Download relationships between shell shape, water reserves, survival and

/. MolL Stud (1997), 63,511-529
© The Malacological Society of London 1996
RELATIONSHIPS BETWEEN SHELL SHAPE, WATER
RESERVES, SURVIVAL AND GROWTH OF HIGHSHORE
LITTORINIDS UNDER EXPERIMENTAL CONDITIONS IN
NEW SOUTH WALES, AUSTRALIA
M.G. CHAPMAN
Institute of Marine Ecology and Centre for Research on Ecological Impacts of Coastal Cities, Marine Ecology
Laboratories All, University of Sydney, NSW, 2006, Australia
(Received 22 October 19%; accepted 19 February 1997)
ABSTRACT
INTRODUCTION
The shell morphologies of the highshore littorinids,
Littorina unifasciata Gray and Nodilittorina pyramidalis (Quoy & Gaimard) have previously been shown
to vary at a variety of spatial scales, including among
replicate sites at the same height, from height to
height and from shore to shore. In this study, the
relationships between morphology of the shell, the
reserves of water held within the shell, the size of the
foot and survival on different shores and rates of
growth in different habitats were examined for L.
unifasciata and, to a lesser extent, N. pyramidalis.
Reserves of water were not consistently related to
size or shape of the shell, but did increase as relative
weight of shell increased. This may be due to the relatively smaller body providing more internal volume
for extra-corporeal water. Water reserves and the
amount of free water held in the shell were also not
related to loss of water or survival during extended
periods of emersion. Although the shape of the shell
on sheltered and exposed shores was correlated with
size of the foot, with the snails on an exposed shore
having larger apertures and feet than those on a
sheltered shore, transplant experiments did not show
differential mortality between morphs from the
different shores. All translocated and transplanted
snails disappeared from the exposed shore at a
greater rate than from the sheltered shore, but this
was probably due to the snails dispersing out of the
experimental areas rather than due to mortality.
Therefore, many of the large-scale models that have
previously been used to describe patterns of shell
shape in intertidal gastropods do not appear to be
important in these highshore littorinids. Finally, field
experiments on growth of juvenile L. unifasciata
indicated that rate of growth, largely governed by
opportunity to feed rather than type and quantity of
food, is the most likely explanation for the small- and
large-scale patterns of shell shape that have been
previously described in this species.
The morphology of many gastropods varies
considerably among different habitats. Highshore species may be more globose or have
relatively smaller apertures or longer spires
than lowshore species (Vermeij, 1973, 1978).
Shell shape may also vary intraspecifically,
giving rise to distinct morphs at different
heights on the shore (Johannesson &
Johannesson, 1990; Johannesson, Johannesson
& Rolan-AJvarez, 1993; Chapman, 1995).
Thickness of the shell may also differ from
height to height on a shore, although patterns
of difference vary among species (Seapy &
Hope, 1973; Seed, 1973; Vermeij, 1978; Chapman, 1995). In many species, shell morphology
also varies intraspecifically, frequently correlated with gradients of wave exposure. Shells
from sheltered shores may have smaller
apertures and be more elongated than those
from exposed shores (Kitching, Muntz &
Ebling, 1966; Newkirk & Poyle, 1975; Heller,
1976; Raffaelli, 1978; Van Marion, 1981;
Crothers, 1984). Shells from sheltered shores
may also be thicker and heavier than those
from exposed shores (Kitching et al., 1966;
Kitching, 1976; Geller, 1990).
The shape or weight of a shell or the relative
size of the apertura] opening of highshore
morphs (or species) may increase reserves of
water, reduce rate of loss of water or assist
regulation of temperature during emersion
(Vermeij, 1973; Britton, 1995). Most hypotheses relating shell shape, reserves of water
and survival or fitness of highshore gastropods
have not been tested in the field, although
Wolcott (1973) showed that loss of water may
be influenced more by behaviour than shape in
512
M.G. CHAPMAN
limpets. In addition, littorinid snails, which
characterize highshore levels on many shores,
may close the aperture and not lose water
when emersed (Vermeij, 1978; McQuaid &
Scherman, 1988; Britton, 1995). Predation by
crabs may select for thicker, heavier shells at
lowshore levels where predation is important
(Kitching, 1976), or decalcification many thin
shells in highshore animals (Vermeij, 1978).
Crabs, which are common predators on
sheltered shores, have greater success preying
on thinner, broader shells from exposed shores
than on shells from sheltered shores (Kitching
et al, 1966; Kitching & Lockwood, 1974;
Kitching, 1976; Seed, 1978; Hughes & Elner,
1979; Johannesson, 1986). It is proposed that,
in areas where it is common, such predation
will favour survival of snails that produce thick
shells with narrow apertures. Size of aperture
can be positively correlated with size of foot
(Grahame & Mill, 1986; Etter, 1988; Frid &
Fordham, 1994), which may, in turn, be correlated with tenacity (Branch & Marsh, 1978;
Grenon & Walker, 1981). Animals with thinner
shells may accommodate larger bodies and,
hence, larger feet, which may be advantageous
in wave exposed sites (Kitching et al., 1966;
Seed, 1978; Hughes & Elner, 1979). In support
of these models, field experiments have shown
decreased dislodgment and increased predation on Nucella lapillus with larger apertures
and thinner shells (Kitching et al., 1966).
Spatial variability in shell morphology is often
described in species with limited dispersal and
isolated breeding populations, although Boulding & Hay (1993) showed little correlation between variation in shell shape and breeding
strategy. Broadcast fertilizers often have no
particular genetic relationship among generations on a shore, so morphological differences
among these populations may reflect phenotypic responses of individuals to environmental
variability within their life-times, rather than
genetic differences among populations.
Most models to explain variation in shell
morphology invoke influences acting at
relatively large spatial scales, such as between
lowshore and highshore levels or between sheltered and exposed shores. Shape and/or weight
of the shell can also be affected by rates of
growth (Kitching & Lockwood, 1974; Kitching,
1976; Kemp & Bertness, 1984) or alternating
periods of growth and quiescence (Van Marion,
1981; Palmer, 1990), so that differences in
morphology may be a by-product of feeding
regime. Small-scale spatial variability in quality
and/or quantity of food (Underwood, 1984;
MacLulich, 1987) or opportunity to feed and
limited dispersal of adults, could potentially
give rise to small-scale variability in shell
morphology.
Four species of small to medium littorinids
are found at different levels on intertidal rocky
shores in New South Wales, Australia (Underwood, 1981a; Chapman, 1994a,b). Two of
these, Littorina unifasciata Gray (often placed
in the genus Nodilittorina) and Nodilittorina
pyramidalis (Quoy & Gaimard), differ in
density, size and shape and weight of the shell
within and among shores (Chapman, 1994a, b,
1995). Because they have dispersive larvae,
differences among populations are likely to reflect phenotypic responses to local conditions,
rather than genetic differences among populations. Differences in shape and weight of the
shell among shores were not clearly correlated
with wave exposure; large and small snails from
the same habitats did not show the same
patterns from shore to shore (Chapman, 1995).
There was a general trend for L. unifasciata to
be more elongate with a smaller aperture high
on the shore compared to midshore levels on
all shores, but there was also considerable
small-scale variability in shell morphology
among replicate sites at the same level on the
shore.
This study examines models relating shell
morphology of L. unifasciata (and to a lesser
extent, N. pyramidalis) to reserves of water,
relative size of the foot, survival on exposed
and sheltered shores and growth under different experimental conditions. If greater reserves
of water are advantageous in areas which are
emersed for long periods, morphs from
sheltered shores should have greater reserves
of water than those from wave exposed shores.
Similarly, if the shape or weight of the shell is
correlated with relative size of the foot and
reduced risk of dislodgment by waves, morphs
from wave exposed shores should, first, accommodate a larger foot than snails of similar size
from sheltered shores and, second, show
reduced dislodgment by waves. In addition, the
quality of food or opportunity to feed may vary
from height to height and/or shore to shore,
causing differences in rates of growth and
potentially, differences in shell morphology or
weight in different places. Therefore, the effects
of food, height on the shore and intraspecific
density on rates of growth and development of
shell shape in L. unifasciata (which lives over
a wider vertical range than N. pyramidalis)
were also examined in a manipulative field
experiment.
