AMERICAN JOURNAL OF PHYSIOLOGY
Vol. 222, No. 2, February
1972. Prtnted
in U.S.A.
Metabolism
during
in the laughing
VANCE
A. TUCKER
Department of Zoology,
flight
gull,
Larus atrida
Duke Uniuersity,
Durham,
North
metabolic
respiratory
rate;
rate;
oxygen
consumption;
wind tunnel
respiratory
quotient;
WILD
BIRDS
PERFORM
prodigious
feats of exercise by human
standards. For example, ducks can fly at speeds in excess of
17 m/set (25), probably for hours, and small birds migrating
over deserts and oceans are thought
to fly nonstop
and
without feeding for distances exceeding
1,000 km and times
exceeding 50 hr (8, 14, 15). Some birds appear to be more
tolerant
of high altitudes
than any other vertebrates,
as
they have been seen in flapping flight at altitudes in excess
of 6,000 m (2 1). What are the metabolic
rates of the birds
during such feats?
The metabolic rate in flight is a function of both the speed
and the angle of flight, and these variables can be controlled
if a bird is trained to fly in a wind tunnel. Metabolic
rates
durin flight in a wind
tunnel
at controlled
speeds and
angles are available for two animals, the budgerigar
(Melopsittacus undulatus) (22) and the laughing
gull (Larus atricilla
(23)
In’ an ideal wind tunnel, the behavior of a stationary
object exposed to moving air in the test section is identical
to
the behavior
of the object moving
through
stationary
air.
Actual wind tunnels are not ideal, and in addition,
a bird
flying in one can be attached to instrumentation
which its
wild counterpart
is not. In this study I will quantify some of
the differences that occur between the metabolic
rate of a
gull flying in a wind tunnel and one flying in still air in
nature.
237
UNITS
AND
27706
ACCURACY
The International
System of Units (6, 13) based on the
meter, kilogram,
and second has been used throughout
this paper. The power unit of watts can be converted to the
more familiar kilocalories
per hour by multiplying
by 0.860.
The force unit of newtons can be converted
to kilogramsforce by multiplying
by 0.102. All measurements
in this
study were made at air temperatures
between 23 and 30 C
unless otherwise specified. All gas volumes are corrected to
a temperature
of 0 C and a pressure of 760 mm Hg.
The terms velocity and speed have different meanings in
this paper. Velocity is a vector and has both magnitude
and
direction.
Speed is the magnitude
of velocity.
The accuracies of the various measurement
processes are
described in terms of systematic error and imprecision
as
recommended
by Eisenhart
(4) and Ku ( 10). I have called
imprecision
negligible
if the standard error of measurements
made during an experiment
under controlled
conditions
is
less than one-sixth of the systematic error, or if the standard
error of measurements
made during calibration
is one-sixth or
less of the standard error of measurements
made during an
experiment
under controlled
conditions.
I have called systematic errors negligible
if they are less than one-half of the
standard
errors of measurements
made during
an experiment under controlled
conditions.
AERODYNAMIC
RELATIONS
If an object in free flight is confined to the test section of a
wind tunnel, then both the mean acceleration
of the center
of mass of the object and the sum of the force vectors acting
on the center of mass must be zero. The forces acting on the
object comprise a gravitational
force (weight,
W) and an
aerodynamic
force, which must be of equal magnitude
but
opposite in direction
to We The aerodynamic
force is conventionally
resolved into the two orthogonal
components,
lift (L) and drag (D) (Fig. 1), whose directions are defined
by the direction of the air velocity (V) relative to coordinates
originating
on the object that is generating
the aerodynamic
force. Lift is directed upwards and at right angles to V. Drag
is parallel to and in the same direction
as V. Lift and drag
need not be vertical
and horizontal,
respectively,
for V
need not be horizontal.
Additional
information
on the aerodynamic relations discussed in this section can be found in
Goldstein’s
texts (7).
The description
of the aerodynamic
forces on an object
.
in free flight in a wind tunnel can be complicated,
for, with
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
TUCKER,
VANCE A. Metabolism
during flight
in the laughing gull,
Lams
atricilla.
Am.
J. Physiol.
222(2):
237-245.
1972.-The
mean
metabolic
rates of two laughing
gulls (mean
body
mass,
0.30 kg) flying in a wind tunnel
were between
23 and 38 w/kg0*325
depending
on flight
speed and angle of flight.
A change
in the
intensity
of air turbulence
in the tunnel
from 0.88 to 1 .44y0 had
little effect on the mean metabolic
rate in flight.
The drag of the
mask and tube carried
by the gulls increased
metabolic
rates by as
much as 10 %. The metabolic
rates reported
for budgerigars
flying
in a wind tunnel
in a previous
study are reduced
as much as 18 y0
after correction
for the drag of the mask and tube. The metabolic
rates of the gulls in level flight,
after correcting
for the drag of the
mask and tube, were 6-8 times the metabolic
rates of the gulls
resting at temperatures
between
25 and 35 C, and 12-14 times the
mean basal metabolic
rate of nonpasserine
birds of the same size.
The rate of wing flapping
was almost
constant
at 3.8 flaps/set,
irrespective
of flight speed, and the mean respiratory
frequency
in
flight
was 2.5 breaths/set.
Carolina
238
V.
A. TUCKER
tudes of T and V. Thus, changes in P, at a given V when
the tunnel tilts can be calculated
from equation 3 and the relation :
Al’,
Since the power
rate, the partial
of the wings is:
/ccic
HOR
FIG. 1. Aerodynamic
and gravitational
forces on body
flying to left at air speed V along a path inclined
at angle
zontal. W = weight,
L = lift, D = drag, T = thrust.
of a bird
8 to hori-
L = -w
cos 0
(I>
and
T=
-
(D + W sin 9)
(3
where D is the drag of the bird’s body exclusive of the wings
and 8 is positive when V has a downward
component
(Fig.
