Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31

Contents lists available at ScienceDirect

Journal of Experimental Marine Biology and Ecology

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j e m b e

Feeding preferences and food searching strategies mediated by air- and water-borne
cues in the mud whelk Terebralia palustris (Potamididae: Gastropoda)
Sara Fratini ⁎, Marco Vannini, Stefano Cannicci
Dipartimento di Biologia evoluzionistica "L. Pardi”, via Romana 17, 50125 Firenze, Italy

A R T I C L E I N F O A B S T R A C T

Article history: The gastropod T. palustris is one of the major species responsible for leaf consumption and degration within
Received 12 October 2007 the Indo-Pacific mangrove forests, and it strongly competes with herbivorous sesarmid crabs in consuming
Accepted 19 May 2008 fallen leaves. This snail feeds at high and low tides and it is able to locate food items by means of chemical
cues. The aim of this study was to assess the food preferences of T. palustris and to define its feeding
Keywords:
strategies at low and high tides, by conducting field trials on water-borne mediated food location at high tide,
Food choice
Mangrove leaves
grazing rate and the chemical attraction exerted by different mangrove leaves. The results showed that T.
Molluscs palustris was able to perceive underwater grazed leaves. In addition, we demonstrated that T. palustris
Olfaction consumes all the mangrove species (preferentially the Rhizophoraceae leaves) but Xilocarpus granatum.
Palatability Moreover, this snail is differentially attracted to different mangrove species: the major attractive power is
wielded by the rhizophoracean species and Pemphis acidula, while X. granatum does not attract this snail at
all. The efficacy and adaptive value of a chemically mediated food searching strategies is unquestionable
since by using this ability T. palustris can locate and reach the leaves it preferentially consumes. Moreover, T.
palustris is the only macrobenthic species of East Africa mangroves able to search, detect and consume
mangrove leaves at both high and low tides. Such an expanded feeding window permits T. palustris to occupy
temporal niches left empty by the sesarmid crabs.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction mangrove gastropods in nutrient dynamics has been largely over-

The mud whelk Terebralia palustris (Linnaeus) is widespread
throughout the Indo-Pacific area and is one of the most abundant
inhabitants of mangroves, reaching remarkable densities in some
locations, e.g. 150 adults m− 2 in New Caledonia (Plaziat, 1984) and
approximately 20 adults m− 2 in Kenya (Fratini et al., 2004). It colonizes
the entire mangrove forest, from the lower subtidal fringe to the upper
littoral fringe, adapting itself to very different intertidal belts and
mangrove microhabitats. Fratini and co-workers (2004) studied the
population structure and distribution of this snail in a Kenyan
mangrove forest and demonstrated that, in contrast to previous
reports (Wells, 1980; Crowe, 1997; Crowe and McMahon, 1997), the
snail's zonation depends mainly on the organic content of the soil
with the adults (reaching a shell size of 10–12 cm) segregating the
juveniles (with a shell size of less than 5–6 cm) into less favourable
areas (i.e. those with more acidic soil and less frequently flooded).
This snail has amphibious features and feeds both during low and
high tides (Fratini et al., 2004), with the adults grazing on fallen
mangrove leaves and the juveniles eating mud and leaf debris
(Nishihira, 1983; Houbrick, 1991; Dahdouh-Guebas et al., 1998; Fratini
et al., 2004). Although classically the role of T. palustris and of other
⁎ Corresponding author. Tel.: +39 0552288204; fax: +39 055222565.
E-mail address: sarafratini@unifi.it (S. Fratini).
looked, recent studies have demostrated their central ecological role.
Adult T. palustris contribute massively to leaf degradation and
consumption: Fratini et al. (2004) showed that, when fed ad libitum,
this whelk is able to consume a huge amount of fallen leaves, much
more than is naturally available in the wild. Moreover, thanks to
analyses of carbon and nitrogen isotope ratios in various mangrove
macroinvertebrates, it has been shown that molluscs massively
contribute to entrapping primary production within the mangrove
ecosystem, since they ingest a much higher amount of mangrove
leaves than sesarmid crabs (Bouillon et al., 2002a,b, 2003).
As widely recognised for many gastropods (Croll, 1983), T. palustris
uses olfaction to locate suitable food resources. In particular, in its food
search it relies on chemical cues released by broken (scraped)
mangrove leaves or propagules, as experimentally demonstrated by
Fratini et al. (2001). This means that this snail is able to locate a
mangrove leaf feeder thanks to the cues released by the leaves when
grazed, while conversely it is not attracted to snails that are not
feeding nor to food items that have not been scraped (Fratini et al.,
2001). As a result of this feeding recruitment process, T. palustris
actively competes with the leaf feeding sesarmid crabs. When the
crabs encounter a leaf on which a substantial number of snails have
settled, they are not able to grasp and capture the leaf for themselves
(Fratini et al., 2000). Moreover, while at low tide this snail bases its
food-finding ability on air-borne cues (Fratini et al., 2001), it can

