m m-i JOURNAL OF SHELLFISH RESEARCH VOLUME 17, MUMBER 4 DECEMBER 1998 The Journal of Shellfish Research (formerly Proceedings of the National Shellfisheries Association) is the official publication of the National Shellfisheries Association Editor Dr. Sandra E. Shumway Natural Science Division Southampton College. Long Island University Southampton, NY 11968 Dr. Standish K. Allen, Jr. (1998) School of Marine Science Virginia Institute of Marine Science Gloucester Point, VA 23062-1 1346 Dr. Peter Beninger (1999) Laboratoire de Biologic Marine Faculte des Sciences Universite de Nantes BP 92208 44322 Nantes Cedex 3 France Dr. Andrew Boghen (1999) Department of Biology University of Moncton Moncton, New Brunswick Canada ElA 3E9 Dr. Neil Bourne (1999) Fisheries and Oceans Pacific Biological Station Nanaimo, Briti.sh Columbia Canada V9R 5K6 Dr. Andrew Brand (1999) University of Liverpool Marine Biological Station Port Erin. Isle of Man Dr. Eugene Burreson (1999) Virginia Institute of Marine Science Gloucester Point. Virginia 23062 Dr. Peter Cook (1998) Department of Zoology University of Cape Town Rondebosch 7700 Cape Town. South Africa EDITORIAL BOARD Dr. Simon Cragg (1998) Institute of Marine Sciences LJniversity of Portsmouth Ferry Road Portsmouth P04 9LY United Kingdom Dr. Leroy Creswell (1999) Harbor Branch Oceanographic Institute US Highway 1 North Fort Pierce. Florida 34946 Dr. Lou D'Abramo (1998) Mississippi State University Dept of Wildlife and Fisheries Box 9690 Mississippi State, Mississippi 39762 Dr. Ralph Elston (1999) Battelle Northwest Marine Sciences Laboratory 439 West Sequim Bay Road Sequim. Washington 98382 Dr. Susan Ford (1998) Rutgers University Haskin Laboratory for Shellfish Research P.O. Box 687 Port Norris. New Jersey 08349 Dr. Raymond Grizzle (1999) Randall Environmental Studies Center Taylor University Upland, Indiana 46989 Dr. Robert E. Hillman (1998) Battelle Ocean Sciences New England Marine Research Laboratory Duxbury. Massachusetts 02332 Dr. Mark Luckenbach (1999) Virginia Institute of Marine Science Wachapreague, Virginia 23480 Dr. Bruce MacDonald (1997) Department of Biology University of New Brunswick P.O. Box 5050 Saint John. New Brunswick Canada E2L 4L5 Dr. Roger Mann (1998) Virginia Institute of Marine Science Gloucester Point. Virginia 23062 Dr. Islay D. Marsden (1998) Department of Zoology Canterbury University Christchurch. New Zealand Dr. Kennedy Paynter (1998) 1200 Zoology Psychology Building College Park. Maryland 20742-4415 Dr. Michael A. Rice (1998) Dept. of Fisheries. Animal & Veterinary Science The University of Rhode Island Kingston. Rhode Island 02881 Dr. Tom Soniat (1998) Biology Department NichoUs State University Thibodaux, Louisiana 70310 Susan Waddy (1998) Biological Station St. Andrews. New Brunswick Canada, EOG 2XO Dr. Gary Wikfors (1998) NOAA/NMFS Rogers Avenue Milford, Connecticut 06460 Journal of Shellfish Research Volume 17, Number 4 ISSN: 00775711 December 1998 Joiirmil of Shellfish Research. Vol. 17, No. 4. 9()?-910. 1998. PREDATION OF JUVENILE SEA SCALLOPS (PLACOPECTEN MAGELLANICUS) BY CRABS {CANCER IRRORATUS AND HYAS SP.) AND STARFISH {ASTERIAS VULGARIS, LEPTASTERIAS POLARIS, AND CROSSASTER PAPPOSUS) MADELEINE NADEAU AND GEORGES CLICHE Ministere de I'Agriciiltiire. des Pechehes et le I'Alimeiilalloii dii Quebec Direction de I'liiiuntition et des Technologies Cup-aux-Meules. Quebec. Canada GOB I BO ■ -^ .-^y ABSTRACT A scallop fishermen's association has seeded sea scallops, Placopecleii iiianellaniciis. off Iles-de-la-Madeleine (Quebec, Canada) since 1993. When the .scallops reach the bottom, they fall prey to starfish and crabs. Two size classes are available during the iles-de-la-Madeleine seeding period: small scallops from collectors (15 to 25 mm shell height) and large scallops grown-out in pearl nets (35 to 45 mm). To evaluate predation on both classes shortly after seeding, experiments were performed under controlled laboratory conditions. Testing involved three species of slstrfish. Asterias vulgaris ( 120 to 200 mm diameter), Lepiasterias polaris ( 120 to 200 mm diameter), and Crossasler papposus ( 1 10 to 130 mm diameter), and two species of crab. Cancer irroratus (85 to 1 20 mm carapace width) and Hyas sp. (90 to 1 10 mm carapace length). Scallops from both size classes were presented separately (nonchoice treatment) and together (choice treatment) to each predator species. Starfish and Hyas sp. consumed less than one scallop per predator per day as compared to C. irroratus. which consumed as many as 12 scallops per predator per day. Starfish and crabs did not show a clear prey size preference in both treatments. However, large .scallops tend to be consumed faster by both crab species in choice treatments. These results indicate that bottoms w ith high densities of starfish or crab should be avoided in seeding. In addition, bottom seeding should be done if possible with scallops greater than the size classes used in this study (over 50 mm), according on these results and those of other studies. KEY WORDS: Placopeclen nwfiellanicus. scallop, enhancement, predation INTRODUCTION Sea scallops, Placopecten mageUanicu.s (Gmelin), have been seeded commercially off Jles-de-la-Madeleine (Quebec, Canada) in the Gulf of St-Lawrence since 1993. The profitability of these operations hinges on scallop survival rates 4 years after seeding, until scallops reached the commercial shell height of 90 mm. Pre- dation is an important factor affecting the survival of juvenile scallops between seeding and hai-vest. In Nova Scotia, Hatcher et al. ( 1996) showed that 50% of their seeded scallops were killed by crabs and starfish within 2 weeks. Experimental summer and win- ter seedings performed by Barbeau et al. ( 1996) indicated survival rates of 1 and 10%, respectively, after only 8 weeks. Crab (Cancer irroratus Say) predation was identified as the prime factor of scal- lop mortalities. Cliche et al. (1994) determined during an experi- mental seeding performed in Iles-de-la-Madeleine that 1 1 .5% of seeded scallops were killed by crabs within 44 days. Haugum et al. (1997) lost all of their seeded scallops to crab (Cancer paguni.s) predation within 3 weeks during an experimental bottom seeding in Norway. Many factors may affect the impact of predation after seeding. Size of seeded scallops is generally considered important (Elner and Jamieson 1979, Morgan et al. 1980, Lake and Jones 1987, Minchin 1991. Barbeau and Scheibling l_994a, Barbeau et al. 1994. Arsenault and Himmelman 1996). In iles-de-la-Madeleine. two size classes are seeded in late fall. Scallops kept in collector bags attain 15 to 25 mm shell height after 1 year, and those transferred to peari nets, 35 to 45 mm. Production costs (labor and equipment costs) of scallops grown-out in pearl nets are five times higher than scallops grown directly in collector bags. Thus, it is essential to estimate the survival rate of both classes when choosing an optimal growing strategy. Previous surveys around Iles-de-la-Madeleine showed that star- fish are abundant (-0.5 starfish/m") in natural scallop grounds. \N\\h Asterias vulgaris (Verrill) and Lepiasterias polaris (Miilleret Troschel) predominating and Crossasler papposus (L.) regulariy observed. Low densities of two crab species (-0.05 crabs/m"). C. irroratus and Hyas sp., are also reported. Predation of juvenile scallops by A. vulgaris and C. irroratus has been documented by Elner and Jamieson (1979). Jamieson et al. (1982). Lake et al. (1987), Barbeau and Scheibling (1994a), Barbeau and Scheibling (1994b), Barbeau et al. (1994), and Arsenault and Himmelman (1996). However, few studies have examined predation involving sea scallops (P. magellanicus) of 15 to 50 mm shell height (Elner and Jamieson 1979). Barbeau and Scheibling ( 1994a) and Barbeau and Scheibling ( 1994b) and Barbeau et al. ( 1994) used scallops of 5 to 28 mm shell height, and Jamieson et al. (1982) used scallops of 40 to 55 mm and 80 to 110 mm. In addition, very little docu- mentation exists on the predatory capacity of L polaris. C. pap- posus and Hyas sp. on scallops (Arsenault and Himmelman 1996). Scallop vulnerabihty is high the first few days after seeding (Cliche et al. 1994, Barbeau et al. 1996). Stress induced during handling, exposure to air, and transportation to the seeding site may affect scallop vitality (Fleury et al. 1996). After seeded scal- lops reach the bottom, the time required to turn up their superior valve and find refuge may have an impact on their survival. Ar- senault and Himmelman (1996) concluded that refuge use de- creased the risk of predation for smaller scallops. Thus, studies conducted to increase the seeded scallops survival must target this period in particular. The objective of this study was to evaluate the impact or pre- dation by starfish. A. vulgaris. L polaris and C. papposus and crabs. C. irroratus and Hyas sp.. on small scallops from collector bags and larger ones from pearl nets under controlled conditions. This information is important when planning a seeding bottom strategy for a profitable commercial operation. MATERIALS AND METHODS Experiments were conducted in 2,000-L (I.I in x 2.3 m x 0.7 m) and 1,500-L (I.I m x 1.6 m x 0.7 m) fiber glass tanks with 905 906 Nadeau and Cliche circulating sea water. Temperature was maintained at 1 1.3 ± 3.0°C (mean ± SD), oxygen at 8.4 ± 0.7 mg/L, and salinity at 30.3 ± 0.9 ppt. A 12:12 light:dark regime was simulated by fluorescent lights over each tank (-250 lux). Red bulbs were used to allow video recording in the dark, as used in Barbeau and Scheibling's ( 1994b) experiments. The cultivated juvenile scallops were taken from commercial operations of the scallop fishermen's association of Iles-de-la- Madeleine. Specimens measuring 15 to 25 mm shell height came from spat collectors, and those measuring 35 to 45 mm from intermediate culture in pearl nets. The shell height was measured from the middle of the dorsal hinge to the farthest point of the ventral shell edge. A plastic tag (glue-on flexible polyethylene shellfish tag. Hallprint Ltd) of 4 x 8 mm was glued with cy- anoacrylate glue on the upper valve. C. inonitus and Hyas sp. were caught with rock crab or lobster traps. C. inoratits averaged 109.8 ± 6.6 m (mean ± SD. n = 45. range = 95 to 120 mm) carapace width, and Hyas sp. were 96.4 ± 7.7 mm (n = 6, range = 85 to 1 10 mm) carapace length. Similar to Barbeau and Scheib- ling's ( 1994a) experiments, only male crabs were used to eliminate potential sex-related biases (differences in morphology and preda- tion behavior). Starfish A. vulgaris of 153.3 ± 15.9 (n = 47, range = 120 to 200 mm) diameter. L. polaris of 151.9 ± 21.2 (n = 45, range = 120 to 200 mm), and C. papposus of 1 18.5 ± 6.6 mm (n = 6, range = 1 10 to 130 mm diameter) were collected by divers and scallop drags. The size classes of each predator species were selected according the most abundant predator class evaluated at the actual seeding sites (Roussy et al. 1994). Scallops, crabs, and starfish were maintained in 650-L seawater tanks for 2 to 10 weeks before testing began. Each species was kept in separate baskets. Scallops were fed on phytoplankton (Monochiysis httheri and Thalassiosira pseudonana) at concentra- tions of 5x10 cells/mL. Each predator was fed once a week on two living mussels (Mytilus edulis) weighing 10 to 20 g. Before experiments, predators were starved for 72 hours to standardize hunger levels. Two series of experiments were performed. The first was con- ducted between August and October 1994 in two 2,000-L tanks, divided using plexiglass separators to obtain four experimental sections of 0.8 m"* (Table 1). Predatory activity of C. irroratiis. A. vulgaris, and L. polaris was evaluated and compared. Three preda- tors of the same species were placed with 16 scallops of either 15 to 25 mm or 35 to 45 mm (nonchoice treatment) or eight scallops of each class together (choice treatment). Predator density was 2.4/m~. Scallop density was similar to the commercial seeding target density. 10/m". Control treatment involved eight scallops of each class without predator. Each treatment was repeated three times over a maximum of 5 days. In crab treatments, replicates were stopped earlier if all scallops had died. A video camera (Panasonic, Lunar Lite) fitted over the tanks filmed only one replicate of each experimental treatment because of logistical limitations. Starfish behavior was recorded up to 88 hours and crab behavior, up to 22 hours, because the latter was more active. During frame analysis, the time each predator devoted to searching for prey and the number of encounters between preda- tor and prey were noted. Prey search by starfish was defined as displacement on the tank floor toward .scallops with arms tips tumed up. Because prey search by crabs was more difficult to determine, all ciab movement was considered searching behavior. Any contact between predator and prey was considered an encoun- ter. The capture occurred when starfish arms attached a scallop with their tubefeet or when crabs chelae grabbed a scallop. The number of active scallop escapes after encounters was also counted. An active escape was noted when scallops jumped or TABLE I. Laboratory experiments performed from August to October 1994 and .June to November 1995. Tank or n prey/replicate Predi ator/Replicate Predator Experimental Section II Size (mm) Species Year Treatment Size (m') Replicate Small Large n Mean ± SD A. vulgaris 1994 Nonchoice 0.8 3 16 3 148.2 + 12.6 Nonchoice 0.8 3 16 3 Choice 0.8 3 8 8 3 1995 Nonchoice 1.8 2 24 T 160.9 ± 17.5 Nonchoice 1.8 2 24 2 Choice 1.8 2 12 12 2 Choice 1.2 4 8 8 2 L. polaris 1994 Nonchoice 0.8 3 16 3 146.4 ±2 1.6 Nonchoice 0.8 3 16 3 Choice 0.8 3 8 8 3 1995 Nonchoice 1.8 3 24 2 163.5 ± 15.4 Nonchoice 1.8 3 24 2 Choice 1.8 3 12 12 2 C. papposus 1995 Choice 1.2 3 8 8 2 118.5 + 6.6 C. irroratiis 1994 Nonchoice 0.8 3 16 3 108.8 ±6.7 Nonchoice 0.8 3 16 3 Choice 0.8 3 8 8 3 ■1995 Nonchoice 1.8 2 24 2 112.1 ±6.0 Nonchoice 1.8 2 24 -) Choice 1.8 2 12 12 2 Choice 1.2 3 8 8 2 Hyas sp. igg.s Choice 1 1 3 S S 1 964 ± 7.7 Predation on Seeded Sea Scallops 907 swam away from starfish or crabs. Passive escape, noted when a scallop closed its valves without displacement in a predator con- tact, was impossible to detect, because the video camera was too far from the subject. The number of retractions defined by a retreat of a predator after an encounter was noted. The second series of expenments was performed between June and November 1995. Some of the trials aimed to repeat the 1994 experiments in two 2,000-L tanks. However, improvements were made to the 1994 experimental design. To increase the surface to 1.8 m", no .separators were used. Predator (C iironirus, A. vul- garis, and L polaris) density was reduced to 0.8/m- to simulate natural densities more closely. Observations were extended to 1.^ days to collect more information on starfish predation. Scallops available from commercial operations were larger (15 to 30 mm and 35 to 50 mml than in the 1994 tests. Experimental treatments consisted of putting two predators of the same species with 24 scallops from one or both size classes (12 of both classes). A control treatment was conducted with 12 scallops of 15 to 30 mm and 12 scallops of 35 to 50 mm v\ithout predator. Number of replicates ranged from two to three. Second, the predatory activities of crab Hyas sp. and starfish C. papposus were compared, respectively, with those of C. iironirus and A. vuli;aris in two 1.500-L tanks that offered surface of 1.2 m"\ Predator density was 1.1 /m" and scallop density. lO/m". For each replicate, two predators were placed in a tank with eight .scallops of 15 to 30 mm and eight scallops of 35 to 50 mm. Control treatment u.sed eight scallops of 15 to 30 mm and eight .scallops of 35 to 50 mm without predator. Testing ran a maximum of 13 days but was stopped earlier if all scallops had died. Each treatment was repeated three or four times. Sequence of experiments and tank allocation for each replicate were random. To simulate seeding conditions, predators were im- mersed 24 hours before the scallops. During daily observations, dead scallops were removed. The chi-square test was used to com- pare daily mortality in both size classes. Contingency tables (treat- ment X mortality) were prepared, and cells with insufficient data for test validity were grouped. Fisher's exact test was performed if the number of data was still lower than five (Sherrer 1984). Pre- dation rates (number of scallops consumed per predator per 24 hours) were evaluated. All statistical analyses were performed us- ing SAS (1982), version 6.03 software. Data collected during video recording (encounter, capture, escape, and retraction rates) were considered more as an indication, given the absence of rep- licates. RESULT Scallop mortality was related to predation, because no mortality occurred in control treatments during 1994 and 1995 experiments. Starfish Video recordings from 1994 showed that starfish spent less than lO'vf of their time searching for prey in both treatments, remaining immobile on the tank walls the rest of the time. En- counter rates (or A. vulgaris and prey were between 0.9 to 4.1 per day (Fig. 1). Contacts tended to be higher with large prey in both treatments. L polaris encountered 0. 1 to 3.3 prey per day. Contacts between L. polaris and small scallops occuired more often. Active scallop escapes were higher for treatments involving A. vulgaris (0.8 to 4.0 per predator per day), and larger prey tended to escape more frequently. Predator retraction was noted with L. polaris (0.1 to 1.5 per predator per day) and occurred primarily after encoun- Treatment Figure 1, Behavior of small (S) and large (L) scallops in the presence of Asterias vulgaris, Leptaslerias polaris, and Cancer irroratus in choice and nonchoice treatments during 1994 experiments. ters invoh'ing smaller prey in choice treatment. All of these factors resulted m low capture rates during recording periods (0 to 0.06 per predator per day). In nonchoice treatments in 1994. A. vulgaris consumed more larger scallops after 4 days 0.22 ±0.13 (mean ± SD) per predator per day than smaller ones, 0.03 ± 0.13 (p = .03) (Fig. 2). How- ever, when size classes were presented together, mortality rates were similar (0.13 ± 0.12 per predator per day) for both (p = 0.99). In contrast, in 1995, only small scallops were consumed in nonchoice treatments after 1 3 days, with a predation rate of 0.37 ± A vulgaris - Non-choice treatment 1994 f 10- o |05- _ 1 1 T 1 0.0- ■ • — A 1995 > 1.0- < 05- 00- ; 12 3 1 Choice treatment 1,5 ^ 1.0- g TO a> 05- 1994 • Small prey ■ Large prey T I- • 1 r 1 1995 10- 05- 00- '1 1 ' 1 2 3 Replicate 1 2 Replicate Figure 2. Mean predation rates (n scallops consumed/predator/24 hours) (± 95% CI,) o\ Asterias vulgaris on small and large scallops in choice and nonchoice treatments per replicate during 1994 and 1995 experiments. 908 Nadeau and Cliche 0.45. Statistical analysis revealed a difference between mortality rates for both size classes (p = .001). In choice treatments, mor- tality rates were similar for both classes. 0.04 ± 0.05 per predator per day. Comparison between C. papposiis and A. vulgaris in 1995 choice experiments (Fig. 3). showed that A. vulgaris consumed more scallops (p = .04) after 13 days. However, both starfish demonstrated similar predation rates on 15 to 30 mm and 35 to 50 mm scallops (p > .05). During this experiment. A. vulgaris con- sumed more small scallops. 0.35 ± 0.37 per predator per day. than larger ones. 0.1 ± 0. 12 (p = .002). C. papposus induced similar mortality rates (p = .06) on both classes, with predation of 0.05 ± 0.02. Experiments performed in 1994 showed that L. polaris con- sumed only larger scallops in nonchoice treatments after 4 days, with a predation rate of 0.42 ± 0.33 (Fig. 4). Predation on both size classes was statistically different (p = .03). In choice treatments, predation rates on both classes were similar. 0.10 ±0.1 1 (p> 0.99). In 1995. nonchoice treatments showed that predation rates for small scallops (0.56 ± 0.34) were higher than those for larger scallops (0.08 ±0.13) after 13 days (p = .001). In choice treat- ments, smaller scallop mortality was significantly higher (0.24 ± 0.18 per predator per day) than larger scallop mortality (0.08 ± 0.04) (p = .004). Crabs Video analysis showed that C. irroratus spent 70 to 90% of the recorded time searching for prey. Encounter rates between preda- tor and prey were high ( 3 1 to 66). Some 7 to 1 1 active prey escapes occurred per predator per day. Retraction rates were noted 13 to 56 times per predator per day. Encounter and retraction rates tended to be higher with small scallops, but capture rates were higher for larger scallops. In both experimental years, C. irroratus consumed almost all juvenile scallops available in a few days. Consequently, mortality rates were similar for both classes (p > .05) at the end of each replicate. Statistical comparison was performed 1 day after seeding to identify a preference for a particular class. In the 1994 experi- ments, predation rates after 24 hours in nonchoice treatments in- volving larger scallops were 4.22 ± 0.84 as compared to 2.89 ± 1.07 for smaller ones (Fig. 5). Predation rates in choice treatments L, polans Non-choice treatment A. vulgaris C. papposus 25- « Small prey 20- ■ Large prey 1,5- 1,0- 0,5- ao- I i \ 2 3 4 Replicate 1 2 3 Replicate Figure 3. Mean predation rates (n scallops consumed/predator/24 hours) (± 95% C.I.) of Asterias vulgaris and Crossaster papposus on small and large scallops in choice treatments per replicate during 1995 experiments. £..\j n 1994 15- 03 to 1 1 c g 00 10- 1 . 0} ■ 0.5- ^ T 00- ♦ . h 1995 1.5- r 1.0- ( 0.5- 1 1 T J 1 ■ OO- — ■ — 1 — 1 _j 1 2 3 Choice treatment 1995 15- • Small prey ■ Large prey 10- ■ \ 05- 00- & i 1 12 3 12 3 Replicate Replicate Figure 4. Mean predation rates (n scallops consumed/predator/24 hours) l± 95% C.I.) of Leptasterias pularis on small and large scallops in choice and nonchoice treatments per replicate during 1994 and 1995 experiments. were 2.67 for larger scallops and 1.44 ± 0.77 for smaller ones. Statistical analysis showed that larger scallops were consumed more often than smaller ones in both treatments (p < .05). In 1995, C. irroratus consumed more smaller scallops after 24 hours. 1 1 ± 0.71 per predator, than larger ones (3.75 ± 1.06) in nonchoice treatments (p = .001). However, larger scallops were consumed more quickJy in choice treatments, with a predation rate of 5 ± 0.7 1 as compared to 2.25 ± 2.47 for smaller ones (p = .001). Small scallop mortality was statistically lower with Hyas sp. (p = .0001 ). but larger scallop consumption was similar for both crab species (p > .05). Hyas sp. consumed more larger scallops than smaller ones (p = .014) in 13 days (Fig. 6). C. irroratus had consumed all of the scallops presented at the end of each replicate. However, I day after seeding, C. irroratus consumed all large scallops available, four per predator. Large scallops were con- sumed more often than smaller ones. 2.67 ± 1 .04 per predator (p = .002). DISCUSSION In this study, starfish consumed less than one scallop per preda- tor per day. In neither test year did starfish show a clear size- related preference in choice and nonchoice treatments. However, Barbeau and Scheibling (1994a) showed that, in aquaria. A. vul- garis of 30 to 150 mm diameter consumed more scallops of 5 to 8.5 mm shell height than those of 10 to 15 mm or 20 to 25 mm in Predation on Seeded Sea Scallops 909 C irroratus - Non-choice treatment 12 C irroratus Hyas sp S 6 nj ■a a 3- 1994 "T r- 12 3 12 Choice treatment 12 0) ra c o ra ■o (U 1994 9 • Small prey ■ Large prey 6- 3- ■ Ji ■ • • 0- 1 r 1 ' .12 3 Replicate Figure 5. Predation rates (n scallops consumed/predator/24 hours) of Cancer irroratus after 1 day on small and large scallops in choice and nonchoice treatments per replicate during 1994 and 1995 experiments. choice and nonchoice treatments. Under field conditions. Barbeau et al. (1994) confirmed that A. vulgaris consumed more scallops of 5 to 9 mm and 10 to 15 mm than 20 to 25 mm. In addition, tests performed by Arsenault and Himmelman (1996) showed that vul- nerability of the scallop Chlamys iskmdica (between 10 to 60 mm shell height) to L. polaris, C. papposus. and A. vulgaris decreased with increasing prey size in choice treatments. Crossaster papposus had lower predation activity (0.05 prey per predator per day) than A. vulgaris (0.2 prey per predator per day) on juvenile scallops. However, A. vulgaris and L polaris induced comparable mortality rates for juvenile scallops. These last predators spent less than 10% of the recorded time searching for prey. Active scallop escapes and predator retractions, espe- cially involving L. polaris. kept the consumption rate under one scallop per predator per day. These findings were confirmed by Barbeau and Scheiblmg (1994a). in whose tests A. vulgaris spent 1 1 to 27% of time searching for prey, resulting in less than one scallop of 20 to 25 mm being consumed per starfish per day. Differences between experimental years and replicates could be related to various factors. Differences in years may be partly as- sociated with the experimental design modifications. In 1994, predator density was twice as high and tank sections half as large as in 1995. Thus, active escapes induced after starfish encounters were probably limited by the walls of the smaller tanks used in 1994. The 1995 experimental design more closely simulated natu- ral conditions. Replicates may also have been affected by the 10 6- 2- 13 days Small prey Large prey i i t 2 3 Replicate 2 3 Replicate Figure 6. Predation rates (n scallops consunied/predator/24 hours) of Cancer irroratus after 1 day and mean predation rates (± 95% C.l.) of Hyas. sp. after 13 days on small and large scallops in choice treatments per replicate during 1995 experiments. variability in individual behavior and in physiological needs dur- ing the year. Crabs were very effective predators, able to ingest as many as 12 scallops per predator per day. In both experimental years and treatments, C. irroratus consumed all scallops available within 24 to 72 hours after seeding. In nonchoice treatments, this crab spe- cies tended to consume all available prey rapidly. In choice treat- ments, the mortality of larger scallops after 24 hours was higher than that of smaller ones. Although Hyas sp. also selected larger specimens in choice treatments, the predation rates were lower than those of C. irroratus. In fact, the predation rates of H\as sp. (0.5 scallop per predator per day) were closer to the starfish. Bar- beau and Scheibling (1994a) showed that C. irroratus consumed more scallops of 10 to 15 mm shell height than of 5 to 8.5 mm or 20 to 25 mm in nonchoice treatments, and more scallops of 20 to 25 mm in choice treatments. In contrast, their field testing (Bar- beau et al. 1994) on scallops of 7 to 28 mm indicated that prey size had little effect on crab predation success. Video tapes showed that C. irroratus spent 70 to 90% of the time searching for prey. This result may be an overestimate, be- cause all crab displacements were related to searching behavior. Barbeau and Scheibling ( 1994a) estimated crab searching behavior at less than 11% of the recorded time. In the present study, C. irroratus encountered as many as 3 1 to 66 scallops per day and a reaction of retraction followed in 58.4 to 86.6% of the time. In Barbeau and Scheibling's (1994a) study, no predator retraction was observed, but passive escapes represented 58 ± 41% of total escapes. However, passive escape was impossible to detect on recording image. Therefore, this escape may have been confused with the retraction behavior. Wilkens (1991) postulated that scal- lops often respond to crab encounters by closing their valves (pas- sive escape). This behavior may result from scallops detecting crab movement with the use of eyes on the mantle edge. Elner and Jamieson (1979) showed that C. irroratus had a specific scallop (P. magellaiiicus) size preference in multiple prey choice treatments under controlled conditions. Larger predators ( 120 to 130 mm carapace width) preferred larger scallops (40 to 50 mm shell height), and smaller predators (90 to 100 mm) tended to prefer smaller scallops (20 to 30 mm). Based on these results, it is possible to hypothesize that the intermediate predator size (100 to 910 Nadeau and Cliche 120 mm), used in the present study, would prefer the intermediate scallop size (30 to 40 mm). Lake at al. (1987) also showed that Cancer pagurus predation was influenced by predator size and scallop size. The number of scallop Pecten maximus (30 to 60 mm shell height) consumed increased with predator size (60 to 140 mm carapace width) and decreased as prey size increased. However, their study did not use scallops smaller than 30 mm. In contrast. Arsenault and Himmelman (1996) showed that C irromnis and Hyas araneus. size 1 12 ± 2 mm (mean ± SD) carapace width and 97 ± 1 mm carapace length, respectively, preyed more on scallops (Chlamys islandica) of 10 to 30 mm shell height than on scallops of 30 to 60 mm in multiple prey choice treatments. Furthermore, the feeding rate of H. araneus was about twice that of C. irroratus. Effective predators (A. vulgaris. L. polaris, C. papposiis. C. irroratus, and Hyas sp.) live on the sea bottom, where scallops seeding activities occur in the Iles-de-la-Madeleine. In this study, predation rates may have been biased by the absence of refuge or alternative prey. In addition, interaction between predator species was not evaluated. Nevertheless, we can assume that these preda- tors will cause high mortality in seeded scallops under natural conditions. Based upon our results and those obtained from other studies, bottom seeding should be done, if possible, with scallops greater than the size classes used in this experiment (over ?0 mm). On bottoms with high crab density, both size classes of scallops will be vulnerable. Because crabs are highly effective scallop predators, these bottoms must be avoided. Alternative solutions must be evaluated to minimize the impact of predation. Control of seeding density and predator elimination (Ventilla 1982) are solutions used in Japan. As observed by Bar- beau et al. (1994), predation increases with scallop density and crab density. Barbeau and Scheibling (1994b) also demonstrated that the seeding season may have an important impact on preda- tion, because temperature affects feeding activities of crabs (C irroratus) and starfish (A. vulgaris). Seeding at lower temperatures was effective in reducing predation, particularly for starfish. Bio- chemical analyses of Pecten nia.