1). For the remainder
of this paper, the symbols V, L, D,
and T when applied to birds will be used strictly with the
meanings that they have in the first sentence of this paragraph and in equations I and 2.
By tilting the longitudinal
axis of a wind tunnel in which a
bird is flying, one can quantitatively
add to or substract
from the magnitude
of T without
changing
the magnitude
of L appreciably.
When
the tunnel
tilts by angle 0
from horizontal,
V also tilts by 0, and the magnitude
of T
changes by an amount:
AT
= W sin 6
of L changes
AL
= W(1
(3)
by an amount:
-
co&)
(4)
Since cos 0 for the angles used in this study is within
1%
unity, AL is essentially zero as the tunnel tilts.
The rate at which the thrust of the bird’s wings does work
(power output, PO) is by definition the product of the magni-
(s>
where APi is the difference
between
the metabolic
rates
during
level flight and during
flight when the tunnel is
tilted.
Another way of interpreting
the effect of tilting the tunnel
is as a duplication
of level, ascending,
or descending
flight
through
still air. When the direction
of V in an ideal wind
tunnel
is horizontal,
the aerodynamic
and gravitational
forces on a bird are exactly the same as they would be for a
bird flying a level path through
still air. If the tunnel is
tilted through angle 8 so that the air velocity vector has an
upward
component
(8 negative),
the aerodynamic
and
gravitational
forces on the bird are exactly the same as if the
bird flew downward
in still air along a path inclined at 8 to
horizontal.
Ascending
flight is duplicated
if the tunnel is
tilted to positive values of 8.
The aerodynamic
forces acting on an object under specific conditions
can be used to predict the aerodynamic
forces under other conditions
if the Reynolds number for
the specific conditions
is known. Reynolds number is nondimensional
and is defined as:
Re = pVd/~
(7)
where p is air density ( 1.18 kg/m3 in this study), V is the air
velocity relative to the object, d is some specified linear dimension of the object, and p is the dynamic viscosity of air.
The ratio p/p in this study has the value 65,200.
Lift and drag forces on an object are often described in
terms of nondimensional
lift and drag coefficients (CL and
CD , respectively)
because these coefficients are functionally
related to Re and are almost constant over certain ranges of
Re values. Thus:
and
CD = 2D/(pW)
where
S is some specified
\
(9)
surface area of the object.
METHODS
Experimental
Animals
The two gulls used in this study were hand reared from
chicks and kept in an outdoor cage 6.1 m wide, 6.1 m long,
and 3.7 m high. They were fed on raw fish and canned cat
food. Their masses varied between 0.26 and 0.36 kg unless
otherwise
noted.
The gulls were trained to fly freely in an open-circuit
wind
tunnel that was mounted
on a cradle so that its long axis
could be tilted up to 8” from horizontal.
The wind tunnel
and the characteristics
of the airflow in the test section are
described in detail in the study of Tucker and Parrott (24).
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
the exception of gliding or diving, flight requires that some
parts of the object move relative to other parts. Thus, the
vectors V, lift, and drag all will have different values for
different parts of the object; and, for a given part, all will
vary with time. For example, in flapping
flight, V, lift, and
drag at any instant vary from the wing tip to the wing base;
and at any one wing region, they vary throughout
the stroke
cycle (see ref. 3 for a detailed discussion).
On the average, over a stroke cycle, bird wings produce a
mean aerodynamic
force that can be resolved into a lift
component
(L) and a drag component
(D), where the directions of the components
are defined by the direction
of V
relative to coordinates
originating
on the bird’s body exclusive of the wings. The drag component
is negative in sign
and is commonly
called thrust (T). Thus, for a bird flying
in a wind tunnel in which V is inclined
to horizontal
by
angle 0:
(5)
of a flying bird is the metabolic
of the work done by the thrust
EP = AP,/APi
I ZONTAL
W
The magnitude
input (Pi)
efficiency
= ATV
METABOLISM
OF
FLYING
GULLS
The airspeed in the test section could be controlled in steps
in the early experiments bychangingpulleys
on the fan drive,
but in later experiments the fan was fitted with a variablespeed motor. The air temperature in the wind tunnel was
between 23 and 31 C.
The first gull was trained to fly in the test section by lifting the bird from the floor with a rod until it learned to
avoid the rod by fluttering momentarily into the air. After
several days of training, the gull had learned to avoid the
rod for several minutes by flying, and after several weeks
of training, it would fly for 1 hr or more in the tunnel in
what appeared to be a normal manner.
The second gull was trained more rapidly by placing it in
the test section with the previously trained bird. The naive
bird was treated as described above, but within 15 min it
learned to fly for a few minutes when accompanied in flight
by the trained gull. After 2 weeks of training, the naive gull
would fly alone in the test section for 1 hr or more.
Metabolic rates of resting and flying gulls were determined from oxygen consumption
and carbon dioxide
production, which, in turn, were calculated from measurements of gas concentrations
and flow rates. The caloric
equivalent of oxygen was assumed to be 20 kJ/liter
(4.7
kcal/liter) .
Resting birds were confined in a Plexiglas cylinder 0.76
m long and 0.28 m in outside diameter with a wall thickness of 3.2 X IO+ m. The cylinder was kept in a dimly
lighted cabinet, the temperature of which was regulated
within 0.2 C of 20, 30, or 35 C. A gull was put into the
cylinder in midmorning and remained there without food
until early evening.
Rates of oxygen consumption and carbon dioxide production were calculated from the concentrations
of these
gases in the air entering and leaving the cylinder, and the
flow rate of air through the cylinder. The lowest rate of
oxygen consumption sustained for 1 hr or more during the
experiment was used to calculate the resting metabolic
rate. Between experiments, the gulls were exposed to summer weather in the outside cage.
