Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
DYNAMICS OF AGGREGATIONS OF A GASTROPOD PREDATOR/SCAVENGER ON A NEW ZEALAND HARBOUR BEACH ALAN D. ANSELL* Scottish Association for Marine Science, PO Box 3, Oban PA34 4AD, Argyll, UK (Received 25 August 2000; accepted 25 January 2001) ABSTRACT Observations of a New Zealand harbour beach revealed that aggregative feeding behaviour was common in the buccinid gastropod Cominella glandiformis (Reeve, 1847). The majority of such aggregations are associated with either the tellinid bivalve Macomona liliana (Iredale, 1915) or the venerid cockle Austrovenus stutchburyi (Wood, 1828). Marking experiments indicate that aggregations persist for one tidal cycle. This information, together with the density of aggregations associated with each ‘prey’ species allows the calculation of rate of predation by Cominella glandiformis and of the mortality of M. liliana and A. stutchburyi at different stations on the beach. Averaged over spring periods, when predation and mortality are reduced, and summer, the calculated rates suggest that between 2 to 16% of M. liliana and 3 to 9% of A. stutchburyi fall prey to Cominella each year. The majority of bivalves at the centre of aggregations appeared to be tightly shut and undamaged, suggesting that the gastropods are predators attacking live prey, although late arrivals may be attracted as scavengers by leaking body fluids. Aggregated feeding may provide the gastropod with both competitive and metabolic benefits. INTRODUCTION The buccinid gastropod Cominella glandiformis (Reeve, 1847) is an abundant member of the fauna of sheltered harbour beaches in New Zealand, appearing either as solitary individuals or in aggregations of tens or sometimes hundreds. Morton & Miller (1968) described C. glandiformis as the universal scavenger of mudflats which ‘converge in scores on dead animal remains, and may also be active carnivores, thrusting the long proboscis through the narrow siphonal fissures of Chione and other bivalves’. The short term dynamics of C. glandiformis aggregations on the shore, and the nature of the relationships between C. glandiformis and the bivalves with which aggregations are most often associated, have not previously been studied. It is interesting to establish whether these aggregations are scavenging or preying on the animal around which they are all assembled. This paper attempts such a study on Cominella aggregations on a typical harbour beach in the North Island of New Zealand. MATERIAL AND METHODS Study location The study was carried out on the beach at Matua, within Tauranga Harbour, a large natural harbour inlet on the east * Sadly Alan Ansell died in 1999 before this paper was completed. The manuscript was completed by Drs M. Barnes and E.M. Harper and Mrs L. Robb. J. Moll. Stud. (2001), 67, 329–341 coast of the North Island of New Zealand (Fig. 1). At this location, protective gabions restrict the upper intertidal zone so that the area exposed at low tide lies between low water (of spring or neap tides) and a little below high water of neap tides. At its seaward edge, the intertidal region is bounded by the deeper water channel draining the Wairoa River. The topography of the shore shows slight seasonal changes depending on deposition and erosion from localized channels and sand banks. Observations were carried out over two periods: in summer, 4–13 February 1998; and in spring, 5 September–5 October 1998. At low water of spring tides during the summer period, the exposed intertidal area that was accessible comprised an outer sand bar bordering the channel, separated by a shallow trough from an inner sand bar which in turn was separated from a final shallow-sloping inshore area of muddier sand by a second trough. Much of this inner trough, which drained at low tide apart from some residual shallow pools, supported a relatively strong growth of eelgrass (Zostera sp.) and consequently the sediment was muddier and wetter than that of the bars on the upper intertidal area. The outer trough retained a shallow flow of water at low tide and had an abundant mix of shell debris with some attached growth of algal sporelings. The overall extent of the intertidal was about 250 m, of which the inshore flat and inner and outer bars each comprised some 70 m. This situation remained largely unchanged during the spring period apart from some deepening of the inner trough. The inshore flat and inner and outer bars remained as consistent fairly uniform areas. On each tide the inshore flat remained immersed for 7 h and the inner and outer bars for up to 9h. Tides in Tauranga Harbour are semi-diurnal with a vertical extent of about 1.9 m at spring tides. Details of physical conditions within the Harbour are given by Park & Donald (1994). © The