\inuis performed by Fleury et al. ( 1996) showed a very low glucide content in the fall. Lower energy reserves may affect the survival rates of seeded scallops. In addi- tion, stress induced by the seeding operations can reduce their vitality. Fleury et al. (1996) found that seeding stress represented high energy expenditures for scallops (Pecten maximus). Finally, weak animals escape predators less often and less rapidly (Hatcher et al. 1996). Seeding strategies will have to deal with all these factors. Seeding operations in the Iles-de-la-Madeleine will be financially profitable if 20 to 30% of the scallops seeded are caught by scallop fisherman. An effort must be made to ensure such survival rates. ACKNOWLEDGMENTS The authors thank the technical staff of the Station Tech- nologique Maricole des Iles-de-la-Madeleine of MAPAQ for their valuable participation in this project and the local scallop fisher- mens' association, who provided the juvenile scallops. A special thanks to Michel Giguere of the Department of Fisheries and Ocean from Ste-Flavie for assistance. REFERENCES Arsenault, D.J. & J. G. Himmelman. 1996. Size-related change.s in vul- nerahiiity to predators and spatial refuge use by juvenile Iceland scal- lops Chlamys islandica. Mar. Ecol. Prog. Ser. 140:115-122. Barbeau, M. A. & R. E. Scheibling. 1994a. Behavioral mechanisms of prey-size selection by sea stars Asterias vulgaris (Verrill) and crabs Cancer irroratus (Say) preying on juvenile sea scallops Placopecten magellanicus (Gmelin). / Exp. Mar. Biol. Ecol. 180:103-136. Barbeau. M. A. & R. E. Scheibling. 1994b. Temperature effects on preda- tion of juvenile sea scallops Placopecten magellanicus (Gmelin) by sea stars Asterias vulgaris (Verril) and crabs Cancer irroratus (Say). J. Exp. Mar. Biol. Ecol. 182:27-48. Barbeau, M. A., B. G. Marcher. R. E. Scheiblmg. A. W. Hennigar. L. H. Taylor & A. C. Risk. 1996. Dynamics of juvenile sea scallops Pla- copecten magellanicus and their predators in bottom seeding trials in Lunenburg Bay, Nova Scotia. Can. J. Fish. Aquat. Sci. 53:2494-2512. Barbeau. M. A.. R. E. Scheibling. B. G. Hatcher. L. H. Taylor & A. W. Hennigar. 1994. Survival of tethered juvenile .sea scallops Placopecten magellanicus in field experiments — effects of predators, .scallop size and density, and site and season. Mar. Ecol. Prog. Ser. 115:243-256. Cliche. G.. M. Giguere & S. Vigneau. 1994. Dispersal and mortality of sea scallops. Placopecten magellanicus (Gmelin 1791) seeded on the sea bottom off iles-de-la-Madeleine. J. Shellfish Res. 13:565-570. Elner, R. W. & R. N. Hughes. 1978. Energy maximization in the diet of the shore crab Carcinus maenas. J. Anim. Ecol. 47:103-1 16, Elner. R. W. & G. S. Jamieson. 1979. Predation of sea scallops Pla- copecten magellanicus by rock crab Cancer irroratus and American lobster Homarus americanus. J. Fish. Res. Board Can. 36:537-543. Fleury. P. G.. C. Mingant & A. Castillo. 1996. A preliminary study of the behavior and vitality of reseeded juvenile great scallops of three sizes in three seasons. Acpiacull. Int. 4:325-337. Hatcher, B. G., R. E. Scheibling. M. A. Barbeau, A. W. Hennigar, L. H. Taylor & A. J. Windust. 1996. Dispersion and mortality of a population of sea scallop Placopecten magellanicus seeded in a tidal channel. Can. ./. Fi.ih. Aquat. Sci. 53:38-54. Haugum. G. A.. O. Strand. A. Svardal & S, Mortensen. 1997. Survival and behavior of scallops Pecten maximus L. after transfer to the seabed — effect of emersion treatment. Proceedings Eleventh International Pec- linid Workshop, La Paz, Baja California Sur, Mexico. 30 pp. Jamieson. G. S., H. Stone & M. Etter. 1982. Predation of sea scallop Placopecten magellanicus by lobster Homarus americanus and rock crab Cancer irroratus in underwater cage enclosures. Can. J. Fish. Aquat. Sci. 39:499-505. Lake. N. C. H.. M. B, Jones & J. D, Paul, 1987, Crab predation on scallop Pecten maximus and its application for scallop cultivation. J. Mar. Biol. Ass. U.K. 67:55-64. Minchin, D. 1991. Decapod predation and the sowmg of the scallop, Pecten maximus (Linnaeus, 1758). pp. 191-197. In: S. E. Shumway and P. A. Sandifer (eds.). An International Compendium of Scallop Biology and Culture. World Aquaculture Workshops. No. 1. World Aquaculture Society, Baton Rouge, Louisiana. Morgan. D. E.. J. Goodsell. G. C. Matthiessen, J. Garey & P. Jacobson. 1980. Release of hatchery-reared bay scallops (Argopecten irridians) onto a shallow coastal bottom in Waterford. Connecticut, Proc. World MaricuL Sac 11:247-261. Roussy, M., G. Cliche & M. Giguere. 1994. Caracterisation et eradication des pr6dateurs du petoncle geant (Placopecten magellanicus) aux Iles- de-la-Madeleine. MAPA-Pecheries. D.R.S.T Doc. Rech. 94/15. 85 pp. SAS Institute. 1982. SAS user's guide: Statistics. SAS Institute Inc., North Carolina. Sherrer. B. 1984. Biostalistique. Gaetan Morin Edileur. Chicoutimi. Que- bec. Canada. 850 pp. Ventilla. R. F. 1982. The scallop industry in Japan, pp. 309-382. In: J. H. S. Blaxter, F. S. Russell and M. Yonge (eds.). Advances in Marine Biology, vol. 20. Academic Press, San Diego, California. Wilkens, L. A. 1991. Neurobiology and behavior of the scallop, pp. 429- 469. In: S. E. Shumway (ed.). Scallops: Biology, Ecology, and Aqua- culture. Developments in Aquaculture and Fisheries Science, vol. 21. Elsevier Science. New York. .louniiil i>l Shrllfish Reseanh. Vol. 17. No. 4, 911-917, 199S. GROWTH, PRODUCTION, AND REPRODUCTION IN BAY SCALLOPS ARGOPECTEN IRRADIANS CONCENTRICUS (SAY) FROM THE NORTHERN GULF OF MEXICO PAUL A. X. BOLOGNA University of South Alabama Dauphin Island Sea Lab Dauphin Island. Alabama 36528 ABSTRACT The bay scallop. Argopeclen inadians. is a commercially and recreationally important fisheries .species on the Atlantic and Gulf Coasts of the United States. Surprisingly, little information exists on northern Gulf of Mexico populations. This research assessed growth, production, and reproduction in a population from St. Joseph Bay. Florida. Specifically, scallops exhibited high initial growth rates (Gw day''), with rates declining as individual size increased. Additionally, significant interannual variability existed for both scallop growth and production. These differences may be attributable, in part, to a reduction in salinity (2l9rr) associated with Tropical Storm Alberto, which resulted in significantly lower growth rates and a mass mortality event. Reproductive assessment oi A. irradians showed significant peaks in spawning condition, gonad weight, and gonadal-somatic index (GSI) during the winter (De- cember. January. February) compared to other seasons. However, the salinity minima in July 1994 ( 1 IS?c) significantly reduced gonad weight and GSI from winter 1994/1993. suggesting that a single storm event had a dramatic but short-term reproductive impact on the population. Assessment of gonad condition and GSI, coupled with field observations, showed spawning occurred in the spring and fall as well, but the presence of small individuals (<4 mm shell height) during July, August, and September suggests that reproduction may occur throughout the year in St. Joseph Bay. Florida. KEY WORDS: Argopeclen iiradiuiis. gonad weight. GSI. Gulf of Mexico, salinity effects INTRODUCTION Bay .scallops. Argopeclen irradians (Lamarck 1819) are com- mon members of many shallow water benthic communities along the Atlantic and Gulf Coasts of North America. Clarke (1965) identified three distinct subspecies of A. irradians based on mor- phologic characteiistics, and recent morphologic and genetic stud- ies support these distinctions (Wilbur and Gaffney 1997). Al- though it is recognized that scallops recrtiit to seagrass habitats (Gutsell 1930. Thayer and Stuart 1974. Eckman 1987) and that they cling to leaves to escape benthic predators (Pohle et al. 1991. Ambrose and Irlandi 1992). little information exists on several key aspects of their ecology. Several studies have assessed scallop growth rates under experimental conditions detennining the im- pact of flow (Kirby-Smith 1972, Cahalan et al. 1989. Eckman et al. 1989) or food concentration and stocking density (Rheault and Rice 1996), but there are few field estimates of adult growth rates under natural, noncaged conditions. Eckman (1987) showed greater growth rates for juvenile bay scallops in grass beds with low shoot densities, and Ambrose and Irlandi (1992) showed that juveniles may trade-off decreased growth rates for increased sur- vivorship by altering their placement on seagrass leaves. In addi- tion, it has been shown that habitat configuration has an impact on the growth and survival of juveniles (Irlandi et al. 1995). Although these studies adequately assess growth of juveniles, few published studies exist on growth rates of adult scallops under natural field conditions. Another aspect of A. irradians is the relative lack of estimates of natural mortality (but see Marshall 1963). Many studies have assessed predation mortality (Tettelbach 1986. Pitcher and Butler 1987. Peterson et al. 1989. Prescott 1990. Pohle et al. 1991. Am- brose and Irlandi 1992. Bologna 1998); however, few estimates of natural population mortality exist. Exceptions include clear pat- terns of high postspawn mortality (Gutsell 1930, Capuzzo and Hampson 1984). mortality associated with 2nd-year individuals prior to spawning (Bricelj et al. 1987a). and the severe impacts of nuisance algal blooms (Suiinnerson and Peterson 1990. Tettel- bach and Wenczel 1993). One feature of A. irradians ecology that is relatively well known for many populations is reproductive effort. Given that scallops suffer high moilality after spawning, significant research has focused on reproduction, reproductive conditioning, and changes in biomass associated w ith gonad development (Bricelj et al. 1987h). It has been shown that scallops show distinct peaks in spawning, and these peaks differ temporally based on latitude of the populations (Sastry 1970, Barber and Blake 1983, Crenshaw et al. 1991). Although scallops do show peaks in reproduction, re- cruiting individuals (<10 mm shell height) have been reported throughout the year, suggesting that trickle spawning may occur in some locations (Gutsell 1930, and see Bricelj et al. 1987b). How- ever, in general, bay scallops are considered to be a semelparous or short-lived iteroparous species, and reproductive effort tends to follow this life history trait (Sastry 1970, Sastry 1979, Barber and Blake 1983, Bricelj et al. 1987h). Despite a wealth of knowledge on populations of A. irradians from the Atlantic Coast and southern Florida, little is known about northern Gulf of Mexico populations. The goals of this research were to assess several life history traits from a bay scallop popu- lation from St. Joseph Bay, Florida. U.S.A. Specifically, research focused on the following objectives: ( I ) determine the natural growth rate of adult scallops: (2) estimate natural population mor- tality; (3) estimate production within the population: (4) determine reproductive cycles; and (5) assess reproductive effort of scallops. METHODS Study Site This research was conducted in St. Joseph Bay, Florida, which lies in the northeastern Gulf of Mexico (29° N, 85.5° W). It is a semi-enclosed lagoonal system with little freshwater input. Salin- ity and temperature data were collected from 1992 through 1995 in St. Joseph Bay (Fig. I) and show that temperature follows a sea- 911 912 Bologna sonal trend; whereas, salinity is normally high and ranges from 25 to 35%c annually. However, large storm events (e.g.. Tropical Storm Alberto) can significantly reduce salinity over short periods of time (e.g., to 1 \%c). Shallow regions of the St. Joseph Bay benthos primarily consist of the seagrass Thalassia testiuliniiiii interspersed with Halodide wrightii. Syringodium filifonne, and open sediment. T. testiidiiuim is the dominant species and covers approximately 2.300-2,400 hectares (Savastano et al. 1974, Iverson and Bittaker 1986). GROWTH, MORTALITY, AND PRODUCTION To determine rates of growth and mortality, a series of mark- recapture experiments were undertaken. Scallops were marked by cleaning and drying the ventral valve and gluing on a numbered tag. Scallop shell height and breadth were then measured using vernier calipers to the closest 0.05 inm. Six additional scallops were marked and held in aquaria for 2 weeks to assess the impacts of handling and marking on survival. All six control individuals survived the 2 week trial and were released back into the field. These individuals were not used to estimate growth, mortality, or production. In 1993, 79 scallops (31.3 mm to 44.1 mm shell height) were marked and released on two dates in June into an expansive Thalassia tesnidiniim grass bed. Scallops were marked and re- leased on June 1 1 (n = 31 ) and June 14 (n = 48). During 1993, scallops were relocated and measured on June 14. 23. July 8. 22, August 25, September 19, November 5, and December II. In 1994, 95 scallops were marked in three cohorts. In April, 40 scal- lops (37.75-66.3 mm shell height) were marked and measured in the same manner as above. The second cohort of 33 scallops (41-51.6 mm shell height) was marked on June 23, and the final scallop cohort (n = 22, 41.8-53.8 mm shell height) was marked and released on July 14. All cohorts were released into an exten- sive T. tesrudiiuiin grass bed mosaic and periodically sampled throughout the year. Specifically, scallops were relocated on May 5, June 6, 23, July 7, 14, 20, August 10, 18, 31. and October 12. Estimates of growth and mortality followed the protocol of Cowan and Houde (1990). Instantaneous growth rate was calcu- lated using the change in biomass over time. Scallop biomass was estimated using the following regression equation from Bologna (1998), which provides an estimate of tissue dry weight: (ln[tissue 3 5- I • Temperature ' Salinity 3 2 5 .■• 1992 1993 1994 1995 Figure 1. Weekly temperature (C) and salinity (*?() values collected from St. .Joseph Bay, Florida from January IWI through December 1995. Source: City of Port St. Joe, Florida industrial waste-water treat- ment plant, 1996; ^indicates a salinity minima of ll%fi on July 27, 1994. dry wt. (g)] = -9.779 -I- 0.7909* ln[.shell height] + 2.2124* In- [shell breadth]; n = 161, r" = .92). Growth rate (Gw, growth in weight), expres.sed as day"', was computed using the Eq. 1. Gir = ln(lV,)-ln(W,_|) (1) where: C\v = instantaneous growth rate (day"'); W = estimated scallop dry weight at time t (ln[grams tissue dry weight]); and, t = time in days. Instantaneous scallop mortality rate (|x) was esti- mated from recaptures using Eq. 2. \n{no) - \n(nt) (2) where: |j. = calculated instantaneous mortality rate (day"'); n = is the number of scallops present at a given time t; and t = time. Because scallops were motile (>500 m. pers. obs.. this study), calculated mortality actually retJects mortality plus emigration and, consequently, is an overestimate. To limit the impact of initial emigration bias on mortality estimates by second and third cohort marked scallops (e.g.. field release dates June 14. 1993. June 23 and July 14. 1994). mortality on successive dates (June 23, 1993, July 7 and 20, 1994) was calculated using the number of scallops present from the previous cohort. Subsequent mortality estimates were based on any individual present beyond one sampling date. Based on estimated instantaneous growth and mortality rates calculated from mark-recapture scallops, production (g dry wt day"') for time intervals was calculated using the Cowan and Houde (1990) equation: P = B * Gw where: B = average biomass (g) during a time interval; Gw instantaneous growth rate; and B was calculated as follows. B = : for {Gw > |ji) (3) (Cm- n.]r -J_l^ (^l ■Cw)r ■ for (|a, > Gil) (4) where: B^, = estimated initial biomass of all marked individuals for a given time interval; Gw = instantaneous growth rate; ij, = instantaneous mortality rate; and. t = time. Growth, mortality, and production rates were compared be- tween 1993 and 1994 using a nonparametric two-tail trend analysis (a = 0.05) for rates on similar calendar dates between years. REPRODUCTION Scallop reproduction was assessed by visual inspection of the gonad condition and resultant gonad weight and relative weight ratio (GSI). Scallops were collected on 37 dates from St. Joseph Bay, Florida from November 1992 to October 1996. Scallops were frozen and returned to the laboratory where their shell height and breadth were measured to 0.05 mm. Scallops <22 mm were con- sidered juveniles, and body tissue was dissected and weighed as a whole. Scallops >22 mm shell height had both gonadal and so- matic tissue dissected out. Reproductive and somatic tissue were then dried at 80°C for 48 to 72 hours and weighed (g dry weight). A GSI was calculated for each scallop using the following equation: GSI = (gonad dry weight/total dry weight) * 100. To assess seasonal timing of maxi- Growth. Production, and Reproduction of Bay Scallops 913 mal GSI and gonad weight, collections were groups as winter (samples collected in December. January, and February), spring (March. April, and May), summer (June. July. August) and fall (September. October. November). Data were analyzed using a one- way analysis of variance (ANOVA; a = CO."!). with season as the independent variable and gonad weight and GSI as dependent variables. Multiple comparisons were made using Scheffe's F-test. In addition, scallop gonad weight was analy/.ed against date of collection using one-way ANOVA to determine annual variability in reproductive effort. Visual condition of gonadal material for scallops greater than 22 mm shell height was assessed for each individual beginning in August 1993. Gonad condition was assessed as undeveloped, rip- ening, very ripe, or postspawn. Undeveloped gonads appeared gray in color, but were robust in nature. Ripening gonads were robust and showed visual development of gonadal material (i.e., female portions were, to some degree, red-orange in color and/or male portions were white or whitening). Very ripe gonads were robust, with female portions fully red-orange and male portions entirely white. Gonads form postspawn individuals were either gray and flaccid or gray with tinges of red-orange/white and flaccid (as opposed to robust, which were associated with undeveloped and developing gonads). Evidence of reproduction was assessed by collection of small individuals (<20 mm shell height). This was accomplished by visually locating them in the field and returning them to the labo- ratory for morphological measurement and biomass determination. In addition, in 1994 and 199.'!. artificial seagrass mats (for a full description of seagrass mat construction and characteristics, see Bologna 1998) were placed in the field during the summer to determine the presence and recruitment density of bay scallops. RESULTS Growth, Mortality, and Production Scallop growth, mortality, and production rates are summarized in Table 1. In 1993, scallops showed growth rates that were rela- tively high, declining slowly from June to September and dramati- cally in the fall. However, scallops marked in 1994 showed a rapid decline in growth rate from May to June and maintained very low growth rates throughout the summer and into the fall. When the data were analyzed using trend analysis, calculated compared mean scallop growth rate for 1993 was significantly greater than growth rates in 1994 (P [K « 017, 0.475] = 0.01 17). Scallop mortality rates were initially relatively high for all se- ries of mark-recapture scallops (Table 1 ). This may be attributed to relatively high emigration (dispersal) rates after initial deploy- ment into the field, and these mortality estimates must be treated with caution. The calculated mortality rates for scallops on subse- quent cohort release dates in 1993 and 1994 did not show this, because mortality calculations were based on previous individuals in the field and not the potentially high rates of emigration on initial placement in the field. In 1993. mortality was high during June, July, and August, but relatively reduced in September- November, with major losses by the December collection date. No mortality patterns were present in 1994 data. However, mortality rates were lower in 1994 than 1993 (Table 1 ), albeit not significant (P[K < 317. 0.475] = 0.574). However, the relatively high mor- tality rate calculated for August 10, 1994 corresponds to the sa- linity minima of 1 19ct seen in 1994 from freshwater input associ- ated with Tropical Storm Alberto (Fig. 1). TABLE L Instantaneous growth, mortality, and production rates for mark-recaplure .4. irradians from 1993 and 1994. Growth Mortality Production Date n cd Day ' Day-' g Day ' 6/14/93* 24 a 0.02509 0.07380 0.33560 6/23/93 54(19) b 0.01865 0.02696 0.68474 7/8/93 38 c 0.02027 0.02343 0.49523 7/22/93 26 d 0.01638 0.02711 0.48080 8/25/93 15 e 0.01054 0.01618 0.27.561 9/19/93 13 f 0.01064 0.00572 0.33591 1 1/5/93 9 g 0.00539 0.00782 0.11650 12/11/93 2 0.00224 0.04178 0.01852 1994 5/5/94 21 0.02603 0.03391 0.5998 6/9/94 14 a 0.00544 0.01158 0.1722 6/23/94** 11 b 0.00417 0.01723 0.1049 7/7/94 19(9) c 0.0043 1 0.01433 0.0905 7/14/94*** 18 0.00632 0.00772 0.1889 7/20/94 25 (17) d 0.00442 0.00817 0.1238 8/10/94 17 0.00469 0.01836 0.1589 8/18/94 16 e 0.00650 0.00758 0.1743 8/31/94 14 f 0.00638 0.01027 0.1627 1 (VI 2/94 5 g 0.00343 0.02451 0.0584 Growth and mortality values expressed as day" , production expressed as grams day"'; n indicates the number of recaptured individuals u.sed to estimate growth, mortality, and production; values in parentheses indicate the number of individuals used to estimate mortality when a successive cohort of marked individuals was added to the tleld. thus limiting initial emigration bias of mortality hy newly placed marked scallops; cd indicates calendar dates of comparison for growth, mortality, and production rates between 1993 and 1994; * indicates the date when 48 additional individu- als were marked and released into the field; ** indicates the addition of 33 marked scallops; *** indicates the addition of 22 marked scallops. Scallop production in 1993 showed an increase during the ini- tial part of the study with declining productivity as the year pro- gressed (Table 1 ). Because production was based upon estimates of mortality that include both mortality and emigration, calculated production is a conservative estimate. In 1994, production was considerably lower and remained les than one-third that of 1993. Results from trend analysis showed a significant reduction in scal- lop production from 1993 compared to 1994 (P[K < 017, 0.475] = 0.01 17). This reduction in production between years was likely the result of the significantly lower growth rates in 1994 compared to 1993 (p < .023); and not to changes in relative mortality (Table I ). A scallop mortality event was observed on July 20, 1994 throughout the southern half of St. Joseph Bay. Although the ab- solute salinity minima ( 1 \%c) occurred on July 27 (see Fig. 1), the reduction in salinity attributable to Tropical Storm Alberto on July 3, not to mention the potential for disturbance and burial, may have had a severe impact on the survival of bay scallops. The observed salinity minima, however, corresponded well to the high estimated instantaneous mortality rate calculated for mark-recapture indi- viduals on the following collection date (August 10. 1994. Table 1). On June 24. 1996. another mortality event was observed in St. Joseph Bay. During this time period, numerous sitings of red-tide algal blooms occurred along the coastal regions of the northeast Gulf of Mexico (J. Stout, pers. comm.) and may have been re- sponsible for this mortality event. 914 Bologna REPRODUCTION Reproduction assessment by visual inspection of the gonad condition showed several interesting features regarding reproduc- tive individuals (Table 2). First, scallops as small as 31 mm shell height showed very ripe gonads and scallops as small as 34.1 mm shell height showed gonad condition indicative of postspawn (May 1994. Table 2). Second, scallop gonad condition associated with individuals from the winter of 1994/1995 showed relatively low proportions of individuals in any reproductive condition, as well as few individuals (one of 59) showing a postspawn condition (Table 2). Last, of the 1.106 scallops assessed for reproductive purposes, only one individual (October 1995, Table 2) showed nonhermaph- roditic sex development with only male gonadal tissue present. When scallop reproduction was assessed using GSI. results showed scallops had significantly greater GSI values during the TABLE 2. Visual gonad condition index. Date n Juvenile Undeveloped Ripe Very Ripe Postspawn 8/25/93 9/11/93 9/19/93 10/1/93 11/5/93 11/19/93 24 12/11/93 32 1/9/94 2/25/94 4/15/94 5/5/94 6/8/94 7/20/94 8/19/94 8/24/94 12/1/94 1/21/95 2/1/95 3/10/95 4/9/95 5/17/95 6/20/95 7/5/95 8/6/95 8/30/95 9/26/95 10/28/95 12/5/95 1/20/96 2/15/96 3/13/96 4/19/96 5/22/96 6/24/96 8/8/96 10/18/96 12 27 27 37 40 36 12 29 45 44 23 4? 45 6 25 37 42 34 31 32 17 37 22 6 13 20 28 24 21 21 16 19 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 28.00 17.86 10.71 2.70 0.00 0.00 0.00 0.00 0.00 43.24 0.00 0.00 23.08 0.00 42.86 8.33 0.00 0.00 0.00 0.00 66.67 96.30 37.04 45.95 47.50 16.67 15.63 11.11 4.55 44.83 2.22 95.45 .S2.17 86.67 42.22 66.67 44.00 57.14 71.43 72.97 83.33 100.00 100.00 68.75 47.06 45.9.S 0.00 0.00 0.00 5.00 10.71 62.50 52.38 100.00 68.75 5.26 33.33 3.70 .59.26 40.54 0.00 41.67 0.00 11.11 63.64 24.14 60.00 0.00 0.00 8.89 35.56 0.00 20.00 2 1 .43 10.71 13.51 7.14 0.00 0.00 15.63 52.94 10.81 9.09 50.00 7.69 0.00 10.71 12.50 0.00 0.00 31.25 21.05 0.00 0.00 3.70 13.51 52.50 25.00 50,00 63.89 27.27 13.79 24.44* 0.00 0.00 0.00 2.22 33.33 4.00 3,57 7.14 10.81 9.52 0.00 0.00 15.63 0.00 0.00 90.91V 33.33 69.23 95.00 ,35.71 16.67 4.76 0.00 0.00 73.68 0.00 0.00 0.00 0.00 0.00 16.67 34.38 13.89 4.55 17.24 13.33* 4.55 47.83 4.44 20.00 0.00 4.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 16.67 0.00 0.00 0.00 0.00 42.86 0.00 0.00 0.00 winter as compared to other seasons (F^qgq = 241.2. p < .0001, Fig. 2). In addition, scallop gonad weight was significantly greater for samples collected during the winter as compared to other sea- sons (Figgg = 107.6. p < .0001). However, both GSI and gonad weight varied significantly among dates among years (F,^!,.;^ = 57.5, p < .0001: F^^^s, = 47.7. p < .0001. respectively). Specifi- cally, significant reductions in monthly means occurred in both GSI and gonad weight during the 1994/1995 winter (Fig. 3), which conesponds well to the poor gonad development .seen in the gonad condition inde.x above (Table 2). These reductions in GSI and gonad weight may have been a result of osmotic stress associated with the salinity minima during 1994 (Fig. 1). A clear peak in scallop reproduction existed during the winter, with mean GSI values exceeding 15'7r dry tissue weight (Fig. 2). However, based on gonad condition and GSI. significant repro- ductive effort also occurred in the fall and spring (Table 2. Fig. 3). On May 17, 1995 a bay scallop mass spawning event was ob- served. This event occuired while collecting individuals for repro- ductive assessment and natural population surveys. Specifically, scallops held in a catch-bag (n ~ 50-75) spontaneously released gametes. Consequently, independent of GSI, this observation showed that scallops do spawn in late spring. Support of year-long reproductive effort and success was as- sessed by the presence of small scallops (<20 mm shell height) collected in the field. Based on these data, it seems that scallops showed some signs of recruitment throughout he year (Table 3). In addition, during the summer of 1994. a companion study assessing the effects of seagrass habitat architecture on bivalve recruitment collected recruiting bay scallops (0.5-2 mm shell height) on novel substrata and from Thalassia testudinum (turtle grass). These data showed that scallops were recruiting to novel substrata at densities of 8.09. 