Flying birds were fitted with a mask connected to a
vacuum system (illustrated in ref. 23) by a flexible vinyl tube
3.5 X lO+ m in outside diameter. The mask was made from
a piece of flat celluloid folded up and taped to a piece of
celluloid centrifuge tube 25 X low3 m in diameter (Fig. 2)
and was held on by a rubber band as described in Tucker’s
study (22). The mask had a mass of 4.4 X 10G kg. The part
of the tube supported by the flying bird had a mass of 6.0 X
lOmakg and a length of 0.8 m. Air entered the back of the
mask at 0.25-0.33 liters/set and swept the expired gases into
the vinyl tube before they could escape out the back of the
mask. The air then passed through a flowmeter (a rotameter), and a sample was pumped through a desiccant
(Drierite) to recording carbon dioxide and oxygen analyzers.
When the mask was in place on the gull, the pressure within
it was never more than 2 mm of Hz0 below the static pressure outside the mask. The vacuum system regulated the
pressure downstream
from the flowmeter at 460 mm Hg
absolute. Flights during which metabolic measurements
were made lasted 0.5 hr or more, and usually two or three
measurements of metabolic rate were made each day.
The carbon dioxide analyzer (Beckman infrared analyzer,
model 215) was used in the following manner to determine
the carbon dioxide production of a flying gull. The concentration of carbon dioxide in the air coming from the mask
was recorded while the gull was flying. The respiratory rate
was so rapid that breath-to-breath
variations in concentration were not recorded. After each flight, the mask was removed from the bird, the end of a plastic tube (PE-100) was
placed in the mask, and the opening of the mask was plugged
with a porous ball of cheesecloth. Pure COz flowed through
aflowmeter
and the plastic tube at a rate determined by a
needle valve adjusted until the analyzer record came to ca.
10 % above the recording obtained from the flying bird.
The flowmeter reading was noted and the procedure was repeated; this time the needle valve was adjusted to produce a
record slightly below that recorded from the flying bird.
Then the rate of carbon dioxide production of the bird was
calculated by linear interpolation. This null-balance technique took only a few minutes and did away with the need
to measure the carbon dioxide concentration
in the air
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
Metabolic Measurements
239
240
V.
% turbulence
= 100 %Vrms/Vdc
where Vrms and Vdc are the root-mean-square
a-c voltage
(corrected
for instrument
noise) and d-c voltage, respectively, of a linearized
anemometer.
At eight points spaced
in a three-dimensional
array in the region where the gull
flew, the mean high-turbulence
intensity was 1.45 % (SD =
&0.50), and the mean low-turbulence
intensity was 0.88 %
(SD = ~0.26).
These mean values correspond
to turbulence
factors of 2.1 and 1.7, respectively ( 16).
The turbulence
through which birds fly in nature varies
from zero on calm days to greater than the high-turbulence
intensity
measured
in the wind tunnel.
For example,
I
measured a mean turbulence
intensity of 3 % at a location
20 m above the forest canopy near Durham,
N.C. on a
day when the mean windspeed was 7 m/set.
The imprecision
of the root-mean-square
voltmeter
was
negligible
and the readings were accurate within 5 %. Both
the systematic error and the imprecision
of the d-c voltmeter were negligible.
Body mass and metabolic rate. The effect of changes in body
mass on the oxygen consumption
during flight was investigated in one gull whose mass changed by 27 % during the
course of the experiments.
In the early summer of 1969,
measurements
were made when the bird had a mass between 0.328 and 0.352 kg. In the winter of 1969, the bird’s
mass dropped from 0.420 to 0.383 kg during 3 weeks of experiments.
All measurements
were made at an airspeed of
10.8 m/set at the low level of turbulence.
Body mass was measured wi th a Mettler balance, the syste matic error and imprecision
of which was negligible.
Drag of Mask and Tube
The drag of the tube between the points where it attached to the mask and touched the floor of the test section
was calculated from the angle (+) of the tube axis relative to
horizontal
as the tube hung suspended from the mask and
streamed in the wind at various-airspeeds.
Values of 4 were
measured
with an optical
protractor.
They had a systematic
error of less than l/Z”
and an imprecision
estimated to be l/4”.
The value of 4 at any point results from the direction
of
the vector sum of two orthogonal
forces on the tube, weight
and drag, so that :
D = w cot 4
W)
Both W and D increase proportionately
with length along a
tube of constant diameter and wall thickness, and it follows
that such a tube should be straight when streaming
in the
wind ($ constant at all points along the tube). This condition was observed.
The drag of the mask was measured by mounting
a gull
model (a stuffed skin) on a one-component
flight balance
and measuring
the change in drag as the mask was placed
on the gull model or removed at various airspeeds. The
flight balance was a vertical brass rod 0.8 m long that passed
through the floor of the test section. The base of the rod was
equipped with strain gauges that measured the strain caused
by the drag of the rod and gull model. The gull model was
mounted
on the end of the rod and was appropriately
oriented in the airstream with its wings tightly folded. The
flight balance was calibrated
by hanging
weights from a
thread that ran over a pulley and attached to the gull model
so as to apply force in the same direction
as the drag vector.
The changes in strain were kept so small that they did not
significantly
alter the position in space of the center of mass
of the rod-model
system. Changes in strain caused by alterations in the center of mass when the mask was attached or
removed were measured and corrected for.
The systematic error of the flight balance was negligible,
and the standard error of the value of drag determined
at
each airspeed was h2.3 X 10v3 newtons (N).
The drag of the mask and tube was partially or completely
compensated
for during metabolic
measurements
on flying
gulls by tilting the wind tunnel so that the air velocity in
the test section was inclined upward by 1.5O relative to horizontal (0 = - 1.5 “). This procedure adds a weight component
in a direction
opposite to the drag vector and compensates
for the drag of the mask and tube at a speed of 8 m/set. For
convenience,
flights when 0 was - 1.5” will be referred to in
the text following
as level, unless otherwise noted. Metabolic
measurements
were also made at 8 = -2.5”
and at 0 =
- 6.5”.
Wingbeat
and Respiratory Rates
Wingbeat
rate was measured by synchronizing
a stroboscope with the wingbeats.