Malacological Society of London 2001 ALAN D. ANSELL Figure 1. Location of Matua beach within Tauranga Harbour, New Zealand. The intertidal fauna was typical of North Island protected harbour beaches as described by Morton & Miller (1968) and other authors (e.g. Grange, 1977). The dominant large bivalves present were the tellinid Macomona liliana (Iredale, 1915) and the venerid cockle Austrovenus stutchburyi (Wood, 1828). The crabs Helice crassa Dana, 1851, Macrophthalmus hirtipes (Heller, 1862) and Hemigrapsus crenulatus Milne Edwards, 1837 were common in the muddier areas of the upper beach and troughs. The last was less common in the more loosely compacted and well sorted sand of the bars. The gastropod Zeacumantus lutulentus (Kiener, 1841) was abundant on the surface of the upper beach. At all levels, aggregations of Cominella glandiformis were a distinctive feature of the intertidal fauna, especially during the summer period, usually associated with intact or partially or completely consumed Macomona liliana or Austrovenus stutchburyi. Field methods The observations made aimed to determine the distribution of aggregations and solitary individuals of Cominella glandiformis and of the larger bivalves with which they are commonly associated, and to examine the nature of the relationship between them. Observations were made at three stations: 1) on the crest and nearshore slope of the outer sand bar, 2) on the crest and nearshore slope of the inner sand bar, and 3) on the inshore flat for each of the seasons (Table 1). The density of aggregations of C. glandiformis was determined by counting within replicate marked areas of between 25 and 200 m2 depending on the perceived density. At each station, the density of large Macomona liliana and Austrovenus stutchburyi, and of individual C. glandiformis was determined by sieving five replicate 0.5-m2 samples through a 0.95-cm mesh sieve. To avoid changes in distributions around this sampling influencing other results, quantitative benthic sampling was not carried out until other observations were completed in each period. The size distributions of Macomona liliana and Austrovenus stutchburyi present at each site were determined by measuring animals from the benthic samples. Representative groups were also weighed to provide biomass estimates. Individuals over the entire size range represented from the spring benthic sampling were used to provide relationships between ash-free dry weight of tissue (AFDW) and shell length (L) or height (H) for each species (Equations 1, 2, 3). 330 GASTROPOD AGGREGATIONS Table 1. Characteristics of the stations sampled at Matua beach, Tauranga Harbour. *Figures in parentheses show number of replicates. **Figures in parentheses show mean density where there was no significant difference between summer and spring numbers. CD = Chart Datum. Inshore flat Depth versus CD (m) Time immersed (h) Sea-water temp. (C) Summer Spring 7.0 23–25 0.75 7.0 13–14 Density (No.m–2) Cominella aggregates* 0.005 (4) Cominella individuals** 20[14] Macomona ** 178 Austrovenus ** 372 [320] For Macomona liliana: log AFDW 2.99 log L–2.27 For Austrovenus stutchburyi: log AFDW 2.62 log L–1.80 For Cominella glandiformis: log AFDW 2.89 log H–1.96 Inner bar Summer 7.0 0.035 (5) 8 54 268 0.156 (20) 22 [13] 164 [175] 248 [389] Equation 1 Equation 2 Equation 3 Aggregations were sampled at each station and the following information was recorded: the number of Cominella glandiformis present, their size distribution, and the identity of the ‘prey’ organism present in each case. The persistence of individual aggregations was monitored through the lowand high-water periods, by marking them with canes inserted a short distance shoreward of the group to prevent interference by floating algae that became trapped by the markers. Aggregations were marked on ebbing tides, or at low tide, and revisited as the tide rose and during the subsequent low tide period where possible. Further information on the dynamics of aggregation was obtained by exposing damaged and undamaged Macomona liliana and Austrovenus stutchburyi on the surface of the inshore flat or in the inner trough during the ebb tide and examining the rates of aggregation and disaggregation of Cominella glandiformis attracted to them. Data from these studies was used to calculate the predation rate by C. glandiformis and the equivalent mortality rates of Macomona liliana and Austrovenus stutchburyi at each station during each period. RESULTS Density of solitary and aggregated gastropods and of bivalves During the summer period the density of Cominella glandiformis aggregations differed markedly among Outer bar Spring 0.75 7.0 0.013 (5) 4 186 530 Summer 9.0 0.022 (7) <1 [<1] 36 [42] 8 [6] Spring 0.52 8.5 0 (5) <1 48 4 the three stations with