Content uploaded by Roger P Croll
Author content
All content in this area was uploaded by Roger P Croll on Jul 16, 2018
Content may be subject to copyright.
The Contribution of the Swimbladder to Buoyancy in the
Adult Zebrafish (Danio rerio): A Morphometric Analysis
George N. Robertson,
1
Benjamin W. Lindsey,
1
Tristan C. Dumbarton,
2
Roger P. Croll,
2
and Frank M. Smith
1
*
1
Department of Anatomy and Neurobiology, Dalhousie University, Halifax, Nova Scotia B3H 1X5, Canada
2
Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 1X5, Canada
ABSTRACT Many teleost fishes use a swimbladder, a
gas-filled organ in the coelomic cavity, to reduce body
density toward neutral buoyancy, thus minimizing the
locomotory cost of maintaining a constant depth in the
water column. However, for most swimbladder-bearing
teleosts, the contribution of this organ to the attainment
of neutral buoyancy has not been quantified. Here, we
examined the quantitative contribution of the swimblad-
der to buoyancy and three-dimensional stability in a
small cyprinid, the zebrafish (Danio rerio). In aquaria
during daylight hours, adult animals were observed at
mean depths from 10.1 66.0 to 14.2 65.6 cm below the
surface. Fish mass and whole-body volume were linearly
correlated (r
2
50.96) over a wide range of body size
(0.16–0.73 g); mean whole-body density was 1.01 60.09
gcm
23
. Stereological estimations of swimbladder volume
from linear dimensions of lateral X-ray images and
direct measurements of gas volumes recovered by punc-
ture from the same swimbladders showed that results
from these two methods were highly correlated (r
2
5
0.85). The geometric regularity of the swimbladder thus
permitted its volume to be accurately estimated from a
single lateral image. Mean body density in the absence
of the swimbladder was 1.05 60.04 g cm
23
. The swim-
bladder occupied 5.1 61.4% of total body volume, thus
reducing whole-body density significantly. The location
of the centers of mass and buoyancy along rostro-caudal
and dorso-ventral axes overlapped near the ductus com-
municans, a constriction between the anterior and poste-
rior swimbladder chambers. Our work demonstrates
that the swimbladder of the adult zebrafish contributes
significantly to buoyancy and attitude stability. Further-
more, we describe and verify a stereological method for
estimating swimbladder volume that will aid future
studies of the functions of this organ. J. Morphol.
269:666–673, 2008. Ó2008 Wiley-Liss, Inc.
KEY WORDS: teleost; cyprinid; gas bladder; stereology;
anatomy; morphology
A significant portion of the body mass of teleost
fishes is composed of tissues such as bone and
muscle that are denser than water (Alexander,
1993). Consequently, compensatory mechanisms
are required to reduce the overall density of the
body, in order to decrease the energetic cost of
swimming to maintain vertical position in the
water column. Pelagic teleosts have evolved sev-
eral strategies to overcome the inherent negative
buoyancy related to the density of body tissues.
These include the synthesis of large amounts of
lipid, the development of ‘‘watery’’ muscles, the
reduction of bone mass, and the presence of an in-
ternal gas-filled chamber, the swimbladder
(Alexander, 1972, 1989; Lefrancois et al., 2001). Of
all the mechanisms used by fish to reduce total
body density, the swimbladder has been proposed
to be the most energy efficient (Alexander, 1993).
The volume of low-density gas in the swimbladder
offsets the higher density of body tissues so that
fish possessing swimbladders are very close to neu-
tral buoyancy at a specific depth in the water col-
umn. However, the quantitative contribution of the
swimbladder to the attainment of neutral buoy-
ancy has been established for only a few species of
teleosts, including some cyprinids (Alexander,
1959) and the toadfish (Fine et al., 1995). Here we
investigated the contribution of the swimbladder
to buoyancy in a small fresh-water cyprinid, the
zebrafish (Danio rerio).
The zebrafish is a model species used exten-
sively for investigations of developmental, genetic,
and molecular questions in vertebrate biology
(Grunwald and Eisen, 2002) and has recently
gained popularity in studies of integrative organ
function and behavioral neurobiology (Briggs,
2002; reviewed by Miklo
´si and Andrew, 2006). Fur-
thermore, the zebrafish has been the subject of
recent studies aimed at understanding the control
of buoyancy. We have described the morphology of
the swimbladder together with its innervation, mus-
culature, and vasculature both in adults (Finney
et al., 2006) and during development (Robertson
et al., 2007). The swimbladder is double-chambered
Contract grant sponsor: Canadian Space Agency; Contract grant
number: 046016/001/ST; Contact grant sponsor: Natural Sciences and
Engineering Council of Canada; Contract grant number: 38863-02.