Respiratory
rate during flight was
measured
by recording
from a thermocouple
circuit with
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
drawn into the mask. A similar technique
was used in experiments
with resting birds, except that the pure carbon
dioxide was delivered to the system from an infusion pump.
The oxygen analyzer (Beckman G2, described in ref. 22)
was calibrated
by changing
the total pressure of air flowing
through
the paramagnetic
sensing unit. Oxygen consumption was calculated
using equations
from Tucker’s
study
(22). Equation 2 was used if carbon dioxide production
was
known. Otherwise,
equation 3 and an estimate of respiratory
quotient were used.
All flowmeters
were
calibrated
with
Vol-U-Meters
(Brooks Division,
Emerson Electric Co.) for the particular
pressure and gas with which they were to be used. The systematic error of the flowmeter
readings was negligible,
and
the standard
error was less than 2 % of the value used in
calculations.
The gas analyzers had systematic errors and
imprecisions
that were negligible.
Air turbulence and metabolic rate. Metabolic
measurements
were made for one gull flying at a variety of speeds at two
levels of air turbulence.
High turbulence
was generated
by
a hardware-cloth
screen (13 X 1Oe3 m mesh; wire diameter,
1.2 X low3 m) at the front of the tkst section and by a vertical strut centered in the entrance cone of the wind tunnel.
Lower turbulence
was achieved by removing
the strut and
replacing
the hardware
cloth with a net made of fine
thread (diagonally
stretched mesh, 3.0 X 10e2 m; thread
diameter, 0.14 X 10s3 m).
Turbulence
was measured with a DISA 55D05 hot-wire
anemometer
connected
to a true root-mean-square
voltmeter with a frequency response of 5 Hz to beyond 50 kHz.
The anemometer
measured
the percentage
of turbulence
defined as :
A. TUCKER
LMETABOLISM
OF FLYING
GULLS
241
one junction
placed in the mask. Each time the bird exhaled, the subsequent warming
of the thermocouple
could
be clearly recorded.
Systematic error and imprecision
were
negligible
for both wingbeat
rate and respiratory
rate.
RESULTS
Behavior
Metabolic
Measurements
(SE of estimate
Pi =
1.562 + 0.325 log m
= =tO.OlS),
W)
or:
pi = 36.4
The exponent
I
I
I
I
I
0
I
0.34
24
0.32
I
0.36
MASS,
I
0.38
KG
I
0.40
I
0.42
FIG. 3. Metabolic
rate of a laughing
gull flying level at a speed of
lo.8 m/set with different
values of body mass. Each point is a single
measurement.
Line was fitted as described
in text.
in Fhgh t
Efect of body mass. Thirteen
measurements
were made on
one gull flying at a speed of 10.8 m/set in low air turbulence
with body mass (m) between 0.328 and 0.420 kg (Fig. 3).
The linear least-squares
relation
between
log metabolic
rate (Pi , in watts) and log body mass was:
log
29
has 95 % confidence
mo-325
limits
(12)
of 0.05 and 0.60.
15 ’
6
I
7
I
8
I
I
I
I
9
IO
II
12
SPEED,
M/SEC
I
13
4. Metabolic
rates relative
to body mass of laughing
gulls
flying
under
different
conditions.
Each point
is a mean of 5-l 1
measurements.
Standard
error of each ,mean is between
A 0.2 and
AI 2.2 (mean SE = A 1 .O). Curved
line without
points was fitted to
all means for level flight as described
in text. q : larger
gull, level
flight,
low turbulence;
0: larger
gull, level flight,
high turbulence;
A: smaller
gull, level flight,
low turbulence;
X : larger
gull, 8 =
-6.5 O, high turbulence.
FIG.
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
Both gulls learned to fly in the center of the test section
so that their flight usually appeared
identical
to the steady
flight of wild gulls. However,
at one time or another, both
displayed
aberrations
in flight that were caused by the
wind-tunnel
environment.
They had a tendency
to fly at
the top of the test section, striking the ceiling (which was
covered with a protective
net) with their wings at the top
of the upstroke.
This behavior
occurred
most frequently
during the first 10 min of a flight. However,
even after 0.5
hr of flight, the gulls would fly at the top of the test section
for a few seconds of each minute.
The erratic flight described above increased in frequcney
when the tunnel was tilted to negative values of 8. When 8
would glide for several secwas - 6.5’, a gull occasionally
onds, remaining
motionless relative to the tunnel. Yet most
of the time the bird would flap at the top of the test section,
working harder than necessary to stay airborne.
One gull, when flying at speeds less than 9 m/set, developed the habit of continually
flying into the net at the
front of the test section. Measurements
were discarded when
the gull behaved in this mannerWhen wearing
masks, the gulls appeared
to flap more
laboriously
than when they were unmasked.
In addition,
the gulls without
masks would fly for several minutes at a
speed of 14 m/set, but, when wearing the mask, they would
not fly faster than 13 m/set for more than 1 min. The
masked gulls usually flew with their beaks open, although
their beaks usually were closed in the absence of a mask.
Both gulls exhibited
two distinct types of wingbeat,
depending on the speed of flight. At speeds below 9 m/set, the
wing tips had a slower downward
velocity during the early
part of the downstroke
than they did at higher flight speeds.
This difference
caused the birds to have a characteristic
snapping
motion
of their wings during
the downstroke
when flying at speeds above 9 m/set. The same snapping
motion can be observed in free-living
gulls. The change in
wingbeat
pattern
occurred
at the same speed at which
metabolic rate increased in a stepwise manner as flight speed
increased (Fig. 4).
The mean mass of both birds for all of the metabolic
measurements
during
level flight was 0.321 kg (SD =
ztO.026, n = 112). One gull was smaller (mean- mass =
n = 25) than the other
0.277 kg, SD = hO.014,
(mean mass = 0.322, SD = & 0.020, n = 87). All of
these masses include that of the mask and suspended tube
(1 X lO-2 kg).