4.33. and 6.07 individuals m"" for the months of July, August, and September, respectively. Based on the length of larval development (10-14 days, Sastry 1965, Lu and Blake 1997), these periods of recruitment correspond to potential spawning events during June. July, and August, when minimum values in both gonad weight and GSI existed (Fig. 3). DISCUSSION For species of economic value assessing growth, production, and reproduction in populations are essential for both wise man- -H •a E o c o O Fall Values expressed as percentage of sample classified in each category; n indicates the number of scallops collected for reproductive assessment; * indicates a scallop with 31 mm shell height exhibiting this condition; ** indicates a .scallop with 34.1 mm shell height exhibiting this condition; ¥ indicates a scallop with only male gonadal development. Winter Spring Summer Season Figure 2. Comparison of mean scallop CSI for seasons pooled across years: n = the number of scallops collected upon which the means are based; letters above seasons indicate significant differences among means in GSI (a = 0.05). Growth. Production, and Reprodiiction of Bay Scallops 915 Vi O 30- ■J GSI Index Gonad Weighl ! 2 5- 1 " 2 0- (i) Q 1 5- I 0- 5 - - V\ 1 1 [ 3 a. ITO' 0.4 a. 0.2 £ 1/22 1993 1/23 1994 1/23 1995 1/24 1996 Figure 3. .Seasonal patterns in scallop (JSI and gonad weight from individuals collected between November 1992 through October 1996. Peaks in (I.SI and gonad weight occurred during the winter, and mini- mum values were recorded during the summer. Significant interan- nual variability in gonad weight occurred and was significantly re- duced during the winter of 1994/1995. agement of healthy populations and conservation and enhancement of endangered populations. Results from this research have shown that bay scallops from St. Joseph Bay, Florida exhibit high growth rates throughout the summer but have dramatic declines in the fall (Table 1). This observation of differential growth as body size increases has also been shown for larvae and juveniles (Lu and Blake 1996) and was similar to results found by Barber and Blake (1983) for scallops collected from the eastern Gulf of Mexico. This seasonal trend was evident in both 1993 and 1994, but in 1994. reduction in salinity by 21%c may have had a significant impact on the natural growth rates of scallops. Although other researchers have shown that salinity has an impact on the survival of scallops (Duggan 1975. Mercaldo and Rhodes 1982). this is the first case where coupled measurements of physical parameters can be ap- plied directly to differential natural growth rates. Similarly, estimates of natural "mortality" (mortality plus emi- TABLE 3. Presence of recruiting scallops (<20 mm shell height) during 1994 to 1996. Month 1994 1995 1996 February March t April May June July *-H August *-!- September *-H October t November t December t t NA NA Presence denoted by *; *-(• indicates recruiting individuals (<4 mm shell height) collected from suction dredge samples; t indicates samples were not collected during this month; NA indicates the termination of scallop collection. gration) from field studies ranged from 0.005 to 0.025 for non- placement periods (Table 1 ) and were similar to those of Allison and Brand (1995), who used a similar mark-recapture technique on kequipecten opeirulahs. In addition, the rapid dispersal after initial deployment has been recorded from other studies (Barbeau et al. 1996, Hatcher et al. 1996). However, the dramatic mass mortality seen on July 20, 1994 in the field suggests a direct correlation between scallop mortality and the decline in salinity associated with Tropical Storm Alberto (Fig. 1). The observed mass mortality in the field was also seen in the in.stantaneous mortality for mark-recapture scallops but correlates better with the salinity minima (Fig. 1, Table 1) and provides experimental evi- dence to the observed pattern of high mortality. Based on Mer- caldo and Rhodes (1982) estimates, the drop in salinity to \\7 X X vO o a es « n """"^M. p ?< s - Figure 4A. Egg emission from C. glaucum broodstock fed T-ISO and related water temperature. Figure 4B. Egg emission from C. glaucum broodstock fed a 50:50 mixture of T-ISO and T. suecica and related water temperature. 922 Trotta and Cordisco TABLE 2. Number of emission and total eggs per year laid by cockles fed either T-ISO and a 50:50 mixture of T-ISO/7". stiecica in Experiment 3 T-ISO T-ISO/T. suecica Class interval 4 3 3 Class interval 3 17 20 Class interval 2 25 20 Class interval 1 14 15 Total emis./year 59 58 No emis./week 1.13 1.12 No eggs/specimen/year 681.000 687.000 had ripe gonads (stage III and IV), although those fed the latter microalga appeared undernourished after the biopsy test. Experiment .?. Spawning of Cuckles Fed T-ISO and 50/50 Mixture of T-ISO/T. suecica, and Previously Conditioned for 2 Months to Maturation in a 600-L tank with a T-ISO Feeding Regime The cockles conditioned in a 600-L tank started to spawn im- mediately after stocking in the 40-L tanks. The emission of the gametes occurred for 12 months, during which the water tempera- ture regime ranged from 9 to 24"C (Fig. 4A, B ). The specimens fed either T-ISO or the mixture T-ISO/7". suecica. spawned in total 59 and 58 times in the course of the 12 months, with an average of 1.13 and 1.12 times per week, respectively. The mean annual number of laid eggs was 681,000 and 687.000 per cockle for animals fed T-ISO and T-ISO/F. suecica. respectively (Table 2). C. glaucum Sampled in the Wild From January 1995 to October 1996, cockles were found in all stages of maturation, with individuals occurring in the undifferen- tiated to ripe stage at each time period (Fig. 5). DISCUSSION For many bivalves, the period for hatchery operators to obtain, condition, and spawn animals is short. This is because of different factors, among which the most important are water temperature and feeding regime (Helm et al. 1973, Helm 1977, Rossi et al. 1994, Heasman et al. 1996, Wilson et al. 1996). Manzi and Castagna (1989) make exhaustive mention of difficulties met by bivalve mollusks hatchery operators when they want to prolong (anticipating and retarding) the natural spawning period. Alterna- tive methods for extending spawning imply costly procedures for heating and cooling the water during the conditioning or importing the broodstock from lower or higher latitude sites. Beattie ( 1995) gives an economy view of brookslock management and shows that lower costs result when the parent bivalve mollusks are kept in captivity and conditioned, as compared to the continuous supply of wild caught broodstock. Our study shows that it is possible to get viable gametes from C. glaucum. and thereafter seeds, all year at a relatively wide range of temperatures (Fig. 4A, B). As with most bivalve mollusks, the phase of maturation con- ditioning with selected algae feeding is obligatory (Millican and Helm 1994. Heasman et al. 1996). For C. glaucum with proper conditionmg, as shown in experiment 3, it is possible to maintain viable broodstock for seed production all year. This study confirms that D. tertiolecta is of poor nutritional value for C. glaucum. as it is for other bivalve mollusks (Millican and Helm 1994). However. T-ISO is shown to be an adequate food source for the maturation conditioning phase as well as for prolonged spawning period of C. glaucum. C. glaucum is widely distributed throughout Europe. Brock and Wolowicz (1994) mention that this species enters sexual matura- tion phase once a year in the Baltic Sea and twice a year in the north Mediterranean Sea. Zaouali ( 1975) states that the cockle has developed gonads all year in the lagoons of the coastal African Mediterranean Sea, with the exception of 1 month in autumn. Likewise, C. glaucum from Lesina Lagoon and border areas had sexually developed specimens all year (Fig. 5). Unfortunately, live cockles could not be obtained at each sampling interval in this study because of the high mortality of C. glaucum after prolonged anossic period typical of eutrophic lagoon or the poor transparency of the water caused by wind generated resuspension of silt into the water column. From these findings, the reproductive potentiality of this filter feeder seem to be important in the trophic economy of the Lesina Lagoon in all seasons and makes this lagoon more specimen npe ' specimen undifferentiated Figure 5. Maturation phase of gonads in C. glaucum sampled in the wild, during 2 years of survey. To facilitate the figure reading, fertile specimens from stage II to IV were reported as ripe. Maturation Conditioning of Cerastoderma glaucum 923 similar lo the southern Mediterranean wetland ratlier than to the northern one. CONCLUSION By conditioning cockles on T-ISO in the laboratory, cockles ripen and are able to produce viable seeds all year. In light of the fluctuating presence of live cockles and the abundance of empty shells in several places of the lagoon, it seems that, despite the high nutritional value of the suspended organic matter (organic particles and microphytoplankton) and the frequent finding of veliger stages in the plankton sampling (Cordisco 1996), the lagoon does not properly support the reproductive potentiality of this organism. Thus, this bivalve mollusk might play an important ecological role in eulrophic lagoons by converting the energy of organic particles into food for such cockle predators as sea bream (Barbaro et al. 1982). This cockle is found in the Lesina Lagoon in places with different salinity (10 ppt and ."iO ppt). m the ponds near Margherita di Savoia salterns (60 ppt), in the Bay of Cadiz (Spain) (50-60 ppt). It also inhabits such frontier as the Fortore River outlet, which undergoes frequent daily salinity changes (0 ppt to 36 ppt). Considering the temperature range tolerated by these cockles for thriving and spaw ning. in this study, it is intriguing to consider this bivalve mollusk as useful for restocking and production recovery of eutrophic lagoons and embayments with considerable loads of particulate organic matter. ACKNOWLEDGMENTS The authors thank Mr. A. D'Amato, Mr. P. Cammarino, and Mr. N. Dentale for their continuous help and Miss E. Carlino for her fruitful collaboratiim. The work was supported by the 4th Triennal Plan "Aquaculture and Fishery" of the Italian Ministry of Agriculture. Food Resources, and Forestry. REFERENCES Barbaro, A., A. Francescon & R. Guidastri. 1982. Alimentazione di Spams auram in ambiente vallivo. Boll. Miis. Isl. Biol. Univ. Genovci 50:372. Beattie, H.J. 1995. Serial spawning of the geoduck clam (Panopea abnipta). J. Shellfish Res. 14:227. Boyden, C. R. 1971. A comparative study of the reproductive cycles of the cockles Ceraslodernia ediile and C. t^laitciim. J. Mar. Biol. Ass. U.K. 51:605-622. Boyden, C. R. & P. J. C. Russell. 1972. The distribution and habilal range of a brackish water cockle Cardium (Cerasioderma) in the British Isles. J. Anim. Ecol. 41:719-734. Brock, V. & G. Christiansen. 1989. Evolution of Cardium (Cerasioderma) edule, C. lamarcki, and C glaucum: studies of DNA-variation. Mar. Biol. 102:505-511-. Brock, V. & M. Wolowicz. 1994. Comparisons of European populations of the Cerasioderma i^laucmn/C. lamarcki complex based on reproductive phisiology and biochemistry. Oceanologica Acta 17:97-103. Cordisco, C. 1996. Catene trotlche di ambienti vallivi e costieri con par- ticolare riguardo al moUusco bivalve Cerasioderma glaucum. Docu- mento consuntivo relativo all' atdvita di formazione svolta. nell'ambito del Programma del Fondo Strutturale Europeo (FSE). marzo 1995- febbraio 1996. Glude, J. B. 1984. The applicability of recent innovation to mollusc culture in the western Pacific island. Aquaculture 39:29—13. Heasman, M. P., W. A. O'Connor & A. W. J. Frazer. 1996. Temperature and nutrition as factors in conditioning broodstock of the commercial scallop Pecleu fiimalus Reeve: Aquaculture 143:75-90. Helm, M. M„ D. L. Holland & R. R. Stephenson. 1973. The effect of supplementary algal feeding of a hatchery breeding stock of Oslreu edulis L. on larval vigour. / Mar. Biol. Ass. U.K. 53:673-684. Helm, M. M. 1977. Mixed algal feeding of Osirea edulis larvae with Iso- chrysis galbanu and Tetraselmis suecica. J. Mar. Biol. Ass. U.K. 57: 1019-1029. Labourg. P. J. & G. Lasserre. 1980. Population dynamic of Cerastoderma glaucum in an artiUcial lagoon of the Arcachon region. Mar. Biol. 60:147-157. Mann, R. & J. H. Ryther. 1977. Growth of six species of bivalve molluscs in a waste-recycling-aquaculture system. Aquaculture 11:231-245. Manzi. J.J. & M. Castagna (eds.). 1989. Clam mariculture in North America. Elsevier, New York. 461 pp. Millican. P. F. & M. M. Helm. 1994. Effect of nutrition on larvae produc- tion in the European flat oyster, Osirea edulis. Aquaculture 123:83-94. Rossi, R.. D. Campioni. A. G. Conte, F. Paesanti & E. Turolla. 1994. Riproduzione artificiale del Tartufo di mare (Venus verrucosa): primi risultati. Rivista Italiana Acquaculturu 29:53-62. Rygg, B. 1970. Studies on Cerasioderma glaucum (Poiret). Sarsia 43:65- 80. Trotta, P. 1981. A simple and inexpensive system for continuous mono- xenic mass culture of marine microalgae. Aquaculture 22:383-387. Wilson, J. A., O. R. Chaparro & R. J. Thompson. 1996. The importance of broodstock nutrition on the viability of larvae and spat in the Chilean oyster Osirea chilensia. Aquaculture 139:63-75. Wolowicz. M. 1987. A comparative study of a reproductive cycle of cock- les Cardium glaucum (Poiret, 1789) and C. hauniense (Petersen, Rus- sell 1971) (Bivalvia), from the Gdansk Bay. Pol. Arch. Hydrohiol. 34:91-105. Zaouali, J. 1975. Study of the sexual cycle of Ceraslodenna glaucum in two Tunisian hyperhaline lagoons: Tunis Lake and Bou Grara Sea. Rapp. P.-V. Reun.. Conint. Int. E.xplor. Sci. Mer. Medilerr.. Monaco 23:79-80. Jdiinml of Shellfish Research. Vol. 17. No. 4. 925-929, 1998. THE OCCURRENCE OF JUVENILES OF THE GRAPSID CRAB CHASMAGNATHUS GRANULATA IN SIPHON HOLES OF THE STOUT RAZOR CLAM TAGELUS PLEBEWS JORGE L. GUTIERREZ AND OSCAR O. IRIBARNE Depurtamento de Biologi'a (FCEyN) Universidad Nacional de Mar del Pkila CC573 Correo Central Mar del Plata (7600), Argentina ABSTRACT Previous samplings for megalopae and juvenile instars of the southwestern Atlantic burrowing crab Chasmagnathus gramdata showed that they occur at high densities in adult crab burrows. Nevertheless, this study documents the occurrence of juveniles of this species in siphon holes of the stout razor clam Tagehis plebeius after a settlement event. Carapace width of the crabs collected ranged between 1.9 mm and 4.5 mm (2nd to 8th instari. A significantly higher proponion of crabs were found inhabiting inhalant instead of exhalant siphon holes. The siphon holes occupied by crabs showed a significantly larger diameter than those not occupied by them, and a significant positive correlation was found between siphon hole diameter and crab carapace width. A lateral chamber connected to the inhalant siphon gallery was observed in siphon holes inhabited by larger crabs. No correlation was found between the density of pairs of siphon holes and the proportion of pairs of siphon holes inhabited by crabs. However, the proportion of pairs of siphon holes occupied by crabs was higher in a site with low density of adult burrows in relation to adjacent sites with high density of adult burrows. We propose that the presence of clam siphon holes allows crab larvae to colonize poorly structured habitats where adult crab burrows are absent or at low densities, making possible the growth of areas inhabited by crabs. KEY WORDS: Habitat structure. Chasmagnathus gramilalu. Tagehis plebeius, settlement, siphon holes INTRODUCTION In shallow maiine environments, high densities of juvenile life stages or small-sized species of decapod crustaceans are com- monly as.sociated with habitats that show a complex tridimensional structure, such as cordgrass marshes (Zimmerman et al. 1983). seagrass meadows (Thomas et al. 1990), mangrove roots or cano- pies (Wilson 1989), tubiculous polychaete reefs (Gore et al. 1978), beds of mollusk shells (Gunderson et al. 1990). woody debris (Everett and Ruiz 1993). or cobble (Wahle and Steneck 1991 ). The habitat structuring elements serve as obstacles for predator activi- ties and also act to minimize the injurious effects of the different disturbance sources and environmental extremes (Kneib 1984). Thus, mortality risk for these organisms is lower in structurally complex sites with respect to adjacent flat areas (e.g.. Heck and Thoman 1981, Fernandez et al. 1993a). Although differential mor- tality after settlement may account for the higher density of deca- pods in structurally complex habitats (Johns and Mann 1987. Fernandez et al. 1993a). decapod larvae may actively select com- plex habitats as settlement sites (Botero and Atema 1982, Fernan- dez et al. 1993a, Fernandez et al, 1994). This preference may also be shown by juveniles (Johns and Mann 1987. Fernandez et al. 1993a). The grapsid crab Chasmagnathus granidata Dana is one of the most abundant macroinvertebrates in saltmarsh and estuarine en- vironments of the southwestern Atlantic (Boschi 1964). Its distri- bution ranges from Rio de Janeiro (23°S. Brazil) to the San Mati'as Gulf (41°S. Argentina; Boschi 1964). This gregarious species ex- cavates and maintains semipermanent open burrows in the inter- tidal, from the soft bare sediment flats to areas vegetated by the cordgrass Spartina densiflora (Spivak et al. 1994. Iribame et al. 1997). and behaves as deposit feeders in mud flats and as herbivo- rous-detritivorous in 5. densiflora marshes (Iribame et al. 1997). Because of their high density (up to 30 adults ■ m""; J. Gutierrez pers. obs.) and the effect of their deposit feeding and burrowing activities on sediment composition (Botto and Iribame 1997) and benthic community structure (Botto and Iribame 1996). this spe- cies is likely to play a key role in determining the structure of SW Atlantic marshes and estuarine soft-bottom communities. The stout razor clam Tagelus plebeius Solander is a deep burrowing bivalve species that constnicts permanent burrows and siphon holes (Hol- land and Dean 1977). This species is distributed in estuarine en- vironments of the western Atlantic coast from North Carolina (34°N, USA; Holland and Dean 1977) to the San Mati'as Gulf (41°S. Argentina; Iribame and Botto 1998). attaining densities of 200 ind. ■ m'- (Holland and Dean 1977. Iribame et al. 1998) and coexisting with C. gramdata in intertidal mud flat areas in most of their distribution range. Although there is no information, this coexistence pattern may produce several direct and indirect inter- actions. Recently settled megalopae and juvenile Chasmagnathus granidata are found at high densities in association with adult crab burrows, and dense burrow beds seemed to be the major nursery habitat for this species (Spivak et al. 1994). However, during ob- servations performed after a settlement event, we noticed the pres- ence of juvenile C. gramdata dwelling in siphon holes of Tagelus plebeius. Given that this type of shelter may be important, the purpose of this work is to describe the pattern of siphon hole use by juvenile C. granidata. focusing on; (1) size and molting instar of the crabs using the holes; (2) type of siphon hole occupied by crab (inhalant or exhalant); (3) differences between siphon holes occupied and not occupied by them; and (4) the percentage of holes occupied and density of crabs dwelling in siphon holes in crab beds with different densities of adult burrows. MATERIALS AND METHODS This study was carried out in the Mar Chiquita coastal lagoon (Argentina. 37°32' to 37°45' S. 57° 19' to 57°26' W). a 46 km= body of brackish water (Fasano et al. 1982) affected by low am- plitude ( .05). The percentage of pairs of siphon holes occupied by crabs was significantly higher at the LBD site (LBD: x = 31.24%, SD = 5.38; HBD: x = 8.52%, SD = 2.34; Mann-Whitney test; Z = -3.55, p < .001; Fig. 4). How- ever, the density of crabs occupying holes did not differ between 12 F 10 E HI 8 \- LU ? < b Q LU — 1 4 o X I ° I , a: 2.5 LU Q 2 1.5 < o LU _l o X cr O z in o I o 0.5 B — \ — 1 A CRAB NO CRAB Figure \. Median quantile box plots showing: (.\) the diameter of the siphon holes occupied by juvenile Chasmagnathus granulata and the coniplementaries of each pair: and (B) the ratio between the diameter of the major and minor hole for pairs of siphon holes with and without crab. Juvenile Crabs Dwell in Clam Siphon Holes 927 2 3 CARAPACE WIDTH (mm) Figure 2. Carapace width of juvenile Chasmagnathiis graniilata against diameter of the siphon hole occupied by them. LBD HBD sites (LBD: x = crabs ■ m~'^, SD 3.87 crabs ■ m = 1.51; f-test: t SD = 0.83; HBD: x = 3.4 0.79, df = 16, p > .05). DISCUSSION Larval decapod settlement is primarily an active process in which larvae choose settlement sites based on sediment character- istics or chemical cues (Castro 1978. Botero and Atema 1982). Flume experiments demonstrated that megalopae of Chasmag- nalhiis gramdata actively swim in current conditions similar to those encountered in the field, and it was also proposed that they are able to select settlement sites (Valero 1998). In addition, the higher densities of recently settled megalopae and juveniles occur in association with adult crab burrows (Spivak at al. 1994), and metamorphosis of C. gramdata megalopae occurs earlier in the presence of chemical cues of adult conspecifics than with sea water alone (Gebauer et al. 1998). All these data suggest that megalopae of this species settle in response to adult-released chemical cues. In this context, the occurrence of juvenile C. grami- lala dwelling in siphon holes of stout razor clams cannot be ex- plained by means of selective settlement of megalopae. An alter- native hypothesis may be stated on the basis of competition be- tween cohorts. Settlement of C. gramdata megalopae occurs in the lagoon between December and June, with peaks in intensity (Spi- vak 1994). Thus, there is a potential for interaction between co- horts. Eariy cohorts of the Dungeness crab Cancer magister re- duces the abundance of subsequent cohorts in intertidal shell habi- CHAMBER NO CHAMBER Figure 3. Median quantile box plots showing carapace width of juve- nile Chasmagnatlius gramilata occurring al siphon holes with and without lateral chambers. Figure 4. Median quantile box plots showing the proportion of pairs of siphon holes occupied bv juvenile Chasmagnatlius granulata at the site of low (LBD) and high density of adult burrows. tats, and smaller juveniles migrate to the open fiats in response to high density of conspecifics (Fernandez et al. 1993b). Eariy juve- nile instars of C. granulata are subject of cannibalism by larger juveniles (Luppi et al. 1995). and it may be possible that a density- dependent migration from adult burrows to siphon holes has been carried out by crabs settling in late summer. Nevertheless, the possibility of direct settlement of C. granu- lata megalopae in siphon holes cannot be discarded. Muddy sand substrata (such as those occurring at both study sites) also pro- motes an eariier metamorphosis of C. granulata megalopae when compared with coarser sediment types. Despite the fact that a combination of muddy substrata and adult chemical cues deter- mines the shortest time to metamorphic molt, muddy sand sub- strata alone have an important effect (Gebauer et al. 1998). If we assume that muddy sand substrata alone may also induce settle- ment of C. granulata megalopae, settlement in sites other than adult burrows is plausible. In addition, Gebauer et al. (1998) found that artificial substrata of the same grain size do not have the same effect on metamorphic molt, suggesting that characteristics other than grain size are relevant for the induction of metamorphosis (see Pawlik 1992). As previously mentioned, sediment from adult burrows and sediment depressions in association with adult bur- rows are organically richer than those of the surrounding flat. Taking into account that this species behaves as a deposit feeder in unvegetated sediment fiats, it is likely that high levels of organic matter may function as a cue for the settlement of megalopae of this species. Although no data are available about the small-scale distribution of organic matter around bivalve siphon holes, in- creased levels may occur as a result of the enhancement of local particle deposition caused by flow convergence toward the inhal- ant siphon (see Ertman and Jumars 1988). Moreover, larvae may respond to physical factors (see Pawlik 1992). Negative bottom roughness elements (such as burrows and siphon holes) behave as hydrodynamically quiet microhabitats (DePatra and Levin 1989). Weak current conditions are commonly associated with habitats that have a complex tridimensional structure (e.g.. oyster beds, Wright et al. 1990). which may provide refuge for juvenile deca- pods. Dungeness crab C. magister megalopae are able to select positively for weak current conditions (e.g., Fernandez et al. 1994). In addition, metamorphosis of the blue crab Callinectes sapidus megalopae is accelerated by textural cues caused by the presence of eelgrass (Forward et al. 1994). Burrows or siphon holes may produce similar effects on C. gramdata megalopae. 928 Gutierrez and Iribarne Siphon holes occupied by juvenile Chasinagnatluis graimlata appeared modified with respect to the complementary of the pair. These crabs enlarge the diameter of the siphon holes occupied by them and larger crabs also construct lateral chambers. This may occur as a result to the provision of space for both the crab and the extended siphon. The thalassinidean shrimp Jci.xea noctiirna in- habit burrows of the echiurid Maxmiielleria lankesteii. modifying them by the excavation of semicircular side branches (Nickell et al. 1994). In addition, it was proposed that / noctiirna probably ben- efit by the irrigation activities of the echiuran. which supply both oxygen and food (Nickell et al. 1994). Pinnotherid crabs Pinnixa schmiui and Scleroplax granulata behave as symbionts in burrows of the suspension feeding burrowing shrimp Upogebia pugettiensis of the Northwestern Atlantic (Kozloff 1987). Thus, the higher proportion of juvenile C. gniniihita occurring in siphon holes cor- responding to the inhalent siphon may be the result of selectively favoring a site with a high inflow of food particles. However, we do not know to what extent clam attributes are affected (e.g., decreased filtration and oxygen consumption rates in mussels with pea crabs in the mantle cavity; Biembaum and Shumway 1988), to determine if the relationship is commensalistic of parasitic. It is also interesting to notice that the percentage of holes oc- cupied by juvenile Chasmagnathiis gninutata was higher in the site with a low density of adult burrows than in the site with a high density of adult burrows. This pattern indicates that where habitat structure is poor, small scale sediment features may be important as settlement sites and/or refuges for this species. We also believe that the strategy of settling into holes of bivalves or other burrow- ing species may be a fairly common phenomenon. For example, new settlers of the Dungeness crab Cancer inagister have been found in holes of burrowing shrimps in the Grays Harbor estuary (Washington, USA: O. Iribarne, pers. obs.) but their importance has never been quantified. ACKNOWLEDGMENTS Support for this project was provided by grants from the Universidad Nacional de Mar del Plata. CONICET (PIA No; 6097) and IFS {Sweden, No. A/2.'iOI-l/2). We very much appre- ciate the comments and conections made by two anonymous reviewers. LITERATURE CITED Biernbaum, R. M. & S. E. Shumway. 1988. Filtration and oxygen con- sumption in mussels, Mytilus ediilis. with and without pea crabs. Pin- notheres maculatus. Estuaries 11:264—271. Boschi, E. E. 1964. Los crustaceos decapodos brachyura del litoral honaer- ense (R. Argentina). Bol. Inst. Biol. Mar. (Mar del Plata. Argentina) 6:1-99. Botero, L. & J. Alenia. 