Ocean acidification impairs olfactory discrimination
and homing ability of a marine fish
Philip L. Munday
a,b,1
, Danielle L. Dixson
a,b
, Jennifer M. Donelson
a,b
, Geoffrey P. Jones
a,b
, Morgan S. Pratchett
a
,
Galina V. Devitsina
c
, and Kjell B. Døving
d
a
Australian Research Council Centre of Excellence for Coral Reef Studies,
b
School of Marine and Tropical Biology, James Cook University, Townsville, QLD
4811, Australia;
c
Ichthyology Department, Faculty of Biology, Moscow MV Lomonosov State University, Moscow 119992, Russia; and
d
Physiology Program,
Institute of Molecular Bioscience, University of Oslo, N-0316 Oslo, Norway
Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved December 29, 2008 (received for review October 6, 2008)
The persistence of most coastal marine species depends on larvae
finding suitable adult habitat at the end of an offshore dispersive
stage that can last weeks or months. We tested the effects that
ocean acidification from elevated levels of atmospheric carbon
dioxide (CO
2
) could have on the ability of larvae to detect olfactory
cues from adult habitats. Larval clownfish reared in control sea-
water (pH 8.15) discriminated between a range of cues that could
help them locate reef habitat and suitable settlement sites. This
discriminatory ability was disrupted when larvae were reared in
conditions simulating CO
2
-induced ocean acidification. Larvae be-
came strongly attracted to olfactory stimuli they normally avoided
when reared at levels of ocean pH that could occur ca. 2100 (pH 7.8)
and they no longer responded to any olfactory cues when reared
at pH levels (pH 7.6) that might be attained later next century on
a business-as-usual carbon-dioxide emissions trajectory. If acidifi-
cation continues unabated, the impairment of sensory ability will
reduce population sustainability of many marine species, with
potentially profound consequences for marine diversity.
climate change
larval sensory mechanisms
population connectivity