Efect of speed. Since equation 12 indicates that metabolic
rate in flight at a constant speed is proportional
to body
mass raised to the 0.325 power, all metabolic
rates subsequently
measured
during
flight were expressed per mass
raised to the 0.325 power. This procedure
yields a value that
should be insensitive
to variations
of body mass in an individual bird. Over a speed range from 6.1 to 12.6 m/set, the
mean metabolic
rates for each of the two gulls flying level
(one gull flew in both high- and low-turbulence
air) varied
from 23 to 38 w/kg o*325 (Fig. 4). At each speed, the metabolic rate of the smaller gull was lower than that of the
242
V.
larger gull. Berger et al. (1) reported
a metabolic
rate of
28.3 w/kg0a325 for a flying ring-billed
gull with a mass of
0.427 kg.
A curve (Fig. 4), fitted by least squares to the means of all
measurements
during level flight had the equation:
p/
i mO.325
=
(v
-
6.0)2/(5.00
+ 27.3)
Measurements
Pi = (13.4 -
at Rest
Five metabolic
measurements
were made on each of two
resting gulls at temperatures
of 20, 30, and 35 C. Metabolic
(SE of estimate = hO.76,
The mean respiratory
taneously
with
resting
( SD = ~tO.050, n = 30).
0.14 T,)m
(14
n = 30).
quotient
at rest, measured simuloxygen
consumption,
was 0.70
Drag of Mask and Tube
The drag of the mask increased
creased according
to the equation:
D = 0.0013
V -
linearly
0.0065
as airspeed
in(15)
Actual measurements
of drag were within 2 X 10m3 N of the
values given by this equation
between 6 and 11 m/set, but
were 15 X 1Ow3 N high at 13 m/set.
The tube streamed
in the wind at increasingly
acute
angles to horizontal
as airspeed increased from 6 to 13 m/
set, and its total drag increased
according
to the linear
least-squares
equation :
D = 0.0128
V -
0.0185
u6)
(SE of estimate
= =tO.O0235, n = 5). This equation can be
combined
with equation 9 to yield an expression for the drag
coefficient per unit length of the tube as a function of airspeed. When equation 9 is applied to a cylinder, D is defined
as drag per unit length and S is defined as the outside diameter of the cylinder. Thus, for the tube:
CD =
7.65/V
-
11.1/V2
WI
The drag coefficient for the tube at the lowest speed (CR =
that measured by
0.97, v = 6 m/set, 4 = 45”) approaches
others for a long cylinder
oriented
with its long axis perpendicular
to the direction
of airflow. Such a cylinder has a
CD of 1.0 at a Reynolds number of 2,300 (7), which is the
appropriate
Re value for the tube attached
to the mask.
(Reynolds
number
for a cylinder
is defined by equation 7
with d equal to the outer diameter of the cylinder.)
The total drag added to a flying gull by the mask and tube
is given by the sum of equations 15 and 16:
D,
Wingbeat
= 0.0141
V -
0.025
WI
and Res@‘ratory Rates
The wingbeat
rate, measured at speeds between 8.6 and
11.2 m/set and 8 between 0 and - 6.5”, was virtually
con(SD = hO.096,
stant with a mean value of 3.78 beats/set
12 = 2 1). The respiratory
rate over the same range of speed
and angles was more variable, with a mean value of 2.54
(SD = JrO.232, n = 36). The ratio of the mean
breaths/set
wingbeat
rate to the mean respiratory
rate is 1.5. Respiratory rates were first measured at lo-15 set after the onset of
flight, and did not change consistently
before the termination of the experiments,
which lasted 2-3 min.
Berger et al. (2) measured a midrange
wingbeat frequency
and mean respiratory
frequency of 4.6 and 2.0, respectively,
during
flight in the ring-billed
gull (Larus delawarensis,
mass = 0.376-0.475
kg). Th e ratio of these rates is 2.3.
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
Metabolic
rate during rest was assumed to be directly proportional
to
body mass, and was expressed as watts per kilogram
to correct for variations
in body mass. The linear least-squares
equation
fitted to the relation between metabolic
rate and
air temperature
(TJ was:
(13)
(SE of estimate = zt3.69, n = 14). Prior to the least-squares
fitting procedure,
the curve was linearized
by using (V 6.0)2 as the independent
variable. Equation 13 represents a
parabola
with its vertex at 6.0 on the abscissa, and 27.3 on
the ordinate.
The rate of change of metabolic
rate with speed was
slightly but consistently
higher between
8 and 10 m/set
than at other speeds (Fig. 4). Although
I averaged out this
phenomenon
in the mean curve described above, it occurred
in both gulls and in both level and descending
flight (e =
-6.5”).
It occurred at the same range of speeds where the
wingbeats
changed from one pattern
to another
(see Behavior).
Effect of duration. Usually, the metabolic
rate of a flying
gull was high at the onset of flight and declined 15-20 % to a
plateau in 15 to 20 min, where it remained
until the termination
of the experiment
after 10 or 20 min more. Occasionally,
the metabolic
rate was almost constant over the
measurement
period. The decline in metabolic
rate with
time correlated
with the observation
that the gulls flew erratically for the first 10 min of flight (see Behavior).
E$ect of angle. Measurements
were made only during
level flight or flight at negative values of 9, as neither bird
would fly long enough at positive values of 0 for measurements to be made. One gull, flying at the high-turbulence
level, always decreased its metabolic
rate at a given speed
when the tunnel was tilted from level to - 6.5 O (Fig. 4).
However,
the behavior
of the gull suggested that the observed decrement in metabolic rate would have been greater
had the gull not struggled at the top of the test section (see
Behavior).
Ten measurements
were made on the other gull flying at
11.7 m/set in air at the low-turbulence
level in level flight
and at an angle of -2.5”.
At the latter angle, the bird appeared to fly in the same manner as in level flight, and its
mean metabolic
rate decreased by 2.9 w/kg”.325. The standard errors of the means for level and descending flight were
=tO. 16 and rtO.39, respectively.