the greatest density on the inner bar and the lowest on the inshore flat (Table 1, Fig. 2). At this time, isolated individuals of C. glandiformis were most abundant on the inshore flat and inner bar. During the spring period the overall density of aggregations was much reduced except on the inshore flat which at this time had the greatest density recorded at any location or time. In spring isolated individuals were seen on the sediment surface at all stations, with the greatest density on the inshore flat. The distributions and abundances of Macomona liliana and Austrovenus stutchburyi were similar during both the summer and spring experimental periods at all sites except for the former taxon on the inshore flat where the summer density was at least triple that recorded in the spring. In all other cases the quantitative samples were combined to provide the mean data given in Table 1. Comparing the distribution of the two bivalve taxa, A. stutchburyi was more abundant than M. liliana on the inshore flat and inner bar but less abundant on the outer bar. Size distributions of Macomona liliana and Austrovenus stutchburyi Figure 3 shows the distribution of the size classes (for individuals of over 0.95 cm) of individuals of Maconoma liliana and Austrovenus stutchburyi for each of the three study sites, with the exception of A. stutchburyi. Maconoma liliana from the outer bar had a greater modal size than those from the inner bar and inshore flat, whilst there were no differences among stations in the modal sizes of A. stutchburyi from the inner bar and the inshore flat. Using the biomass data (Table 2), it was apparent that the greatest biomass of 331 ALAN D. ANSELL M. liliana was found on the inner bar and the lowest on the outer bar in summer, whilst for A. stutchburyi the greatest biomass was on the inshore flat and the lowest on the outer bar in summer. In spring the greatest biomass of both M. liliana and A. stutchburyi was on the inner bar. Composition of aggregations Of those Cominella glandiformis that occurred in aggregations, in summer, on the inshore flat, more were associated with Austrovenus stutchburyi than with Macomona liliana; on the inner bar the numbers of aggregations associated with each was about equal; on Figure 2. Distribution of Cominella glandiformis (solitary and aggregated), Macomona liliana and Austrovenus stutchburyi at three sites on Matua beach, Tauranga Harbour, New Zealand, in spring and summer. IF inshore flat, IB inner bar, OB outer bar. 332 GASTROPOD AGGREGATIONS the outer bar most aggregations were associated with M. liliana (Table 3). In spring, most aggregations on both the inshore flat and the inner bar were associated with A. stutchburyi; on the outer bar the numbers of aggregations associated with each was about equal. In both summer and spring, the numbers of C. glandiformis per aggregation varied greatly both among stations and depending on the ‘prey’ bivalve (Table 3). Overall, the numbers in aggregations associated with M. liliana were greater than those associated with A. stutchburyi with the largest number in any one aggregation reaching 206. In some aggregations the ‘prey’ bivalve appeared to be undamaged; others showed all stages of consumption to a completely empty shell. In some cases the associated C. glandiformis had clearly begun dispersing and were distributed at short distances in all directions around the empty shell valves. The modal sizes of individuals of C. glandiformis associated with M. liliana were generally greater than those associated with A. stutchburyi (Fig. 4). Behaviour of Cominella glandiformis in aggregations In aggregations exposed at low tide, the gastropods formed a loose group with the lower individuals buried, especially in large aggregations, whilst many of those on top were completely uncovered with dry shells. Where aggregations could be observed in shallow water, the gastropods were seen to change position frequently around the ‘prey’ nucleus. Although it was difficult to observe individuals without disturbance to the group, but on many occasions the proboscis was seen to be extended into the shell cavity of a tightly closed bivalve, most frequently through the siphonal aperture of Macomona liliana or through the dorsal commissure of Austrovenus stutchburyi. Persistence of aggregations Re-examination of marked aggregations showed that, in summer, at all stations, the aggregations persisted throughout that period of low tide for which they were accessible, but that in most cases the aggregations had dispersed completely when revisited on the next low tide. In spring a greater proportion of marked aggregations survived a single tidal cycle, but all had dispersed before the second low tide period following marking, indicating a slightly longer average