Mollusk Habitats and Fisheries in Kiribati: An Assessment from the Gilbert Islands1

Frank R. ThomaslThomas2

Abstract: Biological and ecological attributes of24 species ofedible bivalves and gastropods from thethe Gilbert Islands Group, Kiribati, Micronesia, were assessed for their resilience by examining size at maturity, intertidal burying, adjacent subtidal populations, benthic mobility, and larval type. Foraging for mollusks is largelylargely confined to the intertidal and shallow subtidal regions, although modern diving gear and outboard motors now provide human foragers access to offshore resources. Changes brought about by human demographic pressures have re­ sulted in overexploitation of a number of molluscan resources. It isis suggested thatthat the sustainable use of invertebrates and other marine species for food and nonfood purposes in Kiribati rests on a remodeled form of marine tenure.

ATOLLS PRESENT A variety of marine envi­ from resilient resources, and their effect on ronments and molluscan resources that have subsistence. provided subsistence living to generations of Can traditional ecological knowledge be Pacific islandersislanders (Tebano(Tebano et al. 1993, Taniera wedded with modern technology to revive thethe 1994). Kiribati (Figure 1) is but one of more fisheries? Because the thrust of this paper isis thanthan 20 nations in the Pacific, each of which limited to prey biology and ecology in thethe possessed a rich inshore fisheries tradition context ofmollusk fisheries, details of analytic and lore.lore. InIn recent years, however, traditional methodology related to foraging efficiency ecological knowledge (TEK) has suffered the are not discussed here, but can be found inin inevitableinevitable decline that characterizes similar Thomas (1999). bodies of knowledge throughout most of Data on molluscan resources of the Gil­ Oceania. High human population density, bert Islands date back toto thethe 1950s when urban drift, more efficient extractive tech­ Banner and Randall (1952) described thethe in­in­ nologies, and expanding market opportu­ vertebrates of Onotoa Atoll, but identifica­identifica­ nities all have inevitably affected the ocean tions were poor and there were few detailed resourcesresources of these islands. In this study, I descriptions of exploitive strategies. A major examined the prey biology and ecology of interdisciplinary environmental survey on mollusk fisheriesfisheries in Kiribati, Gilbert Islands, the main atoll ofTarawa, Republic ofKiribati lookinglooking at factors distinguishing vulnerable (Abbott and Garcia 1995), expanded species identification and habitat description, and provided an assessment of the impact of 1 Financial support from the University of Hawai'i's human activities (Figure 2). The Tarawa La­ Arts and Sciences Advisory Council; funds provided by goon Project of 1992-1994 (Paulay 1995) is a Alison Kay throughthrough a University of Hawai'i Foundation major study of the lagoon, and data from thatthat Grant and Research Assistantship; Sigma Xi (The Scien­ tifictific Research Society) Grant-in-Aid of Research nos. project are included in this study. 2710 and 8326; thethe Hawaiian Malacological Society; two grants fromfrom Conchologists of America, Inc.; and travel fundsfunds provided by Temakei Tebano at the University of MATERIALS AND METHODS thethe South Pacific. Manuscript accepted 1 February 2000. 22 Ogden Environmental and Energy Services, 680 Several communities engaged in subsistence lwileiIwilei Road, Suite 660, Honolulu, Hawai'i 96817. and commercial exploitation ofmollusks were investigated intermittently between 1993 and 1998. Research focused on four atolls in the Pacific Science (2001), vol. 55, no. 1:77-97 1998. Research focused on four atolls in the © 20012001 by University of Hawai'i Press central Gilberts: Abaiang, Tarawa (both ur­ All rightsrights reservedreserved banized and rural sectors), Maiana, and Abe-

77 78 PACIFIC SCIENCE· January 2001

...~;;;/Makin~",;Makin 180° .\Po.... U.S.·Admlnl.t.~dU.S.·Admlnlst.~dIslsnds'.'.nd. .s;~ ButaritariButarilari ·:=1;Sularilari';:;;Sutarilari In the Central Pacific < TatluaeranTat>uaeran 0, 100 200300 ..... KiritimaliKlrIllmal' Maiana!tarawa\!tarawa HowlandHOwland \ !I I !I ~ IslandISland'-=- STATUTE MilES ~Abemama '= '!.'!IAbemama Il'I.!.' Baker Island N JarvIsJarvIS Island.,ISland.~, <:.:~ 0° 1>,~ BanabaSanaba •.f.\Nonouti\Nonouti Tabileuea\;tlli~STabi(euea\:t~~S .",.,,-., ' Marakei 2°N PHOENIX ~"/o~?IO ISLANDS ~ I "'.J""~ Abaiang Tamana.:lTamana ~ ~ Arorae O/'''O/....,/~.":l!;Enderbuty../~.':$Enderbury KIRIBATI ~ :f) 0., ~Il':~ t: ~Rawaki~ Rawaki '\...... Tarawa GILBERT ISLANDS ., 0., to Malden ::..':. NikumaroroNikumarOtO...... ~lf.~tl< ~Manra 160 W ~.SlartlUCk~.Slart>uCk ;)Maiana.:)Maiana 'Ii>'ti> Abemama ~ .",:--Il"~,:,\ GILBERT ISLANDS <:) ° Kuria~~KUria~'~~ 'I.... 0 50 100 o ° 'DO "\"\KanlanKanlan VO'lok"VOSlok" lJIU' .,.\' 10 00 Aranuka I! !I II CarolinelOrr 0 STATUTE MilES 3 S Flint·'Flinl" PHOENIX ISLANDS 'Enderbury 0 50 '00'DO .y,,, I I I NonOuli '~«i~;k'?'~\ • Birnie NonOuli STATUTE MILES McKean • RawskiRaWSki

!~. .• Nikunau 4° r8biteuea":t:JTabiteuea":.~ Beru~~~Beru~\ , \:'Orona -ManraaManra Nikumaroro, ':'Orona Onolo/::> . 1~40W 172° 2°S

0 173 E 175°0 Tamana·Tamana' '75 Arorae~Arorae'l\)

FIFIGUREG URE 1. The Republic ofKiribati (modified from Motteler 1986:1986:16;16; reprinted with permission from Bishop Museum Press, Honolulu, Hawai'i).

mama, and one atoll in the Southern Group, along the southern portion of Tarawa Atoll, Tabiteuea North (Thomas 1999). Data on with samples taken between December 1992 various aspects of mollusk gathering were and February 1994. This information greatly obtained by participant observation, inter­ expanded the sample size of foraging events, views, and from the Shellfish Gatherer Sur­ with 83 foraging trips covering 26 days and vey (SGS) of the Tarawa Lagoon Project approximately 191 hr of foraging time (257 (TLP), in cooperation with BioSystems, Inc., forager-hours). Mean group size was 1.3 for­ of Santa Cruz, California. agers (SD = 0.7, n = 112). I accompanied foragers, both individuals Prey type identifications and basic habitat and groups, on 73 foraging trips covering 69 information are based on Cernohorsky (1971, days, for a total of 139.63139,63 hr of direct obser­ 1972), Eisenberg (1981), Abbott and Dance vation (286.92 forager-hours). Mean group (1982), Paulay (1995), and personal observa­ size was 2.6 foragers (SD = 0.9, n = 146).146), tions. For each prey type, five factors were Fifty-nine different individuals took part in examined to assist in distinguishing vulnera­ these activities. When it was not possible to ble from resilient resources: (1) size at matu­ record foraging activities directly, the rele­ rity; (2) intertidal burying; (3) adjacent sub­ vant information was elicited via interviews. A tidal populations; (4) benthic mobility; and total of 65 foraging trips was recorded over a (5) larval type (see Catterall and Poiner 1987). period of 51 days, for an estimated total of In a number of cases, the description of rele­ 88.5 hr offoraging effort (161 forager-hours). vant properties was limited, reflecting gaps in Mean group size was 3.1 foragers (SD = 0.7, the data. n = 148).148), A total of 19 different individuals was interviewed. Atoll Zonation and Molluscan Patch Types Information was also derived from data sheets made available by personnel of the From windward to leeward (Figure 3), the TLP. These data originated from landings basic zones and molluscan patch types char- Mollusk Habitats and Fisheries in Kiribati . Thomas 79

FIGURE 2. Tarawa Atoll (modified from Paulay 1997: fig. 1; reprinted with permission from Gustav Paulay).

acteristic of the open atolls of the Gilbert Is­ THE SEAWARD REEF: The seaward reef lands are the: (1) reef slope; (2) submarine slope descends steeply from the upper limit terrace; (3) reef front; (4) reef margin; (5) of vigorous hermatypic corals at a depth of moat; (6) reef crest; (7) reefflat; (8) shoreline; about 18 m to the deepest seaward portion of (9) lagoon reef flat including mangrove, sand the reef between 4000 and 6000 m. In the flats, and sea grass beds; (10) lagoon slope and Gilberts, little is known of the biota below floor; and (11) leeward reef. 40 m, but mollusks commonly found in the 80 PACIFIC SCIENCE· January 2001 r-

-. -:-:.'..