SHELL SHAPE IN LITTORINIDS
MATERIALS AND METHODS
Measurement of reserves of water
Reserves of water were measured from snails
collected form a number of shores within 50 km of
Sydney, New South Wales, Australia (described in
detail in Chapman, 1995). In summary, Shores 1, 2
and 4 were steeply sloping with considerable wave
action and approximately 2-5 m vertical height
between high- and midshore levels (i.e. across the
vertical range of L. unifasciata). Shores 3 and 6 were
broad, gently-sloping platforms, with less than 2 m
vertical height between high- and midshore levels.
(Shore 5 in Chapman (1995) was not included in this
study). On each shore, L. unifasciata were collected
from mid- and highshore sites, i.e. towards the top
and bottom of their range. N. pyramidalis were collected from the same level on the shore as highshore
L. unifasciata (Chapman, 1995). The maximal length
of each shell and maximal internal length of the apertural opening were measured to the nearest 0.05 mm
using calipers. The shape of the shell was calculated
as the ratio of shell length:aperture length (Crothers,
1984; Basingthwaite & Foulds, 1985; Boulding &
Hay, 1993; Chapman, 1995). The relative weight of
the shell was the weight of the dried shell/weight of
dried body (Chapman, 1995).
Reserves of water were measured within 12 hours
of collection of each specimen. To determine total
water-holding capacity of the snails (hereafter referred to as water reserves), they were placed in fresh
seawater in the laboratory. The snails immediately
became active and moved around the container before clustering above the surface of the water. When
carefully removed from the container, the snails
withdrew slightly into their shells and closed their
opercula. They were then left undisturbed for a few
minutes to ensure that they did not withdraw further
into the shell, expelling free (extra-corporeal) water.
The outside of each shell was then dried with tissue
paper (with minimal handling). Any snails which reemerged or withdrew further into the shells during
handling were discarded. The snails were then
weighed to the nearest 0.05 mg (wet weight). If the
opercula were prodded with a blunt probe, the snails
withdrew into their shells, expelling water from the
mantle cavity and shell chamber (= expelled water;
Boyle et al., 1979; Garrity & Levings, 1984), but they
did not expel all free water. The amount retained
depended on the extent of withdrawal which could
not be standardised because it depended on individual responses of the animals. The amount of expelled
water was therefore not a reliable measure of the
amount of free water held in the shell and so total
water reserves (tissue water and extra-corporeal
water) were measured.
To measure reserves of water, each shell was carefully broken open after the wet weight was measured
and the shell and body were separated. The surface
of the body was then dried lightly, ensuring no rupture and loss of body fluids and it was then reweighed
(wet body weight). The body and shell were dried
and separately weighed. Reserves of water were
513
calculated as the total weight of water (= wet weight
- weights of dried shell and dried body) divided by
the wet weight of the body without the shell. This
scaled the amount of water in the tissues and held
within the cavity of the shell to the weight of the
body of the animal independently of the weight of
the shell. This was done because of large intraspecific
variation in shell weight (Chapman, 1995).
Water reserves of experimental snails were
measured in the same way, except that the snails
were not rehydrated first. As soon as they were
collected, they were placed with their apertures
embedded in plasticine to prevent them from
emerging and losing water. They were kept cool until
the measurements were made.
Relationships between reserves of water and the size,
shape and relative weight of the shell
In a preliminary study, water reserves were tested
against shell length, shell shape and relative shell
weight to examine the specific models that larger snails,
morphs with higher spires and relatively smaller apertures (characteristics of many highshore specimens)
or morphs with relatively lighter shells have greater
reserves of water. Comparisons were done for each
species from two or three different shores (Chapman,
1995). Snails were randomly collected along a transect
of approximately 20 metres at high- and midshore
levels on each shore.
Water reserves were separately compared against
shell length, shell shape or relative shell weight
across all shores by analyses of covariance for each
species in turn. Cochran's test was used to test for
heterogeneity of variances among samples. If nonsignificant (P > 0.05), slopes were compared. Each
population was considered separately if slopes were
heterogeneous. If the slopes did not differ, a common
slope was calculated and tested for significance, i.e. a
significant relationship with the covariate. Differences among intercepts were tested to identify mean
differences in water reserves among samples, having
adjusted the data to a standard length, shape or
weight, respectively.
Reserves of water of snails living on different shores
Although reserves of water varied according to some
aspects of shell morphology on some shores, these
differences could not easily be correlated with wave
exposure on the different shores. Water reserves may
be influenced by the behaviour of the animals, in
addition to shell morphology, if snails from different
shores withdraw into the shell to a different extent
when they become emersed. To examine the model
that reserves of water would be larger in snails living
on sheltered than on exposed shores, water reserves
were compared for large and small individuals among
replicate exposed (Shores 1 and 2) and sheltered
shores (Shores 3 and 6; Chapman, 1995). Small
animals (« 150 mg wet weight) were the average size
of L. unifasciata at midshore levels on these shores.
514
M.G. CHAPMAN
Large animals (400-700 mg wet weight) included the
largest highshore L. unifasciata found in this study.
The two sizes of each species were not found at each
level on each shore. Large L. unifasciata and N. pyramidalis were not found at midshore levels on any
shore and large L. unifasciata were not found at highshore levels on Shore 2. Therefore, water reserves
were compared inter- and intraspecifically using
selected subsets of data. Size of snails, wave exposure
and height on the shore were treated as fixed factors;
shores were a random factor nested within wave
exposure.
were continuously emersed. To test the specific hypothesis that the shape (or behaviour) of morphs found
on sheltered shores reduces loss of water during long
periods of emersion, the shell shape and water
reserves (without rehydration) of 10 transplanted
and 10 resident snails of each species were calculated
at the end of this 10 day period. No distinction was
made between undisturbed and disturbed animals
from Shore 3 because effects of disturbance were
unlikely to persist for many weeks (Chapman, 1986).
To examine further the model that water reserves
enhance survival during prolonged emersion on
sheltered shores, survival of fully-hydrated individuals was compared to that of specimens which had
Relationship between the shape of the shell and the
been forced to expel extra-corporeal water by
size of the foot
repeated prodding of the opercula. Twenty N. pyraLittorina unifasciata and N. pyramidalis from Shore 1 midalis and 20 high- and midshore L. unifasciata of
each treatment were marked and placed in the sun
(a wave exposed shore) had relatively larger aperabove the level of high tide on Shore 3. To test
tures and shorter spires than specimens from Shore 3
whether environmental conditions during the experi(an extremely sheltered shore; Chapman, 1995).
ment were harsh enough to cause mortality of midThese snails were therefore selected to examine the
shore snails, survival of fully-hydrated specimens of
model that morphs from wave exposed shores have
Bembicium nanum (Lamarck), Nerita atramentosa
relatively larger feet than those from sheltered shores.
Random samples of 19 L. unifasciata and N.pyrami- Reeve and Austrocochlea porcata A. Adams (= A.
dalis of a range of sizes were collected from the high- constricta (Lamarck)) was also measured. These
species co-exist with midshore L. unifasciata, but selshore levels on each of these shores. The shell length
dom extend to highshore levels. During the following
and aperture length of each were measured and the
week, the seas were calm and the temperature
SL:AL calculated. The aperture of each snail was
exceeded 30°C each day. The experimental site did
then photographed. The snails were placed in water
not get wet by waves or splash during this period. The
and immediately climbed the vertical glass sides of
survival of the snails was monitored after 1) 2 and 4
the container. The extended foot of each snail was
days by observation of undisturbed animals in the
photographed 2-3 times during locomotion. The area
field and after 7 days by placing them into seawater
of the aperture and the area of the extended foot
and counting those emerging and becoming active.
were determined from these photographs using a
digitiser. The average area of the foot (from the
To examine the model that midshore morphs of L.
replicate photographs) and the area of the aperture,
unifasciata (which have relatively large apertures
relative to the size of the snails, were compared
compared to highshore morphs) would lose more
between shores using analyses of covariance for each
water during periods of emersion than would highspecies separately.
shore morphs (with relatively smaller apertures;
Chapman, 1995), water reserves of mid- and highshore morphs on Shore 3 were compared after an
Relationship between shell morphology and water
extended period of emersion. First, twenty mid- and
reserves or survival on a sheltered shore
highshore morphs were placed in water to hydrate.