0022-0981/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jembe.2008.05.008
 S. Fratini et al. / Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31
probably use water-borne cues as well since it has been shown that
foraging activity also occurs during high tide (Fratini et al., 2004).
Finally, while no data are available on the existence of a preference
towards specific mangrove plants, T. palustris is known to discriminate
between different food items, being more attracted to fresh mangrove
leaves than to senescent or fallen leaves or propagules (Fratini et al.,
2001).
Within this conceptual framework, the main aim of this study was
to use field trials to test the possible feeding preferences of T. palustris
and the attraction exerted by different leaves via air-borne cues. In
addition, we also investigated whether at high tide (i.e. under water)
the food searching strategy of T. palustris is chemically mediated as it
is at low tide.
2. Materials and methods
2.1. Study site
The research was conducted at Dabaso shore, within Mida Creek
(03° 21′ S; 39° 59′ E), North Kenya Coast, in Febraury 2005 and July
2007. This mangrove forest is characterised by a very narrow land-
ward Avicennia marina belt followed by an extended Rhizophora
mucronata - Ceriops tagal forest. In particular, the experiments were
performed in the muddy small creeks within the R. mucronata
dominated forest. In these creeks the density of adult T. palustris is
about 5 individuals/m2 (Fratini et al., 2001, 2004).
2.2. Field experiments
The goal of assessing the feeding preferences and strategies of T.
palustris was achieved by means of four different sets of experiments,
designed to address three specific issues.
2.2.1. Is the perception of the food items chemically mediated
underwater?(experiment 1)
In order to verify whether at high tide, the snail's attraction
towards food items was chemically mediated as it is at low tide, we
tested the underwater attraction power of minced R. mucronata green
leaves. The minced leaves were placed into net bags thus avoiding the
scattering of leaf debris by the tidal currents, but allowing the odours
to disperse. The bags were fixed with an iron wire to the muddy soil, in
order to prevent them from being transported away by the tide.
During each trial, we used five bags with leaves and five empty net
bags as controls, randomly distributed approximately 5 m apart along
a 50 m transect. The transect was parallel to the direction of the
current. Four trials were performed at diurnal high tides close to the
spring tide, on different days, in Febraury 2005. The trials began as the
tide was entering the creek and the snails were covered by at least
10 cm of water. We counted the numbers of snails present within a
30 cm radius around the bags as soon as they were placed in position
and one hour later. The difference between the final and initial snail
number (ΔN) represents the number of snails attracted towards the
bags.
2.2.2. Do the leaves of various mangrove species promote different
grazing rates?(experiment 2 and 3)
In order to assess wheter the snails consume different amount of
the different mangrove plants we performed two experiments:
experiment 2 established the grazing rate in “free conditions”, i.e.
the snails (at least the first one per leaf) encountered the leaf by
chance moving within the experimental area, while experiment 3
measured the leaf grazing rate in “obliged” conditions, i.e a fixed
number of snails was offered the different mangrove species. The
grazing rate as calculated by experiment 2 is affected by the feeding
recruitment process, and the number of snails on a single leaf was not
established a priori.
27
Expt. 2: a total of 480 fresh green leaves collected among the
branches of the eight mangrove species present in the study area
were placed on the ground at the beginning of a diurnal low tide
and left in place for 3 hours. The species used were: R. mucronata,
C. tagal and Bruguiera gymnorrhiza (Rhizophoraceae), A. marina
(Avicenniaceae), Sonneratia alba (Sonneratiaceae), Pemphis acidula
(Lythraceae), Lumnitzera racemosa (Combretaceae) and Xylocarpus
granatum (Meliaceae).
The trials were performed six times close to a spring tide on
different not consecutive days, three during the dry season (February
2005) and three during the wet season (July 2007). The leaves had
been previously marked with a red numbered tag tied to the petiole.
They were photographed with a digital camera, immediately before
and after the trial in order to estimate the percentage of leaf consumed
by the snails (Fratini et al., 2004). We also counted the number of
snails grazing on a single leaf at the end of each trial. The smaller size
of the A. marina, P. acidula, L. racemosa and X. granatum leaves
compare with the rhizophoracean and S. alba leaves could affect the
probability of a snail finding the leaf. In order to compensate for this
size difference we joined together 2–4 leaves using adhesive tape and
in the calculation of grazing rate we treated these combined leaves as
a single leaf.
Expt. 3: fifteen fresh leaves of the 8 mangrove species (for a total of
120 leaves) were put within experimental cages, distributed 20 cm
one from each other on the mud platform in front of the Rhizophora
fringe. The cages were plastic cylinders about 30 cm high and of 20 cm
in diameter, open up and down, thus the leaves were put down on the
mud. Two adult snails (with a shell size of 8–10 cm), taken within the
experimental area, were led, at random, on each leaf. Before each trial
the leaves were marked with a tag and weighted using an electric
precision balance (±0.01 gr.). After two hours, the snails were removed
and the leaves were gently collected, washed (for removing the mud),
dried and then weighted. The difference between the weight before
and after each trial rapresents the amount (in percentage) consumed
by the snails.
The trials were performed 3 times around spring tide, in different
not consecutive days, during the wet season (July 2007). As in
experiment 2, we compensated the minor size of A. marina, P. acidula,
L. racemosaand X. granatum leaves joining together 2–4 leaves.
2.2.3. Are there differences in the attractiveness of the leaves of various
mangrove species?(experiment 4)
In order to test whether T. palustris was differentially attracted by
the various mangrove plants, we tested the attraction effect of green
leaves belonging to the above-cited eight mangrove species. To
answer this question, we used the “cylinder experiment” approach
as performed by Fratini et al. (2001). Briefly, this method consists of
putting the potentially attractive cues (i.e. the different mangrove
leaves, minced) on a petri dish placed inside opaque plastic cylinders,
about 30 cm high and 20 cm in diameter. The cylinders were raised
about 0.5 cm above the soil in order to exclude the visual component
Table 1
A two factor mixed-model ANOVA based on the number of snails found around the two
different treatments (control and minced mangrove leaves) at different times
Experiment 1
Source SS DF MS F P
Treatment 846.4 1 846.4 15.50 0.03
Time 218.6 3 72.87 2.40 0.09
Time X Treatment 163.8 3 54.6 1.80 0.17
RES 971.6 32 30.36
TOT 2200.4 39
Data (△N: the difference between the final and the initial number of snails around the
tested cues) were not transformed (Cochran's Test: C = 0.3672, P = n.s).
28 S. Fratini et al. / Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31

