Blackwell Science, LtdOxford, UK

FISFisheries Science0919-92682004 Blackwell Science Asia Pty Ltd

701February 2004
764
Selection of food sources by bivalves
A Kasai
et al.
10.1046/j.0919-9268.2003.00764.x
Original Article1120BEES SGML

FISHERIES SCIENCE 2004; 70: 11–20

Selection of food sources by Ruditapes philippinarum

and Mactra veneriformis (Bivalva: Mollusca) determined
from stable isotope analysis
Akihide KASAI,* Hisako HORIE AND Wataru SAKAMOTOa

Graduate school of Agriculture, Kyoto University, Sakyo, Kyoto 606-8502, Japan

ABSTRACT: Carbon and nitrogen isotopic ratios in the body of the bivalves Ruditapes philippinarum
and Mactra veneriformis were measured to estimate the role of the bivalves in the ecosystem in an
estuary. The isotopic signatures were compared with those obtained from particulate organic matter
(POM) in the middle reach of the river, in the estuary, and off the estuary. Both the bivalves showed
constant d13C and d15N values of approximately -17‰ and 10‰ that were significantly enriched
compared with those for POM in the habitat (-25‰ and 4‰). The relative contributions of terrestrial
organic matter to the POM in the habitat and to the bivalve diet were estimated: the former was
approximately 90% while the latter approximately 10%. This indicates that the bivalves select marine
POM from the organic matter available in their habitat, and play a minor role in the direct removal of
terrestrial particulate matter. Temporal changes in d13C and d15N for both bivalves were small, but a
distinct decrease in d13C was observed following a period of heavy rain. During rainfall the concen-
tration of terrestrial material increased in the habitat, and the bivalves would then have to change
their food sources temporarily.

KEY WORDS: bivalve, ecosystem, food source, Mactra veneriformis, particulate organic
matter, Ruditapes philippinarum, stable isotope.

INTRODUCTION diet of bivalves, because organic matter in tidal

Tidal flats have been recognized as playing an
important role in the ecosystem in shallow coastal
areas. They serve as nursery grounds for migrating
fish larvae and as habitats for benthic animals. In
addition, retention and removal of organic and
inorganic matter in water by benthic filter feeders
may be a particularly important mechanism in
nutrient and carbon cycling, and may have impor-
tant consequences for the flux of particulate mate-
rial in shallow coastal areas.1 Among various
benthic organisms, bivalves have attracted atten-
tion because of their large biomass. They are
believed to purify the water, removing organic
material that flows into the sea through rivers from
the land and those formed in the sea, so that they
are thought to function as sewage disposal facili-
ties.2 However, it is still uncertain of the specific
*Corresponding author: Tel: 81-75-753-6314.
Fax: 81-75-753-6468. Email: kasai@kais.kyoto-u.ac.jp
Received 30 April 2003. Accepted 15 August 2003.
a
Present address: Kinki University, Fisheries Institute, Shira-
hama, Wakayama 649-2211, Japan.
flats contain various matter, which have different
origins.
The primary objective of the present study was
to assess the role of two infaunal suspension-feed-
ing bivalves Ruditapes philippinarum and Mactra
veneriformis, both of which are ubiquitous and
abundant in Japanese coastal areas, in removal
of terrestrial organic matter. The life history and
ecological aspects of R. philippinarum were well
reported. It is distributed from intertidal zones to
a water depth of approximately 10 m. It lives in the
top 10 cm of the sediments and extends a siphon
to take in water including organic materials.
In contrast, reports on M. veneriformis are rare
because it is not important in industry. Because
M. veneriformis inhabits similar places to R.
philippinarum and their size is also similar, it is
supposed that M. veneriformis has similar biologi-
cal characteristics to R. philippinarum.
There are several techniques for understanding
the diet of organisms, including direct observation
of feeding behavior, gut content analysis, and
examination of chemical constituents such as
stable isotopes or nutrients within the tissues of
12 FISHERIES SCIENCE
organisms compared with their potential food
sources. Direct observation of the feeding behavior
of bivalves has primarily involved either measure-
ment of filtering rates of artificial seston3 or exam-
ination of filtering morphology,4 but it is difficult
to determine the diet of bivalves in their natural
environment. Gut contents of bivalves have
been observed for some species. For example, it
was reported that the digestive diverticula of
R. philippinarum contained phaeo-pigments.5
However, this method cannot distinguish ingested
material that is not assimilated, and thus provides
only approximate estimates of the percentages of
various materials ingested by the bivalves. Algae,
for example, can survive passage through the
digestive tract, and fecal material can have a high
nitrogen content due to ingested material bypass-
ing the digestive diverticula and moving directly to
the intestinal tract.6
In contrast, natural abundances of stable car-
bon and nitrogen isotope ratios (d13C and d15N)
have increasingly been used to identify the poten-
tial food sources and its trophic level in supporting
food webs. Stable carbon analysis provides the
most convincing results especially for coastal
marine ecosystems on tidal flats and salt marshes,
where terrestrial and marine vegetation sources
coexist.7 The d13C values of primary producers are
significantly different between land plants and
marine algae8 and thus the isotopic signatures will
be different in the particulate matter that is pro-
duced.9,10 Therefore, it is feasible to determine
which source of plant food is selected by consum-
ers, or the relative share of each primary producer
biomass in the total amount of detrital organic
matter ingested. From d13C and d15N values,
Deegan and Garritt estimated the importance
of several organic sources in estuarine food webs,
and noted that consumers relied most heavily on
locally produced organic matter.11 Riera et al.
showed that organic matter inputs carried by river
inflow can contribute significantly to the feeding of
juvenile brown shrimp in the Rincon Bayou marsh,
although the main food sources were marine detri-
tus and benthic diatoms.12 The oyster Crassostrea
gigas exhibited significantly different d13C varia-
tions between sites along an estuarine gradient,
indicating that the terrestrial organic matter might
be one of the food sources in the upper estuary,
although C. gigas exhibited poor growth perfor-
mances there compared with that in the mouth of
the estuary.13
The d13C analysis of tissues of R. philippinarum
was investigated previously by Koike and
Nakajima14 and Koike et al.15 These studies sug-
gested that the diet of R. philippinarum was based
on the benthic diatom assemblages. By contrast,
A Kasai et al.
Numaguchi indicated from observation of phy-
topigments and organic substances in the sedi-
ment that terrestrial organic matter was important
in the diet of R. philippinarum.16 Terrestrial
organic matter was neglected as food sources for
R. philippinarum in Koike and Nakajima14 and
Koike et al.15 but it could be one of the main com-
ponents of the diet of R. philippinarum because
they inhabit river mouths, where a large amount of
terrestrial organic matter flows in. The relative con-
tribution of terrestrial matter is, however, still
unidentified in the total amount of assimilated
organic matter for R. philippinarum in addition to
M. veneriformis.
Dual isotope measurements of d13C and d15N
may be more helpful in resolving marine versus
terrestrial sources of organic matter than a single
isotope analysis. Within regions of freshwater
influence, the differences in d15N between primary
producers are small, while terrestrial organic
matter has significantly lighter isotope values than
marine organic matter.7 The purpose of the present
study was therefore to identify and compare the
food sources supporting R. philippinarum and
M. veneriformis, using nitrogen and carbon stable
isotope analysis. We payed particular attention to
the contribution of terrestrial organic matter in the
bivalve diets. To differentiate between terrestrial
and marine organic sources, the carbon and nitro-
gen isotopes of particulate organic matter (POM)
in a river, in and off a tidal flat were also measured.
In addition, the environmental conditions, which
cause a temporal change in the isotope values for
bivalve tissues, are discussed.
MATERIALS AND METHODS
Sampling was carried out in the Miya Estuary
(Fig. 1). The Miya River and the Isuzu River empty
into Ise Bay, which is a gulf-type bay connected to
the Pacific Ocean at its southern end. There are
tidal flats in and around the river mouths. The
study site is at the mouth of the Isuzu River, and
dominated by muddy sediments. Few macroalgae
were observed during the sampling period.
Live bivalves were collected by hand at station A
from April to September 2001. Sampling dates are
listed in Table 1. Bivalve samples were stored at -
40∞C until analysis. Shell length, shell height, shell
width and wet weight were measured and then the
foot muscle was excised for each sample. The mus-
cle tissues were dried in an oven at 60∞C, ground to
a fine powder with a mortar and pestle, and lipids
were removed with a chloroform : methanol (2:1)
solution. Fifty-four individuals of R. philippinarum
and 42 individuals of M. veneriformis were
Selection of food sources by bivalves FISHERIES SCIENCE 13