' ' ..., " . " " ..... 0"' " , ... ", '.':.,',','", . ', .. ' . .. " ::-... " '.. " ,', ..

..~"

:.:.:GR.o.V(l.

c..c..~ ....la.polUllolU '0010,"'4' ..10, 4 .t...lI>uc.lc.l.. Aft" REffREfF fl([rfl(fr n." CRESftRESf RU,.RUF S-OP5LOP fAO/oITfR~T

FIGURE 3. Schematic diagram of atoll features of the Gilbert Islands (modified from Richmond 1993:17, fig.fig, 5; re­ printed with permission from the South Pacific Applied Geoscience Commission [SOPAC),[SOPAC]).

upper reaches of this zone include Pinctada demarcates the seaward limit ofthe reefflat; it and Trochus (Wells 1957), which also occur supports the gastropod Turbo setosus. The reef on the submarine terraces between depths of flat itself is commonly exposed or barely 15 and 18 m. The dominant mollusks on the awash at lowest tide. It consists of a truncated reef front, the shoreward face of the seaward floor zoned by the tides, which effect varying reef slope where coral growth is sparse and degrees of exposure and temperature between heavy surf dominates, are gastropods such as seaward edge and inshore margin. The sur­ Patella and Trochus (Demond 1957). The face of the reef flat supports Cypraea, Thais, seaward reef margin, distinguished by spur­ Trochus, Turbo, and other species of gastro­ and-groove formations, which disperse and pods adapted primarily as crab predators and channel ocean water, supports compact and to the force of waves (Banner and Randall encrusting corals and coralline algae such as 1952). Porolithon. Mollusks found in this zone in­ The shoreline of islets (moms)(motus) on both clude Trochus, Turbo, Cypraea, and Drupa lagoon and ocean sides consists of both (Wells 1957, Zann 1982). Shoreward of the cemented and uncemented topographic fea­ seaward reef margin are moats, areas of shal­ tures. Beachrock, lithified beach sand and low water normally awash even at extreme coral debris, and conglomerates, cemented low tide. In many respects, moats resemble storm deposits between 1 and 2 m above lagoon habitats, except for the hard bottom of low-tide level, are occupied by dense con­ coralline rock or gravel. Thick-branching centrations of the gastropod Nerita plicata, Acropora and Heliopora are the dominant adapted to withstand heat and desiccation, as corals, and the mollusks include Cypraea, well as rain (Cloud 1952). Moderately to Conus, Lambis, Hippopus hippopus, and Tridacna poorly sorted medium sand, gravel, and cob­ gigas in the deeper sections (Wells 1957). bles that rise with a relatively sharp change in The reef crest, a high-energy environment, slope above the reef flat harbor the bivalves Mollusk Habitats and Fisheries in Kiribati . Thomas 81

Asaphis violascens and the small beach clam at Nonouti and Tabiteuea. The deepest Ataetodea striata. The latter is also locally lagoon, at Butaritari, at 33m is shallower abundant in other intertidal sandy areas. than atoll lagoons in the Marshalls and THE LAGOON: Lagoon beaches tend to be Tuvalu. Deep reef passes are restricted to well to moderately sorted, from fine sand to leeward lagoon margins. There is an increase gravel; highest elevation is generally less than in live coral coverage and decrease in sedi­ 1 m above the highest tide; and sand spits are ment coverage toward the leeward margin. a common feature, especially near interislet The lagoon slope, predominantly covered channels and as lagoon-directed seaward­ with sand or coral rubble, begins at the produced growth at the ends of the atolls. deepest edge of the sea grass bed, approxi­ Asaphis violascens and Ataetodea striata are mately 0.5 m deep at low tide. It slopes gently common in this zone (Paulay 1995). to depths of 8-10 m at 1 km from the sea The lagoon reef flat is typically a deposi­ grass margin, if present. This area supports tional surface built up by ocean-to-lagoonocean-to-Iagoon scattered colonies or clumps of primarily sediment transport. Sediments tend to be Porites spp. and Pocillopora damicornis in South poorly sorted with some silt. The reefflats are Tarawa. Mollusks such as Anadara uropyg­ generally better developed along the sections melaname/ana and psammophilic crustaceans and facing the prevailing winds than elsewhere echinoids predominate (Wells 1957, Paulay and are zoned as are the windward reef flats. 1995) on fringing or patch reefs often domi­ The reef flat is primarily intertidal, grading nated by Acropora on the lagoonal slope in into subtidal zones toward the center of the North Tarawa. lagoon. Except during periods of westerly The lagoon floor is the more or less level storms, waters are generally calm. Man­ bottom ofa lagoon. Sandbars, pinnacles, coral groves, mainly Rhizophora mucronata, may rim knolls, and patch reefs (dead and living) the shoreline. The sediments associated with present sharp irregularities in the bottom. mangroves are characteristically soft, of sand Lagoon depth, sedimentation rates, and other mixed with rich organic matter. Fiddler crabs, physical factors affect the growth of coral and Uca sp., and stomatopod burrows occur in thus species density and diversity. Mollusks transitional areas between the mangrove are uncommon and constitute only a small sediments and the sand flats. The bivalve fraction of the biomass on Tarawa. The lee­ Gafrarium peetinatum is occasionally found in ward reef platform is characterized by irreg­ the mangroves. ular outer reefs and few or no islets. Leeward Sand flats are the most extensive habitat on reefs are commonly submerged over most of the lagoon reef flat. The fauna varies with their area, even at low tide, and show rela­ depth of the water, size of sand particles, tively vigorous coral growth. This zone sup­ amount ofwave action, and depth ofthe sand. ports Lambis and Tridacna. The seaward slope This habitat supports the economically im­ is generally much steeper than on windward portant Anadara uropygmelana,uropyp;melana, Strombus lu­ reefs. As with the seaward (windward) reef huanus, and Gafrarium peetinatum (Paulay slope and submarine terrace, Trochus and 1995). The low intertidal and subtidal regions Pinetada may be abundant (Wells 1957). of the lagoon border may support extensive TIDES AND CURRENTS: The greatest tidal beds of Thalassia hemprichii that stabilize and fluctuations (spring tides) occur on the days bind fine sand and silt rich in organic matter. of full and new moons; the smallest fluctua­ The sea grass beds are productive grounds for tions (neap tides) are about halfway between Anadara uropygmelana,uropyp;melana, Trachycardium anfJUla­an[!;Ula­ spring tides, during the first and last quarters. tum, Gafrarium peetinatum,pectinatum, Timoclea marica, illIn Tarawa Lagoon, semidiurnal tides have a Strombus luhuanus, and other mollusks (Pau­ mean amplitude of 1.5 m (neap range, 1.2 m; lay 1995). spring range, 1.8 m) (Gillie 1993). Fluctua­ Atoll lagoons in the Gilberts encompass a tions in sea level (on the order of 10-20 em)cm) much larger area than the landmass. They follow seasonal changes in the location and vary from an almost enclosed lagoon at strength of the trade winds. During the ElEI Marakei to poorly defined shallow expanses Nino event in 1982-1983,1982 -1983, the monthly mean 82 PACIFIC SCIENCE· January 2001

sea level was 28 cm above the long-term The higher in-patch returns offered by mean in 1982, but 21 cm below mean in late present-day lagoons compared with ocean 1983. Repeated cycles of submergence and patches in general seem to provide a likely exposure and variation along these lines as a explanation for the overwhelming importance result of lunar periodicity make the intertidal of lagoonal gathering. Although foragers are a zone of high environmental stress. quite willing to walk relatively long distances on the exposed intertidal lagoon flat (often in excess of overland distances to the ocean Foraging side), they also have more opportunities to increase their net energy acquisition rate The indigenous Micronesian population con­ The indigenous Micronesian population con­ across the various lagoon patches as the tide siders many species of mollusks edible, but siders many species of mollusks edible, but comes in and they are compelled to move only a handful appear to be specifically tar­ only a handful appear to be specifically tar­ closer to shore to the less-productive patches geted. Only 24 of the perhaps 1000 or more geted. Only 24 of the perhaps 1000 or more (but still more productive relative to the in­ species that may occur in the Gilbert Islands species that may occur in the Gilbert Islands tertidal ocean patches). (see Kay and Johnson [1987] for the numbers estimated from the Marshall Islands) com­ posed the major "crop" in this study. RESULTS A foraging bout usually takes place within The Mollusk Harvest a single patch type, for example, a sand flat, sea grass bed, or nearshore biotope. Most Twenty-four species of mollusks are the households on Tarawa and other atolls are principal objects of foraging in the Gilbert concentrated on the lagoon side of the motu Islands, with three species composing the and harvest mollusks and other resources bulk of the harvests. The relative dietary from this habitat. Where sea grass beds oc­ contributions of lagoonal shellfish from the cur, they are the focus of the most intensive intertidal beach to the lagoon slope on South gathering pressure (Paulay 1995). Women Tarawa, derived from the SGS, are listed in and children concentrate on resources in the Table 1. Thomas (1999) converted gross intertidal and shallow subtidal areas where weights into edible meat weights using ratios they can easily located telltale marks of their presented. Protein and energy were calcu­ prey. Offshore diving is an activity carried out lated by multiplying the values per 100 g, by male divers only. Limited ocean-side illustrated in Thomas (1999) by the total gathering does take place, however. weight of the harvested mollusks (Table 1).