Effect of air turbulence. Metabolic
measurements
were
made on one gull flying at the high- and low-turbulence
levels at various speeds. At the high-turbulence
level, the
metabolic
rates were more variable
and were sometimes
lower than at the low-turbulence
level (Fig. 4).
Resfiiratory quotient in flight. Oxygen consumption
and carbon dioxide production
were measured
simultaneously
in
one bird flying at various speeds. The mean respiratory
the
quotient
was 0.74 (SD = ~tO.040, n = 24). Between
onset of flight and the first determination
of respiratory
quotient,
1 or 2 min elapsed. There was no consistent
change in respiratory
quotient from the beginning
to the end
of an experiment,
which lasted 20 to 30 min.
A. TUCKER
METABOLISM
OF FLYING
GULLS
DISCUSSION
results describe the energetic cost of flight
The preceding
for laughing
gulls in a wind tunnel, but does this description apply to birds in nature? To answer this question, sevenvironment
and
era1 differences between the wind-tunnel
that in nature must be evaluated. The differences that I will
discuss are : effects due to pressure gradients and proximity
of the walls to the test object in the wind tun .nel (boundary
effects), air turbulence
in the wind tunnel, and drag of the
mask and tube.
Boundary Efects
Extremely
high levels of turbulence
can increase the energetic cost of flight. The mean metabolic rate of budgerigars
flying at 10 m/set in a highly turbulent
closed-circuit
wind
tunnel (20) was almost twice as high as the mean metabolic
rate measured at the same speed in a less turbulent,
opencircuit tunnel (22). The turbulence
intensities in these tunnels, measured
with the techniques
described
here, were
43 % and 4%, respectively.
Drag of Mask and Tube, and Tunnel
Gulls. The flying gulls were hindered
to some extent
during metabolic
measurements
by the aerodynamic
drag of
the mask and were aided by the tilt of the wind tunnel. In
this section I will develop equations that correct the measured power expenditures
of the flying gulls for these two
effects.
When a bird wears a mask, its drag increases, and it must
increase thrust to maintain
its position in the test section.
The bird also must change its thrust when the wind tunnel
is tilted from the horizontal
by angle 8. If D, is the drag
added by the mask and its attached tube, the total increase
in the magnitude
of thrust to compensate for mask and tube
drag and tunnel tilt is given by:
AT
Turbulence
The metabolic
rate during flight was virtually
unaffected
by changes in the level of turbulence
in the wind tunnel.
This finding is interesting
because the aerodynamic
forces on
model aircraft or wings in wind tunnels can be profoundly
affected by changes in turbulence
levels, particularly
when
the models have Re values (about 100 X 103) similar to
those for the birds in this study. The sensitivity
of such
models to turbulence
can be reduced by roughening
the surface of the model. Roughening
of the surface can cause transition of the boundary layer from laminar to turbulent
flow,
and thereby prevent separation
of the boundary
layer (7, 9,
17, 18, 19).
The drag coefficient
of wing models shaped like bird
wings can decrease by 20 % or more when transition
of the
boundary
layer occurs at the appropriate
region (7, 18).
In addition,
the lift coefficient increases as the drag coefficient decreases, so that the ratio of lift to drag can increase
by 50 % with transition
( 18). Feldman
(5) measured a 50 %
increase in lift-to-drag
ratio when transition
occurred
in
an appropriate
region on the wings of a plaster model of a
gull mounted in a wind tunnel.
It is not surprising
that the metabolic
rates of flying gulls
are insensitive to turbulence
intensities up to 1.45 %. Gulls
in nature fly in air with different turbulence
levels, and one
might expect that they have evolved surfaces that keep the
energetic cost of flight as low as possible. The roughness of
the feathers on gull wings probably is enough to cause transition of the boundary
layer, whether the bird flies in turbulent or nonturbulent
air. The overlapping
feathers that cover
the wing form a series of sharp valleys and rounded peaks,
and the relief of these Surface features is between 0.5 X IO-3
and 1 )( 10-S m. This amount of roughness is more than
enough to cause transition
of the boundary layer at the leading edge of a Aat plate exposed to a nonturbulent
airflow of
10 m/set (7).
Tilt
=D,+Wsine
The increase in power output
increase at flight speed V is:
AP,
(PJ
u9)
required
= ATV
to provide
this
mo
The amount that a bird must increase its metabolic
rate
(power input, Pi) to achieve this increase in P, depends on
the partial efficiency (E, , equation 6), so that
APi
= (D, -I- W sin @V/E,
(24
If E, can be determined,
equation 21 can be solved, and APi
can be subtracted
from the measured power input of the
gulls flying in the wind tunnel to get a value for the power
input of a gull in level flight and unhindered
by the drag of
the mask.
Partial efficiency can be determined
from measurements
of Pi of gulls flying in a level and a tilted tunnel. I encountered
difficulty
in making
appropriate
measurements
because the gulls changed
their flight pattern when the
tunnel was tilted by several degrees (see Behavior). However,
one gull flew at 11.7 m/set in what appeared to be an unchanged manner as the tunnel was tilted from 8 = - 1.5’
toe = - 2.5”. The change in power output of the bird as
the tunnel tilted was:
APO = WV
where
A sin 6
(22)
A sin 8 = sin 1”. The change in power input was 2.9
w/k o*325.These data indicate an E, of 0.3 for a bird of mean
mass in this study (0.322 kg).