population replenishment
O
cean acidification caused by the uptake of additional car-
bon dioxide (CO
2
) at the ocean surface is now recognized
as a serious threat to marine ecosystems (1–4). At least 30% of
the anthropogenic CO
2
released into the atmosphere in the past
200 years has been absorbed by the oceans, causing ocean pH to
decline at a rate
100 times faster than at any time in the past
650,000 years (1, 4). Global ocean pH is estimated to have
dropped by 0.1 units since preindustrial times and is projected to
fall another 0.3–0.4 units by 2100 because of existing and future
CO
2
emissions (1, 5–6). Considerable research effort has focused
on predicting the impact that reduced carbonate-ion saturation
states that accompany ocean acidification will have on calcifying
marine organisms, particularly corals and other invertebrates
that precipitate aragonite skeletons (2–3, 6). However, the
effects that ocean acidification will have on other marine or-
ganisms, including fishes, remain almost completely unknown,
especially for conditions of atmospheric carbon dioxide and
seawater pH that could occur in the near future (4, 7–9).
The persistence of most coastal marine species depends on the
ability of larvae to locate suitable settlement habitat at the end
of a pelagic stage that can last weeks or months. Accumulating
evidence for reef fishes suggests that both reef sounds (10) and
olfactory cues (11–13) are used by larvae to locate reefs. The
olfactory organs of many reef fishes are well-developed by the
end of the larval phase (14–15), and it has recently been shown
that larvae of some species can discriminate the smell of water
from their natal reef compared with water from other reefs (13),
which provides a mechanism to explain high levels of self-
recruitment in some reef fish populations (16–19). It is well
known that coral reef fish larvae can use olfactory cues to
identify suitable settlement sites once they are in the vicinity of
reef habitat. Settling larvae have been shown to respond to
olfactory signals from preferred microhabitats (12, 20), resident
conspecifics (21–23), or symbiotic partners such as anemones
(24–25). Any disruption to the ability of larvae to detect and
discriminate between olfactory cues that guide them to reefs, or that
enable them to select preferred settlement habitat, would have
far-reaching implications for the sustainability of adult populations.
We tested if elevated CO
2
and reduced seawater pH consistent
with ocean acidification predictions could affect the ability of
orange clownfish (Amphiprion percula; Pomacentridae. Fig. 1)
larvae to respond to olfactory cues that are used to locate reef
habitat and distinguish preferred settlement sites. Specifically,
we tested the ability of settlement-stage larvae to respond to
olfactory cues that are preferred during the settlement process
compared with olfactory cues that are likely to be avoided when
searching for reefs and settlement sites. Orange clownfish mostly
live on oceanic reefs surrounding vegetated islands and recent
research has shown that the larvae can discriminate between
seawater from reefs surrounding vegetated islands versus sea-
water from reefs without islands (26). Furthermore, the larvae
are positively attracted to water-borne cues from tropical rain-
forest trees (26) that should provide a reliable cue to the
presence of vegetated oceanic islands. We tested the response of
larval clownfish to olfactory cues from a range of tropical
vegetation types when reared in seawater simulating 2 future
CO
2
-induced acidification scenarios (seawater pH 7.8 and 7.6)
compared with current-day controls (pH 8.15). For larvae reared
in each treatment we tested preference or avoidance of olfactory
cues from the leaves of 3 vegetation types: (i) a tropical rainforest
tree (Xanthostemon chrysanthus) that is a positive cue for settling
clownfish (26), (ii) a swamp tree (Melaleuca nervosa) that
contains pungent oils in the leaves and is avoided by settling
clownfish (26), and (iii) a tropical savannah grass (Megathyrsus
maximus) that is not expected to provide a reliable cue for the
presence of trees on islands.
It is well known that anemonefishes are positively attracted to
olfactory cues of host anemones (24–25). Therefore, we also
tested the ability of larval clownfishes to respond to anemone
olfactory cues when reared at the 3 pH levels. Finally, the
presence of adult populations should be a good signal of
favorable habitat and previous studies have found that larvae of
some reef fishes are attracted to olfactory cues from conspecific
adults (21–22). However, in species such as the orange clownfish
where many larvae recruit to natal reefs (19), it would also be
advantageous for juveniles to be able to discriminate between
their parents and other adults to avoid inbreeding that could result
Author contributions: P.L.M., D.L.D., J.M.D., G.P.J., and M.S.P. designed research; P.L.M.,
D.L.D., J.M.D., and G.V.D. performed research; K.B.D. contributed new reagents/analytic
tools; P.L.M. analyzed data; and P.L.M., D.L.D., J.M.D., G.P.J., M.S.P., G.V.D., and K.B.D.
wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1
To whom correspondence should be addressed. E-mail: philip.munday@jcu.edu.au.
© 2009 by The National Academy of Sciences of the USA
1848–1852
PNAS
February 10, 2009
vol. 106
no. 6 www.pnas.orgcgidoi10.1073pnas.0809996106