persistence at this time (Table 4). Attraction to damaged and undamaged bivalves Damaged Macomona liliana or Austrovenus stutchburyi exposed in shallow water during the ebb tide or placed on the wet surface of the inshore flat rapidly attracted Cominella glandiformis from downstream (Fig. 5). Generally, in excess of 20 gastropods reached the bivalve ‘bait’ within 15 min of its exposure; the tissues were consumed completely within 20 min for Austrovenus and between 20 and 35 min for Macomona depending on the bivalve’s size. The gastropods then dispersed rapidly. Intact bivalves (Macomona) placed on the surface attracted very few or no Cominella glandiformis and generally reburrowed. Rates of predation by C. glandiformis and mortality of Macomona liliana and Austrovenus stutchburyi The observations made here on the persistence of aggregations allow an estimate to be made of the minimum rate of mortality resulting from ‘predation’ by Table 2. Biomass estimates for the bivalves Macomona liliana and Austrovenus stutchburyi on Matua Beach, Tauranga Harbour. Inshore flat Summer Macomona no.m–2 mean wet weight (g) total wet weight (g m–2) Austrovenus no.m–2 mean wet weight (g) total wet weight (g m–2) Total bivalves wet weight (g m–2) 178 916 372 837 1753 Inner bar Spring 54 5.14 278 268 2.25 603 881 Summer Spring Summer Spring 164 186 5.68 1057 36 48 11.0 528 932 248 433 1365 333 Outer bar 530 1.75 926 1983 396 8 14 410 4 1.75 7 535 ALAN D. ANSELL Cominella glandiformis. Aggregations present throughout the low tide period had dispersed in summer following the subsequent high tide period of 7 or 9 h, suggesting that the maximum time needed for aggrega- tion followed by complete consumption of either bivalve species is one tidal period. The minimum ‘predation’ rate (R no. ‘prey’. m–2 .day–1) may be calculated on this basis as: R Fagg 24/12.5 % species Figure 3. Modal sizes of Macomona liliana and Austrovenus stutchburyi at three sites on Matua beach. 334 GASTROPOD AGGREGATIONS A/100, where Fagg is the number of C. glandiformis aggregations m–2 and % species A is the percentage of the total aggregations associated with prey species A. Predation rates calculated on this basis (Table 5) indicate that in summer the highest rates of predation occurred on the inner bar and affected both bivalve species more or less equally. On the inshore flat in summer, predation rates on Austrovenus stutchburyi exceeded those on Macomona liliana while on the outer bar the reverse was the case. In spring, predation rates were generally much reduced except on the inshore flat. Figure 4. Size distribution and percentage of population of Cominella glandiformis associated with Macomona liliana, Austrovenus stutchburyi and Mesodesma at three sites on Matua beach. 335 ALAN D. ANSELL Table 3. Percentage of all aggregations at each station associated with each type of ‘prey’ organism and mean numbers of gastropod in each type of aggregation. Range given in parenthesis. Inshore flat Summer % with Macomona Austrovenus Other 20.0 60.0 20.0 Inner bar Spring Summer 8.0 74.0 18.0 No. of Cominella per aggregation associated with Macomona 9.0 42.2 (8.96) Austrovenus 10.0 14.3 (8–13) (2–38) Outer bar Spring Summer Spring 29.6 37.0 33.3 24.0 72.0 4.0 86.7 6.7 6.7 44.4 33.3 22.2 62.0 (6-206) 28.4 (13–50) 29.3 (8–100) 7.8 (2–29) 58.7 (13–175) 35.0 22.6 (5–66) 17.5 (5–28) Table 4. Number of marked aggregations remaining after one or two tidal periods at stations at Matua beach, Tauranga Harbour. Inshore flat Summer No. marked 8 No. (%) remaining after 1 tide after 2 tides 0 0 Spring Summer Spring 23 28 47 3 (13%) 0 Mortality in the population expressed as the proportion of the bivalve population consumed day–1 was greatest in summer on the outer bar. In spring, mortality was much reduced on the outer and inner bars, but increased slightly on the inshore flat (Table 5). The equivalent rates expressed on an annual basis are summarized in Table 5. Averaged over spring and summer, these rates suggest that between 2 to 16% of M. liliana and 3 to 9% of Austrovenus fall prey to Cominella each year at different stations on the beach. These figures indicate that C. glandiformis is associated with significant rates of natural mortality of both bivalve species in this area. DISCUSSION Qualitatively, the fauna of the Matua intertidal zone conforms closely to the description of New Zealand harbour flats given by Morton & Miller (1968) with the larger macrofauna being dominated by the two burrowing bivalves Macomona liliana and Austrovenus stutchburyi. Quantitatively, there is little published Inner bar 1 (3.6%) 0 12 (25.5%) 0 information on population density of these species with which to compare the data obtained herein. However, Park & Donald (1994) give data for many sites around