*Correspondence to: Frank M. Smith, Department of Anatomy
and Neurobiology, Dalhousie University, 5850 College Street, Hali-
fax, N.S., Canada B3H 1X5. E-mail: fsmith@tupdean2.med.dal.ca
Published online 26 February 2008 in
Wiley InterScience (www.interscience.wiley.com)
DOI: 10.1002/jmor.10610
JOURNAL OF MORPHOLOGY 269:666– 673 (2008)
Ó2008 WILEY-LISS, INC.
in zebrafish; the anterior chamber is believed to be
specialized for audition, having a connection to the
inner ear via the Weberian ossicles (Bang et al.,
2002), whereas the posterior chamber is believed to
be hydrostatic in function. Anterior and posterior
chambers are joined by the ductus communicans, a
narrow passage isolating the contents of the two
chambers. The zebrafish is physostomous: a pneu-
matic duct connecting the posterior chamber to the
esophagus remains patent in the adult, allowing
swimbladder volume to be altered by passing gas
through this duct.
Field observations indicate that zebrafish
inhabit small, shallow streams and feed primarily
upon terrestrial insects from the surface of slowly
running waterways (McClure et al., 2006). Obser-
vations of groups of adult zebrafish freely swim-
ming in our laboratory aquaria suggest that these
animals range from the middle to upper half of the
available tank depth, and do not show marked ten-
dencies to sink or rise when they briefly stop
swimming. Given the presumptive hydrostatic
function of the swimbladder, it would be expected
that the volume of this organ would be regulated
to maintain neutral buoyancy over the preferred
depth range of the zebrafish, as suggested for
other teleosts (Fange, 1983). Furthermore, it
would be predicted that the swimbladder is posi-
tioned within the coelomic cavity to help the fish
maintain a horizontal attitude, so that locomotory
energy is not expended unnecessarily in counter-
acting tendencies to pitch or roll. In addition, the
pressure inside the swimbladder must also be
maintained above ambient for this organ to fulfill
its acoustic and hydrostatic roles. Thus it has been
hypothesized that the pressure in the swimbladder
is regulated at a fixed value above ambient water
pressure at the preferred depth (Alexander, 1959).
Our preliminary observations of zebrafish behav-
ior support these general suppositions about the
hydrostatic role of the swimbladder, but the spe-
cific contribution of this organ to buoyancy has not
been measured in this species. From an ecological
standpoint, the ability to maintain neutral buoy-
ancy in the water column is critical and facilitates
such daily activities as feeding, reproduction, and
predator avoidance (Gee, 1983). To determine the
potential hydrostatic contribution of the swimblad-
der to these behaviors, a detailed description of
swimbladder morphology as well as measurements
of whole animal density and swimbladder gas vol-
ume are required. Here, we present a noninvasive
method of estimating swimbladder volume from a
single lateral view using morphometric measure-
ments and stereology, as well as techniques for
directly measuring gas volume and whole-animal
density. This study is the first morphometric anal-
ysis of the contribution of the swimbladder to
whole-body density and thus to buoyancy in the
adult zebrafish.
Our results showed that the presence of a swim-
bladder permitted zebrafish to attain nearly neu-
tral buoyancy over their preferred range of depth
in an aquarium. We have also shown that the posi-
tion of the swimbladder in the body was close to
optimal for horizontal stability, thus reducing the
locomotory energy required to counteract changes
in pitch and roll. By direct measurement, we found
that the internal pressure of the zebrafish swim-
bladder, while above ambient, was considerably
less than that reported in most other cyprinids.
MATERIALS AND METHODS
Animals
A total of 120 adult zebrafish D. rerio (Hamilton-Buchanan)
of both sexes were used in this study. Fish were purchased from
a local pet store (Aqua Creations Tropical Fish, Halifax,
Canada) and kept in aerated, dechlorinated tap water in 75 l
aquaria maintained at 28–308C on a 14:10 h light:dark cycle for
at least 3 days before experiments were done. Procedures for
fish care and usage followed the Guide to the Care and Use of
Laboratory Animals as established by the Canadian Council for
Animal Care. Institutional approval was obtained from the Uni-
versity Committee on Laboratory Animals at Dalhousie Univer-
sity. Fish were fed Nutrafin staple fish food (Rolf C. Hagen,
Montreal, Canada) two to three times daily. Behavioral observa-
tions were performed in aquaria maintained under the same
conditions as the holding tanks. Experiments on isolated swim-
bladders were performed at ambient laboratory temperature
(21–228C). Local barometric pressure was recorded before each
experiment during the first part of this study, but variations in
ambient pressure (range 99.2–103.2 kPa) were too small to sig-
nificantly affect swimbladder volume, so this factor was disre-
garded in later experiments.