1982. Behavior and substrata selection during lar\al settling in the lobster Homarus americanus. J. Crust. Biol. 21:59-69. Botto, F. & 0. 0. Iribarne. 1996. Efecto del cangrejo cavador Chasmag- nathiis granulata sobre las comunidades bentonicas de las marismas. Paper presented at 3rd Encontro de Especialistas em Decapoda Brachyura, Rio Grande, Brazil. Botto. F. & O. O. Iribarne. 1997. The effect of ihe southwestern Atlantic burrowing crabs Chasmagnathus granulata and ilea uruguayensis on the sediment composition and hedload transport in estuarine environ- ments. Paper presented at 7th Congreso Latinoamericano de Ciencias del Mar, Santos, Brazil. Conover, W. J. 1980. Practical nonparametric statistics, 2nd ed. John Wiley & Sons, New York. 493 pp. Castro, P. 1978. Settlement and habitat selection in the larvae of Echino- ecus pentagonus (A. Milne Edwards), a brachyuran crab symbiotic with sea urchins. J. E.xp. Mar. Biol. Ecol. 34:259-270. DePatra, K. D. & L. A. Levin. 1989. Evidence of the passive deposition of meiofauna into fiddler crab burrows. J. Exp. Mar. Biol. Ecol. 125:173- 192. Ertman, S. C. & P. A, Jumars. 1998. Effects of bivalve siphonal currents on the settlement of inert particles and larvae. / Mar. Res. 46:797-813. Evereu, R. A. & G. M. Ruiz. 1993. Coarse woody debris as a refuge from predation in aquatic communities: an experimental test. Oecologia 93; 475^86. Fasano, J. L., M. A. Hernandez, F. I. Isla & E. J. Schnack. 1982. Aspectos evolutivos y ambientales de la laguna Mar Chiquita (provincia de Bue- nos Aires, Argentina). International Symposium on Coastal Lagoons, Bordeaux, France. September 8-14. 1981. Acta Oceunologica Special Number 285-292. Fernandez. M.. O. O. Iribarne & D. A. Armstrong. 1993a. Habitat selection by young-of-the-year Dungeness crab Cancer magisler and predation risk in intertidal habitats. Mar. Ecol. Prog. Ser. 92:171-177. Fernandez. M.. D. A. Armstrong & O. O. Iribarne. 1993b. First cohort of young-of-the-year Dungeness crab Cancer magisler reduces abundance of subsequent cohorts in inlerlidal shell habitats. Can. J. Fish. Aquat. Sci. 50:2100-2105. Fernandez, M., O. O. Iribarne & D. A. Armstrong. 1994. Swimming be- havior of Dungeness crab. Cancer magisler. Dana, megalopae in still and moving water. Estuaries 17:271-275. Forward, R. B.. D. A. Z. Frankel & D. Rittschof 1994. Effects of environ- mental cues on metamorphosis of the blue crab Callinectes sapidus. Mar Ecol. Prog. Ser 131:165-177. Gebauer, P., I. Walter & K. Anger. 1998. Effects of substratum and con- specific adults on the metamorphosis of Chasmagnathus granulata (Dana) (Decapoda: Grapsidae) megalopae. J. Exp. Mar. Biol. Ecol. 223:185-198. Gore. R. H., L. E. Scotto & L. J. Becker. 1978. Community composition, stability, and trophic partitioning in decapod crustaceans inhabiting some subtropical sabellariid worm reefs. Bull. Mar. Sci. 28:221-248. Gunderson, D. R., D. A. Armstrong, Y. B. Shi & R. A. McConnaughey. 1990. Patterns of estuarine use by juvenile English sole (Parophrys vetulis) and Dungeness crab (Cancer magisler). Estuaries 13:59-71. Heck, K. L. & T. A. Thoman. 1981. Experiments on predator-prey inter- actions in vegetated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125- 134. Holland, A. F. & J. Dean. 1977. The biology of the stout razor clam Tagelus plebeius. 1. animal-sediment relationships, feeding mecha- nism, and community biology. Chesapeake Sci. 18:58-66. Iribarne. O. O. & F. Botto. 1998. Orientation of the extant stout razor clam Tagelus plebeius in relation to current direction: its palaeoecologic implications. / Shellfish Res. 17:165-168. Iribarne. O. O.. A. Bortolus & F. Botto. 1997. Between-habitat differences in burrow characteristics and trophic modes in the southwestern At- lantic burrowing crab Chasmagnathus granulata. Mar. Ecol. Prog. Ser. 155:137-145. Iribarne, O. O., J. L. Valero, M. M. Martinez, L. O. Lucifora & S. Bach- mann. 1998. Shorebird predation may explain the origin of Holocene beds of stout razor clams in life position. Mar. Ecol. Prog. Ser. 167: 301-306. Johns. P. M. & K. H. Mann. 1987. An experimental investigation of juve- nile lobster habitat preference and mortality among habitats of varying structural complexity. / Exp. Mar. Biol. Ecol. 109:275-285. Kneib. R. T. 1984. Patterns of invertebrate distribution and abundance in the intertidal salt marsh: causes and questions. Estuaries 7:392^12. Juvenile Crabs Dwell in Clam Siphon Holes 929 Kozloft. E. N. 1987. Marine invertebrates of the Pacific Northwest. Uni- versity of Washington Press, Seattle, WA. 511 pp. Lanfredi, N. W., C. F. Balestrini, C. A. Mazio & S. A. Schmidt. 1987. Tidal sandbanks in Mar Chiquita coastal lagoon, Argentina. J. Coast. Res. 3:515-520. Luppi, T., E. Spivak & K. Anger. 1995. Canihalismo y depredacion en Cyrtograpsus angulaliis y Chasmagnathus graindaui. Paper presented at 7th Congreso Latinoamericano de Ciencias del Mar. Mar del Plata, Argentina. Nickell. L. A., R. J. A. Atkinson, D. J. Hughes, A. D. Ansell & C. J. Smith. 1994. Burrow morphology of the echiuran worm Ma.xmueUeria lanke- steri (Echiura: bonellidae), and a brief review of burrow structure and related ecology of the Echiura. J. Nat. Hist. 29:871-885. Olivier, S. R., A. Escofet, P. Penchazadeh & J. M. Orensanz. 1972. Estu- dios ecologicos de la region estuarial de Mar Chiquita (Buenos Aires, Argentina). \. Las comunidades bentonicas. Anal. Sol. Cient. Arg. (La Plata. Argentina) 193:237-262. Pawlik, J. R. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Oceanog. Mar. Biol. Aim. Rev. 30:273-335. Rieger, P. J. & C. Nakagawa. 1995. Desenvolvimento juvenil de Chas- magnathus gramilata Dana, 1851 (Crustacea, Decapoda, Grapsidae) em laboratorio. NaupUus (Rio Grande, Brazil) 3:59-74. Spivak. E. 1994. Ciclos de vida de dos grapsidos en una laguna costera templada (Mar Chiquita. Buenos Aires, Argentina). Paper presented at 2nd Taller sobre cangrejos y cangrejales. Mar del Plata, Argentina. Spivak, E., K. Anger, T. Luppi, C. Bas & D. Ismael. 1994. Distribution and habitat preferences of two grapsid crab species in Mar Chiquita Lagoon (Province of Buenos Aires, Argentina). Helgolander Meeresunters 48: 59-78. Thomas. J. L.. R. J. Zimmerman & T. J. Minello. 1990. Abundance pat- terns of juvenile blue crabs (Calliiiectes sapidus) in nursery habitats of two Texas bays. Bidl. Mar. Sci. 46:1 15-125. Underwood, A.J. 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge University Press, Cambridge, UK. 504 pp. Valero, J. L. 1998. Velocidades de nataci6n y comportamiento natatorio de megalopas de dos cangrejos estuariales. Thesis, Universidad Nacional de Mar del Plata, Mar del Plata, Argentina. Wahle, R. A. & R. S. Steneck. 1991. Recruitment habitats and nursery grounds of the American lobster {Homarus americanus Milne Ed- wards!: a demographic bottleneck? Mar. Ecol. Prog. Ser. 69:231-243. Wilson, K. A. 1989. Ecology of mangrove crabs: predation, physical fac- tors, and refuges. Bull. Mar. Sci. 44:263-273. Wright, L. D., R. A. Gammisch & R. J. Byrne. 1990. Hydraulic roughness and mobility of three oyster bed artificial substrata material. J. Coast. Res. 6:868-878. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ. 718 pp. Zimmerman, R. J.. T. J. Minello & G. Zamora. 1983, Selection of veg- etated habitat by brown shrimp, Peneus aztecus. in a Galveston Bay salt marsh. Fish. Bull. 82:325-336. Joiiniul nf Shellfish Research. Vol. 17. No. 4, 431-4.^3. 1W8. POPULATION BIOLOGY OF XIPHOPENAEUS KROYERI (HELLER 1862) (DECAPODA: PENAEIDAE) FROM UBATUBA BAY, SAO PAULO BRAZIL J. M. NAKAGAKI AND M. L. NEGREIROS-FRANSOZO Depaitaiiicnto de Zoologia Instilulo dc Biocieiicias e Centro de Aquivulnira Universidade Esladiial Paulista (UNESP) 18618-000 Botiicani Sao Paulo. Brasil ABSTRACT The population structure and abundance of Xiphopenaeus kroyeri (Heller, 1862) were analyzed during monthly samples from October 1992 to September 199.3 in Ubatuba Bay (23°26'S and 45°02'W). Brazil. Sampling was carried out at two parallel transects: one of them located in its midregion and the other at the mouth of the bay. After trawling, shrimps were separated from other benthic organisms, sexed. and counted. Their total length was also measured, and the degree of gonadal development was assessed. Xiphopenaeus kraxeri. the most common penaeid species in the bay. was recorded in all samples, but its abundance decreased from November to March. Size ranged from 14.3 to 1 18.3 min in males and from 12.7 to 133.6 mm in females, suggesting a slight sexual dimorphism related to body size. Males prevailed during most of the year; whereas, females predominated during summer and midwinter. Based on the percentage of mature females during this study, two main reproductive periods were identified, occurring in spring and autumn. Despite some breeding activity throughout the year, such a trend indicates that the population follows a tropical/ subtropical reproductive pattern. KEY WORDS: Population biology, Penaeidac, Xiphopenaeus kroxeri. reproduction, sex-ratio INTRODUCTION Xiphopenaeu.s kroyeri is the most intensively exploited shritiip species in Sao Paulo State. According to Pires (1992), X. kroyeri and the swimming crab Portumis spinicarpus (Stimpson 1871 ) are the most abundant species of the benthic niegafauna in the conti- nental shelf off the study area. This species represents the second most important fishery resource along the coa.st of Sao Paulo State (Rodrigues et at. 1993), and its trophic relationships may be es- sential in maintaining the stability of benthic communities (Pires 1992), Despite being an extremely abundant species along the Brazilian coast, information on the biology, ecology, and behavior of X. kroyeri is scarce (Vieira 1947, Mota-Alves and Rodrigues 1977, Cortes and Ciiales 1990, Cortes 1991, Rodrigues et al. 1993 and Branco et al. 1994). Because of the complexity of their life cycle, studies on penaeid shrimp populations (e.g., migration and reproduction) are needed to improve fishery management. Boschi (1969) pointed out its importance when he studied Artemesia longinaris Bate 1888 in Mar del Plata, noting a continuous change of its age composition structure. The study of penaeid reproductive cycles is also important, and usually is achieved by means of recording the degree of gonadal development in sampled specimens. In this procedure, a number of development levels are established and described, usually ranging trom three to five (i.e,, immature, in maturation, almost mature, mature, and spawned). Relevant contributions concerning the re- productive biology of different penaeid shrimp species are Vieira (1947), Olguin-Palacios (1967), Perez-Farfante (1969), Mota- Alves and Rodrigues (1977), Motta-Amado (1978), and El Hady et al. ( 1990). This study examines the abundance and the population biology of X, kroyeri in Ubatuba Bay, Ubatuba, Sao Paulo, Brazil, with emphasis on its population structure and reproductive period. MATERULS AND METHODS Monthly trawlings were carried out from October 1992 to Sep- tember 1993 in Ubatuba Bay. Samples were taken along two transects: transect A, located at the shallow midregion of the bay, and transect B located at the deeper bay mouth (Fig. 1 ). Trawlings were accomplished with an otter trawl net ( lO-mm mesh cod end) and were conducted at constant speed for 1 h, covering a 7,400 m" area. The shallow region of the bay is strongly affected by coastal environmental conditions, receiving freshwater drainage from four rivers. Otherwise, the deep stratum is to subject a greater oceanic influence. Some physical factors were monitored at each transect. Bottom water temperature was obtained with a Nansen bottle provided with a thermometer (±1.0°C). Depth was obtained by means of a marked rope attached to the Nansen bottle, and a VanVeen grab sampler was used to obtain sediment samples. Sieving analysis according to Wentworth grades were carried out for grain size classification. Sediments were .sorted using the phi-.scale (<()) in 1.0- and 4.0 <|)- including the fractions under 4.0 i) (Hakanson and Jansson 1983). Sediment organic con- tents were obtained using the loss on ignition method (Hakanson and Jansson 1983), The shrimps were separated, sexed, and measured (total length, TL) with a vernier caliper to the nearest 0. 1 mm. A Student's r-test was used to detect size differences between sexes. Box plots were performed for males and females in each month to analyze the population structure through the study period. Monthly .sex ratios were also obtained. Median size differences between transects were tested in each month using a Mann-Whitney U-test and among months in each transect using a Kruskal-Wallis test (Si'egel, 1956). In each trawl, subsamples from 200 to 400 individuals were separated for examination of gonads. In females, four development .stages were considered according to Motta-Amado (1978): I- immature, Il-developing, Ill-mature, and IV-spent. In males, the presence (or absence) of spermatophores in the terminal ampoule was recorded. Monthly sex ratios and proportions of mature females were statistically compared by means of a Goodman's test ( 1964, 1965), 931 932 Nakagaki and Negreiros-Fransozo ■25' 26' -27 Figure 1. Map of Ubatuba Bay showing the position of sampling transects. This analysis is based on the binomial proportion comparison for contrasts between and within multinomial populations. These re- suhs were analyzed at the 5% significance level. To estimate the mean size of first maturation in each sex. the Gallon's Ogive, Fr = 1 - t-""^'-'' (Fonteles-Filho 1989). was adjusted to fit the total length (TL. independent variable) versus relative frequency of mature individuals (Fr. dependent vanable) scatterplots. Mature individuals were defined as male with full terminal ampoule and females with gonads in stage II. III. or IV. RESULTS Water temperature averaged 2.3.16 ± 2.97°C (ranging from 20 to 28°C). with highest values recorded from February to April. Mean depth at transect A is 7.6 ± 1.3 m. Along this transect sediments are very poorly sorted (ct, = +2.025), mainly consisting of fine sand (Mz = 2.73 ) and low organic contents (2.97 ±0.61%). Xiphopenaeiis kroyeri was very abundant in the area (5.027 individuals in transect A and 5.282 in transect B) during all sam- pling periods, e.xcept from December to February in transect B (Fig. 2). This fact indicates a seasonal abundance variation. Females ranaed from 12.7 to 133.6 mm (74.52 ± 17.47) and ^^1 Transect A I I Transect B /\pr Ma> Jun Months Aug Sep Figure 2. Monthly abundance of .V. kroyeri in both transects. males from 14.3 to 1 18.3 mm (71.82 ± 14.01), indicating a sexual dimorphism, in which females attain a larger size (Student's r-test, t = 8.825, p< .0001). Mann-Whitney comparative analyses of size frequency distri- butions in transects A and B revealed significant size differences in Oct. 1992, Apr., May, June, and Aug. 1993 (Table 1. Figs. 3 and 4), showing that the population is not equally distributed in these areas. During Oct. 1992, April, and June 1993. larger shrimps were sampled in transect A. and during May and Aug. 1993. larger specimens were captured in transect B. Differences in shrimp median size were repeatedly verified among sampled months (Kruskal-Wallis, p < .001) (Table 2, Figs. 3 and 4). However, they were too complex to reveal recruitment pattern. Males were generally slightly predominant, but the sex ratio varied throughout the year. In November 1992. January and July 1993 (Fig. 5). an increase (Goodman's test, p < .05) of the relative number of females was observed, when sex ratios attained 1 : 1 .78; 1:1.28, and 1:1.2, re.spectively. Xiphopenaeus kroyeri breeds all year, but higher reproductive activity was verified in some periods. Based upon the data ob- tained from gonadal analysis in females, it can be concluded that higher reproductive activity occurred in November 1992, and March, August, and September 1993 (Fig. 6). In the case of males, higher proportions of individuals with full terminal ampoule were recorded in two main periods (November 1992 and May 1993) (Fig. 7). Males achieve sexual maturity (68.02 mm) at a smaller size than females (83.19 mm) (Fig, 8). DISCUSSION The abundance of X. kroyeri showed a marked seasonal varia- tion. During the summer period (December to March) this species' abundance is lower, mainly in the deeper portion of the bay. Sig- noret (1974) observed that X. kroyeri follows a similar pattern in the Terminos Lagoon (Mexico), with low abundances from sum- mer to fall. The low abundance of A', kroxeri during summer may be related to the intrusion of a cold current. Castro-Filho et al. (1987) indi- cated the presence of three oceanic currents in the Ubatuba region, the coastal water (CW) (T > 20°C), the South Atlantic central water (SACW) (T < 18°C), and the tropical water (TW) (T > 20°C). Pires (1992), who studied the benthic megafauna commu- nities in the continental shelf of Ubatuba region, observed by means of a cluster analysis that there is a close association between some species abundance and specific environmental conditions. This is the case of positive correlation between X. kroyeri and the prevalence of CW during winter. Despite different sediment features found in each transect, the incoming SACW during summer could be the most important physical factor influencing the distribution of this species. The physical action of the current itself together with low temperature conditions would restrain the population distribution of .Y. kroyeri within the study area. Statistical differences in shrimp median size among sampled months do not support a growth model through time, which could have explained the growth pattern in this population. The great fishery effort in Ubatuba Bay probably affects the species, as observed by Somers et al. ( 1987) in P. escidentus Haswell in the ToiTes Strait (Australia), who suggested a continuous recruitment and/or the existence of a size-dependent source of mortality. In the present study, the comparison of shrimp size in transects A and B, Population Biology of Xiphopenaeus kroyeri 933 TABLE 1. Mann-Whitney analysis in .V. kroyeri size comparisons in eacli montli between transects in Ubatuba Bay. Transect A Transect B Number of Number of Month individuals Ranked sums individuals Ranked sums U October 1992 552 735.56 719 559.57 8.47* November 1992 193 180.08 163 176.63 0.31 December 1992 184 95.00 7 122.28 1.28 January 1993 142 73.79 4 63.25 0.50 February 1993 409 229.59 52 242.08 0.63 March 1993 673 — — — April 1993 1 169 1568.07 1640 1288.34 8.996* May 1993 390 351.24 370 411.34 3.77* June 1993 275 816.52 1235 741.91 2.56* July 1993 281 509.77 695 479.90 1.50 August 1993 466 381.11 401 495.46 6.70* September 1993 293 296.16 296 243. S5 0.16 ' Statistical significant differences at a = 0.01. reveals remarkable differences in the population distribution. which are likely to reflect recruitment and migration processes. The sex ratio variations observed in this study are supported by other results obtained for the same species. According to Signoret (1974), the sexual distribution through the year in X. kroyeri is not homogeneous, with males and females often strongly segregated. According to Wenner ( 1972), mentioning the Fisher theory, the 1:1 proportion is favored by natural selection. Wenner (1972) Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep 1»'2 Months "" o Outlier Ncm-outUer max Ncm-outlier min o Outlier Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep '"2 Months 1'" Figure 3. Series of box plot graphics for monthly size of males (upper) and females (below) obtained in transect A. Oct Nov Ctec Jan Feb Mar Apr May Jun Jul Aug Sep 1992 ., , 1993 Months Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep 1992 ....^. 1993 o Outlier Non-outlier max Non-outlier min o Outlier Months Figure 4. Series of box plot graphics for monthly size of males (upper) and females (below) obtained in transect B. TABLE 2. Results of Kruskal-Wallis analysis. Transect H df Probability 5027 5282 576.25 640.68 11 in p<.001 P<.001 934 Nakagaki and Negreiros-Fransozo A B A B A A A A B A A a b ah " ab ab a ab hr ab ^ Female Oct Nov Dec 1992 Jul Aug 1993 Sep Feb Mar Apr May Jun Months Figure 5. Monthly sex ratio for V. kroyeri. Capital letters inside the bars indicate comparisons w ithin month, and lower case letters outside the bars indicate comparisons among months. The same letters indi- cate no statistical differences. pointed out that sex-dependent mortality, activity, migration, habi- tat utilization and al.so the effect of restricted food resources are important factors explaining departures from the Mendelian pro- portion. Wenner also stated that the 1:1 ratio is an exception rather than the rule in crustacean populations. He concluded that sex ratio can be a function of size for a given species. The temporal sex ratio variation can be related to a seasonal reproductive pattern in ,V. kroyeri. The proportion of males was higher during most of the year, but females were more abundant in November (spring), when it was recorded a 1:1.78 sex ratio. This peak coincides with major reproductive activity. Contrarily, Cortes (1991) observed that during spawning, males outnumbered fe- males in a Colombian Caribbean population. This fact supports the hypothesis of sex-dependent pattern of migration, because Cortes collected the shrimps near the coast at depths ranging from 1 .5 to 3 m. e e Nov Dec mi Jul 893 Aug St'p Mar Apr May Months Figure 6. Bar graph showing gonadal maturity in females. Capital letters inside the bars indicate comparisons within month, and the lower case letters outside the bars indicate comparisons among months. The same letters indicate no statistical differences. Data of transects A and B are grouped. Females in gonad stage I were con- sidered immature, and females in stages II, III, and IV were consid- ered mature. Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Months Figure 7. Terminal ampoule status in males. Analyzing both transects, we can assume that X. kroyeri ex- hibits a tropical/subtropical reproductive pattern (Dall et al. 1990), in which there is a main reproductive period in the spring and a secondary one in the fall. The present results are similar to those observed by Mota-Alves and Rodrigues (1977). Motta-Amado ( 1978), and Cortes (1991). It can be assumed that the presence of juvenile individuals from February to May is attributable to spawning events during spring. The onset of sexual maturity can vary between populations. The size estimates in this study (68.02 mm in males and 83.19 mm in females) are larger than those observed by Rodrigues et al. ( 1993) (62 mm for males and 71 mm for females) in other locali- ties within Sao Paulo State. The determination of this parameter can be important for assessment of the reproductive stock in natu- ral populations (Fonteles-Filho 1989) and for guidance of future governmental fishery control. Establishing a minimum catch size and defining the period of lower relative abundance of juvenile shrimps will help establish rational management of this species' exploitation. ACKNOWLEDGMENTS CAPES (Coordenadoria de Aperfeigoamento Superior) pro- vided financial support as a fellowship to the first author. We are O r^9872 '^f A./ / A / 7 • Female / * Male A, / / / */ »=i^i — 1 r ^ r-=09823 1 1 ' 1 ■■ 1 ' 1 Total Lengh(mm) Figure 8. .V. kroyeri. Frequency of morphologically mature males and females as a function of total length. Population Biology of Xiphopenaeus kroyeri 935 also grateful to Carlos Roberto Padovani tor his statistical assis- tance and to Adilson Fransozo and Augusto A. V. Flores for manu- script suggestions. Thanks are extended to other NEBECC mem- bers for helping during fieldwork. REFERENCES Boschi, E. E. 1969. Estudio biologico pesquero del camaron Arlemesui longinaris Bate de Mar del Plata. Bol. Inst. Biol. Mar. 18:1^7. Branco, J. O., M. J. Lunardon-Branco & A. de Finis. 1994. Crescimcnto de Xiphopenaeus kroyeri (Heller. 1862) (Crustacea: Natantia: Penaeidae) da regiao de Matinhos. Parana, Brasil. An/. Biol. Tecnol. 37:1-8. Castro-Filho, B. M., L. B. de Miranda & S. Y. Miyao. 1987. Condi96es hidrograficas na plataforma conlinental ao largo de Ubaluba: variances sazonais e em media escala. Bolm. Inst. Oceanogr. 35:135-151. Cortes. M. L. 1991. Aspectos reproductivos del camaron Xiphopenaeus kroyeri (Heller) en Costa Verde, Cienaga (Caribe Colombiano). Cal- dasia 16:513-518. Cortes, M. L. & M. M. Criales. 1990. Analisis del contenido estomacal del camaron titi Xiphopeiweus kroyeri (Heller) (Crustacea: Natantia: Pe- naeidae). Na. Inst. Invest. Mar Punta Betin 19-20:23-33. Dall, W., B. J. Hill, PC. Rothlisberg & D. J. Staples. 1990. The biology of the Penaeidae. V.27. //;.• J. H. S. Blaxter and A.J. Southward (eds). Advance. Mar. Biol. Academic Press, London. 489 p. El Hady. H. A., F. A. Abdcl-Ra/ek & A. Ezzai. 1990. Reproduction of Penaeus semisulcatus De Haan en Damman water (Arabia Gulf), King- dom of Saudi Arabia. Arch. Hydrobiol. 118:241-251. Fonteles-Filho, A. A, 1989. Recurso pesqueiro: biologia e dinamica popu- lacional. Imprensa Oficial do Ceara, Fortaleza, Brazil. 296 pp. Goodman, L. A. 1964. Simultaneous confidence intervals for contrasts among multinomial populations. Ann. Math. Statist. 35:716-725. Goodman, L. A. 1965. On simultaneous confidence intervals for multino- mials proportions. Technometrics 7:247-254. Hakanson, L. & M. Jansson. 1983. Principles of lake sedimentology. Springer-Verlag, Berlin. Heidelberg. 316 pp. Mota-Alves, M. I. & M. M. Rodrigues. 1977. Aspectos da reprodu^ao do camarao sete-barbas, Xiphopenaeus kroyeri (Heller) (Decapoda, Macrura) na costa do Estado de Ceara. Arq. Cien. Mar. 17:29-35. Motta-Amado, M. A. P. 1978. Estudo biologico do Xiphopenaeus kroyeri (Heller, 1862), camarao "sete-barbas" (Crustacea, penaeidae) de Mat- inhos, Parana. Master's Thesis. Federal University of Parana, Curitiba, Brazil. 94 pp. Olguin-Palacios. M. 1967. Estudio de la biologi' del camaron cafe Penaeus califoniiensis Holmes. FAO Fish. Repl. 57:331-356. Perez-Farfante, I. 1969. Western Atlantic shrimps of the genus Penaeus. Fish. Bull. 67:461-590. Pires, A. M. S. 1992. Structure and dynamics of benthic megafauna on the conlinental shelf offshore of Ubatuba, southeastern Brazil. Mar Ecol. Prog. Ser 86:63-76. Rodrigues, E. S.. J. B. Pita, R. Gra^a-Lopes, J. A. P. Coelho & A. Puzzi. 1993. Aspectos biologicos e pesqueiros do camarao sete-barbas (Xi- phopenaeus kroyeri) capturado pela pesca artesanal no litoral do estado de Sao Paulo. B. Inst. Pesca 19:67-81. Siegel. S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill. New York. 312 pp. Signoret, M. 1974. Abundancia, tamaiio. y distribucion de camarones {Crustacea. Penaeidae) de la Laguna de Terminos, Campeche y su relacion con algunos factores hidrologicos. Ann. Inst. Biol. Univ. Nal. Anion. Mexico 45. Ser. Zoologia 0:119-140. Somers. L F.. Poiner. I. R. & Harris, A. N. 1987. A study of the species composition and distribution of commercial penaeid prawns of Torres Strait. Ausl. J. Mar. Freshwater Res. 38:47-61. Vieira, B. B. 1947. Observa^oes sobre a malurafao de Xiphopenaeus kroy- eri no litoral de Sao Paulo. Bol. Mus. Nac. 72:1-22. Wenner, A. M. 1972. Sex ratio as a function of size in marine Crustacea. Am. Nat. 106:321-350. Jdiirmil III Slu'llfish Research. Vol. 17. No. 4. 