Table 2
A three factor mixed-model ANOVA based on the percentage of leaf grazed by adults snails at different times, during the dry and wet season

Experiment 2: leaf grazing rate

Source SS DF MS F P
Treatment (mangrove species) 132723.60 7 18960.51 28.88 b 0.001
Season 1999.60 1 1999.60 0.59 0.52
Time 12338.96 2 6169.48 7.16 b 0.001
Treatment X Season 7786.49 7 1112.36 0.37 0.90
Treatment X Time 9190.02 14 656.43 0.76 0.71
Season X Time 6735.74 2 3367.87 3.91 0.02
Treatment X Season X Time 42135.57 14 3009.68 3.49 b 0.001
RES 372128.28 432 861.41
TOT 585038.28 479

Mangrove Average Dry Season Average Dry Season Average Dry Season Average Wet Season Average Wet Season Average Wet Season
species Time 1 ± SE Time 2 ± SE Time 3 ± SE Time 1 ± SE Time 2 ± SE Time 3 ± SE
A. marina 12.5 ± 6.9 35.5 ± 9.9 23.5 ± 4.8 29.7 ± 6.2 30.0 ± 9.5 51.9 ± 9.4
B. gymnorrhiza 43.5 ± 14.1 95.0 ± 2.9 41.1 ± 12.9 65.3 ±14.4 54.0 ± 11.7 86.4 ± 9.9
C. tagal 29.0 ± 11.6 76.0 ± 7.6 61.4 ± 9.4 54.0 ± 13.3 29.4 ± 10.6 44.5 ± 12.5
R. mucronata 43.0 ± 12.1 38.3 ± 11.3 45.5 ± 8.8 36.5 ± 11.7 48.0 ± 10.0 55.6 ± 12.3
S. alba 38.1 ± 10.0 20.6 ± 9.4 40.3 ± 11.9 26.4 ± 10.8 44.0 ± 7.2 44.4 ± 10.2
L. racemosa 6.7 ± 3.2 25.6 ± 7.8 59.3 ± 6.8 29.2 ± 7.7 43.5 ± 8.1 35.0 ± 6.9
P. acidula 32.0 ± 5.1 41.0 ± 6.7 43.5 ± 10.1 55.5 ± 11.3 37.5 ± 9.9 59.5 ± 10.4
X. granatum 0.5 ± 0.5 3.0 ± 2.5 10.8 ± 6.8 3.2 ± 3.0 0.2 ± 0.2 0.0 ± 0.0

The data were not transformed (Cochran's Test: C = 0.0503, P = n.s.).