Mean Observed Depth of Fish in Aquarium
To quantify the mean observed depth of adult zebrafish dur-
ing daylight hours (09:00–19:00), 10 groups of five adult fish
were monitored in a 75 l observation tank with a maximum
depth of 36 cm. Animals of both sexes, ranging in total length
from 20 to 40 mm, were randomly chosen from the holding
aquaria, transferred to the observation tank, and allowed 12 h
to acclimate. Fish depth was recorded with a monochrome CCD
video camera (Honeywell Model HCM574E, Syosset, NY), aimed
into the tank from one side, and connected to a computerized
recording system (Astra 8 Video Surveillance System, Pace Set-
ter Technologies, Dartmouth, NS, Canada). Observations were
made for 1-h periods every second hour during the recording
schedule. The tank was divided into 12 horizontal bins, each
representing 3 cm of depth. Fish position within these bins was
noted in single frames selected at 4-min intervals throughout
each hour of recording. The presence of fish within a particular
bin was noted by recording the middle depth of that bin (i.e.,
fish within the bin extending 0–3 cm in depth were recorded at
1.5-cm depth). For each hour of all trials, the overall mean
depth of all fish in the group was calculated. Given the rela-
tively small depth interval included in each of the bins and con-
sidering that fish were often oriented at an angle spanning the
entire depth of a bin during sampling, we considered these data
to be continuous over the whole depth of the tank for the pur-
pose of estimating mean observed depth. A one-way ANOVA
was performed to test for differences in mean observed depth
among the different hours of recording.
ZEBRAFISH SWIMBLADDER AND BUOYANCY 667
Journal of Morphology
Fish Mass, Length, Volume, and Density
Measurements were made on 48 fish of both sexes, ranging in
total body length from 20 to 40 mm. The sample included some
gravid females. Animals were anaesthetized with 0.02% MS222
(Ethyl 3-aminobenzoate methanesulfonate salt; Sigma Chemical
Co., Mississauga, ON, Canada) and whole-body mass was deter-
mined on an electronic scale (resolution 610 mg) after blotting
the fish with gauze to remove excess water. Fish length was
measured from the protruding tip of the lower mandible to the
end of the tail fin. To measure whole-body volume, a chamber
was constructed by mounting a graduated 2 ml-pipette on the
needle end of a 10-ml plastic syringe; the opening in the
plunger end of the syringe body was then closed with a remov-
able plug (Fig. 1). The empty chamber assembly was placed ver-
tically on the scale pan and tared. A length of polyethylene tub-
ing (PE 50, Clay Adams, NY) was then inserted into the open
end of the pipette and positioned so that the tip of the tubing
entered the chamber body. The chamber was then filled with
water from a syringe attached to the tubing until the meniscus
reached the 1-ml graduation on the pipette, and the tubing was
withdrawn from the pipette. The mass of this volume of water
was recorded. The chamber was then emptied, air-dried, and
the fish was placed inside before replacing the plug and taring
the assembly again. The chamber was then refilled with the
same mass of water used to establish the initial volume. The
difference between initial and final volumes as indicated on the
pipette scale was taken as the volume of the fish. Whole-body
density was then calculated as the dividend of mass and
volume.
Swimbladder Volume
Swimbladder volume was either measured directly or esti-
mated stereologically. For a subset of animals, both techniques
were employed on the same swimbladder.
Measurement of swimbladder volume. Swimbladders
were dissected intact from the coelomic cavity of anesthetized
fish, ensuring that no gas escaped. These organs were then
immersed in normal zebrafish Ringer’s solution (Westerfield,
1995) in a Petri dish under a gas-collecting funnel (Fig. 2). The
funnel was cut from the shoulder region of a glass Pasteur pip-
ette and was connected to a 10-ml syringe by a 20-cm length of
PE 90 polyethylene tubing. The segment of tubing closest to the
collecting funnel was aligned with the edge of a ruled scale ori-
ented vertically. The tubing and collecting funnel were filled
with Ringer’s solution and the open end of the funnel was sub-
merged in the dish. Volume calibrations were performed before
each swimbladder measurement by injecting a standard 50-ll
bubble of air into the funnel from a 100-ll Hamilton syringe.
This bubble was then drawn into the tube and its length
measured. To measure swimbladder volume, both chambers of
the swimbladder were punctured and the total volume of gas
captured by the funnel was then drawn into the tube. Gas
bubble length was measured to calculate the volume of the
swimbladder.
Estimation of swimbladder volume. Volume estimations
were made from photographs of the lateral aspect of swimblad-
ders dissected from anesthetized animals. Each image was par-
titioned into regions representing the five volumes shown in
Figure 3A; the linear dimensions used in the geometrical analy-
sis below are illustrated in Figure 3B. All geometrical formulas
follow the usage of Beyer (1985).
Anterior chamber volume was approximated by that of a ro-
tary prolate ellipsoid:
Volume V1¼4=3pab2ð1Þ
where a50.5 major axis and b50.5 minor axis.
Fig. 1. Schematic diagram of apparatus for measuring zebra-
fish whole-body volume. See text for details.