from settling in their natal anemone. Therefore, we also tested the
response of clownfish larvae to olfactory cues from their parents
and other adult clownfish when reared at the 3 pH treatments.
Clownfish were reared at James Cook University’s experi-
mental aquarium facility where the pH of unmanipulated sea-
water was 8.15  0.07. This is similar to the pH that pelagic
larvae would experience during development in the open ocean
(1). To simulate ocean acidification the pH of treatment sea-
water was adjusted to either 7.8  0.05 or 7.6  0.05 by the
standard method of dissolving additional CO
2
. The equivalent
atmospheric CO
2
levels for the pH treatments in our experiment
was estimated to be
1,000 ppm for pH 7.8 and
1,700 ppm for
pH 7.6. These values are consistent with climate change models
that predict atmospheric CO
2
levels could exceed 1,000 ppm by
2100 and approach 2,000 ppm by the end of next century under
a business as usual scenario (5, 27). Clownfish were reared at 1
of the 3 pH levels from the day that eggs were laid until the larvae
were competent to settle at 11 days after hatching. The olfactory
responses of larvae were then tested in a 2-channel choice flume
(13) where individuals were allowed to choose between a stream
of seawater containing the olfactory cue to be tested and a
stream of water without that cue.
Results
Larvae reared in control seawater spent equal amounts of time
on each side of the chamber in a control test where neither
stream of seawater in the flume contained an additional
olfactory cue (Fig. 2). Larvae exhibited a strong preference for
Xanthostemon (Fig. 2, P 0.001) spending93% of their time
in the stream of water in which leaves of this rainforest tree had
been soaked. In contrast, all larvae completely avoiding the
stream of water in which Melaleuca leaves had been soaked
(Fig. 2, P  0.001). Avoidance of Melaleuca might be expected
because the leaves of these trees contain pungent oils. Larvae
showed no preference or avoidance for olfactory cues from
grass, spending approximately equal amounts of time in the
stream of water in which grass leaves had been soaked and in
the stream of water without olfactory cues from grass (Fig. 2,
P  0.1). As expected, larvae also exhibited a strong prefer-
ence for anemones (Fig. 2, P  0.001), spending nearly all of
their time in the stream of water in which an anemone had been
placed for 2 h.
Larvae reared in seawater at pH 7.8 exhibited significant
differences in olfactory responses to larvae reared in control
seawater for all comparisons. Although the preferences for
Xanthostemon and anemones remained, there was a significant
reduction in the strength of the response compared with controls
(Fig. 2; P  0.001 and P  0.01 respectively). Larvae exhibited
a preference for grass that did not occur in larvae reared in
control water (Fig. 2, P  0.001), and, more dramatically, larvae
now exhibited a strong preference for olfactory cues from
Melaleuca (Fig. 2, P  0.001). Larvae reared in seawater at pH
7.8 spent 80% of their time in the stream of water in which
Melaleuca leaves had been soaked, even though larvae reared in
control water completely avoided this cues. Larvae still spent an
equal amount of time on each side of the chamber in a control
test where neither stream of seawater in the flume contained an
additional olfactory cue (Fig. 2).
Larvae reared in control seawater almost completely
avoided the water stream containing olfactory cues from their
own parents (Fig. 3, P  0.001), but exhibited a strong
preferences for a water stream containing cues from other
adults (Fig. 3, P  0.001). The ability to discriminate between
parents and other adults was lost in larvae reared at pH 7.8.
These larvae exhibited an equally strong preference for olfac-
tory cues from their own parents (P  0.001) and other adult
orange clownfish (Fig. 3, P  0.001). These responses were
confirmed when larvae were presented simultaneously with
cues from their own parents and other adults. Larvae reared
in control water continued to prefer the water stream with
10cm
1cm
A
B
C
Fig. 1. Clownfish larvae use olfactory cues to locate adult habitat at the end
of their pelagic stage. (A) Adult orange clownfish A. percula form breeding
pairs on host anemones. (B) A settlement stage (11 days posthatching) larva of
A. percula beside an Australian 5 cent coin for scale. (C) The Atema flume
chamber used to test the responses of larvae to olfactory cues from leaves,
anemones, and conspecifics in control and acidified water. Water from 2
different sources flowed through the chamber in the direction of the arrows.
A larva was placed in the test section below the central partition and its
position recorded at 5-s intervals for 2 min. The water sources were swapped
and the procedure repeated.
0
10
20
30
40
50
60
70
80
90
100
Seawater Xanthostemon Melaleuca Grass Anemone

P
e
r
c
e
n
t

o
f

t
i
m
e

i
n

t
r
e
a
t
m
e
n
t

w
a
t
e
r
control
7.8
20
20
20
26
10
40
46
46
46
16
Fig. 2. Response of larval clownfish to olfactory cues from tropical plants and
anemones when reared at current-day seawater pH (control, open bars) and
in seawater, where the pH had been reduced using CO
2
to simulate the effect
of ocean acidification (pH 7.8, filled bars). The first pair of columns shows the
mean percentage of time that larvae spent of one side of a 2-channel flume
chamber when neither stream of water in the chamber contained a test cue.
Subsequent columns show the mean percentage of time that larvae spent in
the stream of water containing the cue when one stream contained the cue
and the other stream did not. Numbers above bars are the number of repli-
cates for each test.
Munday et al. PNAS
February 10, 2009
vol. 106
no. 6
1849
E
C
O
L
O
G
Y