937-944, 1998. PRODUCTION AND RESOURCE ALLOCATION IN THE PERIWINKLE, LITTORINA LITTOREA (LINNAEUS 1758), ON PENDLETON ISLAND, NEW BRUNSWICK M. E. CHASE AND M. L. H. THOMAS Biology Department and Centre for Coastal Studies and Aquaculture University of New Brunswick Saint John, New Brunswick, E2L 4L5 Canada ABSTRACT An intertidal population of Litlonna linnifa L. from a rocky shore on Pendleton Island. New Brunswick, Canada was analyzed to determine its demographic structure and the energy allocated to gamete production, somatic growth, and the synthesis of organic shell matrix. Total production, calculated as the sum of somatic growth (Pg), shell growth (Ps). and the production of gametes (Pr), was 151 kJ/m'^/y'. The percentage of production allocated to Pg. Ps, and Pr was 60.0, 5.9, and 36.1%. respectively. Younger cohorts (0 and 1) were responsible for the bulk of shell (41.5 and 20.5%, respectively) and somatic production (58.5 and 79.5%, respectively), but for 0% of the reproductive output. As compared to other populations of L lillorea. the proportion of total production allocated to gamete production by the population on Pendleton Island was higher. KEY WORDS: Liuorina linorea. production, resource allocation, reproduction INTRODUCTION Estimates of production are useful in assessing contributions of marine species to energy flow through the ecosystetn (Rodhouse 1979, Griffiths 1981a, Griffiths 1981b, Vahl 1981 ) in addition to determining the suitability of different habitats to producers (Bayne and Worrall 1980). Among mollusks, bivalves have re- ceived the most attention (Cerastodenna edule, Ivell 1981, Pla- copecten magellanicus. MacDonald and Thompson 1986, Mytilits edidis. Bayne and Worrall 1980, Thompson 1979, and Gardner and Thomas 1987a) largely because of their commercial and ecological importance. The common periwinkle, Littorina littorea (L.), is one of the most widely studied marine gastropods. Extensive literature exists regarding its biology (Gegenbaur 1852, Caullery and Pelsener 1910, Linke 19.33, Fretter and Graham 1962), breeding cycle (Tat- ter.sall 1920, Elmhirst 1923, Moore 1937. Williams 1964, Fish 1972. Chase and Thomas 1995a) and growth (Hayes 1927. Moore 1937, Ekaratne and Crisp 1982. Gardner and Thomas 1987b). However, there is a paucity of data on its production and allocation of resources. Grahame (1973) examined the importance of lepro- duction as a pathway of energy flow when he calculated both the production of .somatic tissue and gamete production of L. littorea in North Wales, Hughes and Roberts (1980) compared the repro- ductive effort of L. littorea to that of L. neritoides, L. nigrolineata and L ritdis in North Wales. In Canada, the only study that has calculated secondary production of L littorea was that of Gardner and Thomas (1987b) on a population at Welch's Cove, Bay of Fundy. In their study, however. Gardner and Thomas (1987b) considered only the partitioning of energy between somatic and shell growth; reproductive output was not considered. No study to date has looked at all components of production (somatic growth. shell growth, and the production of gametes) and the allocation of resources to those components. L littorea is the most dominant mollusk within the midlittoral zone of rocky shores on Pendleton Island and throughout most of the Bay of Fundy (Gardner and Thomas 1987b). Knowledge of the production and allocation of resources in populations of L littorea will provide information critical to the understanding its life history, and importance to the rocky shore community. The objective of this paper is twofold. The first is to calculate secondary production as the sum of all components (somatic, shell, and gamete) to detemune the propor- tion of total production allocated to gamete production. The second is to deterinine the age-specific pattern of energy partitioning be- tween growth (somatic and shell) and reproduction. MATERIALS AND METHODS Study Site Pendleton Island is located in New Brunswick, Canada at 45°02'N and 67''56'W within the Deer Island Archipelago (Fig, 1), The climate and oceanographic setting have been summarized by Thomas et al. (1990). The collection sites were on an approxi- mately 100 m long section of shore in Pendleton Passage, a shel- tered channel with high tidal currents ranging up to 0.47 and 0.60 m/s' for the ebb and flood of spring tides, respectively (Thomas et al. 1990). The upper shore is predominately sedimentary, with scattered rocks and outcrops, and L. littorea are abundant through- out this area. Demographics Monthly samples of L littorea were collected between May and September, 1988 and May and October, 1989. Samples com- prised approximately 1,000 individuals, collected randomly from all tidal levels along three transects. The spire height was measured using computer-assisted vernier calipers. Gardner (1986) has shown that age of Z.. littorea cannot be ascertained accurately from growth ring analysis. Therefore, length frequency histograms were plotted, and the cohort placement of Cassie (1954) was used to analyze the demographic structure of the population. Although it is accepted that cohort identification by such a means may be sub- jective, we feel that regular sampling of the population and the incorporation of a settlement study (Chase and Thomas 1995a) more than compensated for the subjectivity of this method. Production Secondary production was calculated using the following rela- tionship, P = Pg 4- Pr -I- Ps ( I ) where; P = total production; Pg = energy incorporated into so- 937 938 Chase and Thomas Figure 1. Location of Pendleton Island in tlie Deer Island Archipelago, New Brunswick, Canada. matic growth; Ps = energy incorporated into the shell matrix; and Pr = energy expended on gamete production. Secondary production for the somatic and shell parts for the 1988 and 1989 study periods was estimated using the increment summation method outlined by Rigler and Downing (1984). This method calculates production of cohorts as the sum of the change in biomass over specified time intervals. The density of the population was estimated from a population census along each of three transects running perpendicular to the shore line. The number of L. liaorea in every 5th m" were counted, and a random sample of 25 were measured for spire height. Bio- mass was determined monthly for each size class or cohort (from demographic analysis) m grams of ash-free dry weight (AFDW). This was done by interpolating the mean spire height of each cohort onto the graph of the log regression of the AFDW (somatic tissue and shell) on the log spire height of 50 individuals in the population at each sampling time. All regression equations were highly significant (R" > 0.8). This estimate of weight was then multiplied by the cohort density to give an estimate of biomass in gAFDW/m~. Biomass determination formed the basis of produc- tion calculations. Biomass estimates were converted to energy units using the following conversion of Grahame (1973) of 24.7 kJ/mg' AFDW. The reproductive output was calculated directly for females using the direct method of Crisp ( 1984). L. liuorea females, of a range of spire heights, were housed individually in plastic jars, and the spawn from each was collected (Chase and Thomas 1995b). Four samples of approximately 2,000 eggs were incinerated in a muftle furnace at 435°C for 3 h for ash-free determinations. Be- cause there was little variation in the calculated weights, the mean of the samples were used in all calculations. The reproductive output of males could not be calculated di- rectly as in the females. Indirect analysis of fecundity was based on the mean gonad weight before and after spawning was known to have occurred (weight loss following spawning). For comparison purposes, the female reproductive output was also calculated using the indirect method (Crisp 1984). From May to November, 1989 samples of approximately 230 L. linorea were collected from Pendleton Island and taken to the laboratory, where they were maintained without food in running aerated seawater for 48 hours to allow the gut to clear (Grahame 1973). Histological analysis of the gonads of L. littorea in Pendleton Island revealed that the breeding season was between May and October, and spawning generally occurred between June and September (Chase and Thomas 1995a). Samples of approximately 100 each of males and females were killed by boiling for 1 minute. The gonadal region (including the penis and prostrate in the males and the shell gland of the females) was dis.sected out and weighed before and after drying for 48 h at 60"C. In L. littorea. the gonadal tissue is found closely associated with the digestive system and could not be separated for individual measurement. It was assumed that by emptying the gut and consistently taking the same tissue for each sample, any difference in weight would be attributed to changes in the gonad. However, estimates of reproductive output may be more variable as a result of this approach. Samples were inciner- ated in a muffle furnace at 435°C for 3 h for ash-free determina- tions. Covariance analysis of gonad weight versus spire height was used to test for significant changes in gonad weight between sam- pling dates. This may indicate spawning. Further evidence of spawning was obtained through histological examination of the gonadal tissue (Chase and Thomas 1995a). The differences in gonad weight between these data was used in the calculation of Production of Littorina 939 gamete production. Bioniass estimates for gonadal tissue were converted to energy units using the conversion of Grahame { 1973) of 26 kJ/mg' AFDW. RESULTS Demography 1988 Figure 2(a-c) are histograms of the population of L liltorea on Pendleton Island at peak reproductive times (i.e.. settlement); May (Fig. 2a), July (Fig. 2b). and August (Fig. 2c) 1988. Summary of mean spire height (mm ± SD) and density (number of individuals/ m") of each cohort are shown in Table 1. Cassie (1954) analysis revealed that the population was composed of four cohorts from 3 years in 1988. In May, the population was dominated by mature individuals (>13 mm), which represented l\9'c of the population. The percentage of mature individuals in the population declined throughout the season to 46 and i29c of the population in July and August, respectively, as a result of settlement of new cohorts and subsequent reduction of older individuals from the population. Settlement on Pendleton Island seems to in\ olve the recruitment of two distinct cohorts, cohorts 0+ in July (Fig. 2b) and 0++ in August (Fig. 2c). Two cohorts per year remained discemable until the end of the first year, after which only one cohort per year could be identified. This is because older individuals have a very slow growth rate as opposed to the young, faster growing cohorts (Wil- liams 1964), resulting in a merging of the younger cohorts with the older. In 1988, the newly settled cohorts (0+ and 0++) comprised 22.87f of the population at the end of the breeding season. Demography 1989 Figure 2(d-f) are histograms of the population at peak repro- ductive times in May (Fig. 2d), Augu.st (Fig. 2e), and October (Fig. 2f) 1989. Summary of mean spire height (mm ± SD) and density (number of individuals/m") of each cohort are in Table 1. Cassie ( 1954) analysis revealed that the population was composed of four cohorts from 3 years [ 1 a and 1 b, ( 0++ and 0-H of the previous year), 2, and 3]. A cohort four was detected in May and August, but comprised only 5 and 0.3'7r of the population in each month. In 1988 1989 (A) .MAY 4 14 21 28 LENGTH (mm) 14 21 28 35 (D) MAY 23 lb 1 1 200 - 150 ■ la 2 3 Q 100 ' ^^L u 50 ' lj 1 Mr A 7 14 21 LENGTH (mm) (E) ALGUSl 22 2 1> 1 7 14 21 2« 35 LENGTH(mm) (C) AUGUST 25 14 !1 U 35 LENGTH (mml (F) OCTOBER 30 14 21 2J* 35 LENGTH(mm) Figure 2. Length-frequency histograms of the Littorina littorea population at Pendleton Island at sampling intervals in 1988 (2 a-c) and 1989 (2 d-f ). .\rrows indicate the mean spire height (mm) of the component cohorts in the population. 1988: (A) May 4, (B) July 6, and 1989: (C) August 25, (D) May 23, (E) August 22, (F) October 30. 940 Chase and Thomas TABLE 1. Summary of the mean spire height (mm ± SDl and density (number of individuals/m") of the component cohorts in the Littoriiia littorea population at Pendleton Island at sampMng intervals in 1988 and 1989. 1988 May July August Mean Spire Mean Spire Mean Spire Height (mm) Density Height (mml Density Height (mm) Density Cohort (±SD) (#/m-| Cohort (±SD) (#/m-) Cohort (±SD) (#/m') 0++ 2.5 ±0.9 20 0+ 3.7 ±0.6 3 0+ 8.5 ±1.6 4 IB 44 ± 1.0 2 IB 10.0 + 2.4 1 IB 12.6± 1.7 2 lA 10.6 ±2.3 I lA 14.6 ±1.8 4 lA 17.2 ± 1.3 4 1 17.2 ±2.6 7 2 19.4 ± 0.9 6 2 22.5 ± 2.2 18 3 24.2 ± 2.2 18 3 26.4 ± 2.0 1989 19 3 26.0 ±0.8 4 May .August October Mean Spire Mean Spire Mean Spire Height (niml Density Height (mm) Density Height (mm) Density Cohort (±SI)) (#/ni-) Cohort (±SD) (#/m-) Cohort (±SD) (#/m-) 0++ 2.5 ± 0.9 73 0+ 4.9 ± 1.9 12 0+ 7.3 ± 1.4 1 IB 4,1 ± 1.2 2 IB 10.7 ±1.5 2 IB 12.8 ±2.6 2 lA 10.8 ±2.2 1 lA 17.1 ± 1.5 8 lA 20.0 ±2.0 3 2 14.3 ±2.1 2 2 22.7 ± 1.6 12 2 23.8 ±1.0 3 3 24.1 ±2.5 27 3 26.4 ± 1.2 5 3 26.5 ± 0.5 4 4 27,8 ± 1.5 5 May. the population was dominated by mature individuals (>13 mm), which represented 60% of the population. The percentage of mature individuals in the population declined throughout the sea- .son to 54 and 459c of the population in August and October, respectively, as a result of the settlement of new cohorts and the subsequent reduction in the older individuals (cohort 3: 34% in May to 2% in October) and the loss of cohort 4. Settlement of cohorts occurred in August (Fig. 2e) and October (Fig. 2f), 1989. Experimental data have shown that the differential timing may be the result of local temperature variations (Chase and Thomas 1995b). In 1989. the newly settled cohorts (0+ and 0++) comprised 31% of the population at the end of the breeding season. Somatic and Shell Production Table 2 contains the values calculated for production of so- matic tissue (Pg) and shell (Ps). in kJ/m", for the 1988 and 1989 population of L littorea on Pendleton Island. Somatic production (Pg) for the 1988 season was 70.265 kj/nr. Shell production (Ps) for the same time period was 16,859 kJ/m". An analysis of the production per cohort revealed the largest contribution was from the larger, hence older, cohorts 2 and 3. Production of these two cohoits lepresented 91.7% of Pg and 88.6% of Ps. Somatic production (Pg) for the 1989 season was 87.764 kJ/nr. Shell production (Ps) for the same time period was 8.974 kJ/m". Analysis revealed the majority of Pg was from cohorts 2 and 3 (74.5%); whereas, the majority of Ps was from cohorts I and 2 (88.8%). The decrease in Pg of cohort 4 and the negative Ps for cohorts 3 and 4 were probably a reflection of the decrease in density of each cohort (cohort 3: 1988. 18 individuals/m" to 4 individuals/m- in 1989: cohort 4; 5 individuals/m' in 1988 to individuals/m" in 1989). A two-way analysis of variance (ANOVA) on the mean so- matic production (Pg) of each cohort, for the population in 1988 and 1989 revealed no significant year effect (df = 1,3; F = 0.966; p > .5); however, there were significant differences among cohorts (df = 1.3; F = 25.95; p > .02). Production of cohorts and I was higher in 1989. A two-way analysis of variance on the mean shell production (Ps) of each cohort, for the population in 1988 and 1989 revealed no significant effect of year (df = 1.3; F = 0.546; p > .5) or cohort (df = 1.3; F = 1.18; p > .5). Production of Gametes (Pr) Female Reproductive Output The mean (±SE) AFDW of four batches of approximately 2.000 eggs was 1.15 x 10"* ± 0.2 x 10^'' g AFDW per egg. Table 2 shows the production of gametes for females (Prf) using both direct and indirect methods of determination. Estimated Pr in 1989 using the direct method was 46,160 kJ/m". Pr was high at the onset of maturity (>I3 mm, cohort 1) at 13.661 kJ/m". and in cohort 3 (22.038 kJ/m") but decreased in the oldest cohort. This reduction in Pr is probably a reflection of the decrease in density of this cohort in the population. The estimated Prf in 1989 using the indirect calculation was 33.020 kj/m". Production estimates using this method increased with the age of each cohort 3 (20.865 kJ/m") and then decreased in cohort 4 (11.496 kJ/nr). A two- way ANOVA revealed that there was no significant difference effect of method (df = 1.4; F = 0.72; p > .5) or cohort (df = 1,4; Production of Littorina 941 TABLE 2. Efjtimated production (kj/m') of shell (Ps. 1988 and 1989), somatic tissue (Pg, 1988 and 1989), and the production of gametes (Pr, 1989) of Littorina liltorea on Pendleton Island, Production (KJ/m'/v') 1988 1989 Prf (Indirect) Prf (Direct) Prm (Indirect) Pr(m + f) (Total) Cohort Pg Ps Pg Ps Pt 0.565 0.304 1.741 1.235 2.976 1 5.268 1.621 15.637 4.039 0.002 13.661 0.003 0.005 19.681 2 31.469 7.870 26.981 5.716 0.657 4.084 0.438 1.095 33.792 3 3:.%3 7.064 38.290 -0.962 20.865 22.038 15.740 36.606 73.933 4 — — 5.115 -1.054 1 1 .496 6.378 5.410 16.906 20.967 Total 70.265 16.859 87.764 8.974 33.020 46.160 21.591 54.114 151.349 (80.6%) (19.4%) (60.0%) (5.9%) (36.1%) Percentage total production allocated to each component in parentheses (/). Production gametes (Pr) presented for feinales (Prf I and males (Prm) separately and together [Pr(m + t") Prf calculated using both indirect and direct methods (Crisp 1984). F = 5.83; p > .1 ). However, the indirect metho(j seems to tinder- estimate the Pr of the younger cohorts. Indirect estimates seem to be inappiopriate for young cohorts, and fecundity estimates ba.sed on such will underestimate Pr. especially for the tnales. Male Reproductive Output The total annual production of male gametes (Prm) was calcu- lated to be 21.591 kJ/nr (Table 2). Production increased to a maximum at cohort 3 ( 15.740 kJ/m-. 72.9% of the total Pr). A two-way ANOVA on Pr of male and females using the indirect calculated values revealed no significant effect of se.x (df = 1.4: F = 2.81; p > .2); however, cohort was significant (df = 1.4: F = 27.68; p > .01). Pr for the older females (cohorts 2 and 3) was larger than that of the males. The total estimate of Pr for the L litlorea population (males and females) on Pendleton Island based on the indirect calculations of male and female reproductive output was 54.114 kJ/nr. Total Production The total production for the breeding season in 1989. calculated as the sutn of Ps, Pg. and Pr, was 151,349 kJ/nr (36.1% Pr. 60.0% Pg, and 5.9% Ps). Total production could not be calculated for the 1988 season, because no data were available on production allo- cated to reproduction. Figure 3 shows the percentages of produc- tion allocated to each of Ps. Pg. and Pr in each cohort. In the newly settled cohort (0) all of the production was allocated to either Ps (41.5%) or Pg (58.5%). The percentage of production allocated to Ps and Pg in each cohort decreased once maturity was reached (cohort 1 ). The percentage of the production allocated to Pr in- creased with age to age 4. with Pr comprising 76.8% of the pro- duction of that cohort. DISCUSSION Demography The structure of the intertidal population of L. litlorea on Pendleton Island changed seasonally as a result of recruitment and growth (Fig. 2). At the beginning of each of the 1988 and 1989 seasons, the population of L. Uttorea was dominated by mature individuals (>13 mm). However, at the end of the season, the Z o u so a. AGE (years) Figure 3. Percentage of total production in each age class that is al- located to shell growth (Ps). somatic tissue growth (Pg), and produc- tion of gametes (Pr) in 1989. percentages of the population composed of immature and mature individuals were almost equal (% immature:% mature; 48:52 in 1988 and 45:55 in 1989). Recruitment of L. Uttorea was tempo- rally variable but consisted of two pulses per year, in July and August 1988 and in August and October 1989. The newly settled cohorts comprised 23 and 3 1 % of the population at the end of each season in 1988 and 1989. Recruitment was much higher on Pendle- ton Island than recorded in other studies of L. Uttorea (8.4% Gard- ner and Thomas 1987b at Welch's Cove. Bay of Fundy; 1.6% Lambert and Farley 1968 at Ketch Harbour, Nova Scotia; and 0.4-4.4% Smith and Newell 1955 at Whitsable. England). Larger recRiitment densities may be the result of the existence of two pulses of recruitment per year. In addition. L. Uttorea on Pendleton Island were observed to recruit at high shore levels only (80% tidal level) (Chase and Thomas 1995a). Low recruitment densities in the studies of Gardner and Thomas {1987b) and Lambert and Farley (1968) were attributed to recruitment from subtidal regions that was not accounted for in fall estimates of recruitment density. 942 Chase and Thomas Production Secondary production estimated for L litiorea on Pendleton Island in 1988 was measured as the production of somatic tissue and shell. Reproductive output was not included. In 1989. gamete production comprised 36.1% of the total production; thus, any calculations fori., littorea in 1988 were underestimated. Compari- son of the somatic and shell production estimates in 1988 and 1989 revealed that there was no difference between years, despite a difference in the duration of the sampling season. In 1988, pro- duction was measured from early May to late August as compared to early May to late October in 1989. a difference of approximately 8 weeks. A breakdown of production in 1989 into two time peri- ods: May to late August and late August to late October; however, revealed that the majority of shell (64.7%) and all of the somatic production ( 100%) occurred in the May to late August time period. Comparisons of total secondary production values of L littorea are difficult, because no studies have examined all components of total production (somatic tissue, shell growth, and reproductive output). There are. however, .studies where somatic tissue and/or gamete production have been measured, so some comparisons are possible. Gardner and Thomas (1987b) measured somatic tissue production for a population of L. littorea at Welch's Cove in the Bay of Fundy. Their calculated value was approximately 3X larger than our figures; although the populations at each location were very different, which may prevent comparisons between the stud- ies. The study of Gardner and Thomas ( 1987b) was restricted to a small portion of the intertidal zone and had much higher densities than our study (657 to 934 individuals/ni" versus 28 to 86 indi- viduals/m"). In contrast, our study examined the entire intertidal zone. Grahame (1973) calculated somatic production and gamete production of a population of L. littorea at Anglesey, Wales. Total production of these two measures was 852.51 kJ/ni" [somatic pro- duction = 629.510 kJ/m" ( 138 kcal/nr) and gamete production = 223 kJ/m" (46.7 kcal/m~)]. The estimate of gamete production in this study was only 54 kJ/m". only 24% of that reported by Gra- hame (1973). The population in that study, however, contained many larger individuals (spire height > 25 mm). On Pendleton Island in 1989, the percentage of the population comprising indi- viduals with a spire height > 25 mm was less than 3%. Because fecundity increases with spire height in L. littorea (Grahame 1973, Hughes and Roberts 1980, Chase and Thomas 1995b), it would be expected that we would observe a larger Pr for a population com- posed of larger, more fecund individuals. No estimate of the production of the shell matrix for L. littorea was found in the literature. Values from other studies indicate that shell production in bivalves is usually <5'7r of total production. For L. littorea, the organic component of the shell represented 6% of the total production in 1989. As such, it seems that, for L. littorea. the majority of the secondary production, was attributed to the production of somatic tissue and gametes. Considering such a large amount of energy is required for the production of gametes, it would be expected that there would be a close coupling between the reproductive cycle and energy avail- able for growth. Our data are in keeping with the general obser- vation that animals devote a greater share of the production to reproduction as they age (Fig. 3). However, the proportion of total production allocated to reproduction in the L. littorea population on Pendleton Island was higher as compared to other studies. Hughes and Roberts (1980) examined the proportion of total production allocated to reproduction (Pr/Pr -i- Pg) for different age classes of four Littorinid species, including L. littorea. They ob- served, for all species, that the proportion of total production al- located to reproduction increased linearly during a phase of rapid growth, but began to level off abruptly toward its asymptotic value of 80 to 100% (Hughes and Roberts 1980). For L. littorea. spe- cifically, the proportion allocated to reproduction increased from to 100% in 9 years, with the asymptote occurring at age 5 (Hughes and Roberts 1980). In this study, only four cohorts were discem- able. either because older individuals were not present in the popu- lation, or they were so few in numbers so as not detectable using this method of cohort determination. Over the 4 years, however, the proportion of total production allocated to reproduction in- creased linearly once maturity was reached (cohort 1 ). For an age 4 individual, the proportion of total production allocated to repro- duction (Pr) in L. littorea was approximately 60%' in the study of Hughes and Roberts ( 1980). In this study. Pr in an age 4 individual was 76.87r. Alternative estimates of the proportional allocation to repro- duction include the ratio of gamete to somatic production (Pr/Pg). In this study. Pr/Pg was 61.7%. This estimate is high as compared to studies on other gastropods, including L. littorea. where Pr/Pg was calculated; 34% (L. littorea. Grahame 1973). 30 to 40% (La- cuna vincta. Grahame 1982), 25% (Lacuna pallidula. Grahame 1982). and 11.7% (Fissurella barbadensis. Hughes 1971). Developing hypotheses to explain the higher allocation to re- production (Pr) observed in this study is partly hampered by only 1 year of measurement. In many studies, the larger Pr may reflect older populations, because Pr generally increases with the age of an indi\idual (Hughes and Roberts 1980. this study). The popula- tion monitored in this study was made up of young individuals (only four cohorts detected in each year). A high allocation to reproduction in such young animals might be expected if this population is subject to high mortality and lowered life expectan- cy. This reduced life expectancy may be the result of environmen- tal conditions. The population of L. littorea on Pendleton Island is situated in a passageway where there is a very strong current. Average surface currents 3 hours before and after low tide gener- ally exceed 1.3 m/s'. bottom currents have been measured at 0.47 to 0.60 m/s' (Thomas et al. 1990). Many studies examining the effect of wave exposure on the population energetics of gastropods have found greater mortality and a larger amount of energy de- voted to reproduction (e.g.. Hughes and Roberts 1980; Hart and Begon 1982; Etter 1989). In addition to mortality caused by en- vironmental conditions, it was found that at least 25% of the larger animals examined had parasitic infestations of the digenetic treraa- tode larvae. Cryptocotyle lingua (Crepling) (Chase and Thomas 1995b). Similar infestations have resulted in the destruction of the digestive gland, castration, and change in the migration pattern (Fretter and Graham 1962; Lambert and Farley 1968). Such infes- tations may explain the absence of larger/older individuals on this shore. SUMMARY The results of this study revealed that estimates of population production of L. littorea were much lower than values reported in the literature. Rates of secondary production are known to vary widely in nature and are affected by a variety of biotic and abiotic characteristics of the environment (Plante and Downing 1989). Such variation in production may be the result of independent or Production of Littorina 943 combined effects of annual differences in such environmental fac- tors as water temperature and/or food supply, variation in density, age structure, and allocation to growth and reproduction. Density and age structure have been proposed as factors causmg the lower somatic and gamete production values observed in this study. However, despite lower estimates of production, the proportion of total production allocated to reproduction (Pr) measured as both the proportion of total production (Pr/Pg + Pr + Ps) and somatic production (Pr/Pg) was higher in this study by approximately 20% as compared to estimates for L Httorea in the literature. The Bay of Fundy exhibits a great diversity of marine .species, a combination of the influx of larvae, propagules. and adults from the Labrador Current and the Gulf of Marine waters, and the high productivity as a result of the vigorous tidal mixing (Thomas et al. 1990). Prior research on the population of L. liitorea in Pendleton Passage revealed high growth rates, large recruitment densities, and high reproductive outputs (Chase and Thomas 1995a. b). How- ever, analysis of the demography, production, and resource allo- cation of the population of L. Httorea in Pendleton Passage seems to be more characteristic of an environment with very harsh con- ditions; that is, fewer older individuals, faster turnover rates, and lower total production, with a greater proportion of total produc- tion allocated to reproduction. It is likely that the high tidal cur- rents are a major influence on the population of L. Httorea in Pendleton Passage. However, additional information will have to be gathered before we can speculate on the effects of high tidal energy on this population of L. Httorea. ACKNOWLEDGMENTS We are grateful to G. Bacon, C. Hatfield, and R. Bosien for assistance in the field, to W. Morris for computer and drafting assistance. Special thanks to M. G. Topping and B. A. MacDonald for reading drafts and offering constructive criticism. This research was funded through a Natural Science and Engineering Research Council of Canada (NSERC) grant to M. L. H. Thomas and an NSERC postgraduate scholarship to M. E. Chase. Current address for MEC: Department of Zoology. Miami University. Oxford. Ohio 45056, USA. LITERATURE CITED Bayne, B. L. & C. M. Worrall. 