Fig. 2. Schematic diagram of apparatus for measuring vol-
ume of gas in swimbladder. Swimbladder was ruptured under
collecting funnel and released gas was drawn into tubing by
syringe. Volume was proportional to gas bubble length.
668 G.N. ROBERTSON ET AL.
Journal of Morphology
Posterior chamber volume was approximated by the sum of
the volumes of four regular geometric shapes.
Volume V
2
was a portion of a sphere:
Volume V2¼1=6ph1ð3r2
1þh2
1Þð2Þ
where r
1
50.5 d
1
, measured at the widest point of the rostral
portion of posterior chamber.
Volume V
3
was a cone frustum:
Volume V3¼1=3ph2ðr2
1þr2
2þr1r2Þð3Þ
where r
1
50.5 d
1
,r
2
50.5 d
2
,d
2
taken as 0.9 d
1
.
Volume V
4
was a half-cylinder:
Volume V4¼1=2pr2
2h3ð4Þ
Volume V
5
was a cone:
Volume V5¼1=3pr2
3h4ð5Þ
where d
3
was taken as the base of the cone; r
3
50.5 d
3
.
After swimbladder volume had been obtained, the contribu-
tion of the swimbladder to buoyancy was determined for indi-
vidual animals by calculating fish density using the total vol-
ume of the fish with the swimbladder included (D
1
, below) and
comparing this value with that of the same fish after swimblad-
der volume had been subtracted from whole-body volume (D
2
).
D1¼m=vð6Þ
D2¼m=ðvvsbÞð7Þ
where mis the whole-body mass, vis the whole-body volume,
v
sb
is the swimbladder volume.
Swimbladder Internal Pressure
Given that the swimbladder is a relatively compliant organ, if
its internal pressure were substantially greater than the pres-
sure within the coelomic cavity, as has been shown for a num-
ber of cyprinids [‘‘excess’’ pressure, Alexander (1959)], the vol-
ume of the zebrafish swimbladder might have increased when
the coelomic cavity was opened in this study. This would have
introduced a systematic error into our measurements and esti-
mations of swimbladder volume. To determine the magnitude of
this potential error, we used X-ray images of whole fish to visu-
alize the swimbladder in situ to test whether its volume
increased when the coelomic cavity was opened (Chang and
Magnuson, 1968; Bang et al., 2002). X-rays were taken of the
lateral aspect of the bodies of 10 anesthetized, intact fish in
shallow Ringer’s-filled dishes. The coelom was then opened, X-
ray exposure was repeated with the fish in the same position,
and the estimated volumes were compared using the stereologi-
cal method described earlier. X-ray images were made on a
Belray Dental X-ray machine (Model 096; Takara Belmont Co.,
Mississauga, ON, Canada; 70 kVp, 10 mA at 0.1 s exposure).
The swimbladder from each of these fish was then removed,
photographed, and its volume was again estimated stereologi-
cally. The volumes of eight of these swimbladders were then
measured directly by capturing their gas content.
Internal pressures of both the anterior and posterior chambers
of the swimbladder in a separate group of nine fish were meas-
ured directly with a pressure transducer (Statham Model p23Dc,
Hato Rey, Puerto Rico). Swimbladders were exposed by opening
the coelomic cavity and gas in the two chambers was isolated by
ligating the ductus communicans. A fluid-filled 30-gauge needle,
connected to the pressure transducer via a length of PE 10 poly-
ethylene tubing, was inserted through the wall of each chamber
and internal pressure was read from the screen of an oscillo-
scope. We found that rapid penetration of the wall with the nee-
dle was essential to ensure that the wall did not tear. Transducer
calibration was performed with a 10-cm water column.
Estimating Relative Positions of Centers
of Mass and Buoyancy
A variation of the technique of Bone (1973) was used to deter-
mine the position of the center of mass along the rostro-caudal
axis. A shallow dish was fitted with a pin anchored to the dish
bottom and projecting upward at 908to the plane of the bottom.
An anesthetized fish was laid into the dish on its side with its
ventral aspect touching the pin. Sufficient water was added to
cover the fish and the dish was tilted gently to one side to bal-
ance the fish ventrally on the shaft of the pin. The position of
the fish on the pin was adjusted along the rostro-caudal axis
until the body was evenly balanced. The point at which the ven-
tral surface of the fish touched the pin was then marked as the
surface representation of the rostro-caudal center of mass. The
distance from the tip of the lower jaw to the estimated center of
mass was measured and the fish was then transected in the
transverse plane at this point. The two resulting parts were
then weighed to confirm the anterior–posterior distribution of
mass around this point.
To localize the center of mass relative to the center of buoy-
ancy in the dorso-ventral axis, each anesthetized fish was
placed in a beaker of water to observe whether it came to rest
with the dorsal, ventral, or lateral aspect uppermost.
Statistical Analyses
Values are expressed as means 61 standard deviation. The
level of significance for all comparisons of means was set at P
0.05. Pairs of means were compared using a t-test; multiple-
means comparisons were performed using one-way ANOVA.