from the process of food localisation while allowing odours to flow in
and out of the cylinders, at the ground level. During each trial, we used
five cylinders for each species (including the control, an empty petri
dish). Cylinders were randomly distributed 1 m apart along a transect
of approximately 100 m: since a snail starts to perceive a grated leaf
from about 40 cm (pers. obs.), this distance guarantees that the
attraction of a treatment is out of the range of influence of the other
ones. Four trials were performed, on different days, at a diurnal low
tide close to a spring tide (when T. palustris is more active: Fratini
et al., 2000), two during the dry season (Febraury 2005) and two
during the wet season (July 2007). Before each trial, the number of
snails present within a 30 cm radius around the cylinders was counted
and a similar count was performed one hour later. The difference
between the final and initial snail number (△N) represents the
number of snails attracted towards the cylinders. For more methodo-
logical information on the cylinder experiments, refer to Fratini et al.
(2001).
2.3. Statistical analysis
For the sets of data of experiment 2 and 4, a 3–way mixed model
ANOVA was applied, the factors being Time (random and orthogonal)
and Treatments and Season (fixed and orthogonal). A two–way mixed
model ANOVA (factors being Time, random and orthogonal, and
Treatments, fixed and orthogonal) was applied to the data set of
experiments 1 and 3. As treatments we considered: the tested cues
put into net bags (expt. 1) or into cylinders (expt. 4); the percentage of
consumed leaf as well as the number of snails found on a single leaf for
experiment 2; and the aumont of grazed leaf for experiment 3.
Thirty-four of the 480 experimental leaves used in expt. 2 were lost
during the trials: in the data set, for maintaining a balanced ANOVA
design, the percentage of grazed area and the number of snails present
corresponding to the lost leaves were replaced by the mean values for
each corresponding mangrove species within each trial. Data were
transformed only after having assessed their heterogeneity of variance

Table 3
A three factor mixed-model ANOVA based on the number of snails found on a single leaf at the end of the trials at different times, during the dry and wet season

Experiment 2: number of snails per leaf

Source SS df MS F P
Treatment (mangrove species) 125.97 7 18.0 20.8 b 0.001
Season 0.06 1 0.06 0.04 0.8623
Time 29.7 2 14.9 20.6 b 0.001
Treatment × season 13.0 7 1.9 1.2 0.3810
Treatment × time 12.1 14 0.9 1.2 0.2745
Season × time 23.0 2 1.5 2.2 0.1274
Treatment × season × time 22.3 14 1.6 2.2 0.007
RES 312.4 432 0.7
TOT 518.6 479

Mangrove Average dry season Average dry season Average dry season Average wet season Average wet season Average wet season
species Time 1 ± SE Time 2 ± SE Time 3 ± SE Time 1 ± SE Time 2 ± SE Time 3 ± SE
A. marina 1.6 ± 0.7 5.2 ± 2.2 4.9 ± 1.7 2.6 ± 0.6 2.6 ± 0.6 4.2 ± 0.9
B. gymnorrhiza 2.9 ± 0.9 8.3 ± 1.1 5.9 ± 2.0 6.6 ± 2.2 5.0 ± 0.9 7.7 ± 1.3
C. tagal 4.9 ± 2.4 7.8 ± 0.9 15.3 ± 2.5 4.9 ± 1.2 3.2 ± 1.2 5.0 ± 2.3
R. mucronata 6.2 ± 2.3 5.5 ± 1.9 13.2 ± 2.7 3.8 ± 1.5 6.2 ± 1.40 8.7 ± 2.4
S. alba 2.1 ± 0.7 1.1 ± 0.6 2.3 ± 0.7 2.2 ± 1.1 4.2 ± 0.9 5.1 ± 1.6
L. racemosa 6.9 ± 0.2 3.7 ± 1.2 10.6 ± 1.4 4.0 ± 1.2 7.8 ± 1.8 7.4 ± 1.8
P. acidula 8.3 ± 1.7 14.8 ± 3.8 13.7 ± 1.9 11.1 ± 2.8 4.1 ± 1.8 14.6 ± 3.3
X. granatum 0.3 ± 0.3 1.4 ± 1 0.6 ± 0.4 0.5 ± 0.3 0.4 ± 0.6 0.7 ± 0.3

The data (n + 1) were natural-log transformed (Cochran's test: C = 0.11, P b 0.01).