1980. Growth and production of mussels. Mytilus edidis. from two populations. Mar. Ecol. Prog. Sen 3:219-239. Cassie, R. M. 1954. Some uses of probability paper in the analysis of size frequency distributions. Aust. J. Mar. Freshwater. Res. 5:513-522. Caullery. M. & P. Pelsener. 1910. Ponte et developpement du Littorina liitorea. Bull. Scienl. France Belgique 44. Chase. M. E. & M. L. H. Thomas. 1995a. Evidence for double recruilmenl in the periwinkle, Littorina Httorea, on Pendleton Island. News Bruns- wick, Canada. J. Shellfish Res. 14:153-158. Chase. M. E. & M. L. H. Thomas. 1995b. The effect of the rate and onset of temperature increase on the spawning of the periwinkle. Littorina Httorea L. J. E.xp. Mar Biol. Ecol. 186:277-287. Crisp. D. J. 1984. Energy flow measurements, pp. 197-279. In: N. A. Holme and A. D. Maclntyre AD (eds.). Methods for the Study of Marine Benthos. IBP Handbook No. 16. Blackwell Scientific. Oxford. UK. Ekaratne. S. U. K. & D.J. Crisp. 1982. Tidal microgrowth bands in an intertidal gastropod from shell microgrowth measurements, including a comparison of alternate methods. /. Exp. Mar. Biol. Ass. U.K. 64:183- 210. Elmhirst. R. 1923. Notes on the breeding and growth of marine animals in the Clyde Sea area. Rept. Scot. Mar Biol. As.s. 0:19-23. Elter, R. J. 1989. Life history variation in the intertidal snail Nucella lapil- lus across a wave-exposed gradient. Ecology 70:1857-1876. Fish. F. D. 1972. The breeding cycle and growth of open coast and estua- rine populations of L. Httorea. J. Mar. Biol. Ass. U.K. 52:1011-1019. Fretter. V. & A. Graham. 1962. British prosobranch molluscs: their func- tional anatomy and ecology. Ray Society. London. Gardner. J. P. A. 1986. Demography, Growth, and Production of Mytilus eduHs. Modiolus modiolus and Littorina Httorea at Welch's Cove in the Bay of Fundy. M.Sc. thesis. University of New Brunswick, Saint John. New Brunswick. Gardner. J. P. A. & M. L. H. Thomas. 1987a. Growth, mortality, and pro- duction of organic matter by a rocky intertidal population of Mytilus edulis in the Quoddy Region of the Bay of Fundy. Mar. Ecol. Prog. Ser -39:31-36. Gardner. J. P. A. & M. L. H. Thomas. 1987b. Growth, mortality, and pro- duction of a L. Httorea population in the Bay of Fundy. Ophelia 27: 181-195. Gegenbauer. C. 1852. Uber penisdrusen von Littorina. Z. Wiss. Zoo/. 4: 233-235. Grahame. J. 1973. Breeding energetics of Littorina Httorea (L.) (Gastropod prosobranchia). J. Anim. Ecol. 42:391— 103. Grahame. J. 1982. Energy flow and breeding in two species of Lacuna: comparative cost of egg production and maintenance, //if. / Invert. Reprod. 5:91-99. Griffiths. R.J. 1981a. Population dynamics and growth of the bivalve Chromxtilus meridionalis (Kr. ) at different tidal levels. Estuar. Coast. Shelf Sci. 12:101-118. Griffiths. R. J. 1981b. Production and energy flow in relation to age and shore level in the bivalve ChroniMilus meridionalis (Kr. I. Estuar. Coast. Shelf. Sci. 13:477-f93. Hart. M. & M. Begon. 1982. The status of general reproductive strategy theories, illustrated in winkles. Oecologia 53:37—4-2. Hayes. F. R. 1927. The effect of environmental factors on the development and growth of Littorina Httorea. Trans. Nova Scotia Inst. Sci. 17:6-16. Hughes, R.N. 1971. Ecological energetics of Nertia (Archaegastropoda, Neritacea) populations on Barbados. West Indies. Mar. Biol. 1 1:12-22. Hughes, R. & D.J. Roberts. 1980. Reproductive effort of winkles {Lit- torina spp.) with contrasted methods of reproduction. Oecologia 47: 1.30-136. Ivell. R. 1981. A quantitative study of Cerastoderma — Nephthys commu- nity in the Limtjord. Denmark, with special reference to production of Cerastoderma edule. J. MoH. Stud. 47:140-170. Lambert. T. C. & J. Farley. 1968. The effect of parasitism by the trematode Cryptocoryle lingua (Creplin) on zonation and winter migration of the common periwinkle Littorina Httorea. Can. J. Tool. 46:1139-1147. Linke, O. 1933. Morphologic et physiologie des genitalapparates der nor- dsee Littorina. Wiss Meeresuntersuch. Ahl. Helgoland. 19:3-52. MacDonald. B. A. & R. J. Thompson. 1986. Production dynamics and energy partitioning in two populations of the giant scallop. Placopecten magellanicus. J. Exp. Mar Biol. Ecoi 101:285-299. Moore. H. B. 1937. The biology of Littorina Httorea. Part I. Growth of the shell and tissues, spawning, length of life, and mortality. / E.xp. Mar. Biol. Ass. U.K. 21:721-742. Plante. C. & J. A. Downing. 1989. Production of freshwater invertebrate populations in lakes. Can. J. Fish. Acjual. Sci. 46:1489-1498. Rigler, F. H. & J. A. Downing. 1984. The calculation of secondary pro- ductivity, pp. 19-58. In: J. A. Downing and F. H. Rigler (eds.). A Manual for the Assessment of Secondary Productivity in Fresh Waters, IBP Handbook 17. 944 Chase and Thomas Rodhouse. P. G. \9Ti. A note on the energy budget for an oyster popula- tion in a temporary estuary. / Exp. Mar. Biol. Ecol. 37:205-212. Smith, J. E. & G. E. Newell. 1955. The dynamics of the zonation of the common periwinkle. Liuorina liuarea (L.). on a stoney beach. / Anhn. Ecol. 24:35-56. Tattersall. W. M. 1920. Notes on the breeding habits and life history of the periwinkle. Sci. Invest. Fish. Bich. Ire. 37:1-11. Thomas. M. L. H.. J. A. Stevens, A. McAslan & S. R. Clayden. 1990. Shallow marine, littoral, and terrestrial associations of Pendleton Is- land. New Brunswick. Nal. Sci. 9:1-35. Thompson, R. J. 1979. Fecundity and reproductive effort in the blue mussel (Mytihis edulis), the sea urchin (Strongylocentrotus droebachiensis), and the snow crab (Chionoeceles opilio) from populations in Nova Scotia and Newfoundland. J. Fish. Res. Board Can. 36:95-964. Vahl, O. 1981. Age-specific residual reproductive value and reproductive effort in the Iceland scallop, Chlamys islandica (O. P. Muller). Oeco- logia 51:53-56. Williams, E. E. 1964. The growth and distribution oi Liuorina litlorea (L.) on a rocky shore in Wales. J. Anim. Ecol. 337:413—132. Jaiinuil ofShetlthh Research. Vol. 17, No. 4. 945-953, 1998. IDENTIFICATION OF STOCKS OF THE EXPLOITED LIMPETS PATELLA ASPERA AND P. CANDEI AT MADEIRA ARCHIPELAGO BY ALLOZYME ELECTROPHORESIS LAURA I. WEBER,' JOHN P. THORPE,' RICARDO S. SANTOS,^ AND STEPHEN J. HAWKINS' Port Erin Marine Lxiboratory University of Liverpool Port Erin, Isle of Man 'Departamento Oceanografm e Pescas Universidade dos A(^ores Horta. Fcdal. A(;ores ABSTRACT Allozyme electrophoresis was used to investigate stock integrity of Patella aspeni ( = P. iily.ssiponeiisis) and Patella camlet {Molliisca. Patellogastropoda) from the Madeira Archipelago. Samples from the north of the island of Madeira (Pono Moniz) and the north of Deserla Grande were taken for both species, and a sample of P. aspera was also taken from the south of the island of Madeira (Canine). Twenty-one putative loci were resolved and analyzed for both species. Significant differences in allele frequencies were found between locations in each species, suggesting the presence of different genetic stocks. P. candei was found to be slightly structured with a small standardized variance in allele frequencies between samples from Porto Moniz and Deserta Grande (Fsy = 0.036); whereas, P. aspera was highly structured, with an Fsj value over all subpopulations of 0.157. Our results indicate more genetic interchange between the populations from the north of Madeira and Desena Grande than between north and south Madeira. Our findings are consistent with surface water-mass circulation patterns from northwest to southeast along the Madeira Archipelago, determined mainly by the Canary Current. Species with low dispersal and a high degree of spatial genetic structuring among subpopulations are more susceptible to collap.se caused by overfishing than those wnh high dispersal capabilities; therefore, the individual management and control of these biological units is recommended. KEY WORDS: Mollusca, patellogastropoda. Patella, population genetics, stocks, allozymes INTRODUCTION The stock concept was first desciibed by Larkiti ( 1972) as: "a population of organism.s shainng a common gene pool, that is di.screte to warrant consideration as a self-perpetuating sy.stem that can be managed." This concept, primarily applied to fish species, has been widely used with different connotations by the fisheries community, politicians, and biologists. Stock, as a tnanagement unit, may also be defined as the extent of a population or mi.xed populations over which a fishing activity occurs. Therefore, it may comprise more than one biological stock or subpopulation (Hedge- cock 1986). In the 1970s, allozyme electrophoresis emerged as a powerful tool to recognize biological stocks by means of genetic informa- tion (Gall 1986, Hedgecock 1986, Utter 1986, Utter 1991, Shep- herd and Brown 1992. Avi.se 1994). Once the genetic structure of an exploited species is understood, the fishery can be regulated so that harvesting of each subpopulation can be managed and con- trolled individually (Allendorf et al. 1988). However, despite its importance, the "stock-composition" strategy has been practiced rarely in fisheries management (Pella and Milner 1988). P. aspera Roding, 1798 (= P. tilyssipimensis Gmelin, 1791) and P. candei d'Orbigny, 1840 are two palellogastropod limpets distributed throughout the Macaronesian Archipelagos, where they represent an important food resource. In recent years, they have been intensively exploited for subsistence and for commercial pur- poses. In the Azores, both species have been declining for many years (Martins et al. 1987. Menezes 1991. Corte-Real et al. 1996). Recent works on the population genetics of these limpets, using allozyme electrophoresis (Corte-Real 1992, Corte-Real et al. 1996. Weber et al. in preparation), have shown high macroscale struc- turing of their populations. The aim of the present work is to study the structuring of these species within the Madeira Archipelago, to find out if there are different biological stocks. This work forms part of a fisheries management strategy by the regional govern- ment. MATERIALS AND METHODS Sampling Sites Samples of Patella aspera and Patella candei were taken from the Madeira Archipelago (Fig. I ). Samples of both species were taken in July 1994 from Porto Moniz (north Madeira) and from the north of Deserta Grande in March 1995. An additional sample of Patella aspera was taken from Canifo (south Madeira) in July 1994. Locations in Madeira were chcsen from those points where human activity is concentrated. Sample sizes (n) are specified in Tables 2 and 3. All samples were transported live to the Port Erin Marine Laboratory, where they were frozen at -78°C until re- quired for electrophoresis. Electrophoresis Homogenates were prepared by macerating foot muscle in buffer Tris-HCl. pH 8.0, and then centrifuged at 10,000 rpm for 5 min. Supematants were used afterward for electrophoresis. Stan- dard horizontal 12.5% starch gel (Sigma-Aldrich Co, Ltd,) elec- trophoresis was carried out using the following buffer systems: (I) Tris-citrate, pH 8.0 (Siciliano and Shaw, 1976) and fll) Tris- citrate-EDTA. pH 7.0 (Ayala et al. 1972); (III) Discontinuous Tris-citrate-borate, pH 8.2-8.7 (Poulik 1957); (IV) Tri.s-citrate- borate-LiOH, pH 8.26-8.31 (Redfield and Salini 1980); (V) Tris- borate EDTA, pH 9.0 (Ayala et al. 1974). Specific staining pro- cedures for the enzyines analyzed (Table I) followed the tech- niques of Brewer (1970), Shaw and Prasad (1970), Harris and Hopkinson (1976) and Murphy et al. (1990). 945 946 Weber et al. MADEIRA ISLANDS ^ PORTO SANTO 33"N Porto Moniz MADEIRA Figure 1. (*) Sampling sites for P. caitdei and P. aspera (see text for details). Allele frequencies, mean expected heterozygosity, Nei's ( 1978) genetic identities, and x' tests for homogeneity in allele frequen- cies were calculated using the statistical package BIOSYS-1 (Swofford and Selander 1981). For those individual tests of het- erogeneity that showed p < .05, contingency tables were checked using G (log-likelihood ratio)-test with pooling when more than two alleles, and expected frequencies less than five were observed. Yate's correction was also applied for those 2x2 tables, p-values after G-test were reported in Tables 2 and 3 only when they were notably different from those obtained by the common x"- Unbiased inbreeding coefficients, F,s and F,y (Weir and Cockerham 1984) and the genetic variance between populations (F^-,). with their respective 95% confidence intervals, estimated from a bootstrap procedure of 15,000 resamplings, were obtained by the FSTAT Program, version 1.2 (Goudet 1995). Numbers of migrants be- tween populations per generation (n^^) were calculated by using Slatkin"s ( 1993) formula: n^.,„ = (I/F^t-I )/4. Loci were numbered according to increasing anodal mobility. RESULTS Twenty one loci were resolved for both species. Est-D was only resolved for P. candei and LAh-l only for P. aspera. Allele fre- quencies, variability measures, F-statistics. and results of the x" test for the homogeneity in allele frequencies are summarized in Tables 2 and 3 for P. candei and P. aspera, respectively. Variability The locus Gludh was fixed for the same allele in all populations of both species, and Mdh-l was fixed in all the populations of P. aspera. High levels of polymorphism (76-85.7%) and heterozy- gosity were registered for both species. P. aspera populations showed higher values of heterozygosity (all over 21%) than P. candei {around 10 to 12%). Mean number of alleles per locus was also higher in P. aspera (4.4) than P. candei (2.7). Differentiation among Subpopiilations Both species showed significant differentiation in allele fre- quencies between locations (see x~ tests in Tables 2 and 3). P. candei Subpopulations candei samples An FsT value of The genetic identity (Nei. 1978) between P. from Porto Moniz and Deserta Grande was 0.995 0.036 was found for the analyzed P. candei samples, indicating the subdivision of the total population. The bootstrap 95% confidence interval showed that the F^y value was close to but different from TABLE 1. Name, E.C. number, abbreviation, buffer system and number of loci for each enzyme analyzed Enzvme Name E.C. Number Buffer Number* Abbreviation of loci System 2.6.1.1 AAT 2 ni 3.1.1.- EST 2 rv 3.1.- EST-D m 4.1.2.13 FBALD V 1.2.1.12 GAPDH V 1.4.1.- GLUDH V 5.3.1.9 GPI I 1.1.1.42 IDHP 2 I 1.1.1.27 LDH 2 n 1.1.1.37 MDH 2 IV 1.1.1.40 MEP 2 I 5.3.1.8 MPl 1 IV 5.4.2.2 PGM 1 u 1.1.1.44 PGDH 1 I 24.2.1 PNP 1 I 1.15.1.1 SOD 2 m Aspartate aminotransferase Esterase Esterase-D Fructose-biphosphate aldolase Glyceraldehyde-3-phosphate dehydrogenase Glutamate dehydrogenase Glucose-6-phosphate isomerase Isocitrale dehydrogenase (NADP+) L-Lactate dehydrogenase Malate dehydrogenase Malic Enzyme (NADP-h) Mannose-6-phosphate isomerase Phosphoglucomutase Phosphogluconate dehydrogenase Purine-nucleoside phosphorilase Superoxide dismutase ■ 1UBNC(1984); Shaklee et al. (1990). Patella spp., genetic stocks at Madeira 947 TABLE 2. Patella candei: allele frequencies, F-statistics per locus, and over-all loci (with 95% confidence intervals Bootstrap procedure) and contingency x'' test for the homogeneity in the allele frequencies for each in parentheses obtained by locus and over-all loci. Allele frequencies F statistics per locus Homogeneity Test Locus Allele DES (n = 50) MNZ (n = 50) F,,s Aat-1 Aat-2 Est-1 Est-2 Est-D Fbald Gapdh Gpi Idhp-1 hihp-2 Uh-2 Mdh-1 Mdh-2 Mep-I Mep-2 A B All A B All A B C All A B All A B All A B All A B All A B C D All A B C All A B C D All A B C All A B All A B C All A B All A B C D E All 0.1 ?0 0.850 1.000 0.000 0.110 0.770 0.120 1.000 0.000 0.080 0.920 0.960 0.040 0.000 1.000 0.020 0.930 0.040 0.010 0.010 0.990 0.000 0.030 0.670 0.290 0.010 0.060 0.010 0.930 0.000 1.000 0.170 0.820 0.010 0.010 0.990 0.010 0.970 0.010 0.010 0.000 0.020 0.980 0.950 0.050 0.110 0.820 0.070 0.980 0.020 0.060 0.940 1.000 0.000 0.010 0.990 0.010 0.980 0.000 0.010 0.010 0.980 0.010 0.030 0.450 0.500 0.020 0.050 0.020 0.930 0.010 0.990 0.000 1.000 0.000 0.010 0.990 0.010 0.920 0.040 0.000 0.030 -0.010 0.085 -0.043 0.171 -0.010 -0.032 0.167 0.152 0.174 0.000 0.000 0.094 -0.000 0.041 -0.005 0.000 0.010 0.026 -0.000 0.031 0.000 -0.034 -0.019 0.014 -0.003 -0.010 -0.007 -0.081 0.149 0.074 0.241 0.232 -0.013 0.000 0.000 0.000 -0.196 -0.000 0.164 -0.000 -0.010 -0.010 10.865 5.128 1.473 2.020 2.083 4.082 1.005 4.464 1 .005 10.237 0.424 1.005 19.780 0.000 0.00191 1 0.00098* 0.05353t 0.02354 0.47878 0.15522 0.14892 0.11434t 0.04335 0.31610 0.21550 0.60499 0.01755t 0.01666 0.80887 0.31610 0.00005* 0.99835 -0.039 -0.027 0.011 5.932 0.20426 continued on next page 948 Weber et al. TABLE 2. continued Allele frequencies F statistics per locus Locus Allele DES (n = 50) MNZ (n = 50) Mpi A 0.020 0.010 B 0.980 0.970 C 0.000 0.020 All Pgdb A 0.000 0.020 B 0.010 0.030 C 0.990 0.950 All Pf;m A 0.000 0.020 B 0.210 0.120 C 0.790 0.860 All Pnp A 0.230 0.170 B 0.770 0.830 All Sod-2 A 0.010 0.020 B 0.990 0.980 All Overall Ho-' 0.117 0.105 loci He'' 0.126 0.109 ^0 9y 76.2 85.7 F,s F., Homogeneity Test 0.192 -0.024 0.109 0.130 -0.007 0.187 -0.014 0.120 0.130 -0.014 -0.006 0.011 0.012 -0.000 -0.007 0.071 0.105 rO.006-0.113) (0.057-0.132) .338 0.31061 3.082 0.21412 4.752 0.09294 1.125 0.28885 0.338 0.56075 0.036 81.240 0.00001* (0.006-0.065) ■* Mean observed heterozygosity (direct-count). ''Mean unbiased heterozygosity based on Hardy-Weinberg expectation (Nei 1978). ■■ Percentage of polymorphic loci: a locus is considered polymorphic if the frequency of the most common allele does not exceed 0.99. * Significant at a = 0.05 after a Bonferroni procedure for multiple tests was applied (a' = a/1 + k-i: for k = number of tests, and i = rank of the values of p when ordered from the smallest to the largest; when pi s a', then the corresponding test indicates significance at the "table-wide" a level (Rice. 1989). t P after G test and Yates correction. Key: F,s Mean heterozygote defficiency within populations; Fn- Heterozygote deficiency in the total population; Fj^, Degree of differentiation between subpopulations: MNZ: Pto. Moniz; DES: Deserta Grande; (n) Sample size. zero (Table 2 1. The results of the x~ test corroborated the presence of different subpopulations although with only slight differences. The significant differentiation found in P. candei between Porto Moniz and Deserta Grande was largely a result of the varia- tion in allele frequencies of two loci: Aat-1 and Mdh-2 (Table 2). These showed a geographic pattern in allele frequencies, with a tendency toward the fixation of the most common allele at Deserta Grande Island (see Fig. 2). with a resulting loss of genetic vari- ability at Deseita Grande (see tiieasures of genetic variability in Table 2). The number of migrants estimated between these sub- populations was 6.7 individuals per generation. F,s values were high for various loci indicating a tendency of heterozygote deficit (see Table 2). P. aspera Subpopulations The genetic identities and distances between samples of P. aspera are shown in Table 4. The identity values ranged from 0.930 to 0.959 (see also UPGMA dendrogram in Fig. 4). The total Fs-p estimated over-all samples was 0.136 (0.026-0.307). and 0.126 (0.013-0.262) between the most clo,sely related locations (Canigo and Deserta Grande), both values being different from zero. The estimated numbers of migrants obtained from the ¥^y values are 1.7 individuals per generation between Caniijo and De- serta Grande, 1.1 between Porto Moniz and Canit^o, and 1.5 be- tween Porto Moniz and Deserta Grande. A X" te^t applied to over-all samples was highly significant, showing that seven of the 21 loci analyzed were responsible for the genetic differentiation among populations (see Table 3). Even be- tween the most closely related pair of samples (Canigo and Deserta Grande), there was highly significant differentiation in allele fre- quencies (X" = 256.50; p < lO"*^). which was mainly attributable to the contributions of 4 loci: Est-1 (p < lO""^), Aat-1 (p < 10^"'); Est-I (p < 10"-*), and Idhp-2 (p < lO""*). It was possible to detect geographic patterns in allele frequen- cies (Fig. 3) as in P. candei. Figure 3 shows the si.x loci that displayed the most significant differentiation between subpopula- tions. We could distinguish three different patterns. The first, de- tected only at the Me-2 locus, showed that the less common allele of Porto Moniz. "B." became the most common one in Canigo and Deserta Grande subpopulations. As a second geographic pat- tern, it was possible to detect a dine in three loci. Aat-1. Est-I. and Spd-2. At \he Aat-1 locus, the allele "C" of Porto Moniz increased its frequency from 23 to 36% in Canijo and to 75% in Deserta Grande; at the Est-1 locus, the most common allele "B" in Porto Moniz, also increased its frequency from 62 to 85% in Cani90 and to 87% in Desena Grande; and at the Sod-2 locus, the most com- Patella spp., genetic stocks at Madeira TABLE 3. 949 Patella aspera: allele frequencies, F-statistics per locus, and over-all loci (Hith 95% confidence intervals. bet\»een brackets, obtained by bootstrap procedure) and contingency X' 2,000 m in the southern part of the basin and to near 1.000 m in the northern part. Because the shelf edge provides a curvilinear bound- ary between shallow-water conch habitats and the deep Exuma Sound, distance along the shelf edge was used to standardize abun- dance patterns for juveniles, adults, and shell middens (see below). The total human population around the periphery of Exuma Sound is < 1 0,000 people, centered primarily in George Town on Great Exuma. As a result, there are few sources of pollution, fishing pressure on queen conch is relatively low, and the ecology of the system is relatively unspoiled. The semienclosed nature of the sound and the presence of suitable conch habitat make this system a natural laboratory for the study of fishery recruitment processes and for analysis of distribution of conch from larva to adult. Along the Exuma Cays, juvenile conch live primarily in sea- grass meadows on the shallow bank, and adult conch live primarily offshore in the deeper waters of the sound to 30 m (Stoner and Schwarte 1994, Stoner and Ray 1996). Adult conch lay eggs from April through October (Stoner et al. 1992). The prevailing current on the shelf near the Exuma Cays runs alongshore from the south- east to northwest (Colin 1995) and plays an important role in transporting conch larvae to the northwest. Larvae are advected through the numerous tidal passes between the cays and onto the bank (Stoner and Davis 1997a), where competent larvae settle selectively and metamorphose in nursery grounds that have been well studied (Davis and Stoner 1994). Juveniles live in aggrega- tions at densities of 0.1 to 2 individuals/m" (Stoner and Ray 1993, Stoner et al. 1996a). As they mature into adults, juveniles migrate back through the tidal passes and out to the deepwater reproductive areas (Stoner and Ray 1996). Most conch fishers free -dive for their catch from small boats that have limited range. After removing the meat, they discard the shells along the shores of Exuma Sound, thereby creating ever-growing piles near the site of capture. These piles of discarded shells, hereafter referred to as middens, provide a time-averaged record of large-scale conch distribution, and a history of the fishery for at least 500 years (Stoner 1998). SAIL ROCKS STOCKING ISLAND GREAT EXUMA 76 SAN .o*- SALVADOR ■v » ISLAND CONCEPTION .'"' ■■, ISLAND 't. .,200m , joo?;- ( (e^ ( t LONG , ISLAND N 10k Figure 1. Map of the Exuma Sound system in the central Bahamas. The periphery of the sound was divided into U sectors (boxed numbers). The values below the sector numbers indicate the volume of queen conch shell middens expressed in m' per km of shelf edge (dashed line). Asterisks indicate the five stations at which newly settled conch were transplanted, four at Cat Island (CT-1 to CI-4) and one station at Shark Rock (SR) near Lee Stocking Island. The letter "P" near the island of Eleuthera indicates Powell Point, referred to in the text. Distribution of Queen Conch 957 METHODS Spatial relationships between the abundance of queen conch larvae, juveniles, adults, and discarded shells in middens were examined in and around the Exuma Sound. Surveys for the various life stages spanned several years, and some components of the results, as noted, have been published in previous studies. The methods and results sections will describe the results for different conch stages in reverse ontogenetic order, beginning with mid- dens, because it is this time-integrated spatial record that reflects the long-term fishery that we wish to explain. Furthermore, we were able to quantify shell middens around the entire rim of the Exuma Sound, thereby surveying the entire system. Quantifying all three living conch .stages was much more labor intensive, and only regional surveys could be accomplished. A similar survey strategy was used by Lipcius et al. ( 1997) in an analogous study of spiny lobster (Panulirus argus) populations in Exuma Sound. The purpose of this investigation was not to follow a single cohort of queen conch from larvae to adult stage or to the fishery in the Exuma Sound. Rather, our intent was to examine the long- term record of fishery yields over a relatively large scale (i.e.. Exuma Sound) and to interpret it in terms of cunent abundance patterns observed for early life stages and adults. Long-Term Record of the Conch Fishery Shell midden data used in this study were modified from Stoner (1948). where the survey methods were described in detail. BrieHy, all of the islands facing the Exuma Sound were searched for shell remains in a clockwise direction from the northern tip of Great Exuma to the southern end of Cat Island between 1989 and 1994 (Fig. I). The shoreline of Great Exuma was not surveyed, because this island has the largest human population on the pe- riphery of the Sound, and middens have been removed or disturbed by development. Small settlements occur around the rest of the sound, but most of the extensive shoreline is undisturbed. For this study, the shelf around the sound periphery was di- vided into sectors that were -20-km long (Fig. I ). Not included in ihe division of the periphery were deep-water passes between Cat Island and Long Island, the open-water pass between Little San Salvador and Eleuthera, and the bank periphery where there were no islands for landing conch (i.e., in the extreme northern sound and between Long Island and Great Exuma). Distances were mea- sured at the edge of the shelf and varied somewhat to separate nurseries for queen conch that are as.sociated directly with the tidal flow fields between islands at the edge of the bank (Jones 1996, Stoner et al. 1996b). Shell middens ranged in size from a few scattered shells to accumulations that were 3 to 4-m high. Estimates of the total volume of individual accumulations were made by measuring their basic dimensions as described by Stoner (1998). Notes were also made on the apparent age of the shell middens. For example, some were composed primarily of very old and eroded shells, and the top layers of others were covered with recently landed shells from which the bright shell nacre had not yet faded. The middens were mapped, volumes were summed for each sector, and the volume of shells per kilometer of shelf periphery was used as a standard index of historic fishery yield from individual bank sectors. Adult Surveys Labor-intensive diving surveys for adults were concentrated in four sectors of the sound chosen on the basis of general geographic positions and known productivity patterns in queen conch revealed in the midden survey. They included: ( I ) the conch-poor area at the southern end of Cat Island (Fig. I. sector 1 1 ); (2) the well-studied area near Lee Stocking Island in the southern Exuma Cays (sector I), where conch productivity is moderate; (3) an area inside the Exuma Cays Land and Sea Park near Waderick Wells (sector 4). where Stoner and Ray (1996) found very high densities of adult conch; and (4) an area between the Schooner Cays and the south- east tip of Eleuthera (sector 7), where the highest concentrations of shell middens were located (Stoner 1998). Stoner (1998) found a positive correlation between middens and juvenile conch abundance, and we hypothesized a similar relationship between middens and adult conch abundance. If such a relationship exists, midden volume could be used as an indicator of living adult conch distribution. Extensive adult surveys were made during the summer 1991 near Lee Stocking Island (Stoner and Schwarte 1994), and near Waderick Wells (Stoner and Ray 1996), Cat Island, and Eleuthera in 1994. Although surveys for conch were made during two different years, the conch populations in the