130° 140 E Ise Bay

20km

40°
Ise Bay

Stn. C

Miyagawa
30° N

Miya Estuary

Miya Estuary

Stn. A

Miya R.

Isuzu R. 5km
Stn. B
Fig. 1 Location of study area.

Table 1 Carbon and nitrogen stable isotope ratios (‰) for the bivalves Ruditapes philippinarum and Mactra veneri-
formis
R. philippinarum M. veneriformis
13 15
Date n d C d N n d13C d15N
25 April 2001 13 -16.8 ± 0.30 10.0 ± 0.47 2 -16.6 ± 0.26 10.2 ± 0.62
11 May 2001 9 -16.7 ± 0.26 10.1 ± 0.28 10 -16.4 ± 0.28 10.7 ± 0.39
23 May 2001 8 -17.1 ± 0.51 9.9 ± 0.32 7 -16.7 ± 0.12 10.2 ± 0.54
4 June 2001 5 -16.9 ± 0.31 10.6 ± 0.21 5 -16.6 ± 0.22 10.2 ± 0.35
21 June 2001 6 -17.6 ± 0.23* 10.0 ± 0.20 5 -17.4 ± 0.41* 10.3 ± 0.56
22 July 2001 8 -17.0 ± 0.48 10.1 ± 0.15 8 -16.7 ± 0.23 10.6 ± 0.20
18 September 2001 5 -16.8 ± 0.26 10.6 ± 0.19 5 -16.4 ± 0.21 11.2 ± 0.19

n, no. samples analyzed.