 S. Fratini et al. / Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31 29

Table 4 3.2. Do the leaves of various mangrove species promote different grazing
A two factor mixed-model ANOVA based on the amount of leaf grazed (in percentage) rates? (experiment 2 and 3)
by adults snails at different times

Experiment 3 Expt 2: a total of 327 leaves out of 446 remaining leaves was
Source SS DF MS F P grazed. Damage due to the snail radula is unmistakable (Fratini et al.,
Treatment (mangrove species) 104.65 7 14.95 8.62 b 0.001 2004), so we could be sure that the leaves were brushed by T. palustris.
Time 2.78 2 1.39 1.29 0.28 The proportion of leaves brushed or left intact at the end of the trials is
Time X Treatment 24.28 14 1.73 1.61 0.09 highly different for the different mangrove species (χ2 = 101.89, df = 7,
RES 103.45 96 1.08 P b 0.001) (A. marina: 82.15% vs. 17.85%; B. gymnorrhiza: 81.1% vs.
TOT 235.18 119
18.9%; C. tagal: 73.6% vs. 26.4%; R. mucronata: 81% vs. 19%; L. racemosa:
Mangrove species Average ± SE 82.7% vs. 17.3%; P. acidula: 91.7% vs. 8.3%; S. alba: 79.6% vs 20.4%; X.
A. marina 9.2 ± 1.3 granatum: 18.3% vs. 81.7%).
B. gymnorrhiza 16.1 ± 1.9 The three factor mixed model ANOVA indicates that T. palustris
C. tagal 16.4 ± 3.5 consumes different amounts of the various mangrove species
R. mucronata 15.6 ± 1.9 (Table 2), but it also reveals a significant interaction between the 3
S. alba 16.5 ± 2.3
L. racemosa 10.4 ± 2.5
factors. Although for almost all the species the percentage of grazed
P. acidula 13.5 ± 1.9 leaves was very variable in the three temporal replicates (Table 2),
X. granatum 0.4 ± 0.4 the snails always grazed very small quantities of X. granatum and
The data were square-root transformed (Cochran's Test: C = 0.23, P b 0.01). large amounts of B. gymnorrhiza, C. tagal, R. mucronata and P. acidula
(in average the grazed rate is about 50%). A. marina, S. alba and L.
racemosa are also consumed, but less than the three rhizophoracean
species and P. acidula (in average about 30%). The SNK test confirms
by means of the Cochran test. The SNK test was used, where this differences between X. granatum and the other species.
appropriate, to compare the relative effects of different treatments. The number of snails found on a single leaf varies depending on the
GMAV5 software (University of Sydney, Australia) was used for the mangrove species (Table 3). P. acidula was always able to attract the
analyses. higher number of snails; in average, the three rhizophoracean species
and L. racemosa attract a high number of snails, followed by A. marina
3. Results and S. alba. Otherwise the number of animals attracted by X. granatum
was always very small, as confirmed by the SNK test. In the ANOVA,
3.1. Is the perception of the food items chemically mediated underwater? the factor Time as well as the interaction among the three factors are
(experiment 1) also significant: the SNK test indicates that this is imputable to a great
variability among the three trials recorded for C. tagal and L. racemosa
The number of snails attracted towards the net bags containing the during the dry season, and for R. mucronata during the wet season
minced R. mucronata fresh leaves (ΔN = 9.3 ± 1.85) is consistently (Table 3).
higher than that which moved towards the controls (ΔN = 0.1 ± 0.38). Expt 3: the two factor ANOVA showed that the percentage of
The ANOVA clearly indicates a significant difference between the two consumed leaf is significantly different for the different species
treatments (Table 1). The interaction between the two factors was also (Table 4). The SNK test revealed that X. granatum are grazed less than
significant, and this results from the difference in the number of snails all the other species, and A. marina less than all the other species but
attracted towards the minced leaves in the two experimental trials C. tagal. The factor Time and the interaction between the two factors
(Table 1). did not affect the grazing rate (Table 4).

Table 5
A three factor mixed-model ANOVA based on the number of snails found around the different treatments (control and various mangrove leaves) during dry and wet seasons at
different times

Experiment 4

Source SS DF MS F P
Treatment (mangrove species) 84.57 8 10.57 5.93 0.01
Season 65.64 1 65.64 15.09 0.16
Time 0.22 1 0.22 0.22 0.64
Treatment X Season 24.31 8 3.04 3.93 0.03
Treatment X Time 14.28 8 1.78 1.77 0.09
Season X Time 4.35 1 4.35 4.31 0.04
Treatment X Season X Time 6.19 8 0.77 0.77 0.63
RES 145.22 144 1.01
TOT 344.75 179