Exuma Sound seem to be relatively stable over the long term. Annual surveys for adult conch conducted at selected sites off Lee Stocking Island between 1988 and 1994 (Stoner and Sandt 1992, unpubl. data) revealed that maximum variation from the mean population size and density was just 19%, and the population was only 4% above average in 1991. This stability is probably a function of low fishing pressure, particularly in depths below the reach of the average free-diving fishers (>I0 m). and a queen conch life span of at least 12 years (Coulston et al. 1987). Depth-stratified surveys for adult conch were conducted in each of the four sectors described above. Seven depth intervals were examined: to 2.5 m (where present), 2.5 to 5 m. 5 to 10 m, 10 to 15 m, 15 to 20 m, 20 to 25 m, and 25 to 30 m. The deepest interval was not surveyed at either Eleuthera or Cat Island because of a very steep grade in depths >25 m that did not seem to support adult conch. The intervals were surveyed along nine offshore transect lines pei-pendicular to Lee Stocking Island, six lines per- pendicular to Waderick Wells, four lines perpendicular to Eleuth- era, and three lines pei-pendicular to Cat Island. Because of ex- tremely low conch densities in the shallow waters (900 |xm SL), which were at or near metamor- phic competence. Abundance was calculated as numbers of ve- ligers per unit \olume of water sampled ( veligers/100 m") for each age class. Data are reported as the mean of two tows for each station for individual cruises and as mean of means for 1 993 when cruises were pooled. Relationships Among Different Ontogenetic Stages The abundance and distributional data for middens, adults, ju- veniles, and larval conch were collected to examine the relation- ships among distinct ontogenetic stages in different geographic sectors around Exuma Sound. Tests of correlation were performed between conch midden volume and adult conch density, and be- tween midden volume and juvenile conch density. We hypoth- esized that, if a positive relationship exists between the fishery yield and benthic stages, then midden volume would retlect living adult conch populations and could be used as an index of living conch abundance around Exuma Sound. We also tested the rela- tionship between midden volume and density of early-stage conch veligers along the shelf periphery to determine if newly hatched larvae reflected large-scale distribution of the reproductive stock. To explore the potential importance of larval supply to distribution of benthic populations around the sound, we also tested for cor- relations between midden volume and density of late-stage larvae and between late-stage larvae and juvenile conch. Individual sec- tors were often represented by more than one plankton sampling station, providing increased confidence in the values used in the regressions. This varied with year and sampling strategy; however, all plankton stations inside a sector boundary and within 5 km of the shelf edge were included in a mean value. Transplant Experiment We observed very low densities of both juvenile and adult conch in the shallow shelf environment at the southern end of Cat Island (see Results). Because late-stage larvae were relatively ubiquitous throughout the sound, it is unlikely that such low den- sity is explained by a low supply of settlement stage larvae to that location. Therefore, we hypothesized that the habitat in this area was unsuitable for juvenile growth and survival. To test this habi- tat-limitation hypothesis, we conducted a transplant experiment during the summer of 1995 to measure postlarval growth. If the Cat Island habitat was suitable for newly settled conch growth, then postlarvae transplanted there should grow at rates similar to those transplanted in a conch nursery area near Lee Stocking Is- land called Shark Rock, where a well-studied juvenile aggregation has persisted for over 10 years (Stoner and Waite 1990, Stoner and Ray 1993, Stoner et al. 1994). Although growth rates in enclosures give no indication of predation-induced mortality, they do provide a good index of habitat suitability in terms of food quality and availability (Stoner and Sandt 1991). Three queen conch egg masses were collected from a repro- ductive site on the shelf off Lee Stocking Island on 18 June 1995. The eggs hatched 5 days later, and the larvae were cultured ac- cording to well-established procedures (Davis 1994). Briefly, lar- vae were held in 20-L plastic buckets filled with seawater collected daily from the bank west of Lee Stocking Island. Natural foods in the seawater were supplemented with cultured Tahitian Isochrysis spp. Metamorphosis was induced on 21 July, at -1 mm shell length (SL). and postlarvae were raised in plastic trays with aerated sea- water on a diet of seagrass detritus {Thalassia testudinuin) col- lected from the field. Five enclosures were deployed at one Shark Rock station in an area of uniform habitat characteri.stics at a depth of 4.1 m MLW (Fig. 1 ). It was our intent to deploy the Cat Island enclosures in similar habitat, and. after extensive surveying, four stations were selected along the southwest shore at 3.2-5.3 m depth (Fig. 1). Sediment and seagrass (Thalassia teslmliiuim) detritus samples were collected, and living seagrass shoot density was counted near each enclosure to characterize the station (Table 2). Enclosures were pvc cylinders (diameter = 16 cm, height = 25 cm) with abundant large holes (diameter = 5.5 cm) cut from each to allow for water circulation (after Ray and Stoner 1995). Each cylinder was lined with a polyester mesh ( 1 mm) sleeve, pushed into the substrata, secured to reinforcement bars driven into 960 Stoner et al. TABLE 2. Mean growth rate and survival of postlarval queen conch transplanted at four stations near Cat Island (CI) and one station near Lee Stocking Island at Shark Rock (SR). Growth Rate Survival Seagrass Density Seagra.ss Detritus Station (mm/da.v) (%) Sediment Type (Shoots/m-| (g) CM 0.12 + 0.01- 90 ±9 Medium sand 570 ± 182 1.35±0.17 CI-2 0.10 ±0.03-' 100 ±0 Coarse sand 950 ±178 0.71+0.18 CI-3 0.16±0.01'' 90± 11 Coarse sand 750 ± 257 0.68 ± 0.28 CI-4 0.30 ± o.o:"- 99 + 2 Medium sand 680 ± 144 0.19 + 0.13 SR 0.28 + 0.02" 92 ± 12 Medium sand 710± 167 4.5 1 ± 0.94 Values are mean + SD; n = 5 at each station except CI-2, where n = 4 for conch growth rale. Growth rale data were homogeneous (Cochran's test, p > .05, and differences in the means were determined by one-way ANOVA (F|4 |^, = 1 14; p < .001 ) followed by Tukey HSD multiple comparison test. Means that are not significantly different (p > .05) are designated by similar lower case letters. Habitat characteristics including sediment type, seagrass (Thalassis testudinwn) shoot density, and seagrass detrital biomass (dry weight) (n = 5) are also given for each station. See Figure 1 for location of each station. the sediment, and covered with a mesh top. Large predators were removed prior to introduction of postlarvae. Prior to transplanting, subsamples of the cultured postlarvae. which were relatively uniform in size, were measured for shell length with dial calipers. Po.stlarvae were introduced into enclo- sures (see below) at Shark Rock on 20 August 1995 at 6.0 mm SL (SD = 0.4, n = 40) and at Cat Island on 24 August at 6. 1 mm SL (SD = 0.4. n = 80). Each enclosure contained 18 animals. Postlarvae were recovered from Shark Rock on 9 September, after 20 days in the field, and from Cat Island on 16 September, after 23 days. They were measured for shell length immediately after recovery. Mean daily growth rates were calculated from the living individuals in each cage using the initial shell length from the appropriate subsample. To test for differences among stations, one-way analysis of variance (ANOVA) was performed followed by Tukey HSD tiiultiple comparison test (Day and Quinn 1989). RESULTS Long-Term Record of the Conch Fishery Surveys of shell middens revealed the highly variable nature of the conch fishery yield around Exutna Sound (Fig. 1). Highest shell concentrations ( 198 m' shells/km) occuiTed in the northeast region (sector 7), located between two regions with low concen- trations (-0.2 m'/km). A high concentration (133 m"" shells/km) was also observed in sector 4 in the north central Exuma Cays near Waderick Wells. All four sectors in the eastern sound, from Eleuthera to Cat Island, had very low concentrations of conch shells (^2 m^ shells/km), corroborating the low productivities of conch reported by fishers interviewed at Cat Island. Shells were abundant all along the Exuma Cays island chain on the western boundary of the sound, except at the extreme north. Most of the shell middens in the Exuma Sound contained both very old and recently collected shells. Important exceptions to this were enormous accumulations of recently collected conch in the vicinity of Powell Point on Eleuthera. These shells had bright color indicating capture over the last few years, and most had the thin shell lips indicative of relatively young adults. The largest indi- vidual accumulations (>1000 m') occurred in sector 4 near Wad- erick Wells, but it was apparent that most of these shells were collected much earlier than those on Eleuthera. This finding was not unexpected, because sector 4 lies within the Exuma Cays Land and Sea Park, where all fishing has been prohibited since 1985. Adult Siineys In general, densities of adult conch were highest at Waderick Wells and Schooner Cays, intermediate near Lee Stocking Island, and very low (except in the 15 to 20-m depth interval) at Cat Island (Fig. 2). Highest density of adults occurred at 10 to 15 m near Waderick Wells (270 conch/ha), at 15 to 20 m near Lee Stocking Island (88 conch/ha), and at 15 to 20 m off Cat Island (84 conch/ ha). Distribution at Schooner Cays was bimodal. with density maxima in shallow water (2.5 to 5 m-228 conch/ha) and in rela- tively deep water (20 to 25 m-93 conch/ha). Most of the conch in the 2.5 to 5-m interval at Eleuthera were very young adults (with thin shell lips), with a high density of large, late-stage juveniles mixed in as noted below. The adults at most other locations and depths were older. The benthic habitat within the to 2. 5-m depth inter\al cotn- prised very little surface area in each of the four regions surveyed. Adult conch in this nanow band were rare, and, therefore, consid- ered to be negligible. At all four sites, the two depth intervals between 20 and 30 m represented relatively small proportions of the total habitat occupied by adults; therefore, densities of conch in the depth intervals with largest surface area (2.5 to 20 m) were used to test for correlations between adults and other ontogenetic stages (see below). Juvenile Sun'eys Surveys for juvenile aggregations were conducted in five sec- tors along the periphery of the Exuma Sound (Table 1 ), The lengths of shelf edge, corresponding to the shallow-water areas surveyed, ranged from 10 km near the Schooner Cays, where conch juveniles were abundant, to 20 km near Cat Island. Scattered juveniles were observed in the bight of Cat Island and along the westernmost third of the south shore, but no aggregations were found during our extensive systematic surveys conducted in 1993 or during numerous visits to the area between 1993 and 1995. Largest aggregations of juvenile conch occuired in the vicinity of the Schooner Cays just north of Powell Point on the island of Eleuthera and on the shelf immediately west of the Point (Fig, I), where young adults were also abundant. In August 1993, a single aggregation in the seagrass bed extending south from the Schooner 400 Distribution of Queen Conch 400 Waderick Wells Sector 4 300 200 961 100 - Schooner Cays Sector 7 u c o o 300 200 100 Lee Stocking Island Sector 1 400 300 200 100 c\i o in o in O (M 900 jam) TONOUE OF THE OCEAN «. Figure 3. Density of (A) early-stage, (B) midstage, and (C) late-stage queen conch veligers collected during five cruises in 1993 (12 June to 23 August) at 13 stations. Plankton tows were made at each station with 202-|iim mesh nets. Values represent the mean of means for each station. very few early stages were collected at the offshore, open-water stations. Although mid- and late-stage larvae were widespread through- out Exunia Sound in 1994, they were usually collected in relatively low densities (Figs. 4B.C). Moderate densities of midstage larvae were found along the shelf edge of the northwest sound near Sail Rocks (45 veligers/100 m'), near Waderick Wells (24 veligers/100 m^), and at one station in the center of the Sound (12 veligers/100 m^) (Fig. 4B). The rest of the sound, including its periphery, yielded a mean density of <10 midstage veligers/100 m\ Highest densities of late-stage larvae were found near Sail Rocks (52 ve- ligers/100 m"^). the pass between Eleuthera and Little San Salvador (34 veliger.s/100 m"*), the outer shelf edge of Cat Island (56 ve- ligers/100 ni"*), and one station in the center of the sound (10 veligers/100 m') (Fig. 4C). The rest of the sound .stations yielded to 6 veligers/100 m'^. Thus, despite high concentrations of early stage larvae near the large reproductive populations in the north- central Exuma Cays, settlement-stage queen conch larvae were found throughout the sound in relatively low densities. Veliger distribution was explored along the Exuma Cays in two subsequent cruises, in July and August 1994. and the spatial pat- terns were remarkably similar to those reported above. For ex- ample, highest densities of early-stage larvae were always most abundant from the middle Exuma Cays to the north, and late-stage larvae were always highest in the extreme northern Exumas. More intensive surveys for late-stage larvae from late August to mid-September 1995 revealed that these settlement-ready stages were ubiquitous throughout the Exuma Sound, except in the ex- treme southern sound, in the opening between Cat Island and Long Island, and at numerous stations on the shelf along the Exuma Cays (Fig. 5). Highest densities (10 to 28 veligers/100 m') oc- curred in the extreme northern sound, near Waderick Wells, and at a station south of Little San Salvador. The rest of the Exuma Sound had late-stage densities of 1 to 10 veligers/100 m\ with the ex- ception of one station in the central basin (12 veligers/100 m" ). Relationships Among Different Ontogenetic Stages When the abundance of shell middens in a sector was compared with the inean density of adult conch on the adjacent shelf <20 m in depth (Sectors 1. 4. 7. and 1 1). there was a highly signifi- cant correlation (r = 0.973, p = .03) (Fig. 6A). Midden abundance was also closely correlated with the abundance of juveniles in adjacent waters (Sectors 1. 4, 5. 7, and II) (r = 0.915. p = .03) (Fig. 6B). The correlations between shell midden volumes and living populations of both juvenile and adult conch Distribution of Queen Conch 963 TONGUE OF THE OCEAN ELEUTHERA Eafly-stage (< 500 )im] TONGUE OF THE OCEAN ELEUTHERA Mld-stagc (500-900 Jim) ...jog,. WAOERICK N ^^ WELLS TONOUE OF THE OCEAN jELEUTHEBA Lats-Stage (> 900 urn) . » 1-10 © 10-50 o SO- 100 o 1O0-2O0 G j >2O0 Figure 4. Density of (A) early-stage, (B) midstage, and (C) late-stage queen conch veligers collected during two cruises in 1994 (5 to 13 June, and 22 to 24 June) at 44 stations. Plankton tows were made at each station with 202-nm mesh nets. Values represent the mean for each station. over the mesoscale validates the use of middens as a proxy indi- cator of living conch abundance around the perimeter of Exuma Sound. Abundance of early-stage, newly hatched larvae at any one location should reflect the size and/or density of the reproductive population in the general vicinity. The most synoptic data for early stages were collected in 1994 (Fig. 4). and there were 10 sectors for which we had both veliger and midden data. Very high con- centrations of larvae were collected in the north-central Exuma Cays and near Lee Stocking Island in the southern Exumas. Un- expectedly, the correlation between early-stage larval densities and midden abundance was low and not significant (r = 0.413. p = .24). The poor correlation was a function of one extreme outlier representing sector 7, near the Schooner Cays, where very large populations of adult conch and large middens were found, but few early-stage veligers. As mentioned above, most of the adult conch at this site were very young adults, which may not have been in reproductive state in the summer of 1994. Also, unlike other in- shore shelf stations around the sound, the stations that we sampled for veligers near the Schooner Cays were swept by very strong tidal cunents; therefore, it is possible that sampling at this site during the flood tide resulted in low larval densities. The flood tide would carry locally spawned larvae onto the adjacent bank and away from the sampling stations. When sector 7 was removed from the analysis, there was a highly significant positive conela- tion between the abundance of early larval stages and middens (r = 0.964, p < .001 ). as was predicted. Best distribution of residuals occurred with a natural log transformation of the data (r = 0.746, p = .02) shown in Figure 7. We also hypothesized that the juvenile abundance pattern (Table 1 ) would reflect densities of late-stage larvae (i.e.. those that are at or near metamorphic competence and ready to settle). However, using the juvenile abundance data available for five sectors (Table 1 ), the correlations were low and not significant (p > .35) in all 3 years in which larval data were collected (r = 0.431 in 1993, r = 0.512 in 1994. r = 0.220 in 1995) (Fig. 8). In 1994, high densities of late-stage larvae were found at the south end of Cat Island (sector 11) (Fig. 4). where juvenile populations were typically very small. In the same year, sectors with large juvenile populations (e.g.. sectors 4 and 5) had low densities of late-stage larvae. In 1995. late-stage larvae were relatively high in sectors 4 and 5. but also common in sector 1 1 near Cat Island (Fig. 5). Larval supply was not a good predictor of juvenile concentration. To complete the analysis of the relationship between middens and mesoscale distribution of conch around the Exuma Sound, we examined midden volume as a function of late-stage, competent larvae. The correlations were negative and not significant in 1994 (r = 0.420, p = .23. n = 10) and in 1995 (r = 0.312, p = .35, 964 Stoner et al. .% c o •o ■a 200 r 150 ■ 100 ■ 50 y = 1.7305x- 38.028 R = 0.973 P = 0.03 • 4 50 100 150 Adult density (no./ha) 200 150 100 ■ y = 3.1467x- 0.039 R = 0.915 P = 0.03 • 4 • 7 Concentration of juveniles (ha/km) Figure 6. Conch midden volume plotted as a linear function of (A) adult conch density at 2.5 to 20-m depth and (B) concentration of juvenile conch. Pearson correlation coefficients (R) and p-values are given for each regression equation. The number above each point represents the sector at which surveys were conducted in Exuma Sound. Distribution of Queen Conch 965 a y = 0.845 x + 1.35 n E o R = 0.746 P = 0.021 ? 6 - • ■■*. d c^ -^ ^/^ >« ^ «rf j< « • • ^/^ c 0) j^ •D 4 -/ ^ ^,.^ • 0) • w ^^ o ^^ > j,^ u • ^^ O) V- '^ ra ^ — *-• w 2 - y^ >. • k. w LU n 1 < 1 1 1 ' 1 Midden volume (m /km) Figure 7. Relationship between mean density of early-stage (<5()(l jim) conch veligers (natural log transformed) and conch midden volume, surveyed at nine sectors in Exuma Sound in 1994. The Pearson cor- relation coefficient (Rl and p-value are given for the regression equa- tion. the production of early-stage larvae around the periphery of Exuma Sound. These newly hatched larvae were most abundant in the nearshore areas where adults live and in regions known for high adult concentrations, such as near the Schooner Cays and in the Exuma Cays Land and Sea Park near Waderick Wells. The positive correlation between adults (i.e., spawner abundance) and early-stage larvae was predictable and not surprising. Ultimately, however, it is settlement-stage larvae that supply and sustain benthic populations, and examples of correlations be- tween larval supply and settlement and/or recruitment to various benthic stages are known for a variety of marine invertebrates (Caffey 1985. Keough. 1988, Bertness et al. 1992) and fishes (Milicich et al. 1992, Doherty and Fowler 1994). Relationships between late-stage larval concentration and juvenile population size have been explored for queen conch in several different lo- cations and on different scales. Stoner and Davis (1997a) have shown that aggregations of juvenile conch were directly associated with local concentrations of late-stage larvae within a tidal current flow field (<10-km long) on the Great Bahama Bank near Lee Stocking Island. Significant positive correlations have also been found between densities of late-stage larvae and juvenile popula- tion size on a 10 to 5()-km scale across multiple nursery grounds in the Exuma Cays and in the Florida Keys, although the pattern was not coherent across the two locations (-500 km) (Stoner et al. 1996a). Because of the semienclosed circulation pattern in the Exuma Sound (Colin 1995, Lipcius et al. 1997). it is probable that conch larvae produced in the sound could be retained in the system for the duration of their developmental period. However, the potential tn 0) 'E > p »^ o c ,o +3 CC k. c 4~~~--~ 20 ^ 5 ^^~~~~~~~ --.^.^^^ 10 .• 1 1 ^^^^ --.-.1^1 —J • — 1 10 15 20 25 30 • 7 y = -1.7072x + 29.717 R = 0.220 1995 ■ P = 0.72 ■ 4 —_____• 5 _• 1 1 11 _] 1 1 1 1 1- < ' 70 60 50 40 30 20 10 0123456789 10 Late-Stage veliger density (no./100 m^) Figure 8. Linear relationship between the concentration of .ju>enile conch and mean density of late-stage (>900 (ini) conch veligers from 1993 to 1995. Pearson correlation coefficients IR) and p-values are given for each regression equation. The number above each point represents the sector al which the survey was conducted in Exuma Sound. Note extended x-axis for 1994 data. for dispersion within the sound over the 2 to 4 week precompetent period (Davis 1998) is very large. Although late-stage conch larvae were consistently abundant in the northern Exuma Sound during 1994 and 1995. they were relatively ubiquitous throughout the sound. Correlations between the concentrations of settlement-stage larvae and either juvenile or midden distributions were never sig- nificant in any of the 3 survey years despite consistent spatial patterns of larval density. These results suggest that the regional pattern of distribution in benthic life stages within Exuma Sound is set by settlement processes and/or early postsettlement processes during the first year of life, and not by differences in larval supply. Lipcius et al. ( 1997) arrived at similar conclusions about the large- 966 Stoner et al. 250 200 150 100 y = -2.5411x + 62.841 R = 0.420 P = 0.23 1994 o > C 250 a> •a ■a 200 150 100 50 :-3.6896x-f 56.161 R = 0.312 P = 0.35 25 30 1995 10 12 14 16 18 Late-Stage veliger density (noVlOO m ) Figure 9. Linear relationship between queen conch midden volume and mean density of late-stage (>900 pm) conch veligers in 1994 and 1995. Pearson correlation coefficients (Rl and p-values are given for each regression equation. Symbols represent sectors surveyed in Exuma Sound for the 2 years. scale distribution of spiny lobster (Painilinis argiis) populations in Exuma Sound. Settlement stage lobster were abundant at Cat Is- land, yet benthic populations were small. The decoupling between larval supply and juvenile and adult stages was attributed to habitat limitation for early juvenile lobster at Cat Island. Miron et al. (1995) pointed out the inherent methodological difficulties in correlating larval supply and larval settlement or recruitment. They noted that competent larvae must be quantified and that they must be sampled properly (i.e., with proper respect to location in the water column and settlement substratum). Although settlement-stage queen conch larvae have never been collected in high densities, compared with densities of other mollusks in tem- perate waters, they are relatively easy to sample, because they occupy near-surface waters in most circumstances (Barile et al. 1994, Stoner and Davis 1997a, Noyes 1996). Furthermore, it is relatively easy to identify competent forms on the basis of size, pigmentation, and other features (Davis 1998). Consequently, we believe that we have sampled the correct larval stages using an appropriate technique. It can also be argued that larval supply is best measured as a rate of delivery of competent larvae to a potential settlement site (Olmi et al. 1990, Yund et al. 1991 ). Measuring this is particularly difficult for queen conch on the Great Bahama Bank because of strong tidal currents (see Stoner and Davis 1997b). However, the difficulty is lower in the Exuma Sound, and we have high confi- dence in the regional patterns of larval abundance reported for two reasons. First, currents in the sound are much weaker (<20 cm/sec; Colin 1995), than those on the Bank (often >100 cm/sec; N. P. Smith, unpubl. data), so the issue of larval flux is less complicated in the .sound than on the nursery grounds of the Bank. Second, and more importantly, multiple visits to selected stations throughout the sound between 1993 and 1994 revealed that the regional pat- terns in veliger distribution were consistent over time. For ex- ample, five cruises over 13 stations in 1993 showed that early- and midstage larvae were always abundant in the northern sound, and always highest on the shelf adjacent to Waderick Wells. Late-stage larvae were always most abundant in the sound offshore from Waderick Wells. Three cruises along the island shelf east of the Exuma Cays in 1994 (Stoner and Mehta. unpubl. data) confirmed the pattern of maximum abundance of early- and midstage larvae in the vicinity of Waderick Wells and to the north, and late-stages were most abundant near Sail Rocks in every case. Larvae of all stages were always rare in the extreme south section of the sound. Therefore, because of the consistency of larval distribution, both within and between years, we believe that the regional patterns of larval abundance reported in this study are representative for the sound. Given that the abundance of late-stage larvae did not explain mesoscale variation in the abundance of juveniles, adults, or fish- ery yields of queen conch in the Exuma Sound, we conclude that the regional distribution of benthic stages is regulated by settle- ment and/or postsettlement processes associated with some ele- ment of the habitat. Similar conclusions have been drawn for a large number of other marine invertebrates (Keough and Downes 1982, Luckenbach 1984, Connell 1985, McGuinness and Davis 1989, Osman et al. 1992. Olafsson et al. 1994. Eggleston and Armstrong 1995, Hunt and Scheibling 1997, Lipcius et al. 1997). Many invertebrates settle and metamorphose in the presence of certain chemical agents in or on the substratum (Morse and Morse 1984, Hadfield and Scheuer 1985, Burke 1986, Butman and Grassle 1992, Pawlik 1992), and Mianmanus (1988) has shown that phycobiliproteins associated with red algae are active agents in conch settlement and metamorphosis. We know from extensive dredge sampling for newly settled queen conch (both live and recently killed) in a tidal flow field near Lee Stocking Island that settlement is not random and that it occurs in specific locations (Stoner et al. 1998). This confirms earlier laboratory experiments showing that competent queen conch larvae settle in response to specific biological cues found within nursery grounds (Davis and Stoner 1994). Queen conch larvae are. in fact, capable of testing the substratum, returning to the water column multiple times, and delaying metamorphosis for long periods of time (for at least 60 days after competence is achieved) (Noyes 1996). Experimental laboratory work shows that the larvae settle and metamorphose only in habitats where sub,sequent growth rates are high (Stoner et al. 1996c). Consequently, it is possible that variation in the abun- dance of queen conch populations on the Great Bahama Bank surrounding the Exuma Sound is related to either the quality or quantity of habitat with appropriate settlement cues and high growth potential for postlarvae. Habitat-limitation is the most plausible explanation for the low abundance of juvenile and adult conch west of Cat Island, because competent larvae were present in substantial numbers. Frequency of settlement was not tested, because the only way to ascertain this is by dredging, which is extremely labor intensive. However, trans- plant experiments provide important insights into the nutritional quality of potential nursery sites. Two lines of reasoning suggest that habitat at Cat Island is limiting for conch. First, the type of habitat that typically supports juvenile conch on the Great Bahama Bank (moderate density seagrass with accumulations of decom- posing detritus and red and green algae) has been studied exten- Distribution of Queen Conch 967 sively (see Stoner 1997). and was uncommon on the bunk west of Cat Island. Seagrass was found in relatively small patches ( 1 to 10 ha), and much of this was exposed to higher physical energy than is typical for conch nurseries. Second, only one of the four sites assumed to be suitable for juvenile conch provided for growth rates similar to those in a known nursery near Lee Stocking Island. Therefore, it is likely that the small queen conch population and the poor fishery for conch near Cat Island is habitat-hmited. Differential mortality of young conch could also explain re- gional variation in recruitment to the age- 1 year class. There are a host of predators on juvenile conch (Randall 1964), including a large variety of recently discovered micropredators such as xanthid crabs and certain polychaetes that feed on newly settled conch (Ray-Culp et al. 1997). It is now recognized that mortality rates in newly settled invertebrates can be very high (Osman and Whitlatch 1995. Gosselin and Qian 1997), and queen conch are no exception (Ray et al. 1994. Stoner and Glazer 1998). Although we did not test for regional variation in mortality of juvenile conch, this is a possible explanation for the population patterns observed. Conclusions and Fishery Management Implications Genetic analysis of queen conch collected from 22 populations throughout the greater Caribbean region, including the Bahamas and south Florida indicate a high rate of gene flow among the populations (Mitton et al. 1989, Campton et al. 1992), and certain populations may depend entirely upon upstream reproductive sources (Stoner et al. 1997c). It is clear, therefore, that sound fisheries management will demand good knowledge of larval drift and associated metapopulation dynamics (Berg and Olsen 1989. Appeldoom 1994, Stoner 1997). However, the direct correlation between the quantity of larvae supplied to the nurseries and the subsequent abundance of juvenile queen conch in the benthic population that occurs at a local .scale (Stoner et al. 1996a) seems to break down at the large scale. In Exuma Sound ( 180-km long) the abundance of early-stage larvae was positively correlated with regional abundance of adults. However, the distribution of juve- niles, adults, and fishery yields was independent of the abundance of competent larvae, and processes of settlement and postsettle- ment seem to regulate benthic population size. We have shown in the past that conch nursery grounds have unique physical and biological features that enhance larval settlement, provide high nutritional qualities, and promote high survivorship (Stoner et al. 1995, Stoner 1997). It is now clear that high abundance of com- petent lar\ae does not guarantee high queen conch production, and that fisheries management for the species must consider both qualitative and quantitative elements of habitat for young conch. Because vast shallow-water areas within the biogeographic range of queen conch are. in fact, not suitable for production of the species, both local- and large-scale mechanisms of population dy- namics and habitat use need to be understood, and the ecological integrity of key nursery habitats needs to be preserved. ACKNOWLEDGMENTS This research was supported by grants from the National Un- dersea Research Program of NOAA (U.S. Department of Com- merce) to the Caribbean Marine Research Center. Many people participated in the field work and plankton sorting for this long- term study including P. Bergman. L. Cowell, M. Davis, G. Dennis, L. Hambrick, R, Jones, C. Kelso, S. O'Connell. H. Proft, V. Sandt, J. Waite, and E. Wishinski. Crews from the research vessels Un- dersea Hunter. Chiillenger, Shadow. Sea Diver, and Cyclone pro- vided ship support, and B. Bower- Dennis drafted the maps. LITERATURE CITED Appeldoom, R. S. 1994. Queen conch management and research: status, needs, and priorities, pp. 301-319. In: R. S. Appeldoom and B. Rod- riguez (eds.). Queen Conch Biology Fisheries and Mariculture. Fun- dacion CienuTica Los Roques. Caracas, Venezuela. Barile, P. J.. A. W. Stoner & C. M. Young. 