Statistical calculations were done using Minitab 14 (Minitab,
PA) or SPSS 14.0 (SPSS, Chicago, IL).
RESULTS
Mean Observed Depth of Fish in Aquarium
During the day, fish were located most fre-
quently in the upper middle region of the 36-cm
deep observation tank, although they occasionally
Fig. 3. Photographs of lateral view of swimbladder (anterior
to left, dorsal toward top) illustrating geometric components
used in estimating volume. (A) Anterior chamber (AC) volume
approximated by rotary prolate ellipsoid (volume V
1
). Posterior
chamber (PC) volume approximated by sum of volumes V
2–5
.
(B) Linear dimensions used in estimating volumes V
1
–V
5
(see
text for explanation). Scale bar represents 1 mm.
ZEBRAFISH SWIMBLADDER AND BUOYANCY 669
Journal of Morphology
swam to the surface or bottom. The mean depth of
fish ranged between 10.1 66.0 and 14.2 65.6 cm
from the surface (Fig. 4) with no significant differ-
ences in the mean observed depth at different
times during the day (one-way ANOVA). The over-
all mean depth over the entire daily observation
period was 12.4 66.3 cm.
Mass, Density, and Volume
Body mass and volume were linearly correlated
(r
2
50.96, n548) over a wide range of body size
(0.16–0.73 g, Fig. 5). Mean whole-body density was
1.01 60.09 g cm
23
, indicating that fish were
nearly neutrally buoyant. This group included
both males and females; 13 of the females were
gravid (represented by dashed line in Fig. 5).
Mean density of the gravid females was not signifi-
cantly different (t-test) from that of the other fish.
Mean density of body tissues in the absence of the
swimbladder was 1.05 60.04 g cm
23
for both
gravid and nongravid fish, a value significantly
greater (paired t-test) than mean whole-body den-
sity of intact fish. The gas volume in the swim-
bladder therefore made a significant contribution
to zebrafish buoyancy, bringing body density to
within 1% of that of the surrounding water.
For the same swimbladders, volumes estimated
from linear dimensions of photographs of lateral
views of both chambers and measured directly by
gas recovery were highly correlated (Fig. 6: r
2
5
0.85, slope 0.91). This finding indicated that esti-
mations of swimbladder volumes obtained stereo-
logically were very close to the line of unity slope
(Fig. 6, dashed line).
On the basis of such estimations, we calculated
that the swimbladder occupied 5.1 61.4% of
whole-body volume (n527). However, this calcu-
lation may have been confounded by the possibility
that the internal swimbladder pressure was suffi-
cient to have caused this organ to expand when
the coelomic cavity was opened. Alexander (1959)
found that swimbladders of many cyprinid species
contain gas at pressures estimated to be from 20
to more than 80 mmHg in excess of ambient pres-
sure. We tested this possibility in the zebrafish
using two methods.
In a group of 10 zebrafish, swimbladder volumes
were estimated stereologically from X-ray images
Fig. 4. Mean observed fish depth from surface in a 75-l tank
(maximum depth 36 cm) for 1 h samples taken every second
hour throughout the daylight period. There was no significant
difference in mean observed depth over the sampling period.
Fig. 5. Plots of co-relationship between whole-body mass and
volume for male and nongravid female fish (closed circles, solid
line) and for gravid females (open circles, dashed line). Both co-
relationships were linear and there was no significant differ-
ence between these lines; r
2
50.96 for pooled data.
Fig. 6. Plot of co-relationship between measured and esti-
mated swimbladder volumes; this relationship was linear (r
2
5
0.85, solid line) with a slope of 0.91. The dashed line has a slope
of unity.
670 G.N. ROBERTSON ET AL.
Journal of Morphology
of the lateral aspect of the swimbladder in intact,
anesthetized fish (Fig. 7A) and again after opening
the coelomic cavity of the same fish (Fig. 7B).
Next, swimbladders were removed from these ani-
mals, photographed, and volumes were estimated
stereologically from these images. Finally, gas vol-
umes of eight of these swimbladders were meas-
ured directly. There were no significant differences
(ANOVA) among estimated and measured volumes,
indicating that the swimbladder did not expand
significantly when the coelom was opened. This
result suggested that internal swimbladder pres-
sure was not greatly in excess of ambient pressure.
Direct measurements of swimbladder pressure
confirmed this: the mean pressure of the posterior
chamber was 8 61 mmHg (n59) and pressure in
the anterior chamber was 7 61 mmHg (n57)
above ambient pressure. Thus, pressures in both
chambers of the zebrafish swimbladder were rela-
tively low and would not have likely introduced
significant errors into our volume estimations.