TABLE 1 Relative Dietary Contribution ofMollusks during 83 Lagoonal Gathering Trips on South Tarawa (in Rank Order According to Edible Meat)

Species No. of trips Edible meat (kg) Protein (kg) Kcal

Strombus l.I. 59 93.89 18.31 79,807 Anadara 51 61.73 7.35 34,569 Gafrarium 42 5.13 0.68 3,748 Asaphis 5 1.00 0.08 393 AtactodeaAtaetodea 3 0.66 0.10 498 Quidnipagus 2 0.62 0.04 255 Trachycardium 11 00400.40 0.04 186 Strombusv. 5 0.07 0.01 48 Cymatium 15 0.06 0.01 56 Other 12 0.05 Spondylus 2 0.03 0.003 15 Timoclea 1 0.004 0.001 3 Pitar 2 0.004 0.001 3 Mollusk Habitats and Fisheries in Kiribati . Thomas 83

Strombus luhuanus Linnaeus, 1758. The Anadara uropygmelana (Bory de Saint­ strombid, te nouo, is common throughout Vincent, 1824). This arcid, te bun, is the most much of the low intertidal and shallow sub­ often-targeted mollusk in the central Gil­ tidal areas of Tarawa Lagoon. It also occurs berts, notably on South Tarawa where it oc­ on sandbars with mixed sand and coral rubble curs in low intertidal and subtidal regions of that rise to within 0-2 m below the surface. the lagoon (Figure 4). Unlike other anadarine Like other strombids, te nouo is an herbivore bivalves, A. uropygmelana is attached to both and feeds on algae attached to hard substrata hard and soft substrata by a slender byssus, (Abbott 1960). It often occurs in large col­ even in its adult stage, although byssal threads onies. Maturity is reached within 2 yr after were often lacking among te bun examined by settlement, at which time shell length sta­ Tebano (1990). Byssal attachment may be bilizes at 40-60 mm, and the shell lip thick­ crucial in the early stages of its life history, ens. Adults are distinguished from juveniles with the root system of Thalassia hemprichii by a distinct stromboid (anterior) and poste­ appearing to provide suitable surface. Indi­ rior notch (Poiner(poiner and Catterall 1988). Aver­ viduals from depths of 2 to 8 m tend to be age edible weight is estimated at 2 g. larger than those in the intertidal and are Adults are usually epibenthic; juveniles abundant in habitats with fine, soft substra­ tend to be buried in sediment. However, up tum. Te bun reaches maturity at 38 mm in to 30% of adults at Heron Island, Great Bar­ females and 42 mm in males (Tebano and rier Reef, were also buried, with an increase Paulay 1995). Reports on traditional gather­ in this behavior during periods of inclement ing practices from various geographical areas weather (Catterall and Poiner 1987, Poiner describe minimum shellfish length collected and Catterall 1988). Large numbers of ma­ as 20-30 mm (Catterall and Poiner 1987). ture individuals occur in deep water, with For South Tarawa the value is closer to migration to shallower depths related to 10 mm (Paulay 1995). Te bun yields about breeding. Like other strombids, te nouo can 4 g of edible meat in the lagoon reefflat, with rapidly migrate from subtidal to intertidal a mean length of45 mm, and 8 g offshore for areas by means of its powerful foot. In an average length of about 55 mm. southern Papua New Guinea and eastern These bivalves live half buried in sea grass Australia, this strombid breeds from August beds where young individuals may escape de­ to March and deposits up to 200,000 eggs. tection in thick undergrowth. Like most bi­ The planktonic larval stage may last between valves, te bun is sedentary in its adult stage; 2 and 3 weeks (poiner and Catterall 1988). consequently, replenishment of heavily ex­ Commercial exploitation occurs on Ta­ ploited intertidal areas by reproductive in­ rawa, but at a smaller scale than that of the dividuals from the subtidal regions cannot arcid te bun, Anadara uropygmelana. Te nouo take place. Little is known about larval biol­ is gathered by visually searching for exposed ogy or recruitment in this species. Tebano individuals, making no distinction between (1990) suggested that spawning may be con­ large and small shells or thin- and thick­ tinuous, with peaks during full moon. How­ lipped shells. The annual harvest is estimated ever, there is high spatiotemporal variability at 400 t (Paulay 1995). Total edible weight in recruitment. Thalassia may be important in stands at 85.2 t, with 16.61 t of protein, and larval settlement and development by provid­ over 72 million kcal. Like most gastropods, te ing a refuge from predation and acting as a nouo are boiled in large metal pots over an buffer against wave action and currents, pre­ open fire to facilitate meat extraction. The venting settled recruits from being swept large, pointed operculum allows for relatively away. easy extraction with the fingers, whereas The bivalves are located in the intertidal other gastropods usually require the use of a zone by a telltale slit in the sand marking twig or metal pin. In the absence of stone, their presence. Offshore diving for te bun is empty shells are occasionally utilized in earth confined to Tarawa. There are reports from ovens to heat up certain foods. South Tarawa of te bun artificially buried in 84 PACIFIC SCIENCE· January 2001

FIGURE 4. Anadara midden, Tebanga Village, Maiana (photo by F. Thomas). sand and surrounded by large rectangular numbers than those reported in the SGS, coral enclosures near the lagoon shore, simi­ where sampling error is suspected. Paulay lar to those of the tridacnid gardens (Zann (1995) noted an increase in te bun densities 1985). from sea grass (42,000 specimens per ha, Yearly harvest of te bun gathered offshore SD = 68,000) to lagoon slope (78,000 speci­ from Tarawa Lagoon, including export to mens per ha, SD = 136,000). For the highly Majuro and Nauru, has been estimated at mobile te nouo, mean densities of 14,000 1000 t, with an additional 370 t taken from specimens per ha, SD = 59,000, and 67,000 sand flats and sea grass beds (Paulay 1995). specimens per ha, SD = 191,000, were re­ The figures translate into 217.83 t of edible ported for sea grass and lagoon slope, re­ meat, 25.92 t of protein, and more than 121 spectively. million kcal. It should be noted that these Te bun with perforated umbo were recov­ figures are not consistent with the rank order ered archaeologically and may have func­ in Table 1, where Strombus luhuanus, te nouo, tioned as paring knives (Takayama et al. exceeds te bun in all categories. Diver counts, 1990). I was informed that shells with holes average weight of the daily catch, and the knocked through the top of the valve are estimated number of days when diving is pos­ currently used as supports for the alga Eu­ sible (about 275 days per year excluding the chema. Perforations may also result from traditional day of rest, Sl.lnday,Sunday, and the aver­ predators, such as the gastropod Cymatium age duration of stormy weather), however, muricinum (Tebano and Paulay 1995). The indicate that te bun is harvested in greater opportunistic use of discarded valves to scale Mollusk Habitats and Fisheries in Kiribati . Thomas 85

FIGURE 5. Digging for Asaphis, Kariatebike, Abemama (photo by F. Thomas). fish and for grating coconut meat was also of gathering on South Tarawa. Nevertheless, noted. Like te nouo, the shells of te bun are it remains an important secondary catch, with used as heating surfaces in earth ovens. an annual harvest estimated at 45 t (Paulay Gafrarium pectinatum Linnaeus, 1758. 1995), representing 3.92 t of edible meat, This venerid, te koumara, is found in a variety 0.52 t of protein, and close to 3 million kcal. of settings, from shallow, silty environments Te koumara was at the time of my inves­ of the intertidal to sea grass beds and sand­ tigations the prominent mollusk gathered by bars. These gastropods mature at a shell the people of Tebanga Village on Maiana, length of between 16 and 20 mm, reaching a where systematic digging took place. maximum length ofjust over 35 mm (Morton Asaphis violascens (Forslcil,(Forslcll, 1775). The 1990). It yields just under 1 g of edible meat psammobiid, te koikoi, is mainly an upper at about 27 mm in length (Yamaguchi et al. intertidal species, particularly common in 1993). Te koumara is found buried up to gravely sands on both lagoon and ocean sides. 15 cm deep in the sand, has sedentary sub­ The closely related or conspecific Asaphis de­ tidal populations, and reproduces at a low ftorata,florata, found in the western Atlantic, pro­ level throughout the year, with seasonal peaks vides a model for the biology of te koikoi. during spring and fall in Hong Kong man­ The bivalve matures at a length of 25 mm or groves (Morton 1990). The planktonic larval greater and may attain a length of 77 mm stage probably lasts 3 to 4 weeks (see Catterall (Berg and Alatalo 1985). Average edible and Poiner 1987). Except in the Bonriki­ weight is estimated at 9 g. This psammobiid Temaiku area, this species is rarely the focus burrows in sand to a depth of 30 cm. Mass 86 PACIFIC SCIENCE· January 2001