Values are mean ± 1SD; *statistically significant difference (Bonferroni/Dunn test P < 0.05).

analyzed in total. Their shell lengths were 17.30–
42.15 mm and the majority of them were estimated
at 1 year or 2 years old.
For sampling the estuarine particulate organic
matter (EPOM) and terrestrial particulate organic
matter (TPOM), 1 L of water was collected at
station A and at the middle reach of the Isuzu River
(stn B; Fig. 1), respectively. In addition, as repre-
sentatives of marine-produced organic matter
(MPOM), 1 L of surface water was collected at the
center of Ise Bay (stn C; Fig. 1). In general, most of
the terrestrial organic matter is deposited within
10 km of the river mouth.10,17 Therefore, MPOM
collected from station C is expected to include little
terrestrial matter. The POM was defined as the par-
ticles collected on a precombusted Whatman GF/
F glass microfibre filter (Whatman, Kent, UK). The
filter samples were put in a desiccator with HCL
14 FISHERIES SCIENCE A Kasai et al.

fumes for the first 24 h to eliminate any CaCO3, R. philippinarum

with NaOH fumes for the next 24 h to neutralize the 12.0
25 Apr
(a)
acid, and were then freeze dried. 11.5 11 May
Stable isotope ratios of carbon and nitrogen 23 May
were measured by a continuous-flow isotope-ratio 11.0
4 Jun
mass spectrometer with an elemental analyzer

d N
21 Jun

15
10.5
(Carlo Erba, Lakewood, USA) connected to a mass 22 Jul
spectrometer (Finnigan MAT, Bremen, Germany). 10.0 18 Sep
The concentration of organic carbon and nitrogen 9.5
were measured simultaneously. Isotope ratios, d13C
and d15N, are expressed by the standard d unit 9.0
-18.5 -18.0 -17.5 -17.0 -16.5 -16.0 -15.5
notation as follows:
d13C
dX = [(Rsample / Rstandard ) - 1] ¥ 103 , (1)
where X is 13C and 15N, and R is 13C/12C for carbon M. veneriformis
12.0
and 15N/14N for nitrogen. Pee dee belemnite (PDB) (b)
and atmospheric nitrogen were used as the isotope 11.5
standards for carbon and nitrogen, respectively.
11.0
Salinity was measured at station A simulta-
d N
neously with the bivalve sampling, using a temper-
15

10.5
ature–salinity meter (ACT20-D, Alec Electronics,
10.0
Kobe, Japan). As for precipitation, data obtained
from the Miyagawa Automatic Meterological Data 9.5
Acquisition System (AMeDAS) station were used
9.0
(Fig. 1). -18.5 -18.0 -17.5 -17.0 -16.5 -16.0 -15.5

d13C

RESULTS Fig. 2 d13C–d15N bivalve map.

Muscle tissues without lipids of both species had

rather constant d13C and d15N values: -17.0 ±
0.44‰ (SD) and 10.1 ± 0.37‰ for R. philippi-
narum, respectively; and -16.7 ± 0.38‰ and (a)
-15
10.5 ± 0.51‰ for M. veneriformis, respectively -17 R. philippina
(Fig. 2). The results are summarized in Table 1. -19
M. venerifo
MPOM
There were no significant differences in d13C or d15N
-21 EPOM
d13C

values among the species (t-test, P < 0.05). This TPOM

-23
indicates that there was no difference in food pref-
erence between the species. Figure 3 displays time -25

sequences of isotope ratios. Temporal changes in -27

d13C and d15N for the bivalves were also small. How- -29
1-Apr 1-May 1-Jun 1-Jul 1-Aug 1-Sep 1-Oct
ever, Figs 2,3(a) show that d13C values on 21 June
2001 were explicitly depleted for both species. 12
(b)
A statistically significant difference in d13C was
10
detected only on 21 June (Bonferroni/Dunn test
8
P < 0.05). Temporal changes in d13C and d15N for
6
d15N

POM were larger than those for the bivalves and, in

particular, MPOM displayed the wider range of 4

d13C values (Fig. 3). 2

Isotope ratios of POM were significantly differ- 0
ent among the sampling sites: the average d13C -2
1-Apr 1-May 1-Jun 1-Jul 1-Aug 1-Sep 1-Oct
values for TPOM, EPOM and MPOM were
-25.5 ± 1.38‰, -25.0 ± 0.93‰, and -18.9 ± 2.08‰, Fig. 3 Time change in stable carbon (d13C) and nitrogen
respectively, while the average d15N values were (d15N) isotope ratios for Ruditapes philippinarum,
0.6 ± 1.73‰, 4.3 ± 1.00‰, and 9.0 ± 0.89‰, respec- Mactra veneriformis, marine-produced organic matter
tively (Fig. 4; Table 2). The stable isotope composi- (MPOM), estuarine particulate organic matter (EPOM)
tion of the bivalves and POM are superimposed on and terrestrial particulate organic matter (TPOM).
 Selection of food sources by bivalves FISHERIES SCIENCE 15