Tauranga Harbour and Hewitt, Thrush, Cummings & Pridmore (1996) for Manukau Harbour, which suggest that densities of adult bivalves are of the same order as those found at Matua. Both species were found throughout the exposed intertidal zone, but at different densities, both being progressively less abundant from the inshore flat to the inner and outer bars with the density of Austrovenus stutchburyi showing a greater decrease than that of M. liliana. A. stutchburyi is a suspension feeder that burrows only superficially. This is reflected in the high proportion of individuals that bear epizoans (chiefly the anemone Anthopleura aureoradiata Stuckey, 1909, the spionid worm Boccardia acus (Rainer, 1973), the barnacle Elminius modestus Darwin, 1854 and the limpet Notoacmea helmsi (Smith, 1894) on the posterior region of the shell. One result of this shallow burrowing habit is that individuals may be readily exhumed during strong tides (which particularly affect the outer sand bar) leaving them exposed to epifaunal predators. One unresolved question 336 GASTROPOD AGGREGATIONS regarding the relationship between A. stutchburyi and Cominella glandiformis is whether gastropods also feed on all or any of these regular epizoans associated with this species. The deposit feeding M. liliana, by contrast to A. stutchburyi, occupies burrows of up to 10 cm depth and is thus less likely to be exposed by strong tidal currents. The exact nature of the relationship between Cominella glandiformis and these bivalves is difficult to define. Field observations of individuals of C. glandiformis feeding on dead crabs, and the willingness of others to attack damaged bivalves in these experiments show that they are at least facultative scavengers. However, the predominance of bivalves with undamaged shells at the core of the aggregations, the frequency of aggregations around still apparently tightly closed Figure 5. Aggregation and dispersion time of Cominella glandiformis around damaged and undamaged bivalve ‘bait’. Macomona, damaged— series 1, 2, 3; Macomona, undamaged—series 4; Austrovenus, damaged—series 5,6. Arrow heads indicate dispersion after all bait consumed. Table 5. Estimates of the mortality rates (% of population day–1) of Macomona liliana and Austrovenus stutchburyi from ‘predation’ by Cominella glandiformis, and of predation rates (no.eaten day–1m–2) by C. glandiformis on Macomona liliana and Austrovenus stutchburyi at Matua beach, Tauranga Harbour. Inshore flat Inner bar Outer bar Summer Spring Summer Spring Summer Spring % Mortality Macomona Austrovenus 0.0017 0.0018 0.0100 0.0155 0.0507 0.0285 0.0015 0.0046 0.0880 0.0476 0.0009 0.0044 Predation rate Macomona Austrovenus 0.0019 0.0058 0.0054 0.0497 0.0887 0.1108 0.0060 0.0180 0.0366 0.0028 0.0004 0.0000 Equivalent annual % mortality Macomona 0.6205 Austrovenus 0.6570 3.6500 5.6575 18.5055 10.4025 0.5475 1.6790 32.1200 17.3740 0.3285 1.6060 Average annual mortality Macomona Austrovenus 2.1353 3.1423 9.5265 6.0408 337 16.2243 9.49 ALAN D. ANSELL bivalves, and the observation of the gastropod proboscis extended into the mantle cavity of the bivalve through the slight siphonal gape (Macomona) or through the dorsal commissure (Austrovenus) all point to a predator-prey relationship, although it is probable that as flesh is progressively removed, leaking body fluids may attract scavenging individuals. Despite reports of drilling behaviour by Cominella (Peterson & Black, 1995), no indication of drilling predation by Cominella glandiformis was found in Tauranga Harbour (Ansell, 2000). The ability to both scavenge and actively predate suggest that C. glandiformis is an opportunist feeder. Two factors suggest that the rates of predation and their impact on the bivalve populations as calculated here might underestimate the actual rates. First, the observation of the rate at which C. glandiformis is attracted to damaged bivalves and their subsequent dispersal following consumption of the soft tissues show that aggregation, consumption and dispersal can take place in a much shorter time (1 h) than the 7–9 h for which the intertidal stations were immersed. Morton & Britton (1991) made similar observations for Cominella eburnea (Reeve, 1846) and C. tasmanica (Tenison Woods, 1878) on Australian beaches. This would imply that aggregation, consumption and dispersal could occur repeatedly through the high tide period, greatly increasing the estimated predation impact, although attractive cues may be difficult to follow on a rising or high tide which might lessen this effect (J.C. Britton, personal communication). A similar problem has been noted in assessing the diets of other intertidal predatory gastropods by Fairweather & Underwood (1983). In this study, natural aggregations were rarely associated with