When each had emerged from the water and closed
Although morphs from sheltered shores did not
its aperture, it was marked with non-toxic paint and
generally have larger reserves of water than those
ten of each morph were placed in each of two highfrom exposed shores, their shell shape may reduce
shore sites. In each site, there were also similarlyloss of water during emersion and enhance survival
marked undisturbed and disturbed resident snails
in habitats subjected to prolonged emersion. To evaland snails which were translocated between the two
uate this model, 50 N. pyramidialis and highshore L.
sites. These were not rehydrated and were included to
unifasciata from Shore 1 (an exposed shore) were
test the hypothesis that moving snails into unfamiliar
marked (with small dots of non-toxic paint) and
areas does not affect behaviour in such a way that
transplanted to a similar height on Shore 3 (a shelwill influence water reserves. The snails were left
tered shore). Fifty specimens of each species on
undisturbed for two weeks, during which there were
Shore 3 were marked in situ (undisturbed animals)
calm seas and no rain. A thick layer of salt encrustaand 50 others were similarly marked and translotion around highshore rock pools at the end of this
cated into the experimental site from a different site
period indicated little if any wetting of the substraon the sheltered shore (Chapman, 1986; Chapman &
tum during the experimental period. Water reserves
Underwood, 1992). All snails were left for three
were then measured and compared among the five
months during summer and their survival perioditreatments and two sites to test the hypotheses that
cally monitored. Survival of different morphs were
highshore morphs would retain more water than
compared using x2 tests. Towards the end of the
midshore morphs and that disturbance and movestudy, there was a period of 10 days of hot, dry
ment of snails into unfamiliar areas would not affect
weather with calm seas, during which time the snails
SHELL SHAPE IN LITTORINIDS
515
water reserves (all four treatments of highshore
and 4 days (in the first experiment) and after 1 and 5
snails would have similar reserves of water).
days (in the second experiment). These data provide
As a final test of the capability of L. unifasciata and information on distances moved by the snails in
N. pyramidalis to survive long periods of emersion, sur- different treatments, in addition to their rate of loss.
vival of 20 fully-hydrated specimens of L. unifasciata
and N. pyramidalis and 20 specimens of each species
Experiment to test the effects of type or quantity of
which had been forced to expel extra-corporeal
food or opportunity to feed on growth and shell shape
water, were monitored after 2 months of emersion in
of L. unifasciata
the laboratory. These snails were exposed to sunlight
each day.
This experiment was done twice with slight modification. In each experiment, 48 cores, 7 cm in diameter
and between 5 and 10 cm deep, were drilled out from
Comparative survival of different morphs on a wave Shore 3 (which has large densities of L. unifasciata;
exposed and sheltered shore
Chapman, 1994b). Twenty four cores were taken
from the middle and 24 from the top of the range of
To test the model that shell morphology enhances
L. unifasciata, thereby providing two sources of food
survival of morphs on shores on which they live,
naturally available to this species (Branch & Branch,
survival of different morphs was compared between
1981; Underwood, 1984). A few small pits were
a sheltered and wave exposed shore at Cape Banks
(Shores 3 and 4, respectively). Thirty L. unifasciata drilled into the surface of each core to provide shelter for juvenile snails. Each core was then placed in a
from each of four different sites on each shore were
stainless steel cup, with the surface of the core promarked and reciprocally transplanted between
shores. Nodilittorina pyramidalis was not included in truding approximately 1 cm above the rim of the cup.
A layer of plastic scouring pad was attached to this
this experiment because they were in small densities
rim to contain the snails because a pilot laboratory
in most of the sites on the wave exposed shore. In
each site, 20 undisturbed resident L. unifasciata were experiment had indicated that small littorinids did
marked in situ, 20 were disturbed and replaced in the not readily crawl over this surface. The cores were
arranged in groups of eight (four of each type of
same site and 20 were marked and translocated
core) on each of six wooden beams. These were fasbetween sites on the same shore (Chapman, 1986;
tened to a rock face on a shore in Port Jackson.
Chapman & Underwood, 1992). These treatments
Although a sheltered shore, there was regular wave
are necessary to distinguish between the effects of
splash from ferries using a nearby wharf. The experibeing moved to an unfamiliar place and being moved
ment was not done on the open coast because of
to a different shore. Therefore, each site contained
problems of vandalism.
four treatments; undisturbed resident snails, disturbed resident snails, translocated snails from a difThree of the beams were placed just above midferent site on the same shore and transplanted snails
shore level and the other three just below the level of
from the other shore. The relocated animals were
spring high tide. The lower cores were wetted regucounted after 4, 10 and 15 days. To confirm that the
larly for a few hours each day and the upper cores
snails on each shore were different shapes, but those
were wet for an hour or so during spring tides and by
in each site on a shore were of similar shapes, 10 indi- wave splash and rain at other times. Snails on the top
viduals from each site on each shore were randomly
sets of cores therefore had less opportunity to feed
selected and the relative shell shape (SL:AL ratio)
than those on the bottom set of cores. Either ten or
calculated.
two adult L. unifasciata (> 4 cm shell-length, on
average) and ten juveniles (approximately 1 mm shell
The above experiment showed less recovery of
length) were added to each core for the first experitranslocated and transplanted snails in all sites,
ment. In the second experiment, the large density of
particularly on the wave-exposed shore (see Results,
adult snails was increased from ten to 30 animals per
Fig. 4). Increased loss of these two treatments might
core. These densities are similar to naturally average
have been due to dislodgment by waves or increased
(Experiment 1) or naturally large (Experiment 2) and
dispersal of these snails from the experimental sites
naturally small densities in the field. It was proposed
because translocation and transplantation often increase rates of dispersal of L. unifasciata (Chapman, that these densities would alter availability of food
for small snails to small or large quantities, respect1986). Therefore, it was not clear whether the
ively. Each wooden beam therefore had two replimorphs from the sheltered shore had been dislodged
cates of each combination of origin of core (type of
at a greater rate than those from the exposed shore.
food) and density of adults (quantity of food) and the
To examine this further, two sites, approximately 10
two levels on the shore represented different opporm apart, were selected on Shore 4. In each site, four
tunities to feed. The snails were collected from a
treatments, each of 20 individually marked snails,
moderately exposed shore at Cape Banks where
were established as described above; undisturbed
there were large numbers of juveniles.
residents, disturbed residents, snails translocated
between sites and snails transplanted from the shelThe first experiment was left two months, but there
tered shore. This experiment was set up twice, using
was considerable loss of small snails from all treatdifferent sites each time. The numbers of relocated
ments, particularly from the lower sets of cores. At
snails and the distances and directions displaced by
the end of the experiment, the survival of small snails
each one (Underwood, 1977) were measured after 2
was calculated and the shell length and aperture
Comparison of survival of morphs from a sheltered or exposed shore on a
sheltered shore after transplantation between shores
Comparison of water reserves of morphs from a sheltered or exposed shore
on a sheltered shore after polonged emersion
Comparison of survival (i) of snails which were or were not forced to expel
extra-corporeal water; (ii) with other midshore species; (iii) of midshore and
highshore morphs of L. unifasciata; (iv) after prolonged emersion in the
laboratory
Comparison of survival of morphs from a sheltered or exposed shore on an
exposed shore after transplantation between shores
L. unifasciata and N. pyramidalis from a sheltered shore will survive
better in these habitats than will snails from an exposed shore
L. unifasciata and N. pyramidalis from a sheltered shore will contain
larger water reserves than snails from an exposed shore
Water reserves of L. unifasciata and N. pyramidalis will enhance
survival during prolonged emersion compared to other midshore
snails
L. unifasciata from an exposed shore will survive better on this
shore than will snails from a sheltered shore
Increased loss of transplanted and translocated snails on exposed
shores is due to migration, not dislodgement
L. unifasciata (i) feeding on food from high on the shore, (ii) with
smaller amounts of available food or (iii) with less opportunity to
feed will develop relatively smaller apertures and longer spires than
other snails
Size of aperture and foot were compared between snails from an exposed
and sheltered shore using analyses of covariance
L. unifasciata and N. pyramidalis from an exposed shore will have
relatively larger apertures and feet than snails from a sheltered shore
Two field experiments in which type and amount of food and opportunity to
feed were manipulated and subsequently shell shape and size of juvenile
L. unifasciata compared
Comparisons of rate of loss and rate of dispersal among morphs transplanted
from the sheltered shore and undisturbed, disturbed and translocated
morphs from the exposed shore
Water reserves were compared for large and small highshore L. unifaciata
and N. pyramidalis and small midshore L. unifasciata across replicate
sheltered and exposed shores
Water reserves of snails of a range of sizes and shapes were compared
among shores using analyses of covariance
Experimental tests
Water reserves are similarly correlated with shell length, shell shape
or relative shell weight in L. unifasciata and N. pyramidalis
Water reserves will be greater in large and small L. unifascitata and
N. pyramidalis from sheltered shores compared to wave exposed
shores
Hypotheses
Table 1. Summary of hypotheses being tested and experimental protocols.