Mangrove species Average Dry Season Time 1 ± SE Average Dry Season Time 2 ± SE Average Wet Season Time 1 ± SE Average Wet Season Time 2 ± SE
A. marina 6.2 ± 1.6 0.6 ± 0.6 11.4 ± 4.9 7.0 ± 3.0
B. gymnorrhiza 3.2 ± 1.5 5.8 ± 3.0 15.4 ± 7.5 29.8 ± 5.6
C. tagal 6.8 ± 2.5 4.2 ± 3.2 16.6 ± 6.0 11.6 ± 2.4
R. mucronata 5.8 ± 2.1 3.8 ± 2.6 7.0 ± 2.8 9.2 ± 4.2
S. alba 0.6 ± 0.4 0.8 ± 0.5 9.6 ± 4.2 9.3 ± 1.7
L. racemosa 2.2 ± 1.4 1.2 ± 1.2 10.5 ± 4.8 11.0 ± 2.9
P. acidula 11 ± 4.3 8.4 ± 4.8 13.4 ± 2.9 14.4 ± 3.4
X. granatum 1.6 ± 0.8 0.8 ± 0.6 2.4 ± 0.68 1.0 ± 0.71
Control 0.8 ± 3.7 0.2 ± 0.2 −0.4 ± 1.12 0.2 ± 1.0

Data correspond to △N + 1: the difference between the final and the initial number of snails around the tested cues. The data were square-root transformed with negative values
considered equal to zero (Cochran's Test, C = 0.14, P b 0.05).
30 S. Fratini et al. / Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31

Table 6
Mangrove species ranked for attraction and palatability

Overall results

Mangrove species Family Attraction Palatability Palatability Attraction

Avicennia marina
Bruguiera gymnorrhiza
Ceriops tagal
Rhizophora mucronata
Sonneratia alba
Lumnitzera racemosa
Pemphis acidula
Xylocarpus granatum
(Expt. 2: Number (Expt. 2: Amount (Expt. 3: Amount (Expt. 4: Number
of snails per leaf) of leaf grazed) of leaf grazed) of snails per cylinder)
Avicenniaceae 6 7 7 6
Rhizophoraceae 4 1 3 1
Rhizophoraceae 3 2 2 4
Rhizophoraceae 2 4 4 3
Sonneratiaceae 7 5 1 7
Combretaceae 5 6 6 5
Lythraceae 1 3 5 2
Meliaceae 8 8 8 8

Values from 1 (highest) to 8 (lowest).