Partial efficiency varies with flight speed in the only other
flying animal for which it has been measured, the budgerigar
(22). In this bird, E, varied between 0.19 and 0.28 at flight
speeds between 5.3 and 13.3 m/set (Table
1). Although
I
have no data on changes in E, with flight speed for gulls,
changes in E, such as those measured for budgerigars
have
a relatively
small effect on the values of power input corrected for mask drag and tunnel tilt. The equation that de-
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
The pressure gradients
that exist down the length of a
wind tunnel and the proximity
of the walls of the test section
to the test object can produce forces on the test object that
do not occur when the object moves through a large volume
of still air. These forces have been measured for fixed-wing
aircraft ( 16) but not for flapping-wing
aircraft. However, a
gliding
bird with the dimensions
of the gulls used in this
study would have corrections for boundary effects that would
change its lift and drag by less than 5 % (24). Consequently,
boundary
corrections
are assumed to be insignificant
in this
studv.
243
244
V. A. TUCKER
TABLE
leveljlight
Flight
1. Correction of budgerigar
for mask and tube drag
-
Speed,
Mean Measured
Power Input, w
m/set
5.3
6.7
7.8
9.7
11.7
13.3
Body
metabolism in
mass
CorrIe$33dt
6.34
5.40
4.45
4.27
5.01
6.67
= 0.035 kg.
.27
.28
.20
.19
.19
.28
E, = partial
P;wer
,
6.23
5.22
4.10
3.67
4.12
5.87
pi’/
-
0.0204/E,)
-
efficiency.
V(2.40 - 0.156/E,)
+ 34.5)m0*325
(23)
= (0. 130v2 -
1.86V
j- 34.5)m0.325
+ 31 m V sin 8
(21)
This equation is obtained by adding a term to equation 23 to
account for the change in power input required
for nonlevel flight. The term for nonlevel flight is:
APi
= WV
sin 8/E,
= 9.81 mV sin 8/E,
(0.180
(V -
9.7)2 + 3.52)m
0.035
(26)
(25)
The foregoing
analysis shows that the amount of power
(Pi) that the masked gulls expended
during level flight in
the wind tunnel is close to the amount (Pi”) calculated
for
0 = 0 after correcting
for the drag of the mask. For example, Pi” (equation 24) for a gull with a mass of 0.322 kg is 2 %
greater than Pi (equation 13) at 6 m/set. At 12 m/set, Pi” is
10 % less than Pi .
Budgerigars. The measurements
of drag of the mask and
tube for the gulls make it possible to correct for the drag of
the mask and tube in a previous study of the flight energetics of budgerigars
(22). Since the drag of the gull mask is
small (less than 15 % of the drag of the tube), and the budgerigar mask fitted even more closely than the gull mask,
the drag of the mask in the budgerigars
can be neglected.
The CD values for the tube attached
to the budgerigar
mask should be about the same as those for the tube attached to the gull mask. Although
the Re value (1,400)
for the tube (od, 2.1 X 100~ m; length, 0.15 m) attached to
the budgerigar
mask is only 60 % of that for the gull mask,
the CD values for cylinders do not change appreciably
over
this range of Re values (7). Thus, the drag of the tube for
the budgerigar
mask can be calculated
as a function
of
speed by substituting
equation 9 into equation 17. The increment of power input due to the drag of the tube then can
be calculated with equation 21, since E, values for flying budgerigars have been measured
(22).
The power inputs of budgerigars
in level flight are re-
Comparison of Resting and Flying
Gulls
Metabolic rates. The mean metabolic
rates of a 0.322kg
laughing
gull in level flight at speeds between
6 and 13
m/set are 5.6-8.2 times the mean metabolic
rates of the
gull at rest at temperatures
between
20 and 35 C. These
ratios were calculated
from equations 14 and 23. The metabolic rates in flight are 12-14 times the mean basal metabolic rate (1.64 w) of nonpasserine
birds with masses of
0.322 kg. This basal metabolic
rate (Pib) was calculated
from the equation:
(Modified
from ref. 12.)
Thermal conductance. At rest, virtually
all of a bird’s metabolic rate is lost as heat to the environment
by conduction,
radiation,
and evaporation
of water. The latter avenue
accounts for about 29 % of the heat loss of a resting bird
with a mass of 0.322 kg at temperatures
below 30 C (assuming 2.42 X 10M3 kg water evaporated
per liter of oxygen
consumed
(1 l), 2.43 X lo6 J/kg water evaporated
and
20.1 X lo3 J/liter
oxygen). In flight, perhaps 20 % of the
bird’s metabolic
rate could be transferred
to the environment as work done to overcome drag. The remaining
80 %
is lost as heat and, of this, perhaps 18 % is lost by evaporation (based on figures for the flying budgerigar
(22)).
With these figures in mind, one can calculate
the thermal
conductance
(C) of resting and flying birds with the equation :
total heat loss - heat loss by evaporation
C =
body temp - air temp
(28)
A 0.322-kg gull flying at 12 m/set and 30 C has a metabolic rate of 21.4 w. Of the 17.1 w that are lost as heat,
14.0 w are lost by conduction
and radiation.
If the gull has
a constant body temperature
of 43 C, its thermal conductance is 1.1 w/“C.
The same gull at rest at 30 C has a heat loss of 2.96 w, of
which 2.10 w are lost by conduction
and radiation.
If the
gull’s body temperature
is 41 C, its thermal conductance
is
0.19 w/“C.
Thus, the thermal conductance
in flight is 5.8
times that at rest. The budgerigars
increased their thermal
conductance
in flight to 5.0 times the value at rest (22).
No small mammals
have been shown to increase their
thermal
conductance
values by as much as a factor of 4
(22, 26). The large increase in thermal
conductance
in
flying birds is probably
related
to increased
convection
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
This equation
is a combination
of equations 13, 18, and 21.
A change in E, of 0.1 from an initial value of 0.3 causes a
change of less than 7 % in the power input calculated
from
equation 23 when V = 13 m/set and m = 0.322 kg.