>
X
2
p
o
SHELL SHAPE IN LITTOR1NIDS
length of each surviving juvenile was measured under
a microscope and the shell shape calculated. It was
not possible to know whether loss of juvenile snails
was due to emigration over the scouring paid or
whether they had been washed away. Therefore in
the second experiment, tea-strainers were attached
over the top of each core to prevent loss of snails.
The second experiment was left nearly six months
before the survival of large and small snails and shell
lengths and aperture lengths of all surviving juveniles
and a maximum of five adults per core were measured
and shell shape calculated.
Details of the hypotheses investigated in the study
and the experiments to test these hypotheses are
summarised in Table 1.
517
ata. Therefore, changes in water reserves with
shell shape varied differently among species
and shores.
Reserves of water increased as shells became
heavier (relative to body weight) in all species
on most or all shores (Fig. 1). A common
slope could not be calculated for midshore
(F = 12.30, 1 and 36 df, P < 0.001) nor highshore L. unifasciata (F = 3.32, 2 and 51 df,
P < 0.05). The relationship was steeper on
Shore 3 (sheltered) than on Shore 2 in midshore snails and steeper on Shores 1 and 3
(exposed and sheltered, respectively) than on
Shore 2 (exposed) in highshore snails. Nodilittorina pyramidalis (b = 0.010, F = 2.79, 2 and
54 df, P > 0.05) had a common slope with
RESULTS
significant differences among intercepts
(F = 12.92, 2 and 56 df, P < 0.001). NodilittoRelationships between reserves of water and the rina pyramidalis from Shore 1 (an exposed
size, shape and relative weight of the shell.
shore) had more water reserves than those
from Shores 2 and 3 (an exposed and sheltered
When compared against shell length on each
shore, respectively).
shore separately, reserves of water increased in
larger specimens of N. pyramidalis on Shore 3
In general, therefore, water reserves were
(a sheltered shore; b = 0.15, F = 11.19, 1 and
not clearly correlated with shell length, nor with
18 df, P < 0.001), but water reserves were not differences in shape from short, squat shells to
correlated with shell length in any other com- elongated shells with narrower apertures, even
parison. When data from all shores were though there were significant patterns on some
compared, slopes were heterogeneous in N. shores for one of the species. Snails with relapyramidialis (F = 4.05, 2 and 54 df, P < 0.05) tively heavier shells (and, therefore, relatively
and in midshore L. unifasciata (F = 4.30,1 and smaller bodies), however, had greater reserves
36 df, P < 0.05), indicating no general relation- of water on all shores, although the nature of
ships between water reserves and shell length. the relationship varied between species differIn each species, water reserves increased more ently from shore to shore and could not be
with increasing shell length on Shore 3. easily correlated with wave exposure.
Although a common slope could be calculated
for highshore L. unifasciata (F = 0.92, 2 and 51
df, P > 0.05), it was not significant, i.e. water Reserves of water of snails living on different
reserves did not change with shell length. shores
Therefore, in general, water reserves were not Small and large N. pyramidalis from highshore
correlated with shell length and varied incon- levels were compared across replicated exsistently among shores.
posed and sheltered shores. There were no
When compared against shell shape (shell significant effects of size, shore nor exposure,
length:aperture length), reserves of water were nor any interactions of these variables, on water
relatively smaller in N. pyramidalis with elon- reserves (Table 2a). Although transformation
gated shells and smaller apertures (Le. they could not stabilize the variances (Table 2a), the
analysis was interpreted because analyses of
decreased as the shell length:aperture length
increased) on Shore 3 (a sheltered shore; variance are robust to heterogeneous variances
b = -0.20, F = 12.32, 1 and 18 df, P < 0.001). if there are many independent estimates of
The opposite trend was found for highshore variance within treatments (Box, 1953; UnderL. unifasciata on Shore 4 (an exposed shore; wood, 1981b). Pooling the Residual, the Size x
b = 0.29, F = 5.%, 1 and 17 df, P < 0.05). Shore (Exposure) interaction and Shores
There were no relationships between water (Exposure) (all of which had probability levels
reserves and shell shape on any other shore. P > 0.25) provided more powerful tests of the
Averaged across all shores, the common slope effects of Size and Exposure and their interwas non-significant and water reserves were action. Size x Exposure was significant in this
not correlated with shell shape in N. pyrami- analysis (F = 5.70, 1 and 156 df, P < 0.05).
dalis, nor in midshore or highshore L. unifasci- Averaged over replicate shores within each
518
M.G. CHAPMAN
a
O
O
0.8
0.6
0.4
—i
15
10
20
A A
0.8-
A
• A A A
A
0.6-
|
0.4
—I
—I—
10
15
1-,
20
c
•
0.8•
•
•
0.6it
A
0.4-
—i
A"
*
r * •
A
A
jt
A
A
A
5
10
15
Relative shell weight
20
Figure 1. Relationships between water reserves (as a proportion of wet body weight) and relative shell weight
of (a) N. pyramidalis, (b) highshore L. unifasciata and (c) midshore L. unifasciata on different intertidal shores
in New South Wales; O Shore 1; • Shore 2; A Shore 3; A Shore 4.
level of exposure, small N. pyramidalis had
significantly smaller reserves of water than did
large animals on the exposed shores, but there
were no differences between large and small
animals on sheltered shores (shown by S.N.K.
tests on the means; Fig. 2a). Nevertheless, there
was considerable, albeit non-significant, variation between the replicate shores (Fig. 2a).
Neither large nor small specimens showed
significant differences in water reserves among
shores, either between replicate shores of the
same level of exposure nor among shore of
differing exposure.
The effect of height on water reserves was
examined for small L. unifasciata which were
the only snails found at each level on the shores
(Table 2b). Water reserves differed between
mid- and highshore specimens interactively
among replicate shores (Table 2b). Midshore
animals had less water than highshore animals
SHELL SHAPE IN LITTORINIDS
0.85
519
a
• o
0.75
0.65
0.55
W
CO
0.85
CO
CD
8 0.75
2
0.65
0.55
0.85
0.75
0.65
0.55
Shore 1
Shore 2
Shore 3
Shore 4
Figure 2. Mean water reserves (S.E.) of large ( • ) and small (O) specimens of (a) N. pyramidalis, (b) highshore L. unifasciata and (c) midshore L. unifasciala on two wave exposed (Shores 1 and 2) and two sheltered
(Shores 3 and 6) shores,; n = 20.
on Shore 3, but there were no differences on
each of the other shores (compare small
specimens in Figs 2b and c).
Finally, interspecific comparisons were made
for large and small highshore N. pyramidalis
and L. unifasciata from the three shores on
which large and small specimens of each
species were found (Le. Shores 1, 3 and 6;
shores treated as a random factor). S.N.K. tests
on the significant Species X Size X Shore inter-
action (Table 2c) showed that the water
reserves of N. pyramidalis did not differ among
shores, but large and small L. unifasciata from
Shore 1 had significantly smaller reserves of
water than those from the other shores (Fig.
2a,b). Larger L. unifasciata on Shore 3 and
larger N. pyramidalis on Shore 1 had larger
water reserves than small animals. Otherwise
there were no significant differences between
large and small individuals. Interspecific com-
M.G. CHAPMAN
520
Table 2. Analyses of water reserves in small (*
150 mg wet weight) and large (400-700 mg wet
weight) littorinids from sheltered and exposed
shores; n = 20; ns = P > 0.05; • = P < 0.05; • • =
P < 0 . 0 1 ; » " = P< 0.001.
(a) Large and small N. pyamidalis; Cochran's
test, C= 0.26, P<0.05
Source
Size = Sz
Exposure = E
Shores = Sh(E)
SzXE
SzXSh(E)
Residual
MS
0 061
0.003
0.003
0.015
0.003
0.003
df
1
1
2
1
2
152
19.96 ns
1.00 ns
0.95 ns
5.02 ns
1.10 ns
(b) Highshore and midshore small L unifasciata;
Cochran's test, C = 0.19, P> 0.05
Source
MS
Height = H
Exposure = E
Shores = Sh(E)
HX E
HXSh(E)
Residual
0.05
0.02
0.11
0.04
0.02
0.004
df
1
1
2
1
2
152
1.22 ns
0.17 ns
25.95 • • *
1.89 ns
5.02 • •
(c) Large and small highshore L. unifasciata and
N. pyramidalis; Cochran's test, C = 0.16, P> 0.05
Source
MS
Size = Sz
Species = Sp
Shores = Sh'
Sh XSp
ShXSz
SpXSz
Sh X S p X S z
Residual
0.10
0.05
0.02
0.03
0.02
0.004
0.01
0.03
df
1
1
2
2
2
1
2
228
6.85 ns
1.65 ns
6.68"
9.26***
4.69*
0.33 ns
4.13*
parisons showed that N. pyramidalis either
had larger reserves of water than L. unifasciata
(for large specimens on Shore 1 and small
specimens on Shores 1 and 3) or there were no
interspecific differences in the mean reserves of
water.