the d13C–d15N map (Fig. 4), which provides useful DISCUSSION

information about the transport pathways of
organic matter. The d13C and d15N for EPOM were
between TPOM and MPOM, indicating a mixture
of material originating from terrigenous and
marine sources. However, a shift in the isotope
ratios for EPOM to those for TPOM from MPOM
indicates a higher contribution of TPOM to EPOM.
Isotope ratios for MPOM were, in contrast, clearly
enriched and close to those for the bivalves.
The C/N ratios for POM provide an indication
of the quality of organic matter. The C/N ratio for
MPOM was 6.0 ± 0.36 (Table 2). Because this value
is comparable to that for phytoplankton estimated
from the Redfield ratio (= 6.6), the main compo-
nent of MPOM is regarded as phytoplankton or its
detritus. The C/N ratios for TPOM (9.7 ± 2.38) and
EPOM (9.7 ± 2.58) were higher than that for
MPOM. The similarity of C/N ratios for TPOM and
EPOM shows the high contribution of TPOM to
EPOM, consistent with the indication from the
isotope values.
12
10
8 TPOM
6 TPOM was derived from terrigenous plants in the
d15N
Contribution of terrestrial organic matter to
bivalve diet
Ise Bay is one of the most eutrophic bays in Japan.
Both d13C and d15N for MPOM observed in the
present study (-18.9 ± 2.08‰ and 9.0 ± 0.89‰) are
comparable to the previous results observed in
other eutrophic embayments in Japan such as
Tokyo Bay,10,18 Mikawa Bay19 and Hiroshima Bay,20
and the Delaware Estuary in the USA.21 They are
slightly heavier than those in oligotrophic coastal
areas around Japan.9 Isotope discrimination
during the photosynthetic fixation of CO2 and dis-
solved inorganic nitrogen decreases when the
phytoplankton grows rapidly.22 In these euphotic
areas, growth rates of phytoplankton tend to
increase, shifting d13C and d15N to higher values. In
addition to the growth rates, stable isotope values
of MPOM could be different between species of
phytoplankton and be changed with speed of
decomposition of detritus.23 It is therefore predict-
able that MPOM displayed large variation in the
stable isotope values as shown in the previous
R. philippinarum
M. veneriformis
studies.7,10
MPOM The average d13C and d15N values for TPOM were
EPOM -25.53‰ and 0.62‰, respectively. They are in the
Benthic diatoms range of the previous results,9,10 indicating that

upper catchment basin of the river. Forest plants

4
are characterized by low d13C and d15N. Photosyn-
2 thesis is only partially rate-limited by CO2 diffusion
0 and nitrogen is supplied by atmospheric precipita-
tion and biological nitrogen fixation.9,18
-2 Because fresh river water and oceanic water mix
-28 -26 -24 -22 -20 -18 -16 -14 -12 -10
in an estuary, terrestrial and marine matter should
d13C
coexist in the tidal flat area. In fact, EPOM had
Fig. 4 d13C–d15N map of bivalve and particulate organic stable isotope ratios between those for TPOM and
matter (POM). Error bars indicate standard deviations MPOM in our observation (Fig. 4). As the other
except for benthic diatom. Error bars for benthic diatom component of EPOM, it has recently been recog-
indicate the maximum and minimum of isotope values nized that benthic microalgae contribute to the
presented by previous studies. ecosystem in many coastal areas.24,25 No direct

Table 2 Carbon and nitrogen stable isotope ratios (‰), concentrations and C/N ratios for the bivalves and POM
Material n d13C (‰) d15N (‰) C (Wt%) N (Wt%) C/N ratio
Ruditapes philippinarum 54 -17.0 ± 0.44 10.1 ± 0.37 41.8 ± 3.19 10.3 ± 1.56 4.1 ± 0.47
Mactra veneriformis 42 -16.7 ± 0.38 10.5 ± 0.51 41.3 ± 2.05 11.4 ± 0.84 3.6 ± 0.32
MPOM 7 -18.9 ± 2.08 9.0 ± 0.89 25.3 ± 9.40 4.2 ± 1.54 6.0 ± 0.36
EPOM 7 -25.0 ± 0.93 4.3 ± 1.00 32.4 ± 25.13 3.1 ± 1.66 9.7 ± 2.58
TPOM 7 -25.5 ± 1.38 0.6 ± 1.73 10.7 ± 8.16 1.3 ± 1.03 9.7 ± 2.38

n, no. samples analyzed; MPOM, marine-produced organic matter; EPOM, estuarine particulate organic matter; TPOM, terrestrial
particulate organic matter.
Values are mean ± 1SD.
16 FISHERIES SCIENCE A Kasai et al.

determination of isotopic values for benthic 12

microalgae was made in the present study, but pre-
vious studies showed the d13C values for benthic 10 MPOM
diatom to lie between -17.6‰ and -12.1‰,
and the d15N values between 3.8‰ and 6.6‰ 8
25TPOM : 75MPOM
75MPOM : 25BA

(Fig. 4).13,26–30 In contrast, macroalgae can be 50TPOM : 50MPOM 50MPOM : 50BA

neglected in the present study, although they are PD
25MPOM : 75BA
6

d15N (‰)
also often cited as one of the principal organic
75TPOM : 25MPOM
BM
components in many estuaries. Their effect on 25TPOM : 75BA
EPOM is considered small here because few mac- 4
50TPOM : 50BA
EPOM
roalgae were found in and around the sampling 75TPOM : 25BA