spawning occurs in late summer and the lar­ Trachycardiumangulatum (Lamarck,1822). val stage is about 2 weeks (Berg and Alatalo The cardiid, te koikointari, is found mainly in 1985). Where te koikoi is abundant, harvest­ areas of Thalassia growth and on sandbars. ing involves systematic digging (usually with Adults average 60 mm (Abbott and Dance spoons) (Figure 5). They are also harvested 1982). Mean length recorded in the field is by visually searching the sand for the siphonal 60.4 mm (SD = 10.1, n = 22). This species openings. Te koikoi, and the tellinid Quidni­ yields about 16 g of edible meat. A linear pagus palatam, te nikatona, must be carefully groove in the sand, similar to that of te bun, processed before consumption: the valves are indicates the presence of buried cockles. individually opened with a knife, the sandy However, they may be capable of moving stomach extracted, and the meat is washed in about with short leaps engendered by their fresh water. Discarded valves are still occa­ long foot, as do other Cardiidae (see Kay sionally used as scrapers for removing coco­ 1979). Informants say te koikointari used to nut meat. Te koikoi occurs in areas of the be more abundant at Kariatebike, Abemama, lagoon used for defecation (Paulay 1995) and than it is now, and it is considered rare. It is may be somewhat protected for that reason. not known whether overharvesting caused Nevertheless, on South Tarawa, te koikoi this change or an environmental factor such appears to be below the maximum length that as reduction in Thalassia cover. may be attained by the species, perhaps as a Strombus variabilis athenius Duclos, 1844. result of overharvesting (mean = 53.3 mm, The strombid, te newenewe, is reported from SD = 9.3, n = 101). sandy lagoon flats, but was most often noted Atactodea striata (Gmelin, 1791). The on the seaward reef flat where sand occurs, mesodesmatid, te katura, is abundant in sand and in mangrove-covered passages just below on both leeward and windward beaches. The the low-water mark. The species ranges in shell reaches maturity at about 20 mm and length between 27 and 43 mm (Abbott 1960). may grow to about 30 mm in length (Eisen­ Specimens measured on Tarawa average berg 1981). Te katura is easily accessible, 41.9 mm (SD = 7.9, n = 40). Edible meat often buried at depths of less than 5 cm. The weight is about 2 g. Many of the attributes species has an extended breeding period, with described for Strombus luhuanus such as rapid peaks during the summer months in New maturity, high mobility, a long planktonic Caledonia (Baron 1992). Te katura is among stage, and great capacity for dispersal may the smallest edible mollusk gathered in Kiri­ also be applicable in this species. Numerous bati, with a meat weight of less than 1 g. shells from my archaeological excavations on Because of high densities, however, large Abemama display a broken-off spire, pre­ quantities can be collected in a narrow strip of sumably to extract the meat. The identified sand. pattern may indicate uses other than for hu­ Quidnipagus palatam (lredale, 1929). The man consumption, however. According to tellinid, te nikatona, appears restricted to the local sources, te newenewe damaged in this upper intertidal, both in gravely and silty way may reflect meat extraction for use as fish sands. Mature individuals typically range in bait, notably for bonefish, te ikari, AlbulaA/bula length between 44 and 70 mm (Eisenberg glossodonta. A similar explanation may account 1981). Mean length of specimens measured for the missing spires among other gastro­ on Tarawa and Abemama is 45.5 mm pods, such as te wiaau, Cymatium muricinum. (SD = 7.4, n = 85), with an average edible Cymatiummuricinum (Roding, 1798). The weight of about 3 g. Individuals burrow in ranellid, te wiaau, is found in sandy lagoons at 20-30 cm of sand. Harvesting involves both low intertidal and subtidal depths. It is searching for the siphonal openings, system­ also reported from the deeper portions of atically digging when densities are high, and seaward reefflats (Demond 1957). Maturity is excavating water-covered sand with sweeping reached when the shell attains a length of hand motions. The presence of water re­ about 25 mm (Eisenberg 1981). Average edi­ portedly facilitates digging. ble weight is estimated at 1 g. The larval Mollusk Habitats and Fisheries in Kiribati . Thomas 87 stage may last up to 4 weeks (Clench and erates. The gastropod matures at about Turner 1957). As with most gastropods, this 20 mm in 4 yr and may attain 30 mm (Rich­ prey type is boiled in the shell and the meat ard 1982). The Kiribati samples have a mean pulled from the aperture. However, among length of 12.2 mm (SD = 1, n = 20), with the archaeological remains found on Abe­ an average weight of less than 1 g. Te kaban mama, there is a consistent pattern of shell is most active at night when it feeds upon breakage with the spire missing. This pattern surface algae. During the day, it may find may indicate that the animal was eaten raw or refuge in crevices, where it avoids desiccation. that part of the shell was broken off after it As an archaeogastropod, one would expect was cooked (Thomas 1999). poor dispersal potential and a restricted Spondylus squamosus Schreibers, 1793. range. However, N plicata is very common The spondylid, te koikoinanti, lives firmlyat­ throughout the Indo-Pacific, from East Africa tached to dead coral in intertidal and shallow to as far east as Hawai'i and Clipperton subtidal regions on the ocean side of the (Demond 1957). The widespread distribution motu and in the lagoon where it is found on of a species that would otherwise be expected sandbars. Average adult length is 100 mm to remain within the relatively narrow con­ (Abbott and Dance 1982); mean length of fines of the local environment may relate to local specimens is 75.2 mm (SD = 10.3, n = population dynamics over evolutionary time 6) with an average of 109 of edible meat. scales. For both molluscan and fish pop­ Although sedentary, large te koikoinanti are ulations (Leis 1994, Paulay 1996), it seems often difficult to locate because of the heavy likely that if species find habitat-related ref­ deposition of marine growth on the valves ugia during glacial regressions, their ranges that blends with the environment. The shell would be fragmented. was the preferred species for pendants, which Nerita polita Linnaeus, 1758. Despite were kept as heirlooms (Koch 1986). differences in shell morphology, average size, Timoclea marica (Linnaeus, 1758). The and habitat preference, all neritids are re­ venerid, te koumai, inhabits sand flats and sea ferred to as te kaban. N polita is a high inter­ grass beds on the lagoon reef flat and the la­ tidal species found in sand pockets adjacent to goon floor. Te koumai is usually buried under cemented features and in upper layers of a few centimeters of sand, although some live gravely sands. Adults are between 20 and individuals were found on the sand surface 40 mm in length (Eisenberg 1981). The completely exposed at low tide. The species mean for Kiribati is 22.3 mm in length matures at between 15 and 20 mm (Dance (SD = 2, n = 23), with an edible meat weight 1976). Mean length of Kiribati specimens is of just under 1 g. During the day, the gastro­ 20.6 mm (SD = 1.9, n = 40), with a meat pod remains buried about 2.5 cm deep in weight of less than 1 g. Like te koumara, the sand. At dusk, it emerges from the sand and planktonic larval stage may be 3 to 4 weeks feeds on algae occurring on cemented fea­ (Catterall and Poiner 1987). tures (Demond 1957). Like Nerita plicata, this Pitar prora (Conrad, 1837). The venerid, neritid is widespread in the Indo-Pacific. te nikabibi, occurs in sand flats and sea grass Polinices tumidus (Swainson, 1840). The beds and may extend farther offshore. Adults naticid, te tumara, occurs in intertidal sand average 40 mm (Abbott and Dance 1982). flats and sandbars. Individuals are typically Mean length for Kiribati is 33.8 mm (SD = between 25 and 75 mm in length (Eisenberg 8.8, n = 16), with an average meat weight of 1981). Mean length in field specimens is about 2 g. Attributes of this species' life his­ 31.9 mm (SD = 4.1, n = 40), with an average tory and habitat information may be inferred weight of 3 g. As a mesogastropod, it may from those of other venerids, te koumara and have a relatively long larval stage with good te koumai. dispersal potential. Te tumara is usually har­ Nerita plicata (Linnaeus, 1758). The ner­ vested when the sand is thorougWy exposed itid, te kaban, occurs in the high intertidal at low tide and the gastropod's characteristic zone on beachrock and high-level conglom- track is clearly visible. The animal moves 88 PACIFIC SCIENCE· January 2001 quickly with the aid of a large foot, plowing length of individuals gathered on Abemama is up sand as it feeds. The snail marks its pres­ 37.9 mm (SD = 3.5, n = 20), with an average ence by a ridge at one end ofthe track, where meat weight of less than 1 g. As with most itit can be·be dug less than 5 cm below the sur­ Neogastropoda, it may be assumed that thethe face. The naticid is primarily exploited for its free-swimming larval stage is relatively brief. shell toto make dancing belts (Koch 1986). Turbo setosus Gmelin, 1791. The turbi­turbi­ Polinices melanostomus (Gmelin, 1791). nid, te nimitanin, is mainly found on the reef This naticid, also termed te tumara, occurs in crest (Richard 1982). Adults are usually be­ thethe same habitat as P. tumidus, in intertidal tween 40 and 60 mm long, occasionally ex­ sand flats and sandbars. It differs from P. tu­ ceeding 90 mm, and yield about 4 g of edible midus in having a smaller size range and is meat (Demond 1957). The free-swimming readily distinguishable by its dark brown col­ stage among turbans is short, with larvae re­ umella (the columella is white in P. tumidus maining for less than a week as plankton [Kay 1979]). Adults range between 18 and (Yamaguchi 1989). Subsequent growth isis rel­ 46 mm (Eisenberg 1981), and average weight atively slow, with individuals attaining 95% of isis about 2 g. the shell height in 5 yr (Villiers and Sire Oliva miniacea R6ding, 1798. The olivid, 1985). These gastropods are more accessible tete burebangaki, shares the same habitat as te during low spring tidestides than at neap tidestides tumaratumara in the sand flats and sandbars. This when people can walk greater distances across species isis one ofthe largest ofthe olive shells, the seaward reef flat. At full moon, they re­ with lengths approaching 100 mm. Maturity portedly migrate from their shelter inin coral isis usually reached at 50 mm, but may also cracks to the surface. Shells may also be occur at a length of less than 40 mm among gathered from fish traps, where they find dwarf populations (Petuch and Sargent 1986). suitable habitats between coral slabs. FreshlyFresWy Mean length for Kiribati samples is 53 mm caught individuals are boiled in the shell toto (SD(SD = 4, n = 24), with an average edible release the solid, calcareous operculum. Ban­ weight of 4 g. Larvae probably develop rap­ ner and Randall (1952) noted that thethe shell idly,idly, as would be expected for neogastropods. was also broken and the animal eaten raw. A These sand dwellers also leave well-defined small portion ofthe meat was commonly used trackstracks in the sand and lie just beneath the for fish bait (Cooper 1964). Some of the per­ surface at a spot usually marked by a ridge forated shells excavated on the northernmost with small cracks. Olives are carnivorous and island ofMakin may have been used as paring especially active at night (Eisenberg 1981), knives or net sinkers (Takayama and Taka­ although no night foraging activities for this sugi 1988). species were documented. The shells of te Hermit crabs, te makauro, Coenobita per­ burebangaki are occasionally strung together latus, commonly find refuge in empty tete as dancing belts. nimitanin shells and are gathered for con­ Barbatia foliata (F(Forskal,orskal, 1771775).5). The sumption. The activity is nocturnal, usually arcid, te nikarinei, isis found attached to dead occurring along windward beaches in and coral in shallow areas of the lagoon. There is around strand vegetation. Some hermitted littlelittle information on this species. Adults shells can be distinguished by their eroded average 50 mm inin length (Abbott and Dance appearance both inside and out. However, 1982); in Kiribati mean length is 53.1 mm archaeological sites containing operculae (SD = 16.2, n = 28), with an average of just suggest that at least some of the shells were under 4 g of edible meat. gathered for their resident gastropod, rather Vasum turbinellum (Linnaeus, 1758). The than for crab meat. vasid, te nimakaka, is gathered from the Lambis lambis (Linnaeus, 1758). The windward reef flats and is common in tide strombid, te ang, is the most common and pools. Populations extend at least as far as the one of the most widely distributed species of reef margin. Adults are typically between 44 Lambis, ranging between 92 and 200 mm inin and 85 mm in length (Abbott 1959). Mean length (Abbott 1961). The mean shell lengthlength Mollusk Habitats and Fisheries in Kiribati . Thomas 89 in the Abemama sample, albeit from a small other tridacnids, amount of meat is highly sample, is 260 mm (SD = 30.3, n = 4). Meat variable, with a mean of 1050 g (SD = 250, weight averages 280 g (SD = 61.5, n = 43). n = 7). This species is generally found on Te ang lives in colonies from the low-tide sand among corals on shallow reefs to a depth zone to depths of about 3 m on sand adjacent of about 20 m. Compared with other species to leeward reefs. Yamaguchi (1989) reported of giant clams, te kima grows rapidly to a larval stage of about 3 weeks and juveniles maturity, producing eggs at about 500 mm that grow fairly rapidly on small algae. One (Munro 1993). Large specimens are capable informant on Abaiang claimed that te ang re­ of releasing hundreds of millions of eggs in a mains mostly buried during the first and last single day, but with reductions in the number quarters of the moon phase. Te ang is col­ of mature individuals, the proportion of eggs lected by men, who must dive for it. It is most that are not fertilized increases dramatically often roasted in the shell, which is then bro­ (Tekinaiti 1990). During their week-long ken up and the animal removed, but the dor­ free-swimming period, larvae suffer high sal section of the shell may also be broken mortality rates (Heslingao-Ieslinga and Fitt 1987). with a hammer and the strombid eaten raw. As with te bun and te ang, diving for te Tridacna maxima (Roding, 1798). This kima is a task accomplished by men. For large bivalve, te were, is one offour species of giant specimens, a steel anchor pole may be used to clams found in the Gilberts. It occurs on reef pierce the adductor muscle or an oar prevents structures, partly embedded in coral or firmly the valves from closing before the muscle is attached to coral heads by byssal threads cut with a bush knife, releasing the valves. throughout its adult life. Te were is not fully The meat is scraped out of the shell, which mature until it attains a shell length of 60­ may be left in place or carried home. 80 mm (Munro 1993). Amount of edible Small specimens «40 em)cm) are occasionally meat is more variable than in some of the transferred to shallow lagoon reef flats or de­ other species examined, with an average of posited in passes adjacent to households. In 33 g (SD = 27, n = 30). Spawning may be place, they are allowed to grow until ready for continuous near the equator (Munro 1989). consumption. This custom has been observed Despite very high fecundity, tridacnid larvae in other oceanic communities (see Moir 1989, suffer considerable losses during their brief 6­ Foster and Poggie 1993, Hviding 1993). Like to 12-day planktonic period, with populations fish traps, and areas used for Euchema farm­ maintaining themselves primarily through lo­ ing, giant clam gardens are the property of cal recruitment (Alcazar and Solis 1986). Te individual households, whereas the reef flat were is often firmly attached and is com­ on both lagoon and ocean sides is now monly dislodged by puncturing the adductor regarded as common property. Giant clam muscle and prying the animal free of its bys­ gardens are marked by a circular coral enclo­ sus before the valves shut. Older individuals sure about 40 emcm high, coral rubble, or fish tend to grow out of their coral matrix, easing trap. A few are not marked at all. Concen­ the task of detaching the shells. Like other trations in the gardens range from one or two species of giant clams, they were used tradi­ specimens to larger aggregates; they may tionally to manufacture adzes, but were consist of large individuals, but more com­ quickly replaced by metal tools once they be­ monly small bivalves are used. Because of came available. Empty shells are still used for pilferage in populated areas, giant clam gar­ mashing leaves for local medicines. Giant dens are now confined to relatively isolated clams are regularly exported from Abaiang to locations, such as the motus of the leeward Tarawa. They are also important in the sub­ reef. The gardens are disappearing from the sistence economy ,of Tabiteuea. Kiribati seascape, however, in large part as a Tridacna gigas (Linnaeus, 1758). This gi­ reflection of the erosion of customary marine ant clam, te kima, is the largest living species tenure. Owners are now less inclined to con­ ofbivalve, attaining a shell length ofover 1 m tinue to care for giant clams in designated and weighing in excess of 200 kg. As with lagoon sections or to invest in maintaining 90 PACIFIC SCIENCE· January 2001