Swimbladder Position and Relation to
Center of Mass
The zebrafish swimbladder was consistently
located in the same relative position in the body
irrespective of body size. The mean distance from
the protruding tip of the lower jaw to the rostral
tip of the anterior chamber of the swimbladder
was 22 61% of the total body length (n511, r
2
5
0.94). The mean distance from the tip of the lower
jaw to the ductus communicans was (33 62)% of
total body length (n511, r
2
50.94). Total swim-
bladder length from the rostral tip of the anterior
chamber to the caudal end of the posterior cham-
ber was 25 61% of total body length (n510, r
2
5
0.93, range 27–39 mm).
The mean rostro-caudal center of mass was
located at a distance of 33 62% (n510, r
2
5
0.96) of total body length caudal to the tip of the
lower jaw. Transections through the bodies of eight
of these fish at the estimated center of balance
were found to run directly through the ductus
communicans; in two fish the plane of transection
was through the anterior chamber within 1 mm of
the ductus communicans. The relative masses of
the anterior and posterior portions of all fish that
were transected at the balance point matched
within 2%, thus verifying that estimations of the
center of mass derived from balancing the body on
a pivot provided an accurate index of rostro-caudal
mass distribution. These results demonstrate
that the rostro-caudal center of mass was located
one-third of the distance from the tip of the
lower jaw to the tail and very close to the ductus
communicans.
To evaluate the dorso-ventral center of mass, we
observed the attitude of 12 anesthetized fish
placed in a beaker of water. Ten fish sank onto
their sides, one sank belly up, and one floated
belly up. Each of the 10 fish that sank laterally
was also observed to sink onto the opposite side at
least once in succeeding trials, thus suggesting
that the center of mass was at, or near, the center
of buoyancy in this plane. In this context, meas-
urements from our X-ray images showed that the
ductus communicans was positioned approximately
halfway between the dorsal and ventral surfaces of
the fish (47 65% of the distance from the dorsal
to ventral surface, n510). Data from the experi-
ments on the locations of the center of mass in the
rostro-caudal and dorso-ventral axes are summar-
Fig. 7. Three lateral views of swimbladder from the same
fish. (A) X-ray image of swimbladder in intact animal (AC, ante-
rior chamber; PC, posterior chamber). (B) X-ray image after
opening coelom. (C) Photograph of swimbladder after removal
from the body. Scale bar in A represents 2 mm for all images.
Fig. 8. Region of overlap (dashed oval) of centers of mass
and buoyancy in the rostro-caudal and dorso-ventral axes,
superimposed on lateral X-ray image of whole zebrafish. Scale
bar represents 5 mm.
ZEBRAFISH SWIMBLADDER AND BUOYANCY 671
Journal of Morphology
ized schematically on a lateral X-ray image of
a zebrafish in Figure 8 to show that in both of
these axes the center of mass was likely to be
located in a small ellipse in the region of the duc-
tus communicans.
DISCUSSION
Zebrafish maintained a mean depth of 12 cm
with occasional short forays to deeper or shallower
regions. Most fish did not rise or sink appreciably
when they stopped swimming, suggesting they
were close to neutral buoyancy. This observation is
supported by our measurements for whole-body
density of 1.01. Our results are consistent with
reports that zebrafish live in shallow, slowly-
moving streams, where they feed primarily upon
terrestrial insects that fall on the water surface.
Occasionally fish swim deeper to ingest free-swim-
ming aquatic invertebrates (McClure et al., 2006).
It thus appears that zebrafish density may be
optimized to maintain a depth just below the
surface.
A predicted swimbladder volume of 4.8% of total
body volume would be required to reduce zebrafish
density to 1.00 g cm
23
using Alexander’s (1966)
mathematical model. The actual swimbladder vol-
ume for zebrafish in our study was 5.1% of body
volume, a value very close to the predicted value
and within the range of proportional swimbladder
volumes reported for other cyprinids [5.0%–10.7%,
Alexander (1959); 5.0%–7.1%, Overfield and
Kylstra (1971)].
The position of the swimbladder within the body
of a fish is critical to attitude stability. If the cen-
ter of buoyancy were located rostral or caudal to
the center of mass, the head of a hovering fish at
neutral buoyancy would pitch upward or down-
ward, respectively. Similarly, if the center of buoy-
ancy were located ventral to the center of mass,
the fish would tend to roll to one side. Counteract-
ing either of these tendencies would require
energy-expensive fin movements (reviewed by
Alexander, 1966). Zebrafish swim almost continu-
ously and do not tend to hover; yet even when
they are occasionally nearly motionless these ani-
mals neither appear to have difficulty maintaining
a horizontal attitude nor do they exhibit a tend-
ency to roll. Our analysis of the location of the
swimbladder within the body showed that the cen-
ters of mass and buoyancy were nearly coinciden-
tal in a circumscribed region of the body encom-
passing the ductus communicans. Furthermore,
anesthetized zebrafish tended to sink slowly to the
bottom, coming to lie most frequently on their
sides. It therefore appears that the center of buoy-
ancy in this species may be closer to the center of
mass than it is in other teleosts, which usually
come to rest ventral surface upward when anes-
thetized or dead (Alexander, 1966). Regardless of
such a difference between species, there is evi-
dence to suggest that in some teleosts the location
of the center of buoyancy may be actively regu-
lated in relation to the location of the center of
mass. For instance, Goolish (1992), using exter-
nally applied weights and buoys to artificially
change the center of buoyancy relative to the cen-
ter of mass in Fundulus noted that fish that were
artificially maintained at a head-downward pitch
secreted gas to increase swimbladder volume,
whereas fish pitched head-upward decreased
swimbladder volume.