site during the observation. Zooplankton and 2

higher trophic level animals are present at much
lower abundances than primary producers,31 and
thus their carcasses are also unlikely to be signifi-
cant to the elements of the EPOM in the tidal flat.
From the aforementioned discussion, compo-
nents of EPOM can be divided into three distinct
sources: terrestrial matter, phytoplankton and
benthic microalgae. Because the bivalves appar-
ently capture EPOM, it is meaningful to estimate
the relative importance of each organic matter on
EPOM. In the present study, the concentration-
weighted mixing model should be applied rather
than the linear mixing model to estimate the rela-
tive contribution of three sources that differ in C
and N concentrations.32 Assuming that for each
element (C and N), the contribution of a source to
EPOM is proportional to the included mass multi-
plied by the elemental concentration in that
source, mass balance equations can be written as,
13 13 13 13
Ê (d C T - d C E )[C T ] (d C M - d C E )[C M ]
Á (d15N T - d15N E )[N T ] (d15N M - d15N E )[N M ]
Á from d13C, and is out of the triangle. This local
Ë 1 1
(2)
(d13C A - d13C E )[C A ] ˆ Ê fT ,E ˆ Ê 0ˆ
(d15N A - d15N E )[N A ]˜ Á f M ,E ˜ = Á 0˜
˜Á ˜ Á ˜
1 ¯ Ë f A ,E ¯ Ë 1¯
where d13CT, d13CE, d13CM, d13CA, d15NT, d15NE, d15NM
and d15NA represent the C and N isotopic signatures
for TPOM, EPOM, MPOM and benthic microalgae;
[CT], [CM], [CA], [NT], [NM] and [NA] are the C and N
concentrations in sources TPOM, MPOM and
benthic microalgae; and fT,E, fM,E and fA,E are the
mass fractions of TPOM, MPOM and benthic
microalgae in EPOM, respectively. Isotopic values
and C and N concentrations for TPOM, EPOM,
MPOM used in the model are listed in Table 2. For
benthic microalgae the median of d13C and d15N in
the previous studies are used.13,26–30 It is assumed
that C and N concentrations for the benthic
microalgae are the same as those for MPOM.
Substituting these values into equation 2 gives the
fractions fT,E = 0.89, fM,E = - 0.12 and fA,E = 0.23.
0
TPOM
-2
-27 -25 -23 -21 -19 -17 -15 -13
13
d C (‰)
Fig. 5 Mixing triangle for the concentration-weighted
model. Variations in percentage contribution of terres-
trial particulate organic matter (TPOM), marine-
produced organic matter (MPOM) and benthic
microalgae (BM) are shown along the edges of the
mixing triangles, and serve as labels for iso-diet lines
that intersect the edge. PD, presumed diet for the
bivalves.
Figure 5 shows the d13C and d15N plot of EPOM and
its potential sources, as well as the mixing triangle.
The TPOM primarily occupied EPOM.
It might be ineffective that fM,E takes a negative
value. Figure 5 indicates that d15N for EPOM dis-
plays a slightly enriched value than that expected
enrichment of d15N has been frequently reported in
the previous studies at the river mouth. It is
explained that the denitrification increases d15N for
POM.9,18 Therefore, it might be appropriate to use
only carbon isotope signatures for estimating the
fractions in EPOM. Applying the two-source linear
mixing model with d13C, the mass balance for
carbon isotopes is:
13
CE = fT,E 13CT + fX,E 13CX, (3)
fT,E + fX,E = 1, (4)
where fX,E is fM,E or fA,E, and d13CX is d13CM or d13CA.
The isotopic value for MPOM with these equations
give fT,E = 0.92, and fM,E = 0.08. For benthic microal-
gae, fT,E = 0.95, and fM,E = 0.05. The linear mixing
model also shows that the main source of EPOM is
the terrestrial organic matter, as indicated by the
concentration-weighted mixing model. These
results agree well with the previous studies. It is
reported that organic matter in estuarine sedi-
ments have closer isotopic values to terrestrial
Selection of food sources by bivalves FISHERIES SCIENCE
matter than marine matter, because the large por-
tion of terrestrial matter is deposited at estuary
mouths.9,10,17
In Fig. 5, d13C and d15N values for the presumed
diet of the bivalves are plotted as an open circle,
based on the previous studies that showed that
d13C and d15N of animal tissue are 1‰ and 3‰
heavier, respectively, than those of the diet.33,34 The
presumed diet is considerably enriched in d13C and
d15N compared to EPOM, but is close to MPOM and
benthic microalgae. It is therefore expected that
the bivalves do not assimilate EPOM discretely, but
select the 13C and 15N enriched food. In the same
way as EPOM, the relative contribution of each
component to the bivalves food source can be esti-
mated using the concentration-weighted mixing
model. Assuming that the d13C and d15N for the
bivalves increases 1‰ and 3‰ from the diet,
respectively,33,34
13 13 13 13
Ê (d C T - d C B ¢ )[C T ] (d C M - d C B ¢ )[C M ]
Á (d15N T - d15N B ¢ )[N T ] (d15N M - d15N B ¢ )[N M ]
Á appear to capture particles indiscriminately, but
Ë 1 1 separate suitable and unsuitable matter on the
(d13C A - d13C B ¢ )[C A ] ˆ Ê fT ,B ˆ Ê 0ˆ
(d15N A - d15N B ¢ )[N A ]˜ Á f M ,B ˜ = Á 0˜
˜Á ˜ Á ˜
1 ¯ Ë f A ,B ¯ Ë 1¯
d13C B¢ = d13C B - 1
d15N B¢ = d13N B - 3
where d13CB and d15NB represent the C and N isoto-
pic signatures for the bivalves, and fT,B, fM,B and fA,B
are the fractions of assimilated biomass of TPOM,
MPOM and benthic microalgae by the bivalves,
respectively. Substitute -16.8 and 10.3 (the average
value of d13C and d15N for R. philippinarum and
M. veneriformis) for d13CB and d15NB, respectively,
and the fractions are then calculated as fT,B = 0.12,
fM,B = 0.55 and fA,B = 0.33. Phytoplankton and
benthic microalgae are the main component of the
bivalve diet.
It is noteworthy that fT,E and fT,B, therefore the
other fractions also, are significantly different from
each other. These indicate that the terrestrial
organic matter occupied approximately 10% of the
diet of bivalves, although approximately 90% of
EPOM originated from terrestrial organic material.
The bivalves are considered to select and assimi-
late marine organic materials among EPOM. This
habit is different from some invertebrates, which
ingest in situ organic material.26,35 The Oyster
C. gigas exhibited significantly different d13C varia-
tions between sites along an estuarine gradient,
reflecting the utilization of terrestrial organic mat-
17
ter.13 However, selective feeding has been reported
for some mollusks, which feed on and within
sediments.36 Ruditapes philippinarum and M.
veneriformis inhabit saltier regions than oysters.
This might lead to a preference for marine food
sources of R. philippinarum and M. veneriformis.
A considerably lower contribution of the terrestrial
matter to the food sources of R. philippinarum and
M. veneriformis does not necessarily signify that
they play the minor role in terrestrial water purifi-
cation. Nutrient loads from land through rivers
would contribute to the growth of phytoplankton
and benthic microalgae, which are regarded as the
main food sources of the bivalves. However, it is
confirmed that the bivalves remove little TPOM
directly.
Two different concepts can be proposed to
explain their food selection. The first assumption
is the active selection that implies preferential
ingestion and/or a better assimilation of selected
components from POM. Some species of bivalves
including R. philippinarum and M. veneriformis
(5)
gills and palps.37 The latter is ejected as pseudofe-
ces. The low d13C and d15N for TPOM indicate that
it was derived from forest plants. They contain lig-
nin and cellulose, which have a refractory nature
for biochemical degradation.37 Therefore, a large
(6) portion of TPOM might be selected as unsuitable
matter and ejected. The isotope ratio analysis for
(7) pseudofeces of the bivalves might provide impor-
tant information in future work. In contrast, differ-
ences in biochemical composition of the ingested
food can also affect the true assimilation. Rudi-
tapes philippinarum and M. veneriformis may
have a chemical discrimination mechanism,
which was reported on C. gigas.13 This might
explain the preferential utilization of MPOM,
which contains a larger nitrogen content. Further
laboratory experiments on R. philippinarum and
M. veneriformis are needed to test this hypothesis.
The second assumption is passive selection. Both
R. philippinarum and M. veneriformis obtain par-
ticles through filtration by holding their inhalant
siphon above the sediment surface. However,
there is no guarantee that they always inhale
water. In fact, they live in the sediments and do not
hold their inhalant siphon above the sediment
when the tidal flats are exposed. They could feed
only during the flood tide and/or high water, so
that salty water including marine organic matter
could be inhaled. The careful observation of short
time change in the habit, combined with the anal-
ysis of a stable isotope ratio for EPOM, during a
tidal cycle is necessary to verify the second
hypothesis.
18 FISHERIES SCIENCE A Kasai et al.