Relationship between the shape of the shell and
the size of the foot
Relationships between the area of the aperture
and area of the extended foot were compared
between an exposed and a sheltered shore for
each species separately using analyses of co-
variance (with shell length as the covariate)
after transformation of the areas to natural
logarithms. Variances were homogeneous in all
analyses (P > 0.05; Cochran's test). Littorina
unifasciata showed a similar trend on each
shore for the relationship between shell length
and area of the aperture and between shell
length and area of the foot (F = 1.16, 1 and
35 df, P > 0.05; F = 0.17,1 and 35 df, P > 0.05,
respectively). Snails from Shore 1 (the exposed
shore) had relatively larger apertures and
relatively larger feet than those from Shore 3
(F = 9.11,1 and 36 df, P < 0.05; F = 4.41,1 and
36 df, P > 0.05, respectively; Fig. 3a, b).
Nodilittorina pyramidalis showed different patterns from shore to shore. With respect to area
of the aperture, slopes differed significantly
between shores (F = 9.60, 1, 35 df, P < 0.01),
with a greater increase in area of aperture with
increasing size on the exposed compared to the
sheltered shore (Fig. 3c). The area of the foot,
however, showed a similar trend on each shore
(F = 0.06,1,35 df, P > 0.05), although snails on
the exposed shore had relatively larger feet
than those on the sheltered shore (F = 16.05,1,
36 df, P < 0.001; Fig. 3d). Therefore, in general,
L. unifasciata and N. pyramidalis from the
exposed shore had relatively larger apertures
and larger feet than those from the sheltered
shore.
Relationship between shell morphology and
water reserves or survival on a sheltered shore
There was little emigration or loss of animals in
any treatment during the three month experimental period and 85-95% of marked animals
were relocated each time of sampling. After
three months, there was no significant difference in proportion of snails recovered for
transplanted snails and residents for either L.
unifasciata (x2 = 2.33,2 df, P > 0.05; 90%, 80%
and 88% of undisturbed, disturbed and transplanted snails recovered, respectively) or N.
pyramidalis (x2 = 0.37, 2 df, P > 0.05; 84%,
80% and 84% of undisturbed, disturbed and
transplanted snails recovered, respectively).
The specimens of N. pyramidalis were larger
than L. unifasciata (Cochran's C = 0.31, P >
0.05; F = 45.83, 1 and 36 df, P < 0.001; mean
shell-length (S.E.) = 14.1 mm (0.05) and 10.5
mm (0.03), respectively), but this was similar
for morphs from each shore. Shell shape, however, differed between morphs, not species
(Cochran's C = 0.31, P > 0.05; F = 8.%, 1 and
36 df, P < 0.01; mean shell shape (S.E.) = 2.09
(0.03) and 1.96 (0.03) for the morphs from the
SHELL SHAPE IN LITTORINIDS
4.5
521
b
a
O)
o
£3.5
3.5
o
o
3
c
a.
o
o
2 2.5
2.5
1.5
1.5
13
15
17
45
13
15
17
13
15
17
4.5
d!
o
"jT 3.5
o 3.5
,ouo
r
°O
CD
Q.
a
o 2.5
1.5
1.5
9
11
13
Shell length (mm)
15
17
9
11
Shell length (mm)
Figure 3. Area of aperture (transformed to natural logarithms) and area of foot (transformed to natural
logarithms) of L. unifasciala (a and b) and N. pyramidalis (c and d) from an exposed shore (•; solid line) and a
sheltered shore (O; dashed line); n = 19.
sheltered and wave exposed shore, respectively). Water reserves after a 10-day period of
continuous emersion did not, however, differ
between morphs, but did differ between species
(Cochran's C = 0.32, P > 0.05; F = 37.12,1 and
36 df, P < 0.001; mean (S.E.) = 0.78 (0.01) and
0.71 (0.01) for N. pyramidalis and L. unifasciata,
respectively.
All L. unifasciala and N. pyramidalis
survived 1 week in the field with continuous
emersion, irrespective of whether they were
fully hydrated at the start of the experiment or
had expelled extra-corporeal water. Austrocochlea porcata, B. nanum and Nerita atramentosa suffered 100%, 100% and 50%
mortality, respectively, during the same period.
There was no significant effect of treatment
nor significant interaction between treatment
and site (P values > 0.05) in the water reserves
of mid- and highshore morphs of L. unifasciala
(mean water reserves (S.E.) = 0.56 (0.02), 0.53
(0.03), 0.58 (0.03), 0.56 (0.02) and 0.53 (0.04)
for undisturbed, disturbed, translocated, hy-
drated highshore morphs and hydrated midshore morphs, respectively). The values are
small compared to recently-hydrated snails (cf.
Figs 1 and 2), suggesting that there was loss of
water, but it was consistent across treatments.
All L. unifasciata and JV. pyramidalis survived 2 months of emersion in the laboratory,
irrespective of treatment.
Comparative survival of different morphs on a
wave exposed and sheltered shore
The shape of snails differed significantly between Snores 3 and 4 as predicted, (F = 65.97,
1 and 6 df, P < 0.001) and there were no significant differences among sites on each shore
(F = 1.33, 6 and 72 df, P > 0.05—sites were a
random factor nested within shores). The snails
on Shore 3 were more elongated with smaller
apertures than those from the more exposed
shore. The mean (S.E.) ratio of shell length to
aperture length (S.E.) varied between 2.15
(0.05) and 2.21 (0.03) among sites on Shor^e 3
522
M.G. CHAPMAN
10
Days since start
Figure 4. Mean (S.E.) numbers of L. unifasciata recovered on (a) a wave-exposed and (b) a sheltered shore;
• undisturbed snails, A disturbed snails; • translocated snails; O transplanted snails; n = 4 sites per shore.
and between 1.85 (0.04) and 1.91 (0.05) among
sites on Shore 4.
In the reciprocal transplantation of L. unifasciata among four sites on a wave exposed and
sheltered shore, there was a gradual decline in
he recovery of snails in all treatments (Fig. 4).
After fifteen days, the numbers of relocated
snails differed among treatments (F = 5.01,
3 and 24 df, P < 0.01) and among shores
(F = 7.40, 1 and 24 df, P < 0.05). On each
shore, fewer translocated and transplanted
snails were recovered than undisturbed or disturbed resident snails (although these differences were greater on the wave-exposed than
on the sheltered shore; Fig. 4). Over all treatments, significantly fewer snails were recovered
from the wave-exposed shore (mean no. recovered (S.E.) = 14.8 (1.8)) than from the
sheltered shore (mean (S.E.) 19.6 (1.1)).
In the first experiment to distinguish
between the alternative models that there was
either greater dislodgment or greater rates of
dispersal of translocated and/or transplanted
snails on the wave exposed shore compared to
resident snails, a larger number of undisturbed
snails (mean (S.E.) = 19.5 (0.05)) were recovered after 2 days than all other treatments
(14.0 (1.0), 12.0 (1.0) and 14.5 (1.5) for disturbed, translocated and transplanted snails,
respectively; F = 9.04, 3 and 4 df, P < 0.05).
After four days, similar numbers of all treatments were found (F = 5.40, 3 and 4 df, P >
0.05) indicating no significant differential loss
among treatments, although fewer translocated
(mean (S.E.) 6.0 (2.0)) and transplanted (10.0
(3.0)) snails were recovered than undisturbed
(15.0 (0.0) ) and disturbed (14.5 (0.5)) snails.
The large standard errors for the translocated
and transplanted snails show large variation in
recovery between sites.
In the first experiment, distances displaced
after two days differed interactively among
sites and treatments (Cochran's C = 0.25, P >
0.05; F = 4.09,3 and 80 df, P < 0.01 after transformation of the data to natural logarithms).