Finally, our study addressed whether the zebra-
fish swimbladder contained gas at a pressure in
the range of values reported in other teleost swim-
bladders. In many cyprinids, the internal swim-
bladder pressure is at least 20–30 mmHg greater
than ambient pressure at the depth to which fish
are acclimated [so-called ‘‘excess’’ pressure,
Alexander (1959)]. Such a high pressure in the
zebrafish swimbladder might have caused an
increase in the volume of the swimbladder when
the coelomic cavity was opened, thus potentially
confounding estimations and measurements of vol-
ume. Our estimates of swimbladder volumes from
X-ray images were not, however, significantly dif-
ferent in fish with closed and open coeloms. This
finding suggested that swimbladder pressure in
this species was relatively low, and we confirmed
this with direct pressure measurements. Pressures
in the anterior and posterior chambers of the
zebrafish swimbladder (7–8 mmHg) were consider-
ably lower than those reported for almost all other
cyprinids tested (Alexander, 1959).
This difference between zebrafish and other cyp-
rinids may be due to a difference in mechanisms
for filling the swimbladder. Cyprinids, including
the zebrafish, can fill the swimbladder by gulping
air at the surface and passing small air bubbles
along the pneumatic duct: these species may in
fact employ this mechanism exclusively when near
the surface. However, most cyprinids also possess
a gas gland for filling the swimbladder with gas
from the bloodstream, presumably allowing these
fish to generate internal swimbladder pressures 3-
to 5-fold greater than those we observed in the
zebrafish, albeit at considerable metabolic cost.
As shown by McClure et al. (2006) and confirmed
in this study, zebrafish appear to prefer relatively
shallow depths, an optimal situation for filling the
swimbladder with surface air via the pneumatic
duct. Since this species does not possess a gas
gland (Finney et al., 2006), it might thus
be expected that maximum internal swimbladder
pressure would be limited to the range we
observed.
We found that gravid and nongravid fish had
identical densities after subtraction of the swim-
bladder. Furthermore, we showed that the co-rela-
tionship between fish volume and density was the
672 G.N. ROBERTSON ET AL.
Journal of Morphology
same in both of these groups. Therefore, zebrafish
eggs must have had densities close to 1.05 g cm
23
despite containing a considerable amount of lipid.
Moreover, although we did not determine the spe-
cific gravity of egg masses, we observed that both
immature egg masses and eggs that were laid by
sexually mature zebrafish sank rapidly, confirming
that egg density was greater than 1.00. In this
respect, then, zebrafish eggs are similar to those of
the toadfish (Fine et al., 1995). Nonetheless, gonad
development (Ona, 1990), the increased total mass
of gravid fish and the pressure of eggs on the
swimbladder in the coelom all likely affect buoy-
ancy control and warrant further investigation in
this context.
This study in the zebrafish establishes for the
first time the specific contribution of the swim-
bladder to the overall buoyancy and attitude sta-
bility in a small cyprinid. Moreover, this work
demonstrates a stereological method of estimat-
ing swimbladder volume from linear dimensions
taken from a single lateral view of this organ. We
showed that values obtained stereologically
matched very closely those obtained by direct vol-
ume measurement for the same swimbladders.
Our results thus indicate that the volume of the
swimbladder can be established accurately and
noninvasively in intact, opaque wild-type fish
using X-ray images. This technique should also
be applicable to photographic images of nonpig-
mented stains of zebrafish (Lister et al., 1999) in
which the swimbladder can be observed directly
through the transparent body wall or to those
obtained by magnetic resonance microscopy, as
proposed by Kabli et al. (2006) for noninvasive
studies of internal zebrafish anatomy. Swimblad-
der stereology, as described here, will facilitate
further studies on the effects of ambient pres-
sure, whole body fat content, and gut distension
(Ona, 1990) on buoyancy of intact fish. We have
also shown that accurate estimations of the vol-
ume of isolated swimbladders can be made ster-
eologically without rupturing or otherwise dam-
aging the organ, thus opening the way to further
experiments on whole swimbladders in situ and
in vitro to study physiological and pharmacologi-
cal mechanisms of autonomic control of effectors
within the swimbladder system.