damaged bivalves; many bivalves appeared intact and were still firmly adducted, but proved on examination to have been partially consumed. It is reasonable, therefore, to conclude that aggregation around and consumption of intact bivalves takes place much more slowly than is the case with damaged bivalves. This view is supported by observations made by P. Smythe & R. Black (unpublished observations) for the Australian species C. eburnea. They measured times needed to ‘infiltrate’ and consume the bivalve Katelysia scalarina Lamarck, 1818 in aquaria of 13 h 8.5, 22.2 h 6.9 and 24 h for 20 mm, 15 mm and 10 mm Cominella eburnea, respectively. Secondly, the size of the bivalve may affect the time taken for complete consumption to occur following aggregation. Consumption of Macomona liliana, which are generally larger than Austrovenus stutchburyi, especially on the outer bar, might then be expected to take longer although this may be offset by the generally larger numbers of Cominella glandiformis found in aggregations around the former (Table 3) and their larger size (Fig.4). However, the fact that some aggregations remain intact on the shore at the next low tide, indicating a time for completion of 12.5 h, especially in spring, implies a lower mortality rate than that calculated. Although the aggregations of Cominella glandiformis are a conspicuous feature especially during the summer, they represent only a small proportion of the total population of the gastropod present over much of the intertidal, the exception being the outer sand bar. Aggregations accounted for only some 0.27% of the total over the inner flat in summer (Table 6), for some 22.4% of the total on the inner sand bar, but almost 100% of the total on the outer sand bar, where solitary individuals were not found in the quantitative benthic samples, but were seen on the sand surface or partly buried, during the low tide period. It follows that for most of the time C. glandiformis on the inshore flat are not aggregated while those of the inner bar are aggregated for some 3% of the time. It is difficult to account for these differences unless it is assumed that alternative, presumably smaller, items of prey that do not involve aggregated feeding, are more important in the inner bar and inshore flat areas. Morton & Britton (1991) calculated the area from which C. eburnea were attracted to prey on Australian beaches as: Area Mean no. Cominella per aggregation / mean density of solitary Cominella on the beach. Perhaps more accurately such calculations should include both solitary and aggregated Cominella density (Table 6). Calculations for C. glandiformis yield similar areas of attraction for this species on the inner bar and inshore flat at Matua (Table 6). Observations in shallow tide pools with weak residual tidal currents indicate that C. glandiformis are attracted mainly from the areas downstream of bait, often from considerable distances ( 4m, pers. obs.). The adaptive value of group feeding in vertebrates (e.g. Bernard & Sibly, 1981) has been investigated more than for various invertebrate taxa. Aggregated feeding is a common phenomenon among scavengers (Britton & Morton, 1994) including gastropod molluscs of the families Nassariidae and Melongenidae where individuals are rapidly attracted to carrion, feed rapidly to satiation and are replaced by other individuals until the food resource is exhausted. The aggregated feeding by Cominella glandiformis described herein seems to differ fundamentally from this pattern since groups of individuals are formed around live prey and remain together for a more extended periods until the food resource is exhausted before dispersing. Aggregative 338 GASTROPOD AGGREGATIONS predation by muricid gastropods has been reported in Nucella lapillus (Linnaeus, 1758) (Hughes & Dunkin, 1984), Stramonita haemastoma (Gray, 1839) (Brown & Alexander, 1994) and Thais clavigera (Küster) and T. luteostoma Holten, 1802 (Taylor & Morton, 1996). In a detailed study specifically focussed on group feeding, Brown & Alexander (1994) found although a high proportion (40–58%) of feeding S. haemostoma were found in aggregations around large oysters in the field, laboratory experiments showed that group feeding decreased the flesh yield per capita and that there was no decrease in prey handling time, although the overall amount of prey tissue removed was greater in groups than in solitary predators. They suggested that there were no measurable costs to the individual involved in feeding in groups on prey items with a flesh yield greater than that required to satiate a single predator, and, in any case, an individual had no means of defending its the prey from interlopers. They did suggest that group feeding might provide a mechanism for feeding on larger