After two days, there were no significant differences among treatments in Site 1, although
undisturbed and transplanted snails moved further than the other treatments (Fig. 5a). In Site
2, translocated and transplanted snails moved
further than residents (Fig. 5b). The data
collected after four days were not analysed
because of the large loss of translocated snails
in Site 1 (n = 4), but transplanted snails moved
SHELL SHAPE IN LITTORINIDS
523
80
a
60
1
i
40
1
i
•t
20
o
0
180
f 140
E 100
«j
to
=6 60
ui
CO
20
T
0
• o
Jj 100
80
60
40
20
0
D
TL
TP
Treatments
Figure 5. Mean distance (S.E.) displaced by snails in three sites on Shore 4 at Cape Banks; U - undisturbed
residents, D - disturbed residents, TL - translocated snails from other sites on Shore 4, TP - transplanted
snails from Shore 3; (a) Experiment 1, Site 1, filled and empty symbols are after 2 and 4 days, respectively,
(b) Experiment 1, Site 2, filled and empty symbols are after 2 and 4 days, respectively, (c) Experiment 2,
Site 3,filledand empty symbols are after 1 and 5 days, respectively.
further than residents in each site (Fig. 5a
and b).
In the second experiment, data were only
available from one site because snails in the
second site did not move during the experimental period. After the first day, transplanted
snails dispersed further than translocated snails
which, in turn moved further than residents
(Cochran's C = 0.36, P > 0.05; F = 24.2, 3 and
64 df, P < 0.01; Fig. 5c). In contrast, after five
days, the distances moved by translocated
and transplanted snails were no longer significantly different, although greater than those
moved by residents (Cochran's C = 0.49, P <
0.05; F = 5.62,3 and 56 df, P < 0.01; Fig. 5c).
Therefore, there was no evidence of greater
loss or rates of dispersal on wave-exposed
shores of morphs from sheltered shores compared to morphs from wave-exposed shores
that had been similarly disturbed by being
placed in an unfamiliar site.
Experiment to test the effects of type or quantity
offood or opportunity to feed on growth and
shell shape of L. unifasciata
At the start of the first experiment, juvenile
snails had a mean shell length of 1.60 mm (S.E.
524
M.G. CHAPMAN
0.03) and an average shell length:aperture
Shell length of juveniles was only signifilength ratio of 1.71 mm (S.E. 0.03). After two cantly affected by the height on the shore at
months, there was large loss and/or mortality of which they were confined (F = 27.74,1 and 16
juvenile snails, particularly from the lower df, P < 0.01; Cochran's C = 0.85, P < 0.01 varicores. Survival on the upper cores was also ances could not be stabilized by transformaextremely patchy, with complete loss on many tion; Fig. 6b). There was very little growth high
cores and almost 100% survival on others. Pro- on the shore, whereas juveniles low on the
portional survival of the juvenile snails was shore more than doubled shell length during
compared among all treatments. There was no the experiment period. Mean shell length of
effect of origin of the core, nor density of adults adults was not measured at the start of the
nor any interaction on survival of juveniles. experiment because adults were only included
There was, however, greater survival at high- in the experiment to alter quantity of food.
shore levels (mean proportional survival (S.E.) Nevertheless, at the end of the experiment,
0.57 (0.05) than at midshore levels (mean mean shell length of adults was similar across
(S.E.) 0.09 (0.02); Cochran's C = 0.46, P < all treatments, indicating either no growth or
0.01; F = 65.3,1 and 40 df, P < 0.001). Survival similar rates of growth under all experimental
of adult snails was almost 100%.
conditions.
Because of patchy survival, shell shape and
Shell shape of juveniles was also only
shell-length of juveniles were only compared affected by height on the shore (F = 5.26,1 and
using snails from the higher cores to test hypo- 16 df, P < 0.05; Cochran's C = 0.26, P > 0.05).
theses about the effects of density of adults All juveniles were more elongate with smaller
(quantity of food) and origin of core (type of apertures than at the start of the experiment
food) on growth and shell shape. The data from
(Le. had a larger SL:AL ratio), but this ratio
the replicate cores were combined to provide a was greater at the highshore level than lower
balanced sample size of 29. There were no sig- on the shore (i.e. the highshore juveniles had
nificant effects of either variable on shell length developed longer spires and/or smaller aper(P > 0.05 for all F-ratios; mean shell length tures than midshore juveniles; Fig. 6c). Large
1.58 mm, S.E. 0.02) and no evidence that the snails were more globose than juveniles snails
snails grew in length at all during the two- and their shell shape was similar across treatmonth experimental period. Average shell ments.
shape differed, however, according to the type
of core on which the snails had been feeding
(F = 4.89, 1 and 112 df, P < 0.5; Cochran's
DISCUSSION
C = 0.31, P > 0.05). Snails feeding on cores
collected from highshore areas were less Patterns of shell morphology from shore to
elongate (mean SL:AL 1.65, S.E. 0.02) than shore in these two species of highshore littothose that had been feeding on cores collected rinids have been shown to be complex and varifrom midshore areas (mean 1.72, S.E. 0.02) and able (Chapman, 1995) and, in this study, spatial
less elongate than a random set of animals at patterns in the reserves of water were equally
the start of the experiment.
complex. There were no general patterns of
At the start of the second experiment, juvenile correlation of water reserves (tissue water and
snails had a mean shell length of 1.43 mm (S.E. extra-corporeal water) with differences in the
0.02) and an average shell length:aperture size or shape of the shell. Therefore, there was
no general tendency for larger shells (i.e. charlength ratio of 1.41 (S.E. 0.01). Again, loss of
snails, particularly juveniles, from highshore acteristic of highshore levels) to contain more
levels was great, despite the use of tea strainers (or less) water than smaller shells, or for elonto confine the snails. Proportional survival varied gated shells with narrow apertures (characterinteractively according to height on the shore, istic of highshore levels and some sheltered
density of adults and size of snails (Cochran's shores) to contain more (or less) water than
C = 0.16, P > 0.05; F = 7.31, 1 and 80 df, squatter shells with wider apertures. Nevertheless, significant trends were found for one or
P < 0.01; Fig. 6a), but there was no effect of
origin of the core on survival. Juveniles were other species on some shores. These trends
lost from the treatments at a greater rate than differed between species even though the two
adults and there was greater survival of juveniles species were sampled from the same parts
lower than higher on the shore (Fig. 6a). The of the shore. Therefore, differences between
effect of adult density on survival of adults or species are not simply due to the two species
being collected from different sites. For exjuveniles was minimal.
SHELL SHAPE IN LITTORINIDS
525
a
1
0.8
CO
c
p
0.6
I 0.4
0.2
2
D.
0
B
c
-
4-
UJ
2
c
CO
CD
CD
&
0)
"5
.c
I
U5
LJJ
W
1.5-
.
c
co
CD
LD
SD
Highshore
LD
SD
Midshore
Figure 6. (a) Mean (S.E.) proportional survival, (b) shell-length and (c) relative shell shape of • adult and
O juvenile snails after six months confined to cores at highshore or midshore levels; LD = large density of
adults (n = 30); SD = small density of adults (n = 2).
ample, elongated N. pyramidalis had smaller
reserves of water than squatter specimens on
three shores. The opposite trend was found for
highshore L. unifasciata on one exposed shore
(Shore 4), but there were no trends on the
other shores.
There were also no general differences in
water reserves between specimens from replicated wave exposed or sheltered shores. For
example, water reserves were greater in large
than in small N. pyramidalis on exposed shores,
but not on sheltered shores. There were, however, no differences among the shores for
specimens of either size class. Highshore L.
unifasciata (of each size) from a wave-exposed
shore had more water reserves than those from
both sheltered shores, but large specimens had
more water reserves than small ones on one
sheltered shore. Again, these intra- and interspecific differences were found despite the fact
that the small and large specimens of each
species were collected from the same sites.
Therefore, the factors that influence water
reserves are complex and variable among sizes
of specimens, species and shores.
As a large proportion of the reserves of
water is carried as free (extra-corporeal) water
in the shell, this component may be related to
the shape or internal volume of the shell, but
may also be influenced by the behaviour of the
526
M.G. CHAPMAN
animals. Snails that naturally withdraw deeper
into the shell will retain less water than
those that do not. This behaviour cannot be
standardised among individuals although all
animals were handled similarly to reduce any
effects of experimental disturbances on their
behaviour. Nevertheless, it is likely that much
of the variability in measures of water reserves
within and between species may be due to differences in behaviour. In addition, all individuals could not be tested on the same day because
of the time needed to measure water reserves
and the need to ensure that animals were not
maintained in the laboratory for extended
periods, in case that affected body weight or
behaviour. Specimens from different shores
were tested on the same days to ensure that
differences from shore to shore were not confounded with differences among days. This
meant that there may have been increased
variability within samples (i.e. specimens from
the same shore) because behaviour may have
varied from day to day in response to uncontrollable environmental factors.