ACKNOWLEDGMENTS
The authors thank Kathy McInnis (RTR) and
Audra Hayden for taking and processing the X-ray
images in the Faculty of Dentistry, Dalhousie Uni-
versity. They also thank Mr. Paul Brinkhurst for
his advice on buoyancy. Preliminary results of
parts of this study have been previously presented
(Lindsey et al., 2007).
LITERATURE CITED
Alexander RM. 1959. The physical properties of the swimbladder
in intact cypriniformes. J Exp Biol 36:315–332.
Alexander RM. 1966. Physical aspects of swimbladder function.
Biol Rev Camb Philos Soc 41:141–176.
Alexander RM. 1972. The energetics of vertical migration by
fishes. Symp Soc Exp Biol 26:273–294.
Alexander RM. 1989. Optimization and gaits in the locomotion of
vertebrates. Physiol Rev 69:1199–1227.
Alexander RM. 1993. Buoyancy. In: Evans DH, editor. The Physi-
ology of Fishes. Boca Raton: CRC. pp 75–97.
Bang PI, Yelick PC, Malicki JJ, Sewell WF. 2002. High-through-
put behavioral screening method for detecting auditory
response defects in zebrafish. J Neurosci Methods 118:177–187.
Beyer WH, editor. 1985. CRC Handbook of Mathematical Scien-
ces, 5th ed. Boca Raton, FL: CRC. 982 p.
Bone Q. 1973. A note on the buoyancy of some lantern-fishes
(Myctophoidei). J Mar Biol Assoc UK 53:619–633.
Briggs JP. 2002. The zebrafish: A new model organism for inte-
grative physiology. Am J Physiol Regul Integr Comp Physiol
282:R3–R9.
Chang RKC, Magnuson JJ. 1968. A radiographic method for
determining gas bladder volume of fish. Copeia 1:187–189.
Fange R. 1983. Gas exchange in fish swim bladder. Rev Physiol
Biochem Pharmacol 97:111–158.
Fine ML, McKnight JW, Blem CR. 1995. Effect of size and sex on
buoyancy in the oyster toadfish. Mar Biol 123:401–409.
Finney JL, Robertson GN, McGee CA, Smith FM, Croll RP. 2006.
Structure and autonomic innervation of the swim bladder in
the zebrafish (Danio rerio). J Comp Neurol 495:587–606.
Gee JH. 1983. Ecological implications of buoyancy control in fish.
In: Webb PW, Weihs D, editors. Fish Biomechanics. New York:
Praeger. pp 140–176.
Goolish EM. 1992. Swimbladder function and buoyancy regulation
in the killifish Fundulus heteroclitus. J Exp Biol 166:61–81.
Grunwald DJ, Eisen JS. 2002. Headwaters of the zebrafish—Emer-
gence of a new model vertebrate. Nat Rev Genet 3:717–724.
Kabli S, Alia A, Spaink HP, Verbeek PJ, De Groot HJM. 2006.
Magnetic resonance microscopy of the adult zebrafish. Zebra-
fish 3:431–439.
Lefrancois C, Odion M, Claireaux G. 2001. An experimental and
theoretical analysis of the effect of added weight on the ener-
getics and hydrostatic function of the swimbladder of European
sea bass (Dicentrarchus labrax). Mar Biol 139:13–17.
Lindsey BW, Robertson GN, McGee CAS, Dumbarton TC, Croll
RP, Smith FM. 2007. Development of buoyancy control in the
zebrafish (Danio rerio) swimbladder. (abstract) Society for Inte-
grative and Comparative Biology Annual Meeting, 3–7 January,
Phoenix Arizona.
Lister JA, Robertson CP, Lepage T, Johnson SL, Raible DW.
1999. Nacre encodes a zebrafish microphthalmia-related protein
that regulates neural-crest-derived pigment cell fate. Develop-
ment 126:3757–3767.
McClure MM, McIntyre PB, McCune AR. 2006. Notes on the nat-
ural diet and habitat of eight danionin fishes, including the
zebrafish Danio rerio. J Fish Biol 69:553–570.
Miklo
´si A, Andrew RJ. 2006. The zebrafish as a model for behav-
ioral studies. Zebrafish 3:227–234.
Ona E. 1990. Physiological factors causing natural variations in
acoustic target strength of fish. J Mar Biol Assoc UK 70:107–
127.
Overfield EM, Kylstra JA. 1971. The volume and rate of volume
change of the swimbladder of the goldfish. Respir Physiol
13:283–291.
Robertson GN, McGee CAS, Dumbarton TC, Croll RP, Smith FM.
2007. Development of the swimbladder and its innervation in
the zebrafish, Danio rerio. J Morphol 268:967–985.
Westerfield M. 1995. The Zebrafish Book: A Guide for the Labora-
tory Use of Zebrafish (Brachydanio rerio). Eugene, OR: Univer-
sity of Oregon Press.
ZEBRAFISH SWIMBLADDER AND BUOYANCY 673
Journal of Morphology