FIGURE 6. Giant clamclanI "garden" (Tridacna gigas, Hippopus hippopus), Abatiku, Abemama (photo by F. Thomas). large fish traps (Figure 6). Informants claim giant clam, te neitoro, is found on sandy sub­ that both te kima and te neitoro, Hippopus strata in intertidal areas to depths of about hippopus, from the gardens are harvested only 10 m. Maturity is attained at a length ofabout during periods of inclement weather, when 140 mm or when the shell is no longer at­ fishing is difficult or when no other meat tached to coral rubble. Edible meat weight source is available. Taniera (1988) reported averages 156 g (SD = 316, n = 26). Like te that they were also consumed at feasts. were and te kima, the larval stage is short. Te kima provided the raw material for Settlement of larvae usually occurs within a adzes, chisels, gouges, and scrapers (Taka­ week after fertilization (Alcala et al. 1986). Te yama et al. 1990), and the shell was appar­ neitoro is often found in shallow, nearshore ently used as a container for cooking (Catala patches, where it may be completely exposed 1957). Contemporary uses of empty te kima during spring low tides. Under these con­ valves include troughs for feeding pigs, for ditions, the valves are shut tight, and the ani­ crushing certain foods, and as shrine compo­ mal is nearly camouflaged in the midst of nents. Te kima gathering, whether for im­ dead coral. Neap tides, when the mantle is mediate or delayed consumption, is usually visible, provide the best opportunities to spot incidental to other activities, such as spear­ these bivalves. Like te kiilla,kiina, te neitoro gath­ fishing, and may at times become a risky ering is normally incidental to other fishing venture, if the fisherman is near a leeward activities, but it may become a focal prey type, pass where strong currents and large marine particularly in some sections of the leeward predators may be present. reef platform at low tide where foragers are Hippopus hippopus (Linnaeus, 1758). The able to walk on the reef. The shells were used Mollusk Habitats and Fisheries in Kiribati . Thomas 91 as ax and adze blades, hammerstones, and sion of other species may give the impression containers, the last continuing today (Koch that although foragers are targeting certain 1986, Takayama and Takasugi 1988). Live prey, they essentially gather all those that specimens were also placed in lagoon gardens they consider edible. In her study of Austra­ with te kima. lian Aboriginal subsistence, Meehan (1982) addressed some of the factors leading to a departure from targeting a single mollusk DISCUSSION species. Some of her explanations appear The degree of lagoon closure plays an im­ sensible, for example, differences in foraging portant role in determining the number of ability among participants, but her conclusion species that can be found (Salvat 1969, 1972, that among the Anbarra shell gathering is es­ Kay 1971, 1978, Kay and Switzer 1974). sentiallya random process is in sharp contrast These patterns may not apply in the case of to Bird's (1996) demonstration among the Tarawa and other open atolls influenced by Meriam of Torres Strait that shell-gathering nutrient-rich waters of the equatorial upwell­ activitiescan be explained in terms of the ing area, resulting in unusually high mollus­ variable costs and benefits that foragers face can population densities, together with high while collecting prey. species diversity (Paulay 1995). The significance of size at maturity lies in A substantial rise in mollusk consumption its relationship with size below which indi­ has occurred on South Tarawa, a phenome­ viduals are rejected by foragers. A population non linked to changes in water circulation will always contain reproducing individuals if when causeways were built in the 1960s, en­ size at maturity is less than the rejection size. couraging the establishment of certain species However, the population is at risk if size at and increased fertilization by sewage-driven maturity exceeds the rejection size. Some nutrients. Filter feeders exposed to con­ targeted individuals may escape detection if taminated water may become agents of gas­ they find refuge by burying themselves. A trointestinal diseases, particularly ifconsumed depleted intertidal population may be re­ raw or partially cooked (Danielson et al. plenished through recruitment of individuals 1995). It appears that shellfish contamination from adjacent subtidal areas that are usually resulted in the abandonment of enclosures. inaccessible to traditional gatherers. Repro­ In Kiribati, communities target a limited ductive individuals with partially subtidal dis­ range of marine invertebrates, although mol­ tributions may migrate into the exploited luscan catches often include several species, intertidal zone if they are mobile during their which appear as incidental harvests. Paulay benthic stages. If the duration of the larval (1995) found that Strombus luhuanus, Anadara stage is sufficiently long, then it is possible for uropygmelana,uropyp;melana, and Gafrarium pectinatumpectinamm oc­ larvae spawned by more distant unexploited curred in 67,67,59,59, and 47%, respectively, ofthe populations to settle into the exploited areas. South Tarawa sand flat and sea grass catches, Prey resilience provides an alternative ex­ but 61 % of the catches were dominated (that planation for the apparent balance between is constituting greater than 75% of gross traditional societies and their resources (Cat­ catch weight) by S. luhuanus and 26% by A. terall and Poiner 1987, Poiner and Catterall uropJp;melana.uropygmelana. The current observed and in­ 1988). According to the tenets of human be­ terview data confirm the importance of these havioral ecology (Cronk 1991, Smith and species, with A. uropygmelana,uropyp;melana, S. luhuanus, Winterhalder 1992), if people are behaving to and G. pectinatumpectinamm occurring in 60, 59, and maximize their short-term harvest rate and 33%33 % of sand flat and sea grass catches, re­ vulnerable species fall in the optimal diet, a spectively. Seventy-eight percent of the diminished-returns curve, as assumed in the catches have dominant species (as defined "marginal value theorem" (MVT) (Charnov above, except that weight is for edible meat), 1976), could lead to overexploitation and ex­ with A. uropJp;melanauropygmelana (19%), S. luhuanus tinction ofprey, though this is admittedly less (53%), and G. pectinatum (28%). The inclu- likely than with a step-function curve (where 92 PACIFIC SCIENCE· January 2001