Temporal d13C variation in bivalves (Fig. 6b), implying that a large amount of fresh
water flowed into the study area. On that day the
A small but distinct decrease was detected in d13C concentration of organic carbon and nitrogen in
values for both the bivalves on 21 June 2001 (Fig. 3). the estuarine water was 13.2 mg/L and 1.0 mg/L,
This isotopic variation occurred within a month which were approximately 10-fold changes
(between 4 and 21 June) and it recovered to the (Fig. 6c). In addition, the C/N ratio of EPOM was
previous value within the next month (between 21 13.8, which was almost equal to that of TPOM but
June and 22 July). This indicates that the environ- was considerably larger than MPOM (Fig. 6d;
mental condition of bivalves, especially their diet, Table 2). Figure 3(a) shows that the d13C of EPOM
changed with time, because changes in the stable had a significantly low value (-26.3‰), corre-
isotope ratios of an animal generally result from sponding to that of TPOM (-26.2‰). These facts
dietary shifts. Figure 6 shows changes in various mean that on 21 June 2001 an extremely large
conditions in the study area. Heavy rain of amount of terrestrial organic matter flowed into
>60 mm/day was observed on 19 and 20 June 2001 and occupied the estuary, and there was no room
at Miyagawa, which is in the catchment basin of left for obtaining any marine organic matter.
the Isuzu River (Fig. 1). Following the large precip- One interpretation for the decrease of d13C of the
itation, the lowest salinity was recorded on 21 June bivalves is a reduction of the preferential utiliza-
tion of specific components within POM. In such
condition, the bivalves might have closed their
shell valves and stopped feeding. However, it is
Precipitation (mm/day)

80
(a) reported that starvation results in an enrichment
60 of d13C in animal tissue.38 The starvation is there-
40 fore not appropriate to the present case because
d13C values for the bivalves did not increase but
20
decreased on 21 June (Fig. 3a). It is a more convinc-
0 ing explanation that the bivalves chose their food
16 11 16 21 26
among the terrestrial materials, and decreased the
Day in June
d13C value by the sudden change in the diet. After
30 (b) the event, the C/N ratio and d13C for EPOM
returned to smaller values, indicating that marine
Salinity