Water reserves were, however, clearly correlated with the relative weight of the shell for
each species on each shore. The prediction was
that relatively lighter shells would contain
more water because they may be thinner and
therefore have greater internal volume. In contrast, reserves of water increased with relative
shell weight. As shell weight was estimated
relative to body weight, relative shell weight
can increase because of an increase in shellthickness and/or because of a decrease in size
of the body. The latter may provide more room
inside the shell for extra-corporeal water,
thereby increasing water reserves. This study
suggests that the amount of water held within a
shell may be primarily determined by the relative size of the body, with a smaller body for
the same weight of shell providing more internal volume. Shell weight was not, however,
correlated with height on the shore nor wave
exposure and varied considerably inter- and
intraspecifically within and among shores
(Chapman, 1995). This relationship did not
therefore lead to consistent and significant
differences in water reserves between exposed
and sheltered shores.
Neither did shell morphology appear to influence rate of loss of water, with morphs from
a sheltered and from an exposed shore having
similar reserves of water and surviving equally
after a prolonged period of emersion. It has
been suggested that loss of water during emersion is extremely limited in littorinids because
snails tend to withdraw deep inside the shell,
attach to the rock-surface with a ring of mucus
and aestivate until again submersed (Vermeij,
1973; McQuaid & Scherman, 1988; McMahon,
1990). Littorina unifasciata and N. pyramidalis
showed similar resistance to emersion, unlike
other species of gastropods that co-exist with
L. unifasciata at midshore levels. All L. unifasciata survived prolonged emersion in hot
weather, irrespective of the level of the shore
from which they were obtained. Fifty to one
hundred percent of specimens of other midshore snails died during the same period of
emersion. There was,also no significant difference in survival between fully-hydrated L.
unifasciata and N. pyramidalis and those that
had been forced to expel free water, again indicating that water reserves are not important for
survival of highshore littorinids (McQuaid &
Scherman, 1988; McMahon, 1990). Similarly,
shell morphology did not influence survival on
a sheltered shore.
Morphs of L. unifasciata and N. pyramidalis
from highshore levels on a wave exposed shore
had larger apertures and feet than did highshore morphs from a sheltered shore. This
relationship was, however, only examined for
two shores and it is not therefore possible to
conclude that this is a general pattern, although
results do support the general model that
increased size of aperture in snails living on
exposed shores relates to increased size of foot,
increased tenacity and reduced risk of dislodgment (Grahame & Mill, 1986; Etter, 1988; Frid
& Fordham, 1994). Nevertheless, transplanted
morphs from a sheltered shore with relatively
smaller apertures (therefore with smaller feet)
were not lost from sites on exposed shores at a
greater rate than were snails translocated into
those sites from other sites on the exposed
shore, even though loss of all snails was greater
than from sites on a sheltered shore. Because
translocated and transplanted snails were lost
from all sites at a similar rate, loss appears to
be a function of the familiarity of the snails
with the experimental sites, rather than a function of the morphology of the snails or the
shore on which they had been living. Residents,
which were familiar with the surroundings in
each experimental site, were lost at a smaller
rate (irrespective of whether they were disturbed or not) than introduced snails, irrespective of the shore from which these came. There
is increasing evidence that intertidal molluscs
respond to the familiarity of their normal surrounds and behave abnormally when transplanted into unfamiliar sites, even if such sites
SHELL SHAPE IN LITTORINIDS
are selected to represent similar conditions
(Chapman, 1986; Underwood, 1988; Chapman
& Underwood, 1992).
Although some loss may have been due to
dislodgment by waves, there was increased
emigration of transplanted and translocated
snails from the experimental areas during the
first few days after moving the snails. Translocation and transplantation did .not increase dispersal on sheltered shores to the same extent as
they did on exposed shores, so it is possible
that wave action directly influenced rates of
movement of disturbed snails. Those treatments that moved further were also recovered
in smaller numbers. Therefore, although it is
possible that these snails were lost by dislodgment at a greater rate than were the resident
snails, it is equally likely that increased dispersal moved them rapidly out of the study site
and increased 'loss' was due to the fact that the
snails were not found (despite extensive
searches).
Field experiments to compare the effects of
type of food (using cores from different heights
on the shore), quantity of food (using different
densities of adults to modify this) and opportunity to feed (comparing midshore to highshore
levels) on rates of growth and subsequent shell
shape were inconclusive, but did indicate the
usefulness of field experiments in investigating
factors of this sort. First, despite attempts to
confine the juvenile snails, many were lost from
all treatments. Second, because of continued
problems with vandalism of experiments on
shores around Sydney, this experiment was only
set up on one shore—a very sheltered shore.
Therefore, the results cannot be extended to
explain differences in shell shape from shore
to shore. In the first experiment, juvenile L.
unifasciata survived better at highshore levels,
probably because waves washed them off the
midshore cores. They did not, however, grow
during the experimental period and therefore
differences in shape among the remaining survivors are likely to reflect different survival,
rather than changes in shape in response to
growth. Alternatively, they may be coincidental
because it is difficult to think of an ecological
model that explains increased survival of snails
with relatively larger apertures when feeding
on cores from high on the shore that also
explains a decrease in relative aperture size
higher on the shore.
The juvenile snails in the second experiment
survived and grew faster at mid- than at highshore levels and these were not significantly
affected by the density of adults or origin of the
527
core, suggesting that opportunity to feed may
be more important than type or quantity of
food in survival and growth of L. unifasciata.
All juveniles developed relatively longer spires
and smaller apertures than at the start, which is
characteristic of snails on many sheltered shores.
This is not surprising because these experiments were done on a very sheltered shore.
Highshore snails grew at a slower rate and
were relatively more elongate than snails at
midshore levels. Because all juveniles had been
collected from the same small area of the shore
at the same time (over a few square metres)
and were allocated randomly to treatments,
these differences are unlikely to indicate genetic
differences among treatments. Shell shape of
L. unifasciata appears to be directly caused by
differences in rates of growth, in direct response to opportunity to feed. Whether this
model can explain differences in shell shape
between species, among different sites on a
shore and among different shores as has been
documented (Chapman, 1995) has not been
experimentally investigated. Nevertheless, these
experiments indicate a useful methodology of
attempting to unravel potential causes for
spatial variation in shell shape of these species.
They also suggest that such factors may be
more important than physical factors associated with tenacity, water reserves or loss of
water for this species.
ACKNOWLEDGEMENTS
This research was assisted by grants from the University of Sydney and the Institute of Marine Ecology. I
thank Karen Astles, Graham Housfield and Tony
Underwood for considerable assistance with the field
experiment to examine rates of growth on shell
shape in L. unifasciata. Jillian Grayson and Danielle
O'Connor prepared the figures. Tony Underwood
provided advice and encouragement throughout this
study. Tony Underwood, David Reid and an anonymous reviewer offered useful comments on an earlier
draft of this manuscript.
REFERENCES
BASINGTHWAIGHTE, G. & FOULDS, W. 1985.
The
effect of wave action on the shell morphology of
Littorina unifasciata Gray. Journal of the Royal
Society of Western Australia, 68: 9-12.
BOULDING, E.G. & HAY, T.K. 1993. Quantitative
genetics of shell form of an intertidal snail:
constraints on short-term response to selection.
Evolution, 47: 576-592.
BOYLE, P.R., SILLAR, M. & BYCESON, K. 1979. Water
balance and mantle cavity fluid of Nucella lapillus
528
M.G. CHAPMAN
(L.) (Mollusca: Prosobranchia). Journal of Experimental Marine Biology and Ecology, 40: 41-51.
Box, G.E.P. 1953. Non-normality tests on variances.
Biometrika, 40: 318-335.
BRANCH, G.M. & BRANCH, M.L. 1981. Experimental
approach. Journal of Experimental Marine Biology
and Ecology, 54: 277-308.