thethe expected or marginal harvest rate sud­ in water circulation associated with strong denly drops to zero). However, if a given currents may have disturbed the nutrient-rich patch type has a very high return curve rela­ substratum inhabited by these bivalves and tivetive toto other patch types, and some of the larvae washed away. The extent of these dis­ prey typestypes in that patch type are not resilient turbances on overall te bun stocks isis difficult toto overexploitation, human foragers following to assess, but the paucity of small individuals thethe MVT are quite capable of causing local «20 mm) in other areas surveyed by Tebano extinction. With the most vulnerable prey suggests a decline in population densities. In­In­ gone or greatly reduced in number, most re­ formants at Tebanga, about 8 km from Raw­ maining species presumably possess biological eai, also stated that te bun were considerably attributes that make them resilient to human less abundant compared with a few years ago. predation. Unfortunately, little is known of With less pressure on lagoonal resources, thethe response of many tropical mollusk pop­ Abemama's te bun stocks are not in apparent ulations to gathering practices, in part at least danger of reduced recruitment, and densities because of the paucity of detailed information ofmore than 20 individuals per meter are said on lifelife history and ecology (Catterall and to be fairly common. Tebano (1990) reported Poiner 1987, Wells 1989, Munro 1999). Al­ that exploitation mainly took place during thoughthough it has not been possible to compile periods of inclement weather when fishingfishing detailed information for every prey type, cer­ was hazardous. taintain species are clearly at risk of deple­ Te nouo are conspicuously absent at Mai­ tion,tion, even following a regime of traditional ana and none was located in surface middens. gathering. Older informants did not recall a time when The influx of human populations from the this species was present. Because habitats outer islands to Tarawa, together with a high suitable for this gastropod exist, ititmay be thatthat birth rate, has resulted in the decline (in size the absence of the species is linked toto chance and abundance) of easily accessible intertidal dispersal or local extinction, resulting inin itsits resources, particularly te bun, which is often discontinuous distribution (see Kay 1995). thethe main source of protein for unemployed The resilience of te nouo to traditional and low-incomelow-income families. During the 1993­ gathering tends to rule out overexploitation. 1994 field season, the roadside sale of te bun In contrast to Maiana, the strombid isis fast gathered offshore was putting pressure on becoming the dominant molluscan species inin stocks on Tarawa that were once relatively Tarawa Lagoon, perhaps as a consequence of well protected from harvesting. Divers pos­ the demise of many reefs linked to changes inin sessing a floating device (e.g., canoe, rubber sedimentation (Paulay 1995). The high vari­ innerinner tube) and goggles could collect large ability of species density in sea grass beds re­ quantities of the bivalve on a daily basis and ported by Yamaguchi et al. (1993) (254,000/ sell theirtheir catch in rice bags holding up to ha) and Paulay (1995) (14,OOO/ha)(14,000/ha) may reflect 34 kg of te bun. With approximately 16 mil­ increasing gathering pressure, but could also lionlion tete bun taken annually (before 1995), and be attributed to patchy distribution, perhaps an estimated population of 40 million in the linked to the species' high mobility. Never­ heavily fished Taborio-Bangantebure bed, or theless, a small average shell length for Ta­ 70 million from Nanikai to Bikenibeu, the rawa sand flats (mean = 42.9 mm, SD = 4.9, harvest represented about one-third to one­ n = 206) may indicate harvesting pressure. fifth of the total adult te bun population. On Yap, Kay and Smalley (1989) reported a These figures suggested that harvests were mean length of 47.9 mm (range, 41-54, near, if not exceeding, sustainable rates (Pau­ n = 18). Size range in the Pacific isis 32.2­ laylay 1995). 69.5 mm, with an average of 56.5 mm (Ab­ According to Tebano (1990), reef blasting bott 1960). of Maiana's leeward reef opposite Raweai Current and future leaders will continue toto Village may have contributed to the recent be challenged by the often conflicting goals of disappearance of tete bun in that area. Changes implementing policies conducive to economic Mollusk Habitats and Fisheries in Kiribati . Thomas 93 development and of assuring sustainable use Terry L. Hunt, W. Bruce Masse, and Leslie of the environment. The second goal could E. Sponsel, for their guidance and comments be achieved by applying extant TEK for new on the dissertation from which this paper biological and ecological insights for pro­ stems. tected areas and for conservation education, for development planning and environmental for development planning and environmental Literature Cited assessment, and of course for resource man­ agement (see Berkes 1993). Abbott, R R, and].and J. Garcia, eds. 1995. Man­ The linkage between TEK and resource agement plan for Tarawa Lagoon, Repub­ management, expressed through various lic of Kiribati. Vol. III. Management plan. forms of tenurial systems, has received con­ BioSystems Analysis, Inc., Santa Cruz, siderable attention in the past few decades California. (Crocombe 1989, Johannes 1989). Many Abbott, RT 1959. The family Vasidae in problems remain, however, in implementing the Indo-Pacific. Indo-Pac. Mollusca 1 (1): TEK to serve conservation initiatives and re­ 15-32. lated goals. In addition, there is the challenge ---. 1960. The genus Strombus in the of how TEK and scientific information can Indo-Pacific. Indo-Pac. Mollusca 1 (2): be integrated, and whether such integration is 33-174. desirable, given unequal power sharing in ---.---.1961.1961. The genus Lambis in the Indo­ decision making. The fisheries sector in Kiri­ Pacific. Indo-Pac. Mollusca 1 (3): 147­ bati has been described as both a source of 174. essential livelihood at the subsistence level Abbott, R T, and S. P. Dance. 1982. Com­ and as a means to achieve greater economic pendium of seashells. E. P. Dutton, New independence by promoting the country's York. seemingly vast store ofmarine resources. The Alcala, A. c., E. P. Solis, and S. N. Alcaza. impressive number of feasibility studies, with 1986. Spawning, larval rearing and early an emphasis on finfish (Gillett et al. 1991), growth of Hippopus hippopus (Linn.) (Bi­ should furnish data for making informed valvia: Tridacnidae). Silliman].Silliman J. 33:45-53. choices. Alcazar, S. N., and E. P. Solis. 1986. Spawn­ The sustainabilitysustainability of mollusk fisheries in ing, larval development and growth of Kiribati rests in large part on a remodeled Tridacna maxima (Bivalvia: Tridacnidae). form ofmarine tenure (Johannes and Yeeting SillimanJ.Silliman]. 33:65-73. 1995). To achieve this goal, local fishers Banner, A. H., and].and J. E. Randall. 1952. Pre­ should be encouraged to pursue activities to liminary report on marine biology study, generate income that could at the same time Onotoa Atoll, Gilbert Islands. Atoll Res. provide a means to ensure the long-term use Bull. 13:1-62. of selected high-value resources, such as pearl Baron,].Baron, J. 1992. Reproductive cycles of the shell and giant clam farming, using a combi­ bivalve molluscs Ataetodea striata (Gmelin), nation of traditional knowledge and modern Gafrarium tumidum and Anadara scapha technology. Mollusks are especially suited to (L.) in New Caledonia. Aust. ].J. Mar. efforts aimed at establishing limited entry, Freshwater Res. 43:393-402. because they are mostly sedentary and rela­ Berg, C. J., and P. Alatalo. 1985. Biology of tively predictable. Therefore, those resources the tropical bivalve Asaphis deftorata (Linn., are most economically defendable by various 1758). Bull. Mar. Sci. 37:827-838. cooperative units already engaged in com­ Berkes, F. 1993. Traditional ecological munity projects. knowledge in perspective. Pages 1-9 in ].J. T Inglis, ed. Traditional ecological knowledge: Concepts and cases. Inter­ ACKNOWLEDGMENTS national Development Research Centre, I thank my doctoral committee members, P. Canadian Museum of Nature, Ottawa. Bion Griffin, E. Alison Kay, Harry Ako, Bird, D. W. 1996. Intertidal foraging strategies 94 PACIFIC SCIENCE· January 2001