3.3. Are there differences in the attractiveness of the leaves of various
mangrove species? (experiment 4)
The mean number of adult snails found around the cylinders
containing different mangrove species is reported in Table 5 together
with its standard error. Averagely, the largest numbers of animals
were attracted by cylinders containing the three rhizophoracean
species and Pemphis acidula minced leaves, while the lowest numbers
were attracted to cylinders containing X. granatum leaves.
The ANOVA clearly reveals a significant difference among the
various treatments (Table 5): the SNK test indicates that the control is
significantly different from the rhizophoracean species and P. acidula.
The attraction exerted by X. granatum is different from that of B.
gymnorrhiza, P. acidula and A. marina; A. marina is also differentiated
by P. acidula, R. mucronata and S. alba. Finally, B. gymnorrhiza is dif-
ferent from S. alba. All the other pairwise comparisons were not
significant.
The ANOVA test also reveals a significant interaction between the
factors Treatment and Season (Table 5): the SNK test indicates that for
all the treatments except the control, P. acidula and X. granatum the
number of snails found around the cylinders was higher during the
wet season than the dry season.
3.3.1. Overall results
Table 6 summarises the results of Experiments 2, 3 and 4 in terms
of the attractiveness and palatability of the different mangrove spe-
cies. We have attributed a value, from 1 to 8, to each plant in ac-
cordance to its rank position within each experiment.
4. Discussion
T. palustris is able to locate mangrove leaves underwater utilising
chemical cues released by the leaf itself when broken. Fratini et al.
(2001) demonstrated that this snail is also able to perceive air-borne
cues at low tide. Thus, combining these results, we may conclude that
T. palustris is able to use both air-borne and water-borne odours to
locate food. The efficacy of this attraction towards an underwater leaf
is very high (the average number of snails found around a grazed leaf
at high tide was 9.3 ± 1.8) and comparable to that recorded at low tide
(Fratini et al., 2001 reported an average of 10.0 ± 3.4 snails approaching
cylinders with a grated leaf on a Petri dish). On the basis of these data
and from previous studies (Fratini et al., 2001, 2004), we can declare
that T. palustris is the only macrobenthic species of East Africa
mangroves able to search, detect and consume mangrove leaves at
both high and low tides. Such an enlarged feeding window permits T.
palustris to occupy temporal niches left empty by the sesarmid crabs.
Considering the grazing rate, our results clearly indicate that T.
palustris eats the leaves of all the mangrove species except X.
granatum. However, our data show that this snail tends to eat large
amounts of the three Rhizophoraceae species, that are the most
abundant plants in the Indo-Pacific mangrove forests, of Pemphis
acidula and Sonneratia alba, and a minor quantity of Avicennia marina
and Luminitzera racemosa. These preferences can hardly be explained
in terms of chemical composition and nutrient values, since all the
mangrove leaves have on average a high tannin content and C/N ratio
far exceeding the maximum value for sustainable animal nutrition
(equal to 17) (Russell-Hunter, 1970), i.e. generally speaking they are
food of poor quality (Skov and Hartnoll, 2002). The graze rate rela-
tive to the different species is more variable when measuring in
“free”(expt. 2) than in “obliged” conditions (expt. 3): this is due to the
influence of the feeding recrutiment process and, consequentely, to
the umpredictable number of snails eating on a single leaf.
In addition to the basic constituents of all angiosperms (aminoa-
cids, fatty acid, lipids, carbohydrates, etc.), each mangrove plant is
characterised by specific substances, and the more phylogenetically
related the plants are, the more similar their chemical composition
(Bandaranayake, 2002). Chou et al. (1977) extracted from the X.
granatum leaves a specific compound (N methylflindersine), that is an
insect anti-feedant, thus its presence could be the reason why T.
palustris dislikes this plant. However, the specific effect of this com-
pound on the mud whelk should be tested by conducting further
experiments.
T. palustris is chemically attracted towards the leaves of all the
mangrove species except X. granatum. Moreover, P. acidula and the
Rhizophoraceae leaves wield a stronger attractive power than L.
racemosa, S. alba and A. marina. The evidence that the snails are
neither attracted to, nor do they consume, X. granatum leaves may
indicate that this gastropod avoids eating this plant, and that this
choice is chemically mediated and based on a species-specific cue or
cues. If so, what are the substances that are perceived by the snails?
We are not able to answer this question and attraction field tests using
single specific substances would be necessary. We hypothesise that
the “leaf odour” perceived by the snails is produced by a mixture of
substances, both volatile and water-soluble, that are alternatively
perceived in different tidal conditions. Tropical plants, in general, have
evolved chemical defences against herbivores, and the most likely
candidates for this are low molecular weight compounds such as
monoterpenes, proteins and amino acids, cyanogenic compounds,
alkaloids, and saponins (Coley and Barone, 1996). It may be assumed
that snails base their air-borne olfaction ability on some of these
substances. On the other hand, tannins, typical secondary metabolites
produced by most mangrove species are very heavy (i.e. non-volatile)
and soluble in water, and could be involved in the water-borne odour-
mediated localisation process. Moreover, in light of the differential
attraction exerted by the mangrove plants, we may hypothesise that T.
palustris bases its chemical-mediated food searching strategy on
species-specific cues (as demonstrated for other gastropods with
selective diets: Croll, 1983).