HELLER, J. 1976. The effect of exposure and predation on the shell of two British winkles. Journal of
Zoology, 179:201-213.
analysis of intraspecific competition in an intertidal
HUGHES, R.N. & ELNER, R.W. 1979. Tactics of a
gastropod, Littorina unifasciata. Australian Journal predator, Carcinus maenas, and morphological
of Marine and Freshwater Research, 32: 573-589.
responses of the prey, Nucalla lapillus. Journal of
Animal Ecology, 48: 65-78.
BRANCH, G.M. & MARSH, A.C. 1978. Tenacity and
shell shape in six Patella species: adaptive features. JOHANNESSON, B. 1986. Shell morphology of LittoJournal of Experimental Marine Biology and
rina saxatilis Olivi: the relative importance of physEcology, 34:111-130.
ical factors and predation. Journal of Experimental
Marine Biology and Ecology, 102:183-196.
BRITTON, J.C. 1995. The relationship between
position on shore and shell ornamentation in two JOHANNESSON, B. & JOHANNESSON, K. 1990. Littorina
size-dependent morphotypes of Littorina striata,
neglecta Bean, a morphological form within the
with an estimate of evaporative water loss in
variable species Littorina saxatilis (Olivi)? Hydrothese morphotypes and in Melarhaphe neritoides.
biologia, 193: 71-87.
Developments in Hydrobiology, 111: 129-142.
JOHANNESSON, K., JOHANNESSON, B. & ROLANALVAREZ, E. 1993. Morphological differentiation
CHAPMAN, M.G. 1986. Assessment of some controls
and genetic cohesiveness over a microenvironin experimental transplants of intertidal gastromental gradient in the marine snail Littorina
pods. Journal of Experimental Marine Biology and
saxatilis. Evolution, 47: 1770-1787.
Ecology, 103: 181-201.
CHAPMAN, M.G. 1994a. Small- and broad-scale
KEMP, P. & BERTNESS, M.D. 1984. Snail shape and
patterns of distribution of the upper-shore littorgrowth rates: evidence for plastic shell allometry in
inid, Nodilittorina pyramidalis in New South
Littorina littorea. Proceedings of the National
Wales. Australian Journal of Ecology, 19: 83-95.
Academy of Sciences, 81: 811-813.
CHAPMAN, M.G. 1994b. Small-scale patterns of distriKITCHING, J.A. 1976. Distribution and changes in shell
bution and size-structure of the intertidal littorinid,
form of Thais spp. (Gastropoda) near Baumfield,
Littorina unifasciata (Gastropoda: Littorinidae) in
U.K. Journal of Experimental Marine Biology and
New South Wales. Australian Journal of Marine
Ecology, 23: 109-126.
and Freshwater Research, 45: 635-642.
KITCHING, J.A. & LOCKWOOD, J. 1974. Observations
CHAPMAN, M.G. 1995. Spatial patterns of shell shape
on shell form and its ecological significance in
of three species of co-existing littorinid snails in
thaisid gastropods of the genus Lepsiella in New
New South Wales, Australia. Journal of Molluscan
Zealand. Marine Biology, 28: 131-144.
Studies, 61: 141-162.
KITCHING, J., MUNTZ, L. & EBLING, FJ. 1966. The
CHAPMAN, M.G. & UNDERWOOD, A.J. 1992. Experiecology of Lough Ine XV. The ecological signifimental designs for analyses of movements by molcance of shell and body forms in Nucella. Journal
luscs. In: Proceedings of the Third International
of Animal Ecology, 35: 113-126.
Symposium on Littorinid Biology (J. Grahame,
MACLULICH, J.H. 1987. Variation in the density and
P.J. Mill & D.G. Reid, eds), 169-180. London, The
variety of intertidal epilithic microflora. Marine
Malacological Society of London.
Ecology Progress Series, 40: 285-293.
CROTHERS, J.H. 1984. Some observations on shell
MCMAHON, R.F. 1990. Thermal tolerance, evaporashape variation in Pacific Nucella. Biological
tive water loss, air-water oxygen consumption and
Journal of the Linnean Society, 21: 259-281.
zonation of intertidal prosobranchs: a new
synthesis. Hydrobiologia, 193: 241-260.
ETTER, RJ. 1988. Asymmetrical developmental plasticity in an intertidal snail. Evolution, 42: 322-334.
MCQUAID, C D . & SCHERMAN, P.A. 1988. Thermal
FRID, C.LJ. & FORDHAM, E. 1994. The morphology
stress in a highshore intertidal environment:
of sub-littoral gastropod Gibbula cineraria (L.)
morphological and behavioural adaptations of the
along a gradient of wave-exposure. Ophelia, 40:
gastropod Littorina afneana. In: Behavioural
135-146.
Adaptation to Intertidal Life. Nato ASI Series A,
Life Sciences, Volume 151 (G. Chelazzi & M.
GARRITY, S.D. & LEVINGS, S.C. 1984. Aggregation in
Vannini, eds), 213-224. New York, Planum Press.
a tropical neritid. Veliger, 27: 1-6.
GELLER, J.B. 1990. Consequences of a morphological
NEWKIRK, G.F. & DOYLE, R.W. 1975. Genetic analysis
defense: growth, repair and reproduction by thinof shell shape variation of Littorina saxatilis on an
and thick-shelled morphs of Nucella emarginata
environmental cline. Marine Biology, 30: 227-237.
(Deshayes) (Gastropoda: Prosobranchia). Journal PALMER, A.R. 1990. Effect of crab effluent and scent
of Experimental Marine Biology and Ecology, 144:
of damaged conspecifics on feeding, growth, and
173-184.
shell morphology of the Atlantic dogwhelk Nucella
lapillus (L.). Hydrobiologia, 193: 155-182.
GRAHAME, J. & MILL, PJ. 1986. Relative size of the
foot of two species of Littorina on a rocky shore in RAFFAELLI, D.G. 1978. The relationship between
Wales. Journal of Zoology, 208: 229-236.
shell injuries, shell thickness and habitat characteristics of the intertidal snail Littorina rudis Maton.
GRENON, J.F. & WALKER, G. 1981. The tenacity of
Journal of Molluscan Studies, 44: 166-170.
the limpet. Patella vulgata L., an experimental
SHELL SHAPE IN LITTORINIDS
SEAPY, R.R. & HOPE, WJ. 1973. Morphological and
behavioural adaptations to desiccation in the intertidal limpet Acmaea (Collisella). Veliger, 16: 181188.
SEED, R. 1973. Absolute and allometric growth in
the mussel Mytilus edulis L. (Mollusca: Bivalvia).
Proceedings of the Malacological Society of
London, 40:343-357.
SEED, R. 1978. Observations on the adaptive significance of shell shape and body form in the dogwhelks (Nucella lapillus (L.)) from N. Wales.
Nature in Wales, 16: 111-122.
UNDERWOOD, AJ. 1977. Movements of intertidal
gastropods. Journal of Experimental Marine
Biology and Ecology, 26: 191-201.
UNDERWOOD, A J. 1981a. Structure of a rocky intertidal community in New South Wales: patterns of
vertical distribution and seasonal change. Journal
of Experimental Marine Biology and Ecology, 51:
57-85.
UNDERWOOD, AJ. 1981b. Techniques of analysis of
variance in experimental marine biology and ecology. Annual Review of Oceanography and Marine
Biology, 19: 513-603.
UNDERWOOD, AJ. 1984. Microalgal food and the
529
growth of intertidal gastropods Nerita atramentosa
Reeve and Bembicium nanum (Lamarck) at four
heights on the shore. Journal of Experimental
Marine Biology and Ecology, 79: 277-91.
UNDERWOOD, A.J. 1988. Design and analysis of
field experiments on competitive interactions
affecting behaviour of intertidal animals. In:
Behavioural Adaptation to Intertidal Life. Nato
ASI Series A, Life Sciences, Volume 151 (G.
Chelazzi & M. Vannini, eds), 333-357. New York,
Plenum Press.
VAN MARION, P. 1981. Intra-population variation of
the shell of Littorina rudis (Maton) (Mollusca:
Prosobranchia). Journal of Molluscan Studies, 47:
99-107.
VERMEU, G.J. 1973. Morphological patterns of high
intertidal gastropods: adaptive strategies and their
limitations. Marine Biology, 20: 319-346.
VERMEU, GJ. 1978. Biogeography and adaptation:
patterns of marine life. Cambridge, Massachusetts,
Harvard University Press.
WOLCOTT, T.G. 1973. Physiological ecology and
intertidal zonation in limpets (Acmaea): a critical
look at 'limiting factors'. Biological Bulletin, 145:
389-422.