among the Meriam of the Torres Strait Republic ofKiribati. Vol. III. Management Islands, Australia: An evolutionary ecolog­ plan. BioSystems Analysis, Inc., Santa ical approach to the ethno-archaeology of Cruz, California. tropical marine subsistence. Ph.D. diss., Demond,]. 1957. Micronesian reef-associated University of California, Davis. gastropods. Pac. Sci. 11:275-341. Catala, R L. A. 1957. Report on the Gilbert Eisenberg, ]. M. 1981. Seashells of the world. Islands: Some aspects of human ecology. McGraw-Hill, New York. Atoll Res. Bull. 59:1-187. Foster, K. B., and].]. PoggieJr. 1993. Cus­ Catterall, C. P., and 1.I. R Poiner. 1987. The tomary marine tenure and mariculture potential impact of human gathering on management in outlying communities of shellfish populations, with reference to Pohnpie [sic] State, Federated States of some NE Australian intertidal flats. Oikos Micronesia. Ocean Coast. Manage. 20: 1­ 50:114-122. 22. Cernohorsky, W. O. 1971. Marine shells of Gillett, R, M. Pelasio, and E. Kirschner. the Pacific. Pacific Publications, Sydney. 1991. Kiribati fisheries bibliography. ---. 1972. Marine shells of the Pacific. FAO/UNDP Regional Fishery Support Vol. II. Pacific Publications, Sydney. Programme, Suva. Charnov, E. L. 1976. Optimal foraging: The Gillie, RD. 1993. Coastal erosion problems marginal value theorem. Theor. Popul.Populo in the Gilbert Islands Group, Republic of BioI. 9: 129-13129-136.6. Kiribati. SOPAC Technical Report No. Clench, W.]., and RD.R D. Turner. 1957. The 167. Suva. family Cymatiidae in the western Atlantic. Heslinga, G. A., and W. K. Fitt. 1987. The Johnsonia 3:189-244. domestication of reef-dwelling clams. Bio­ Cloud, P. E. 1952. Preliminary report on the Science 37:332-339. geology and marine environments of Hviding, E. 1993. The rural context of giant Onotoa Atoll, Gilbert Islands. Atoll Res. clam mariculture in Solomon Islands: An Bull. 12:1-73. anthropological study. ICLARM Techni­ Cooper, M. ]. 1964. Ciguatera and other cal Report No. 39. Manila. marine poisoning in the Gilbert Islands. Johannes, R E. 1989. Criteria for determin­ Pac. Sci. 18:411-440. ing the value of traditional marinemarine· tenure Crocombe, R G. 1989. Tomorrow's cus­ systems in the context of contemporary tomary tenure in relation to conservation. marine resource management in Oceania. Pages 21-27 in South Pacific Regional Pages 29-32 in South Pacific Regional Environment Programme. Report of the Environment Programme. Report of the workshop on customary tenure, tradi­ workshop on customary tenure, traditional tional resource management and nature resource management and nature con­ conservation. South Pacific Commission, servation. South Pacific Commission, Noumea. Noumea. Cronk, L. 1991. Human behavioral ecology. Johannes, R E., and B. Yeeting. 1995. I­ Annu. Rev. Anthropol. 20:25-53. Kiribati knowledge and traditional man­ Dance, S. P. 1976. The collector's encyclo­ agement of lagoon resources. In RR Ab­ pedia of shells, 2nd ed. McGraw-Hill, bott and]. Garcia, eds. Management plan New York. for Tarawa Lagoon, Republic of Kiribati. Danielson, R E., RR Abbott, W. ]. Vol. III. Management plan. BioSystems Kimmerer, and K. Etauti. 1995. Micro­ Analysis, Inc., Santa Cruz, California. biological contamination of water and Kay, E. A. 1971. The littoral marine molluscs shellfish in Tarawa Lagoon, Kiribati: Re­ of Fanning Island. Pac. Sci. 25:260-281. mediation strategies for contaminated ---. 1978. Molluscan distribution pat­ shellfish. In RR Abbott and]. Garcia, terns at Canton Atoll. Atoll Res. Bull. eds. Management plan for Tarawa Lagoon, 221:159-170. Mollusk Habitats and Fisheries in Kiribati . Thomas 95