20 organic matter had returned to the estuary. The

10
bivalves then resumed selection of marine organic
matter and the d13C value for the bivalves recovered
0 to a normal level on 22 July 2001. In a previous
report on the temporal variation in d13C of C. gigas,
ay

ay

ul
un
n

ep
pr

Ju

-J
-A

-M

-M

-J

-S
22
4-
26

21

a remarkable decrease of d13C occurred after the

18
11

23

high river discharge.29 The decrease was explained

C-concentration (mg/l)

N-concentration (mg/l)

15
(c) 1.0 by a similar mechanism to that of the present case,
C
10 N although their time and horizontal scales are larger
0.5 than ours.
5

0 0.0
CONCLUSIONS
ay

ay

ul
un
n

p
pr

Ju

e
-J
-A

-M

-M

-J

-S
22
4-
26

21

18
11

23

15
From the results of the present study it is con-
13
(d) cluded that R. philippinarum and M. veneriformis
prefer marine phytoplankton and benthic diatoms
C/N ratio

11
as food sources than terrestrial materials, despite
9
the latter being predominant in the estuary. It is
7
generally believed that benthic organisms in tidal
5 flats have the capacity to purify water but
ay

ay

ul
un
n

ep
pr

our results indicate that R. philippinarum and

Ju

-J
-A

-M

-M

-J

-S
22
4-
26

21

18
11

23

M. veneriformis play a minor role in the direct

Fig. 6 (a) Precipitation at Miyagawa; (b) salinity at removal of terrestrial particulate matter. They
station A; (c) carbon and nitrogen concentration in could assimilate the terrestrial materials as an
estuarine particulate organic matter (EPOM); and (d) exception when the river discharge suddenly
C/N ratio in EPOM. increases and terrestrial materials occupy their
Selection of food sources by bivalves FISHERIES SCIENCE
habitat. In tidal flats there are other kinds of
benthic invertebrates including bivalves, the food
sources for which have been uncertain. Due to
their high abundance, benthic invertebrates play
important roles in carbon and nitrogen cycling in
the estuarine systems. More research is needed to
more precisely identify the contributions of their
different sources. The study will contribute not
only to the management of useful fisheries
resources, but also to the elucidation of the ecosys-
tem in estuaries and coastal areas.
ACKNOWLEDGMENTS
The authors wish to thank Drs J Scorse, H Kennedy,
and C Richardson of University of Wales, Bangor
for helpful discussion and review of the manu-
script. Funding for this research was provided by
the Mikimoto Fund for Marine Ecology, and in part
by a Grant-in-Aid for Encouragement of Young Sci-
entists from the Ministry of Education, Culture,
Sports, Science and Technology.
REFERENCES
1. Prins TC, Smaal AC, Dame RF. A review of the feedbacks
between bivalve grazing and ecosystem processes. Aquat.
Ecol. 1998; 31: 349–359.
2. Sasaki K. Water purification in tidal flats and shallow seas.
Kagaku 2001; 71: 9002–9911 (in Japanese).
3. Kryger J, Riisgard HU. Filtration capacities in 6 species of
European freshwater bivalves. Oecologia 1988; 77: 34–38.
4. Ward JE, MacDonalds BA, Thompson RJ. Mechanisms of
suspension feeding in bivalves. Resolution of current con-
troversies by means of endoscopy. Limnol. Oceanogr. 1993;
38: 265–272.
5. Numaguchi K. Concentrations of chlorophyll a and phaeo-
pigments of bottom sea water, sediment and surface mud
in estuary where short necked clam Ruditapes philippi-
narum (Adams et Reeve) inhabited. Bull. Natl Res. Inst.
Aquacult. 1990; 18: 39–50 (in Japanese).
6. Hawkins AJS, Bayne BL, Clarke KR. Co-ordinated rhythms
of digestion, absorption and excretion in Mytilus edulis
(Bivalva: Mollusca). Mar. Biol. 1983; 74: 41–48.
7. Fry B, Sherr EB. d13C measurements as indicators of carbon
flow in marine and freshwater ecosystem. Contrib. Mar. Sci.
1984; 27: 13–47.
8. O’Leary MH. Carbon isotopes in photoshynthesis. Fraction-
ation techniques may reveal new aspects of carbon dynam-
ics in plants. Bioscience 1988; 38: 328–336.
9. Wada E, Minagawa M, Mizutani H, Tsuji T, Imaizumi R,
Karasawa K. Biogeochemical studies on the transport of
organic matter along the Otsuchi River watershed, Japan.
Estuar. Coast. Shelf Sci. 1987; 25: 321–336.
10. Ogawa N, Ogura N. Dynamics of particulate organic matter
in the Tamagawa estuary and inner Tokyo Bay. Estuar.
Coast. Shelf Sci. 1997; 44: 263–273.
19
11. Deegan LA, Garritt RH. Evidence for spatial variability in
estuarine food webs. Mar. Ecol. Prog. Series 1997; 147: 31–47.
12. Riera P, Montagna PA, Kalkae RD, Richard P. Utilization of
estuarine organic matter during growth and migration by