prey, however, their experiments demonstrated that single individuals were capable of overcoming and feeding on these items. Taylor & Morton (1996) found that aggregations of T. clavigera and T. luteostoma on a boulder shore in Hong Kong formed around relatively small prey items and suggested that group feeding, in this instance, might result from prey shortage. It is possible that these arguments might apply to the formation of aggregates of feeding C. glandiformis. Alternatively, aggregated predation may provide them with a number of definite advantages. First, observations of the aggregation process show that, when sufficient numbers of Cominella have reached a damaged bivalve, other small scavengers such as the crab Hemigrapsus crenulatus (Milne Edwards, 1837) are excluded and prove unable to share the food avail- able. Secondly, as a predation process, aggregation presumably may allow the gastropod to overcome the prey more quickly. Aggregated feeding allows a food resource such as a large bivalve to be shared efficiently between individuals, generally over a single tidal period. Most predatory gastropods are limited in the size of prey they can successfully attack as individuals. Aggregated feeding allows C. glandiformis to escape such a restriction and exploit large prey. A similar strategy is adopted by some other invertebrate predators, e.g., the starfish Anasterias rupicola (Verrill, 1876) feeding on the large limpet Nacella delesserti Hedley, 1916 at Marion Island (Blankley, 1984) and other Anasterias species at McQuarie Island (Simpson, 1976). Perhaps the most spectacular recorded example of aggregated feeding by a gastropod is that of the South African whelk Burnupena sp. at Marcus Island successfully attacking the rock lobster Jasus lalandii Milne Edwards, 1837 (Barkai & McQuaid, 1988). Barkai & McQuaid reported that each rock lobster was killed within 15 min by more than 300 Burnupena that removed all the flesh in less than 1h. The quantitive aspects of aggregated feeding can be examined in terms of the organic tissue available from Macomona liliana or Austrovenus stutchburyi of different sizes (Equations 1and 2, Fig. 3), the organic content of Cominella glandiformis of different sizes (Equation 3), and the nutritional requirements of individual Cominella. Table 7 shows the number of C. glandiformis of mean size for each station that a single Macomona liliana or Austrovenus stutchburyi of the mean size for that station could provide with its daily maintenance requirements, assuming a daily maintenance requirement of 3% of body weight (see Morton & Britton, 1991). The numbers range from 67 to 208 for Macomona liliana at different stations and from 22 to Table 6. Calculation of apparent areas of attraction of Cominella glandiformis to ‘prey’ at Matua beach, Tauranga Harbour. Inshore flat Inner bar Summer Spring Summer Spring Density Cominella Aggr.m–2 Mean no.aggr–1 0.005 11.00 0.035 20.70 0.156 40.90 0.013 7.10 Total number of aggregated Cominella Density solitary Cominella (no m–2) 0.055 20 0.725 8 6.380 22 0.085 4 Total Cominella (no m–2) % aggregated Cominella 20.055 0.274 8.725 8.309 28.380 22.481 4.085 2.081 0.550 2.558 1.859 1.775 Area of attraction (m2) 339 ALAN D. ANSELL Table 7. Calculation of the number of Colminella glandiformis of modal size supported by a bivalve of modal size at stations on Matua beach, Tauranga Harbour, assuming a maintenance requirement of 3% of body weight–1. Inshore flat Inner bar Outer bar Spring Spring Spring 17.5 28.6 17.5 28.6 – – 32.5 178 37.5 273 47.5 554 Modal size Cominella associated with Austrovenus mm shell height 12.5 mg AFDW 16.2 17.5 42.9 – – Modal size Cominella associated with Macomona mm shell height mg AFDW 22.5 88.7 22.5 88.7 22.5 88.7 Number of Cominella supported by Austrovenus by Macomona 59 67 Modal size Austrovenus mm shell length mg AFDW Modal size Macomona mm shell length mg AFDW 59 for Austrovenus stutchburyi; numbers of the same order as those actually found in the field. Much remains to be learned about the dynamics and other components of this process. For example, little is known about how Cominella locate and eventually attack ‘prey’, especially the deeper burrowing Macomona. Nothing is known of the sensory processes involved in the early stages of formation of aggregations, especially since many aggregations can be found that are centred on what appears to be an undamaged bivalve. Possibly secretions of pheromones attract other conspecifics to the aggregation even before the release of possible attractants derived from damage to the ‘prey’. The time involved in these processes remains to be further examined. Aggregated feeding intertidally by gastropods appears to be a particular feature of Australasian beaches. One factor allowing the build-up