---. 1979. Hawaiian marine shells: Reef Veneridae) in a Hong Kong mangrove. and shore fauna ofHawaii. Bishop Museum Malacol. Rev. 23:53-62. Press, Honolulu. Motteler, L. S. 1986. Pacific Island names. ---. 1995. Diversification and differentia­ Bishop Mus. Misc. Publ. 34. tion: Two evolutionary patterns in the Munro, ]. L. 1989. Fisheries for giant clams molluscan fauna of the Pacific Islands with (Tridacnidae: Bivalvia) and prospects for consequences for conservation. Pages 37­ stock enhancement. Pages 541-558 in]. F. 53 in A e. Brugen, S. M. Wells, and Th. Caddy, ed. Marine invertebrate fisheries: e. M. Kemperman, eds. Biodiversity and Their assessment and management. John conservation of the Mollusca: Proceedings Wiley & Sons, New York. of the Alan Solem memorial symposium ---.1993. Giant clams. Pages 431-449 in on the biodiversity and conservation of the A Wright and L. Hill, eds. Nearshore Mollusca. Backhuys, Oegstgeest-Leiden. marine resources of the South Pacific: In­ Kay, E. A, and S. Johnson. 1987. Mollusca of formation for fisheries development and Enewetak Atoll. Pages 105-146 in D. M. management. Institute of Pacific Studies, Devaney, E. S. Reese, B. L. Burch, and P. University of the South Pacific, Suva. Helfrich, eds. The natural history of Ene­ ---. 1999. Utilization of coastal mollus­ wetak Atoll. Vol. II. Biogeography and can resources in the tropical insular Pacific systematics. U.S. Department of Energy, and its impacts on biodiversity. Pages 127­ Office of Energy Research, Office of 144 in L. G. Eldredge,]. E. Maragos, P. F. Health and Environmental Research, Holthus, and H. F. Takeuchi, eds. Marine Ecological Research Division, Oak Ridge, and coastal biodiversity in the tropical Tennessee. island Pacific region. Vol. 2. Population, Kay, E. A, and T. L. Smalley. 1989. On the development and conservation priorities. feasibility of a shell trade for the State of Pacific Science Association, Honolulu. Yap. Pacific Business Center Program, Paulay, G. 1995. Benthic ecology and biota of University ofHawai'i, Honolulu. Tarawa Lagoon: Influence of equatorial Kay, E. A, and M. F. Switzer. 1974. Mollus­ upwelling, circulation, and human preda­ can distribution patterns in Fanning Island tion. In RR Abbott and]. Garcia, eds. lagoon and a comparison of the mollusks Management plan for Tarawa Lagoon, of the seaward reefs. Pac. Sci. 28:275­ Republic ofKiribati. Vol. m.III. Management 295. plan. BioSystems Analysis, Inc., Santa Koch, G. 1986. The material culture of Kiri­ Cruz, California. bati. Translated by Guy Slatter. Institute ---. 1996. Dynamic clams: Changes in of Pacific Studies, University of the South the bivalve fauna of Pacific Islands as a re­ Pacific, Suva. sult of sea-level fluctuations. Am. Malacol. Leis, ]. M. 1994. Coral sea atoll lagoons: Bull. 12:45-57. Closed nurseries for the larvae of a few ---. 1997. Benthic ecology and biota of coral reef fishes. Bull. Mar. Sci. 54:206­ Tarawa Lagoon: Influence of equatorial 227. upwelling, circulation, and human pre­ Meehan, B. 1982. Shell bed to shell midden. dation. Marine Laboratory, University of Australian Institute for Aboriginal Studies, Guam (revised paper, originally published Canberra. 1995). Moir, B. G. 1989. Mariculture and material Petuch, E.]., and D. M. Sargent. 1986. Atlas culture on Takuu Atoll. Ph.D. diss., Uni­ of the living olive shells of the world. versity of Hawai'i at Manoa, Honolulu. Coastal Education and Research Founda­ University Microfilms International, Ann tion, Fort Lauderdale, Florida. Arbor, Michigan. Poiner, I. R, and e. Catterall. 1988. The Morton, B. 1990. The life cycle and sexual effects of traditional gathering on popula­ strategy of Gafrarium pectinatum (Bivalvia: tions of the marine gastropod Strombus lu- 96 PACIFIC SCIENCE· January 2001

huanus Linne [sic] 1758, in southern Papua create a fluctuating resource: The biology New Guinea. Oecologia (Berl.) 76:191­ of Anadara uropygmelana (Bivalvia: Arci­ 199. dae) in Tarawa. In R.R. Abbott and ].J. Richard, G. 1982. Mollusques lagunaires et Garcia, eds. Management plan for Tarawa recifaux de Polynesie fran~aise: Inventaire Lagoon, Republic of Kiribati. Vol. m. faunistique-bionomie-bilan quantitatif Management plan. BioSystems Analysis, -croissance-production. Ph.D. thesis, Inc., Santa Cruz, California. Universite Pierre et Marie Curie, Paris. Tebano, T, T Taniera, T Tabe, and E. Richmond, B. 1993. Reconnaissance geology Taraku. 1993. Akawa ni Kiribati (Tradi­ of the Gilbert Group, western Kiribati. tional fishing methods of Kiribati). Marine SOPAC Technical Report No. 77. Suva. Studies Programme, University of the Salvat, B. 1969. Dominance biologique de South Pacific, Suva. quelques mollusques dans les atolls fermes Tekinaiti, T.T 1990. Status of giant clam stock (Tuamotu, Polynesie): Phenomenes re­ at the four atolls in Kiribati: Abaiang, cents-consequences actuelles. Malacolo­ Maiana, Tarawa, Abemama. Fisheries Di­ gia 9:187-189. vision, Tarawa. ---. 1972. La faune benthique du lagon Thomas, F. R. 1999. Optimal foraging and de l'atoll de Reao (Tuamotu, Polynesie). conservation: The anthropology of mol­ Cah. Pac. 16:31-110. lusk gathering strategies in the Gilbert Smith, E. A., and B. Winterhalder, eds. 1992. Islands Group, Kiribati. Ph.D. diss., Uni­ Evolutionary ecology and human behavior. versity of Hawai'i at Manoa, Honolulu. Aldine de Gruyter, New York. University Microfilms International, Ann Takayama, ].,J., and H. Takasugi. 1988. Ar­ Arbor, Michigan. chaeology on Makin, Kiribati, Central ViIIiers,Villiers, L., and].-and J.-Y. Sire. 1985. Growth and Pacific. Department of Ethnology, Tezu­ determination of individual age of Turbo kayama University, Nara-shi, Japan. setosus (Prosobranchia: Turbinidae), Hao Takayama,].,Takayama, J., H. Takasugi, and K.K Kaiyama. Atoll (Tuamotu, French Polynesia). Pro­ 1990. Test excavation of the Nukante­ ceedings of the Fifth International Coral kainga site on Tarawa, Kiribati, Central Reef Congress, Moorea 5:165-170. Pacific. Pages 1-19 in 1. Ushijima, ed. Ar­ Wells, ].J. W. 1957. Coral reefs. Pages 609­ chaeological research on atoll cultures of 631 in].in J. W. Hedgpeth, ed. Treatise on Micronesia, 1988. University of Tsukuba, marine ecology and paleoecology. Vol. 1. Tsukuba-shi, Japan. Ecology. National Research Council, Na­ Taniera, T 1988. Status of giant clams in tional Academy of Sciences, Washington, Kiribati. Pages 47-48 in ].J. W. Copland D.C. and].and J. S. Lucas, eds. Giant clams in Asia Wells, S. M. 1989. Impacts of the precious and the Pacific. ACIAR, Canberra. shell harvest and trade: Conservation of ---. 1994. Traditional fisheries in Kiri­ rare and fragile resources. Pages 443-454 bati: Survival and sustainability. Pages in ].J. F. Caddy, ed. Marine invertebrate 113-120 in].in J. Morrison, P. Geraghty, and fisheries: Their assessment and manage­ L. Crowl, eds. Science of Pacific island ment. John Wiley & Sons, New York. peoples. Vol. 1. Ocean and coastal studies. Yamaguchi, M. 1989. Report on a baseline Institute of Pacific Studies, University of study for fisheries development in Oce­ the South Pacific, Suva. ania, with special reference to sedentary Tebano, T 1990. Some aspects ofpopulation organisms on coral reefs and lagoons. De­ biology and ecology of the cockle Anadara partment of Marine Sciences, University maculosa in the Gilbert Islands, Republic of of the Ryukyus, Okinawa. Kiribati. M.S. thesis, James Cook Univer­ Yamaguchi, M., B. Yeeting, K.K Toyama, T sity, Townsville, Australia. Tekinaiti, and T Beiatau. 1993. Report on Tebano, T, and G. Paulay. 1995. Variable a preliminary study of molluscan resources recruitment and changing environments in Tarawa Lagoon, with special reference Mollusk Habitats and Fisheries in Kiribati . Thomas 97

to te bun Anadara maculosa. Department of conservation of fisheries in Kiribati and Marine Sciences, University of the Ryu­ Tuvalu atolls. Pages 53-77 in K. Ruddle kyus, Okinawa. and R. E. Johannes, eds. The traditional Zann, L. P. 1982. The marine biology of knowledge and management of coastal Betio Island, Tarawa Atoll, Republic of systems in Asia and the Pacific. Kiribati. Institute of Marine Resources, UNESCO/Regional Office forofor, Science University of the South Pacific, Suva. and Technology for Southeast Asia, ---. 1985. Traditional management and Jakarta.