Overall, our results underline the strong efficacy and the high
adaptive value of the food searching strategy adopted by T. palustris.
 S. Fratini et al. / Journal of Experimental Marine Biology and Ecology 362 (2008) 26-31
Its ability to detect and orient towards distant food sources enables T.
palustris to reach the leaves in a short span of time and in high and low
tide. This is a crucial advantage in an enviroment such as the Indo-
Pacific mangrove forests, where fallen leaves are a limited food
resource due to the high number of leaf-eater species (mainly crabs)
and where intra- and inter-specific interaction and competition arise
(Slim et al., 1997; Fratini et al., 2000). Moreover, thanks to its skill in
discriminating at a distance among different food items, T. palustris
can reach the leaves it prefers and can avoid the non- prefered leaves
(whatever the reasons may be), greatly optimising its feeding activity.
Acknowledgments
We are grateful to Rocco Rorandelli, Angela Sacchi, Nadia Frodella,
Gianni Brescacin, Outi. Lähteenoja and Cecilia Coffa for their help in
the field. A tender thank is due to Guido Fratini, Francesca Zacchi and
Niccolò Biliotti for spending their vacation collecting and weighted
leaves and sharing with us many hours in the mud. Many thanks are
also due to G. Sanders for linguistic revision. Funds are from a FIRB
project (Italian MIUR), from the European Project no. INCO - CT2004 -
510863 (PUMPSEA: Peri-urban mangrove forests as filters and po-
tential phythoremediators of domestic sewage in East Africa) and from
Fondi d'Ateneo to M. V. and S. C. (ex 60% University of Florence). All the
experiments conducted in this study complied with the current laws
of Kenya. [SS]
References
Bandaranayake, W.M., 2002. Bioactivieties, bioactive compounds and chemical
costituents of mangrove plants. Wetlands Ecol. Manag. 10, 421–452.
Bouillon, S., Dahdouh-Guebas, F., Rao, A.V.V.S., Koedam, N., Dehairs, F., 2003. Sources of
organic carbon in mangrove sediments: variability and possible ecological
implications. Hydrobiologia 495, 33–39.
Bouillon, S., Raman, A.V., Dauby, P., Dehairs, F., 2002a. Carbon and nitrogen stable
isotope ratios of subtidal benthic invertebrates in an estuarine mangrove ecosystem
(Andhra Pradesh, India). Estuar. Coast. Shelf Sci. 54, 901–913.
Bouillon, S., Koedam, N., Raman, A.V., Dehairs, F., 2002b. Primary producers sustaining
macro-invertebrate communities in intertidal mangrove forests. Oecologia 130,
441–448.
Chou, F.Y., Hostettmann, K., Kubo, I., Nakanishi, K., Taniguchi, M., 1977. Isolation of an
insect antifeedant N methylflindersine and several Benz[C]phenanthridine alka-
loids from East African plants: a comment on Chelerythrine. Heterocycles 7,
969–977.
31
Coley, P.D., Barone, J.A., 1996. Herbivory and plant defense in tropical forests. Ann. Rev.
Ecolog. Syst. 27, 305–335.
Croll, R.P., 1983. Gastropod chemoreception. Biol. Rev. 58 (2), 293–319.
Crowe, T.P., 1997. The distribution of Terebralia palustris (Linne) with respect to
microhabitat in mangrove forests of Darwin harbour I: description of patterns. In:
Hanley, J.R., Caswell, G., Megirian, D., Larson, H.K. (Eds.), Proceedings of the
Sixth International Marine Biological workshop: the marine flora and fauna of
Darwin Harbour, Northern Territory, Australia. Museum and Art Galleries of the
Northern Territory and the Australian Marine Sciences Association, Darwin,
Australia, pp. 425–433.
Crowe, T.P., McMahon, R.F., 1997. The distribution of Terebralia palustris (Linne) with
respect to microhabitat in mangrove forests of Darwin harbour II: experimental
evaluation of behaviour. In: Hanley, J.R., Caswell, G., Megirian, D., Larson, H.K. (Eds.),
Proceedings of the Sixth International Marine Biological workshop: the marine
flora and fauna of Darwin Harbour, Northern Territory, Australia. Museum and Art
Galleries of the Northern Territory and the Australian Marine Sciences Association,
Darwin, Australia, pp. 435–443.
Dahdouh-Guebas, F., Verneirt, M., Tack, J.F., Van Speybroeck, D., Koedam, N., 1998.
Propagule predators in Kenyan mangroves and their possible effect on regenera-
tion. Mar. Freshw. Res. 49, 345–350.
Fratini, S., Cannicci, S., Vannini, M., 2000. Competition and interaction between Neo-
sarmatium smithi (Crustacea, Grapsidae) and Terebralia palustris (Mollusca,
Gastropoda) in a Kenyan mangrove. Mar. Biol. 137, 309–316.
Fratini, S., Cannicci, S., Vannini, M., 2001. Feeding clusters and olfaction in the mangrove
snail Terebralia palustris (Potamididae: Gastropoda). J. Exp. Mar. Biol. Ecol. 261,
173–183.
Fratini, S., Vigiani, V., Vannini, M., Cannicci, S., 2004. Terebralia palustris (Gastropoda;
Potamididae) in a Kenyan mangal: size structure, distribution and impact on
consuming leaf litters. Mar. Biol. 114, 1173–1182.
Houbrick, R.S., 1991. Systematic review and functional morphology of the mangrove
snails Terebralia and Telescopium (Potamididae; Prosobranchia). Malacologia 33,
289–338.
Nishihira, M., 1983. Grazing of the mangrove litters by Terebralia palustris (Gastropoda:
Potamididae) in the Okinawa mangal: preliminary report. Galaxea 2, 45–58.
Plaziat, J.C., 1984. Mollusk distribution in the mangal. In: Por, F.D., Dor, I. (Eds.),
Hydrobiology of the Mangal: the Ecosystem of the Mangrove Forests. Junk, Boston,
pp. 111–143.
Russell-Hunter, W.D., 1970. Aquatic productivity: an introduction to some basic aspects
of biological oceanography and limnology. Macmillan, New York.
Slim, F.J., Hemminga, M.A., Ochieng, C., Jannink, N.T., Cocheret de la Moriniere, E., van
der Velde, G., 1997. Leaf litter removal by the snails Terebralia palustris (Linnaeus)
and sesarmid crabs in an East African mangrove forest (Gazi Bay, Kenya). J. Exp. Mar.
Biol. Ecol. 215, 35–48.
Skov, M.W., Hartnoll, R.G., 2002. Paradoxical selective feeding on a low-nutrient diet:
why do mangrove crabs eat leaves? Oecologia 131, 1–7.
Wells, F.E., 1980. A comparative study of distribution of the mudwhelks Terebralia
sulcata and T. palustris in a mangrove swamp in Nortwestern. Australia. Malacol.
Rev. 13, 1–5.