The complete
equation
for power input during
flight
without
a mask in laughing
gulls is a function
of flight
speed, body mass, and angle of flight and, taking E, = 0.3,
is given by:
Pi”
=
This equation
fits the data in Table 1 within 4 % or better
except at the slowest speed of 5.3 m/set, where it is 12 %
too high.
scribes the power input of a gull flying level with a body mass
of 0.322 kg and unhindered
by a mask and tube is:
Pi’ = (V2(0.200
duced 2-18 % when the increment
of power input due to
drag of the tube is subtracted
from the measured
power
input (Table
1). After correcting
for tube drag, the power
inputs for a 0.035-kg budgerigar
flying level at speeds between 5.3 and 13.3 m/set
are given by the parabolic
equation :
METABOLISM
OF
FLYING
and exposure of the thinly feathered
and sides of the thorax in flight.
Mrs.
Klopfer
available
245
GULLS
undersides
Marsha
Poirier
provided
technical
reared
and maintained
the laughing
for this study.
of the wings
assistance.
gulls, and
Dr. P. H.
made them
This study was supported
Grant
(no. GB 6160X),
and,
Sciences
Support
Grant
(no.
Grant
(no. 153-5924).
Received
for publication
by a National
Science
Foundation
from Duke
University,
a Biomedical
303-3215),
and a Research
Council
24 May
197 1.
REFERENCES
1. BERGER, M.,
consumption
J. S. HART, AND 0. 2. Rou. Respiration,
oxygen
and heart rate in some birds during
rest and flight.
2. Vergleich.
Physiol.
66 : 201-214,
1970.
2. BERGER, M., 0. 2. ROY, AND J. S. HART. The co-ordination
between
respiration
and wing beats in birds. 2. Vergleich.
Physiol.
66: 190-200,
1970.
3. CONE, C. D., JR. The aerodynamics
of flapping
birdflight.
Virginia
Inst. Mar.
Sci., Spec. Sci. Rept. 52, 1968.
4. EISENHART, C. Expression
of the uncertainties
of final results. In:
Precision
cedures.
Measurement
Natl.
Bu. Std.,
and Calibration.
U.S., Sp ec. Publ.
Statistical
300, vol.
Concepts
1, p. 69-72,
am Model1
and
Pro-
1969.
einer
439-446,
1969.
S. Modern
Developments
in Fluid
Dynamics.
An Account
of Theory
and Experiment
Relating
to Boundary
Layers,
Turbulent
Motion
and Wakes.
New York : Dover,
~01s. I and II, 1965.
7. GOLDSTEIN,
8. JOHNSTON, D. W., AND R. W. MCFARLANE.
Migration
and bioenergetics
of flight in the Pacific Golden
Plover. Condor
69: 156168, 1967.
9. KRAEMER,
K. Fliigelprofile
im kritischen
Reynoldszahl-Bereich.
Forsch.
10. Kv,
and
Gebiete
Ingenieurw.
27: 33-46,
1961.
Precision
Measurement
and Calibration.
Procedures.
Natl.
Bur. Std., U.S.,
Spec. Publ.
H. M.
Statistical
300, vol.
Concepts
1, 1969.
A. CALDER, JR. A preliminary
allovariables
in resting birds. Respirat.
11. LASIEWSKI, R. C., AND W.
metric analysis of respiratory
Physiol.
11: 152-166,
1971.
12. LASIEWSKI, R. C., AND W. R. DAWSON. A re-examination
of the
relation
between
standard
metabolic
rate and body weight
in
birds. Condor
69: 13-23,
1968.
and
Conversion
E. A. The International
Factors
(Revised).
System of Units.
NASA
SP-7012,
Physical
1969.
Constants
14. MOREAU,
R. E. Problems
of Mediterranean-Saharan
migrations.
Ibis 103: 373-427,
1961.
15. NISBET, I. C. T., W. H. DRURY, JR., AND J. BAIRD. Weight-loss
during
migration.
Part I: Deposition
and consumption
of fat by
the Blackpoll
Warbler
Dendroica
striata.
Bird-Banding
34 : 107-l 38,
1963.
Low-Speed
Wind
Tunnel
Testing.
16. POPE, A., AND J. J. HARPER.
New
York:
Wiley,
1966.
Hydroand Aeromechanics.
17. PRANDTL, L., AND 0. G. TIETJENS. Applied
New York: Dover,
1934.
18. SCHMITZ, F. W. Aerodynamik
des Flugmodells.
Duisburg,
Germany:
Carl Lange,
1960.
19. TANI, I. Low-speed
flows involving
bubble
separations.
Progr.
Aeronaut.
Sci. 5: p. 70-103,
1964.
consumption
of a flying bird. Science 154:
20. TUCKER, V. A. Oxygen
150-151,
1966.
V. A. Respiratory
physiology
of house sparrows
in
21. TUCKER,
relation
to high-altitude
flight.
J. Exptl.
Biol.
48: 55-66,
1968.
22. TUCKER, V. A. Respiratory
exchange
and evaporative
water loss
in the flying budgerigar.
J. Exptl.
Biol.
48: 67-87,
1968.
23. TUCKER, V. A. The energetics
of bird flight. Sci. Am. 220: 70-78,
1969.
24. TUCKER, V. A., AND G. C. PARROTT. Aerodynamics
of gliding
flight
in a falcon and other birds. J. Exptl.
Biol.
52: 345-367,
1970.
25. TUCKER, V. A., AND K. SCHMIDT-KOENIG.
Flight speeds of birds
in relation
to energetics
and wind directions.
Auk 88: 97-107,
1971.
regulation
and the effects of water
26. WUNDER, B. A. Temperature
restriction
on Merriam’s
chipmunk,
Eu tamias
merriami.
Camp.
Biochem.
PhyJiol.
33 : 385-403,
1970.
Downloaded from http://ajplegacy.physiology.org/ by 10.220.33.3 on April 2, 2017
5. FELDMANN,
I. F. Windkanaluntersuchung
Mijwe.
Aero-revue,
Zurich
19: 219-222,
1944.
standard
6. GAGGE, A. P., J. D. HARDY, AND G. M. RAPP. Proposed
system of symbols
for thermal
physiology.
J. Appl.
Physiol.
27:
13. MECHTLY,