of gastropod aggregations on these beaches may be the absence of larger more aggressive scavengers. The crabs that are present, for example, are all relatively small when compared with the common shore crabs of European beaches. Some Hemigrapsus crenulatus (Milne Edwards, 1837) were attracted to damaged bivalves exposed on the shore, but the Cominella glandiformis soon formed a compact mass that excluded these relatively small crabs, indicating that the gastropods can outcompete the crabs in exploiting even carrion. Few other preda- 22 103 – 208 tors of Macomona liliana and Austrovenus stutchburyi have been reported. Thrush, Pridmore, Hewitt & Cummings (1991, 1994) report that birds and rays are predators in Manukau Harbour, the former taking mainly small bivalves such as newly-recruited Macomona and the latter larger bivalves. Pits caused by the activities of rays were seen on the beach at Matua during this study, especially during summer, and a variety of birds are seasonally abundant. Gastropod predation, however, seems to be a significant component of bivalve mortality. ACKNOWLEDGEMENTS I am grateful for the support of Drs Phillip and Janet Thwaite and Dr Derry and Mrs Jenny Seddon who provided support and access facilities to the beach at Matua for this study. Sharon Capon (Cambridge Earth Sciences) kindly re-drafted Figure 1 and Geoff Read gave helpful advice about the taxonomy of the New Zealand fauna. This paper is for Jamie Thwaite who already shows a lively scientific curiosity for the beach habitat. REFERENCES ANSELL, A.D. 2000. Drilling and non-drilling predation in gastropod aggregations in the intertidal zone of a New Zealand harbour. Journal of Molluscan Studies, 66: 419–422. 340 GASTROPOD AGGREGATIONS BARKAI, A. & MCQUAID, C. 1988. Predator-prey reversal in a marine benthic ecosystem. Science, 242: 62–64. BERNARD, C.J. & SIBLY, R.M. 1981. Producers and scroungers: a general model and its application to captive flocks of house sparrow. Animal Behaviour, 29: 543–550. BLANKLEY, W.O. 1984. Ecology of the starfish Anasterias rupicola at Marion Island (Southern Ocean). Marine Ecology Progress Series, 18: 131–137. BRITTON, J.C. & MORTON, B. 1994. Marine carrion and scavengers. Oceanography and Marine Biology: An Annual Review, 32: 369–434. BROWN, K. M., & ALEXANDER, J. E. 1994. Group foraging in a marine gastropod predator: benefits and costs to individuals. Marine Ecology Progress Series, 112: 97–105. FAIRWEATHER, P. G. & UNDERWOOD, A. J. 1983. The apparent diet of predators and biases due to different handling times of their prey. Oecologia, 56: 169–179. GRANGE, K.R. 1977. Littoral benthos-sediment relationships in Manukau Harbour, New Zealand. New Zealand Journal of Marine and Freshwater Research, 1: 111–123. HEWITT, J.E., THRUSH, S.F., CUMMINGS, V.J. & PRIDMORE, R.D. 1996. Matching patterns with processes: predicting the effect of size and mobility on the spatial distributions of the bivalves Macomona liliana and Austrovenus stutchburyi. Marine Ecology Progress Series, 135: 57–67. HUGHES, R. N., & DUNKIN, S. D. B. 1984. Behavioural components of prey selection by dogwhelks, Nucella lapillus (L.) feeding on mussels in the laboratory. Journal of Experimental Marine Biology and Ecology, 77: 45–68. MORTON, B. & BRITTON, J.C. 1991. Resource partitioning strategies of two sympatric scavenging snails on a sandy beach in Western Australia. In: The Marine Fauna and Flora of Albany, Western Australia. Vol. 1. (F.E. Wells, D.I. Walker, H. Kirkman & R. Lethbridge eds), 579–595. Western Australian Museum, Perth. MORTON, J. & MILLER, M. 1968. The New Zealand sea shore. Collins, London. PARK, S. & DONALD, R. 1994. The environment B.O.P. Tauranga Harbour regional plan environmental investigations. Ecology of Tauranga Harbour. Environment B.O.P. Environmental Report, No. 94/8: 176 pp. PETERSON, C. H., & BLACK, R. 1995. Drilling by buccinid gastropods of the genus Cominella in Australia. Veliger, 38: 37–42. SIMPSON, R.D. 1976. Physical and biotic factors limiting the distribution and abundance of littoral molluscs on McQuarie Island (Sub-Antarctic). Journal of Experimental Marine Biology and Ecology, 21: 11–49. TAYLOR, J. D., & MORTON, B. 1996. The diets of predatory gastropods in the Cape d’Aguilar marine reserve, Hong Kong. Asian Marine Biology, 13: 141–166. THRUSH, S.F., PRIDMORE, R.D., HEWITT, J.E. & CUMMINGS, V.J. 1991. Impact of ray feeding disturbances on sandflat macrobenthos: do communities dominated by polychaetes or shellfish respond differently. Marine Ecology Progress Series, 69: 245–252. THRUSH, S.F., PRIDMORE, R.D., HEWITT, J.E.& CUMMINGS, V.J. 1994. The importance of predators on a sandflat: interplay between seasonal changes in prey densities and predator effects. Marine Ecology Progress Series, 107: 211–222. 341