juvenile brown shrimp Penaeus aztecus in a south Texas
estuary. Mar. Ecol. Prog. Series 2000; 199: 205–216.
13. Riera P, Richard P. Isotopic determination of food sources
of Crassostrea gigas along a trophic gradient in the estuarine
bay of Marennes-Oleron. Estuar. Coast. Shelf Sci. 1996; 42:
347–360.
14. Koike H, Nakajima T. Distribution and seasonal change of
stable carbon isotope for fauna and flora at a tidal flat of the
Obitsu river mouth. J. Coll. Liberal Arts Saitama Univ. 1989;
25: 21–34 (in Japanese).
15. Koike H, Saito T, Kosugi M, Kakino J. Shell growth and diet
of Tapes Philippinarum from the tidal flats of the Obitsu
River, Tokyo Bay. Fish. Eng. 1992; 29: 105–112 (in Japanese).
16. Numaguchi K. Aspect of sediment as food environment of
Manila clam Ruditapes philippinarum. Fish. Eng. 2001; 37:
209–215 (in Japanese).
17. Mishima Y, Hoshika A, Tanimoto T. Deposition rates of ter-
restrial and marine organic carbon in the Osaka Bay, Seto
Inland Sea, Japan, determined using carbon and nitrogen
stable isotope ratios in the sediment. J. Oceanogr. 1999; 55:
1–11.
18. Wada E, Kabaya Y, Tsuru K, Ishiwatari R. 13C and 15N abun-
dance of sedimentary organic matter in estuarine areas of
Tokyo Bay, Japan. Mass Spectrosc. 1990; 38: 307–318.
19. Takahashi K, Wada E, Sakamoto M. Carbon isotope ratio
and photosynthetic activity of phytoplankton in the
eutrophic Mikawa Bay, Japan. Ecol. Res. 1992; 7: 355–361.
20. Takai N, Hoshika A, Imamura K, Yorozu A, Tanimoto T,
Mishima Y. Distribution of carbon and nitrogen stable iso-
tope ratios in macroalgae in Hiroshima Bay, Japan. J. Ecol.
2001; 51: 177–191.
21. Cifuntes LA, Sharp JH, Fogel ML. Stable carbon and nitro-
gen isotope biogeochemistry in the Delaware estuary.
Limnol. Oceanogr. 1988; 33: 1102–1115.
22. Laws EA, Popp BN, Bidigare RR, Kennicutt MC, Macko SA.
Dependence of phytoplankton carbon isotopic composi-
tion on growth rate and [CO2]aq: theoretical considerations
and experimental results. Geochim. Cosmochim. Acta 1995;
59: 1131–1138.
23. Fry B, Wainright SC. Diatom sources of 13C-rich carbon in
marine food webs. Mar. Ecol. Prog. Ser. 1991; 76: 149–157.
24. Cahoon LB, Cooke JE. Benthic microalgal production in
Onslow Bay, North-Carolina, USA. Mar. Ecol. Prog. Ser.
1992; 84: 185–196.
25. Jahnke RA, Nelson JR, Marinelli RL, Eckman JE. Benthic flux
of biogenic elements on the southeastern US continental
shelf: influence of pre water advective transport and
benthic microalgae. Cont. Shelf Res. 2000; 20: 109–127.
26. Incze LS, Mayer LM, Sherr EB, Macko SA. Carbon inputs to
bivalve mollusks: a comparison of two estuaries. Can. J.
Fish. Aquat. Sci. 1982; 39: 1348–1352.
27. Couch CA. Carbon and nitrogen stable isotopes of meiob-
enthos and their food resources. Estuar. Coast. Shelf Sci.
1989; 28: 433–441.
28. Riera P, Richard P, Gremare A, Blanchard G. Food source of
intertidal nematodes in the Bay of Marennes-Oleron
(France), as determined by dual stable isotope analysis.
Mar. Ecol. Prog. Ser. 1996; 142: 303–309.
20 FISHERIES SCIENCE
29. Riera P, Richard P. Temporal variation of d13C in particulate
organic matter and oyster Crassostrea gigas. In: Marennes-
Oleron Bay (France): effect of freshwater inflow. Mar. Ecol.
Prog. Ser. 1997; 147: 105–115.
30. Riera P, Richard P. Determination of food sources for
benthic invertebrates in a salt marsh (Aiguillon Bay, France)
by carbon and nitrogen stable isotopes: importance of
locally produced sources. Mar. Ecol. Prog. Ser. 1999; 187:
301–307.
31. Lalli CM, Parsons TR. Biological Oceanography. Elsevier
Science, Oxford. 1994.
32. Philips DL, Koch PL. Incorporating concentration depen-
dence in stable isotope mixing models. Oecologia 2002; 130:
114–125.
33. DeNiro MJ, Epstein S. Influence of diet on the distribution
of carbon isotopes in animals. Geochim. Cosmochim. Acta
1978; 42: 495–506.
A Kasai et al.
34. Minagawa M, Wada E. Stepwise enrichment of 15N along
food chains: further evidence and the relation between d15N
and animal age. Geochim. Cosmochim. Acta 1984; 48: 1135–
1140.
35. Hughes EH, Sherr EB. Subtidal food webs in a Georgia
estuary: d13C analysis. J. Exp. Mar. Biol. Ecol. 1983; 67: 227–
242.
36. Hylleberg J, Gallucci V. Selectivity in feeding by the deposit-
feeding bivalve Macoma nasuta. Mar. Biol. 1975; 32: 167–
178.
37. Parsons TR, Takahashi M, Hargrave B. Biological
Oceanographic Processes, 3rd edn. Pergamon Press, Oxford.
1984.
38. Webb SC, Hedges REM, Simpson SJ. Diet quality influences
the d13C and d15N of locusts and their biochemical compo-
nents. J. Exp. Biol. 1998; 201: 2903–2911.