wind tunnel tilt, had a mean of 0·8 cm³ O₂ g^-1 h^-1 with a range of 0·1 to 1·8 cm³ O₂ g^-1 h^-1.

Before a flight the crow’s feet were taped and its head was fitted with the mask and tube. The wind-tunnel was turned on and the crow was launched into the air stream. Each flight was terminated 15−20 min after its start and the bird was permitted to rest for at least 15 min, during which the O₂ analyser trace decreased and stabilized. After three or four such flights at different air speeds and tunnel tilts the bird was weighed to 0·1 g and returned to the aviary.

Resting oxygen consumption was measured in the same system, except that the crow instead of wearing a mask was placed in a darkened, insulated metabolic chamber. Air pressures and flow rates were identical with those maintained during flight. After 3 h of equilibration at a chamber temperature of 25 °C, stable readings of were taken for 1 h.

As done by Tucker (1972), we define power input as the rate of metabolic energy expenditure, which was calculated from the rate of oxygen uptake. We used the commonly accepted value that 1 1 of oxygen consumed corresponds to 20 kJ (4·8 kcal), which means that 1 1 O₂ consumed in 1 h is equivalent to 5·6 W.

The forces acting upon a bird flying in a wind-tunnel can be resolved into components including lift, body weight (W_b), drag on the bird’s body (D_b) and thrust (T). The magnitudes and directions of all but depend upon the air velocity (V) and upon θ. The body drag must be balanced by a thrust of equal magnitude and opposite direction in order for the bird to maintain its position in the wind-tunnel.

The drag exerted by the air on mask and tube (D_m+t) was measured by techniques described by Tucker (1972).

Mask drag (D_m) was measured with the aid of a frozen crow carcass, wearing the mask and with wings tightly folded, mounted on a flight balance and facing into the wind in the wind-tunnel. The drag was determined with the tunnel horizontal at all values of V that were used in measurements of flight metabolism. The procedure was repeated on the unmasked crow carcass, and mask drag determined as the difference between drag on the masked and unmasked carcass. D_m varied between — 4·7 × 10−5 N and 2·3 × 10−3 N, the former value not being significantly different from zero. Mask weight was 4·04 × 10⁻² N.

Tube drag (D_t) was determined in the horizontal wind-tunnel by affixing, at the approximate position of the flying crows, the mask and its freely trailing tube (Tucker, 1972). At each air velocity used in metabolic measurements, the angle ϕ made by the tube, relative to horizontal, was measured with an optical protractor. The mean length of the tube (1·15 m) suspended during flight between the mask and the wind-tunnel floor had a weight (W_t) of 7·3 × 10⁻² N and an outside diameter of 3·5 mm. The drag of the tube (D_t) is given by the product of the tube weight and the cotangent of the angle (ϕ), as W_t cot ϕ.

The drag of the mask and tube carried by the bird must be overcome by additional thrust at the cost of an additional increment of power input ⁠. This value, when subtracted from P_i gives an estimate of power input during flight unencumbered by mask and tube. is calculated in the manner described below.

The drag on the feet of the crow was disregarded. The crows would lower their feet during brief periods of gliding and backward drift, but during steady flapping flight the feet were always retracted. Based on the direct determinations of the drag on the feet of pigeons made by Pennycuick (1968) we calculate that, if the feet of the crows were continuously lowered into the air stream the resulting drag at the highest air speed used in our study would increase the measured power input by 6%. Actually, the feet were lowered during only 22% of the total flight time, and we therefore made no correction for foot drag.

Mean power input (metabolic rate), corrected for mask and tube drag, is plotted in Fig. 2 for level and descending flight at air speeds between 7 4 and 11 0 m s^-1. With increasing angle of descent power input decreased. At a given angle of flight power input varied little with air speed, although at 6 ° there was a minimum at 9·2 m s^-1.

The meaning of these relationships for flight strategy is that, to fly for a particular period of time, it matters little whether the crow flies slowly or fast. Level flight is energetically more costly than descending flight at all air speeds and the energy saving to a descending bird increases with increasing descent angle. It is not surprising that flying downhill costs less than flying level ; this is true also for the budgerigar (Tucker, 1968) and the gull (Tucker, 1972). Obviously, the reason is that the thrust component contributed by gravity increases with increased descent angle.

If partial efficiency is estimated from equations (1) and (5) for 6 ° (θ₁) and 2 ° (θ₂) the values for the range of air speeds studied vary from 0-22 to 0-29. These are within the range of values obtained for budgerigar and laughing gull (Tucker, 1972).

Mean power input (metabolic rate) at rest was 1·48 × 10⁻³ W g^-1.

It is interesting to compare the crow and other birds for which flight information is available. Because of the similarity of body mass of the crows (0.275 kg) and the smaller of the two laughing gulls (0.277 kg) studied by Tucker, we selected this individual gull for comparison. The shape of the power v. air-speed curves for these two birds in level flight are similar (Fig. 3). However, level flight for the fish crow costs about one-third more than for the laughing gull. This means that to fly for a given period of time the crow must expend greater amounts of energy than the gull.

If the total efficiency for flight is similar in the crow and gull, the higher power input for level flight in the crow would reflect greater power output. This could be due to a greater thrust in the crow than in the gull, which would be required if air drag were greater on the crow than on the gull. On the other hand, if thrust as well as efficiency were the same in the crow and gull, then the greater power input by the crow might be due to differences in the effectiveness with which the wing transforms muscular force into thrust. Such differences are suggested by the major differences in wing morphology of the two birds (Fig. 1).

Fig. 3 also shows the power input for sustained level flight of the budgerigar, as well as the power input for a hovering hummingbird. The budgerigar data were obtained in a wind-tunnel different from that used for the crow and gull, and it is therefore not possible to say whether the differences in power curves are due to wind-tunnel characteristics or to differences in flight physiology or aerodynamics.

Table 1 shows the minimum cost for level flight (column c) compared to resting metabolism (column b) for the birds in forward flight in Fig. 3. The body weights of these birds cover approximately a tenfold range (column a). The ratio of power input in level flight to that at rest is about six for all three species.

The energy cost of flight (Fig. 2) divided by air speed gives the cost of flight per unit distance. Such values are plotted in Fig. 4 for the fish crow in level and descending flight. Cost to travel 1 km decreases with increasing air speed at all angles. This means that for a crow to travel a distance of 1 km the energy cost is lowest at the highest speed.

Cost to travel 1 km in level flight for budgerigar, laughing gull and fish crow are compared in Fig. 5. In all three species cost of travel decreases with increasing air speed except that for the budgerigar the cost increases again at the highest air speeds. The minimum cost for level flight for all three species occurs in a narrow range of air speeds, between 11 and 12 m s^-1. The available data do not indicate, however, whether cost of travel in fish crow and laughing gull would decrease further at higher air speeds or increase as in the budgerigar.

Available data were used by Tucker (1970) for an allometric analysis of the relationship between minimum cost of travel and body weight in flying animals. On the basis of Tucker’s equation (7), it can be predicted that the minimum cost of travel for a 0.275 kg flying animal is 7.1 J g^-1 km^-1. This is close to the measured value for the fish crow of 8.0 J g^-1 km^-1 at an air speed of 11 m s^-1. By way of contrast, the cost of travel for a 0.275 kg running mammal, predicted from equation (1) of Taylor, Schmidt-Nielsen & Raab (1970) is 18.0 J g^-1 km^-1, or more than twice as high as for the flying bird of the same size.

This study was supported by NSF Research Grant GB 6160X to Vance A. Tucker, NIH Research Grant HL-02228 and NIH Research Career Award i-K6-GM-2i,522 to KSN.