1994. Phototaxis and vertical migration of the queen conch (Stromhus gigus Linne) veliger larvae. J. Exp. Mar Biol. Ecol. 183:147-162. Berg. C. J., Jr. & D. A. Olsen. 1989. Conservation and management of queen conch (Strombits gigiis) fisheries in the Caribbean, pp. 421^40. In: J. F. Caddy (ed.). Marine Invertebrate Fisheries: Their Assessment and Management. John Wiley & Sons, New York. Bertness, M. D., S. D. Gaines, E.G. Stephens & P.O. Yund. 1992. Components of recmitment in populations of the acorn barnacle Semi- balanus balanoides (Linnaeus). ./. Exp. Mar. Biol. Ecol. 156:199- 215. Burke. R. D. 1986. Pheromones and the gregarious settlement of marine invertebrate iar\'ae. Bull. Mar. Sci. 39:323-331. Butman. C. A. & J. P. Grassle. 1992. Active habitat selection by Capitella sp. I larvae. I. two-choice experiments in still water and tlume flows. J. Mar. Res. 50:669-715. Caffey. H. M. 1985. Spatial and temporal variation in settlement and re- cmitment of intertidal bamacles. Ecol. Monogr. 55:313-332. Campton, D. E.. C.J. Berg. Jr.. L. M. Robinson & R. A. Glazer. 1992. Genetic patchiness among populations of queen conch Strombus gigas in the Florida Keys and Bimini. Fish. Bull. U.S. 90:250-259. Colin, P. L. 1995. Surface currents in Exuma Sound. Bahamas and adjacent areas with reference to potential larval transport. Bull. Mar. Sci. 56: 48-57. Connell, J. H. 1985. The consequences of variation in initial settlement vs. postsettlement mortality in rocky intertidal communities. J. Exp. Mar. Biol. Ecol. 93:11-45. Coulston. M. L.. R. W. Berey. A. C. Demp.sey Si. P. Odum. 1987. Assess- ment of the queen conch (Strombus gigas) population and pre- dation studies of hatchery reared juveniles in Salt River Canyon, St. Croix, US. Virgin Islands. Proc. Gulf Carib. Fish. Inst. 38:294- 306. Davis, M. 1994. Mariculture techniques tor queen conch {Strombus gigas Linne): egg mass to juvenile stage, pp. 231-252. In: R. S. Appeldoom and B. Rodriguez (eds.). Queen Conch Biology, Fisheries, and Mari- culture. Fundacion Cientifica Los Roques. Caracas. Venezuela. Davis. M. 1998. The effects of natural foods, temperature, and salinity on the length of larval life for the tropical gastropod Strombus gigas. Ph.D. dissertation. Florida Institute of Technology, Melbourne. 159 p. Davis. M.. C. A. Bolton & A. W. Stoner. 1993. A compari,son of larval development, growth, and shell morphology in three Caribbean Strom- bus species. Veliger 36:236-244. Davis. M. & A. W. Stoner. 1994. Trophic cues induce metamorphosis of queen conch larvae (Strombus gigas Linnaeus). J. Exp. Mar. Biol. Ecol. 180:83-102. Day. R. W. & G. P. Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monogr. 59:433-463. Doherty. P. & T. Fowler. 1994. An empirical test of recruitment limitation in a coral reef fish. Science 263:935-939. Doheny, P.J. & D. M. Williams. 1988. The replenishment of coral reef fish populations. Oceanogr. Mar. Biol. Annu. Rev. 26:487-55 1 . Eggleston. D. B. & D. A. Armstrong. 1995. Pre- and postseUlement deter- minants of estuarine dungeness crab recruitment. Ecol. Monogr. 65: 193-216. 968 Stoner et al. Gosselin. L. A. & P. Y. Qian. 1997. Juvenile monalily in benthic marine invertebrates. Mar. Ecol. Prog. Ser. 146:265-282. Hadfield. M. G. & D. Scheuer. 1985. Evidence for a soluble metamorphic inducer in Phestilla: ecological, chemical, and biological data. Bull. Mar. Sci. 37:556-566. Hensley, D, A., R. S. Appeldoom. D. Y. Shapiro. M. Ray & R. G. Turin- gan. 1994. Egg dispersal in a Caribbean reef fish. ThaULSsonui bifas- cianim. 1. Dispersal over the reef platform. Bull. Mar. Sci. 54:256-270. Hunt, H. L. & R. E. Scheibling. 1997. Role of early postsettlement mor- tality in recruitment of benthic marine invertebrates. Mar. Ecol. Prog. Ser. 155:269-301. Johnson, D. R. & B. S. Hester. 1989. Larval transport and its association with recruitment of blue crabs of Chesapeake Bay. Estuar. Coastal Shelf Sci. 28:459^72. Jones, R. L. 1996. Spatial analysis of biological and physical features associated with the distribution of queen conch. Stromhii.': giga.'i. nurs- ery habitats. M.S. thesis, Florida Institute of Technology, Melbourne. 98 p. Keough, M.J. 1988. Benthic populations: is recruitment limiting or just fashionable? Proceedings of the Sixth International Coral Reef Sym- posium, pp. 141-148. Keough, M. J. & B. J. Downes. 1982. Recruitment of marine invertebrates: the role of active larval choices and early mortality. Oecologia 54:348- 352. Lipcius. R. N., W. T., Stockhausen, D. B. Eggleston, L. S. Marshall & B. Hickey. 1997. Hydrodynamic decoupling of recruitment, habitat qual- ity, and adult abundance in the Caribbean spiny lobster: source-sink dynamics? yW. Rev. 32:65- 109. Olmi, E. J. III. J. van Montfrans, R. N. Lipcius, R. J. Orth. & P. W. Sadler. 1990. Variation in planktonic availability and settlement of blue crab megalopae in the York River. Virginia. Bull. Mar Sci. 46:230-243. Osman, R. W. & R. B. Whitlatch. 1995. Predation on early ontogenetic life stages and its effect on recruitment into a marine cpifaunal community. Mar Ecol. Prog. Ser. 117:111-126. Osman, R. W., R. B. Whitlatch & R. J. Malatesta. 1992. Potential role of micropredators in determining recruitment into a marine community. Mar. Ecol. Prog. Ser 83:35-43. Pawlik, J. R. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Oceanogr. Mar. Biol. Annu. Rev. 30:273-335. Randall, J. E. 1964. Contributions to the biology of the queen conch, Strombus gigas. Bull. Mar Sci. Gulf Carib. 14:246-295. Ray, M. & A. W. Stoner. 1995. Growth, survivorship, and habitat choice in a newly settled seagrass gastropod, Strombus gigas. Mar. Ecol. Prog. Sen 123:83-94. Ray, M.. A. W. Stoner & S. M. O'Connell. 1994. Size-specific predation of juvenile queen conch Strombus gigas: implications for stock enhance- ment. Aqi/aa/Zmrf 128:79-88. Ray-Culp, M., M. Davis & A. W. Stoner. 1997. The micropredators of settling and newly settled queen conch (Strombus gigas Linnaeus). / Shellfish Res. 16:423-428. Roughgarden. J., S. Gaines & H. Possingham. 1988. Recruitment dynamics in complex life cycles. Science 241:1460-1466. Stoner. A. W. 1997. Status of queen conch (Strombus gigas) research in the Caribbean. Mar Fish. Rev. 59:14-22. Stoner. A. W. 1998. Shell middens as indicators of long-term distributional pattern in Strombus gigas, a heavily exploited marine gastropod. Bull. Mar Sci. 61:559-570. Stoner, A. W. & M. Davis. 1997a. Abundance and distribution of queen conch veligers (Strombus gigas Linne) in the central Bahamas. I. hori- zontal patterns in relation to reproductive and nursery grounds. / Shell- fish Res. 16:7-18. Stoner. A. W. & M. Davis. 1997b. Abundance and distribution of queen conch veligers (Strombus gigas Linne) in the central Bahamas: 2. ver- tical patterns in nearshore and deep-water habitats. J. Shellfish Res. 16:19-29. Stoner. A. W. & R. A. Glazer. 1998. Variation in natural mortality: impli- cations for queen conch stock enhancement. Bull. Mar. Sci. 62:427- 442. Stoner, A. W., R. A. Glazer & P. J. Barile. 1996a. Larval supply to queen conch nurseries: relationships with recruitment process and population size in Florida and the Bahamas. J. Shellfish Res. 15:407-420. Stoner. A. W., M. D. Hanisak. N. P. Smith & R. A. Armstrong. 1994. Large-scale distribution of queen conch nursery habitats: implications for stock enhancement, pp. 169-189. In: R. S. Appeldoorn and B. Rodriguez (eds.). Queen Conch Biology Fisheries and Mariculture. Fundacion CientiTica Los Roques. Caracas, Venezuela. Stoner, A. W., J. Lin & M. D. Hanisak. 1995. Relationships between sea- grass bed characteristics and juvenile queen conch (Strombus gigas Linne) abundance in the Bahamas. J. Shellfish Res. 14:315-323. Stoner. A. W., N. Mehta & T. N. Lee. 1997c. Recruitment of Strombus veligers to the Florida Keys reef tract: relation to hydrographic events. J. Shellfish Res. 16:1-6. Stoner, A. W.. P. A. Pitts & R. A. Armstrong. 1996b. Interaction of physi- cal and biological factors in the large-scale distribution of juvenile queen conch in seagrass meadows. Bull. Mar. Sci. 58:217-233. Stoner, A. W. & M. Ray. 1993. Aggregation dynamics in juvenile queen conch (Strombus gigas): population structure, mortality, growth, and migration. Mar Biol. 116:571-582. Stoner. A. W. & M. Ray. 1996. Queen conch, Strombus gigas. in fished and unfished locations of the Bahamas: effects of a marine fishery reserve on adults, juveniles, and larval production. Fish. Bull. U.S. 94:551-565. Stoner, A, W., M. Ray, R. A. Glazer & K. J. McCarthy. 1996c. Metamor- phic responses to natural substrata in a gastropod larva: decisions re- lated to postlarval growth and habitat preference. J. Exp. Mar. Biol. Ecol. 205:229-243. Stoner, A. W.. M. Ray-Culp & S. M. O'Connell. 1998. Settlement and recruitment of queen conch (Str(tmhus gigas) in seagrass meadows: ■ associations with habitat and micropredators. Fish. Bull. 96:885-899. Stoner. A. W. & V. J. Sandt. 1991. Experimental analysis of habitat quality for juvenile queen conch in seagrass meadows. Fish. Bull. U.S. 89: 693-700. Stoner. A. W. & V. J. Sandt. 1992. Population structure, seasonal move- Distribution of Queen Conch 969 ments, and feeding of queen conch. Stromhus gigas, in deep-water habitats of the Bahamas. Bull. Mar. Sci. 51:287-300. Stoner, A. W.. V. J. Sandt & I. F. Boidron-Metairon. 1992. Seasonality in reproductive activity and larval abundance of queen conch Stromhus gigas. Fish. Bull. U.S. 90:161-170. Stoner, A. W. & K. C. Schwarte. 1994. Queen conch. Strnnihus gigas. reproductive stocks in the central Bahamas: distribution and probable sources. Fish. Bull. U.S. 92:171-179. Stoner. A. W. & J. M. Waite. 1990. Distribution and behavior of queen conch. Stromhus gigas. relative to seagrass standing crop. Fish. Bull. U.S. 88:573-585. Wahle, R. A. & L. S. Incze. 1997. Pre- and postsettlement processes in recruitment of the American lobster. / £v/). Mar. Biol. Ecol. 217:179- 207. Woodin. S. A. 1976. Adult-larval interactions in dense infaunal assem- blages: patterns of abundance. J. Mar. Res. 34:25—41. Yund. P. O.. S. D. Gaines & M. D. Bertness. 1991. Cylindrical tube traps for larval sampling. Liinnol. Oceaiwgr. 36:1 167-1 177. Joiinud of Shellfisli Research. VoL 17, No. 4. 971-477. 199X. ALLOZYME AND MORPHOLOGICAL EVIDENCE SUPPORTING THE SEPARATION OF BABYLONIA FORMOSAE FORMOSAE FROM B. FORMOSAE HABEI AT SPECIFIC LEVEL (PROSOBRANCHIA: BUCCINIDAE) LI-LIAN LIU AND VUH-WEN CHIU Institule of Murine Biology National Sun Yat-Sen Universit}- Kaohsiung, Taiwan 804, Republic of China ABSTRACT Bubylonui Jormoiue is a common specie;, on the west coast of Taiwan. From its shell color pattern and shape, two subspecies have been classified: B. fonnome formosae (Sowerby 1856) and B. formosae habei (Altena and Gittenberger 1981). Recently, both subspecies were collected from the same area, which was not in accordance with previous recorded harvests. Therefore, the present study was undertaken to evaluate the validity of the subspecies. B. areolata occurring in the same area was also examined for comparati\'e purpose. Samples of B. formosae fiirmostu: B. formosae habei. and B. areolalu were collected between September 1990 and March 1991. Si.\ shell characters (shell length, shell width, spire length, apertural length, fasciole ridge width, and shoulder width), radulae. and allozymes were analyzed. The shoulder width of shell could separate species with B. formosae habei > B. areohila > B. formosae formosae. Fixed allelic differences were observed at loci of ark. gol-\. mpi. and pi^m between B. formosae and B. areolala, and at loci of ark. got-\. and mpi between B. formosae formosae and B. formosae habei. Nei's genetic distances (D) were 0.25 for B. formosae formosae vs. B. formosae habei. 0.35 for B. formosae formosae vs. B. areolala. and 0.37 for B. formosae habei vs. B. areolata. The mean heterozygosity among populations were low in B. areolata (Ho = to 0.06). B. formosae formosae (Ho = 0.07 - 0.08). and B. formosae habei (Ho = 0.06 to 0.07). All the above results indicated that the two subspecies deserve to be recognized as full species: B. formosae and B. habei. KEY WORDS: Babylonia, ivory snail, allozyme. shell, radula INTRODUCTION Babylonia formosae (Sowerby 1866) belongs to the family of Buccinidae. Its distribution was limited to the west coast of Taiwan (Altena and Gittenberger 1981). Within this small region, two subspecies have been classified: B. formosae formosae (Sowerby 1866) and B. formosae habei (Altena and Gittenberger 1981 ) (see Table 1). According to Altena and Gittenberger ( 1981 ). B. formo- sae habei was found only in the northeast coast of Taiwan. How- ever, our investigation in late 1990 revealed that B. formosae for- mosae and B. formosae habei both were commonly caught on the southwest coast of Taiwan. Moreover, Lan (1990) mentioned that B. formosae habei sold in Taiwan may be imported from China. Ke and Li (1991) and (Ke and Li 1992) also reported that B. formosae habei is a commercially important shellfish on the south- east coast of China. Meanwhile, studies of reproduction found that B. formosae habei spawns between June to September (Ke and Li TABLE 1, Diagnostic Subspecies Characters of Babylonia formosae Based on Altena and Gittenberger (1981). Diagnostic Characteristics B. formosae formosae B. formosae habei Last body whorl Shoulder on the last body whorl Suture on the last body whorl Spots on the last body whorl Color pattern of the spots Distribution Evenly rounded Narrower Less evenly rounded Narrow Narrow Distinct Bright Narrower Less distinct Dull Northwest to southwest Northeast coast of coast of Taiwan Taiwan 1991) (Ke and Li 1992): whereas, B. formosae formosae spawns between October to January (Chiu and Liu 1994). Although the two subspecies have been studied since 1991 by Ke and Li ( 1991 ) (Ke and Li 1992) and Chiu & Liu (1994), it is still difficult to evaluate the systematic status of the two subspecies at this mo- ment. Hence, the present paper studied the systematic status of the two subspecies using morphological, radula, and allozyme char- acters. B. areolata (Link 1807) was also studied for comparative purpose. B. areolata is the most common Babylonia species in Taiwan. Its distribution is from Ceylon and the Nicobar Islands through the Gulf of Siam, along the Vietnamese and Chinese coasts to Taiwan (Altena and Gittenberger 1981 ). Both B. formosae and B. areolata live in sandy or muddy subtidal areas and can be caught by either bottom trawling or in baited baskets at depths of 15 to 50 m (Lai 1987). MATERIALS AND METHODS Samples of S. areolata. B. formosae formosae. and B. fonnosae habei were collected from coastal waters of Taiwan (Fig. 1) be- tween September 1990 and March 1991. Collected snails were stored at -70°C for later use. Voucher specimens were deposited in the Molluscan Collection. University of Colorado Museum (UCM), with the catalog numbers UCM 37665 for B. areolata. 37666 for B. formosae formosae, and 37667 for B. formosae habei. Species were identified preliminarily using the color-pattern cri- teria (Altena and Gittenberger 1981). Species of B, areolata has three rows of large reddish brown squarish spots on the last body whorl (Fig. 2a,b). B. formosae has four rows of violet-brown spots on the last body whorl. The spots clearly contrast with the light background, and the color is brighter in B. formosae formosae (Fig. 2c,d) than in B. formosae habei (Fig. 2e,f). Shell characters (shell length, shell width, spire length, aper- tural length, fasciole ridge width, and shoulder width) (Fig. 3) and total wet weight of individual snails were determined. Duncan's 971 972 Liu and Chiu CHINA 118 119" 120 121 2S2 24 2# 22 W 122"E Figure I. Sampling localities of Babylonia. KH: Kaohsiung: TK: Tungkang; PH: Penghu; CC: Chungchou; Pt: Pyngtan. multiple comparison tests were used for data analysis (SAS Insti- tute. Inc. 1985). For radula examination, six buccal masses of each species were put in a lOVr KOH solution overnight to resolve muscle and con- nective tissue around the radula. Each radula was then rinsed in distilled water and ultrasonically cleaned for 30 seconds. The |lill|lll!l!lllllli;|ii!'|iiil!li|i|illlji;'ipli!'2!l!. :..'."': '' ^' ^' ' ''' ^' ^' ^ "" -^ Figure 3. Measurements of the shell of Baftv/oma. SH: shell height; Figure 2. SheWs o( Babylonia areolata {a, b), B. formosae foniiosae ic, SW: shell width: SPL: spire length: AL: apertural length; SHW: d), and B. formosae liabei (e, f). shoulder width: FW: fasciole ridge width. Separation of B. Formosae Formosae from B. Formosae Habei 973 TABLE 2. Measurements of Shell Characteristics of Babylonia Species (Mean ± SD). Species/ Shell Shell Spire Apcrtural Fasciole Shoulder Total Sample Length Width Length Length Ridge Width Weight Location n (mnil (mm) (mm) (mm) Width (mm) (mm) (g) Bahxlonict ureoUtttt KH 30 54.7 ± 3.3 B 33.0 ± 1.4 B 29.0 ± 2.2 B 32.6+ 1.2 B 4.7 ±0.4 A 3.0 ± 0.5 C 29.2 ± 3.7 B TK 28 55.3 ± 16.2 3 32.2 ± 8.6 B 29.7 ± 9.6 B 30.6 ± 8. 1 C 4.4 ± 1.2 B 3.3 ± 1.1 C 31.7 ± 19.3 B PH 30 42.9 ± 8.5 C 26.8 ± 4.9 C 21.9 ± 5.3 D 26.5 ± 4.6 D 3.8 ± 0.8 C 3.0 ± 0.6 C 17.7 ± 9.6 C FT 28 65.2 ± 7.7 A 38.2 ± 3.6 A 35.3 ± 4.8 A 37.1 ±3.4 A 5.0 ± 0.7 A 3.7 ± 0.6 B 46.3 ± 11.9 A B. Itinnositf formosae KH 55 45.0 ± 6.0 C 26,1 ±2.9 CD 24.9 + 3.7 C 24.7 ± 2.9 E 3.6 ± 0.5 CD 2.3 ± 0.5 D 15.9 ±5,5 CD TK 50 41.6 ±6.2 CD 24.6 ± 3.3 D 22.3 ± 4.0 D 23.2 ± 3.0 EF 3.3 + 0.6 E 2.0 ± (15 E 13.0 ±5.1 DE B. formostif habei CC 66 43.0 ± 5.3 C 25.5 + 3.2 CD 22.2 ± 3.2 D 24.5 ± 3.0 E 3.9 + 0.6 C 4.3 ± 0.8 A 14.4 ± 5.7 CDE PT 30 38.7 ±4. ID 22.3 ± 2.5 E 20.0 ± 2.4 D 22.1 ± 2.3 F 3.4 ± 0.6 DE 4.1 ±0.5 A 10.2 ± 3.5 E „. .-„.^„i., .,:^., T^ '^ n: sample size. Duncan's test for significant variation is indicated by capital letters (p < .05) cleaned radulae were dried and mounted on SEM stubs. These specimen were then gold-coated and examined v. ith a Hitachi 450 scanning electron microscope at 15 KV. For allozyme studies, foot tissue (0.2 to 0.5 g) was taken and homogenized in a Tekmar tissuniizer with an equal volume of 10 mM Tris-HCl buffer (pH 7.0) containing 1% Triton X-100. Ho- mogenates were centrifuged at 5.000 g for 10 min. and the super- nates were stored at -70 "C. Horizontal starch-gel electrophoresis with buffer systems Tris-citrate pH 8.0. Tris-citrate pH 6.3/6.7. Tris-maleate-EDTA pH 7.4. and lithium hydroxide pH 8.1/8.3 was used. Enzyme-staining methods followed Richardson et al. (1986) and Murphy et al. (1990). Multiple loci encoding the same enzyme (isozymes) were des- ignated by consecutive numbers, with I denoting the slowest mi- grating isozyme. Twelve enzyme loci were scored: arginine kinase (ark, EC 2.7.3.3); esterase (est-\.2. EC 3.1.1.1); glutamate- oxaloacetate transaminase (g«r-l,2, EC 2.6.1.1); isocitrate dehy- drogenase (idh. EC 1.1.1.42); maiate dehydrogenase [mdh. EC 1. 1. 1. 37); mannose-6-phosphate isomerase (mpi, EC 5.3.1.8); oc- topine dehydrogenase {opdh. EC 1.5.1.11); 6-phosphogluconate dehydrogenase [b-pgdh. EC 1.1.1.44); phosphoglucomutase (pgm. EC 2.7.5.1); and sorbitol dehydrogenase (,sy//p. 1.1.1.14). Alleles at each locus were scored by designating the most common allele of B. areohita as 100. All other alleles were numbered according to their relative anodal distance from the reference allele. Chi-square goodness-of-fit tests were computed to determine if there were significant deviations from Hardy-Weinberg equilibrium between observed and expected heterozygote genotype frequencies at each locus (Nei 1978). The mean observed and expected heterozygosity in each population was also calculated (Nei 1978). Nei's genetic distance coefficients (D) were calculated and clustered by the un- weighted pair-group method with arithtnetic means (UPGMA) al- gorithm (Sneath and Sokal 1973). These analyses were performed with BIOSYS-I (Swofford and Selander 1989). Figure 4. Micrographs of the radula of Babylonia: (a) male B. areolala; (b) female B. areolala; (c) female B. formosae formosae: and (d) female B. formosae habei (scale bar = 500 pm). 974 Liu and Chiu TABLE 3. Allele Frequencies of Babylonia Species. Species/ B. fonnosae B. fonnosae Allele B. areolata formosai habei Population KH TK PH PT KH TK CC PT ARK (n) 30 28 30 27 55 50 68 30 189 0.118 0.100 156 0.875 0.900 100 0.917 1 0.950 1 0.007 89 0.083 0.050 56 0.982 0.900 40 0.018 0.100 Ho 0.100 0.100 0.036 0.040 0.221 0.200 He 0.155 * 0.097 0.036 0.182 ** 0.222 0.183 EST-1 (n) 30 28 30 27 55 50 68 30 100 1 1 1 1 1 0.990 1 1 94 0.010 Ho 0.020 He 0.020 EST- 2 (n) 29 28 30 27 55 48 67 30 111 0.017 0.031 0.015 100 0.983 0.982 0.766 0.981 0.991 0.948 0.948 1 89 0.017 0.018 0.217 0.019 0.009 0.021 0.037 Ho 0.003 0.004 0.433 0.037 0.018 0.063 0.075 He 0.003 0.004 0.371 0.037 0.018 O.IOI ** 0.101 ** GOT-1 (n) 30 28 30 27 55 50 68 30 176 0.091 0.070 151 0.609 0.530 128 0.027 0.030 0.956 0.917 112 0.218 0.360 0.044 0.083 100 0.983 1 0.950 0.963 0.055 0.010 94 0.017 0.050 0.037 Ho 0.003 0.100 0.418 0.400 0.088 0.167 He 0.003 0.097 0.073 ** 0.575 ** 0.589 0.085 0.155 GOT-2 (n) 30 28 30 27 55 50 66 30 160 0.030 0.017 100 1 1 1 0.944 1 1 0.955 0.983 40 0.056 0.015 Ho 0.111 0.091 0.033 He 0.107 0.088 0.033 IDH (111 30 28 30 27 55 50 68 30 100 1 1 1 1 1 1 1 1 MDH (n) 30 28 30 27 55 50 68 30 115 0.082 0.060 100 1 1 1 1 0.918 0.940 1 1 Ho 0.091 0.040 He . 0.152 ** 0.114 MPI (n) 30 28 30 27 55 50 68 30 100 1 1 1 1 continued on next page Separation of B. Formosae Formosae from B. Formosae Habei 975 TABLE 3. continued Species/ B. formosae B. formosae Allele B. areolata formosae hahei 93 1 1 76 1 1 OPDH (n) 30 28 30 27 55 50 68 30 100 1 1 1 1 1 1 1 1 6-PGDH (n) 29 28 30 27 55 50 68 30 ISO 0.034 0.015 100 0.966 1 1 1 1 1 0.978 1 67 0.007 Ho 0.007 0.044 He 0.007 0.044 PGM (n) 30 28 30 27 55 50 68 30 327 0.091 0.240 0.176 0.116 303 0.017 0.019 0.809 0.690 0.706 0.800 252 0.091 0.060 0.118 0.067 170 0.033 0.033 0.019 0.010 0.017 100 0.950 1 0.950 0.962 0.009 79 0.017 Ho 0.100 0.100 0.074 0.291 0.400 0.206 0.333 He 0.098 0.098 0.073 0.332 0.467 ** 0.460 ** 0.347 SDH (nl 30 28 30 27 55 50 68 30 100 1 1 1 1 1 1 1 1 (n; sample size. Ho: observed heterozygosity. He: expected heterozygosity. Significant deviation from Hardy- Weinberg proportion at p = .05 and p = .01. respectively). RESULTS Quantitative measurements of shell characters are indicated in Table 2. Shell lengths of the examined species were from 39 to 65 mm. Shoulder width was the onlv character shown significant variation among species: B. formosae habei > B. areolata > B. formosae fonnosae. respectively. No significant difference in the radulae was found among the Babylonia species and between radulae of male and female B. areolata (Fig. 4). The radulae belong to the rachiglossan type. The TABLE 4. Summary of Genetic Variation in Babylonia Species. Species and Mean Number Allelles per Percentage of Polymorphic Mean Heterozygosity Population Number of Loci Locus ± SE Loci* Observed Mean ± SE Expected Mean ± SE B. areolata KH 12 1 .5 ± 0.2 16.7 0.028 ±0.012 0.032 ±0.015 TK 12 1.1 ±0.1 0.0 0.003 ± 0.003 0.003 ± 0.003 PH 12 1.5 ±0.2 33.3 0.061 ±0.036 0.055 ± 0.03 1 PT 12 1.4 ±0.2 8.3 0.019 ±0.01 1 0.024 ±0.011 B. fonnosae formosae KH 12 1.8 + 0.4 25.0 0.071 ±0.040 0.093 ± 0.052 TK 12 2.0 ± 0.4 41.7 0.080 ± 0.044 0.123 ±0.058 B. formosae hahei CC 12 1.9 ±0.3 25.0 0.060 ± 0.023 0.083 ± 0.039 PT i: 1.6 ±0.3 25.0 0.067 ± 0.036 0.068 ± 0.039 * a locus is considered polymorphic if the frequency of the most common allele does not exceed 0.95. 976 Liu and Chiu B. areolata \ B. formosae formosae B. formosae habei . {\K (Tungkang) - PT (Pyngtan) KH (Kaohsiung) PH (Penghu) KH (Kaohsiung) TK (Tungkang) CC (Chungchou) PT (Pyngtan) 0.4 0,3 0.2 0.1 Nei's genetic distance Figure 5. The UPGMA cluster analysis of Nei's (1978) unbiased ge- netic distance (D) among Babylonia species. central tooth has five cusps. The lateral teeth have two curved cusps: the inner cusp short; and the outer cusp is long. Among the 12 loci examined, using 0.95 as the criterion for polymorphism. si.\ were polymorphic: ark, est-2, got-1.2. mdh. and pgm. Detailed allelic frequencies of Babylonia species are shown in Table 3. Fixed allelic differences were observed at ark, goi-\. nipi. and pgm between B. formosae and B. areolata and at ark, got- 1 . and mpi between B. formosae formosae and B. formosae habei. Heterozygote deficiencies among populations and species were found in all the polymorphic loci. Mean heterozygosities among populations of B. formosae formosae and B. formosae ha- bei varied from 0.060 to 0.080. which were higher than the popu- lations of B. areolata (0.003 to 0.061 ) (Table 4). Nei's genetic distance (D) between B. formosae formosae and B. formosae habei was 0.25. Comparison with other marine inver- tebrate genetic distance values, the difference between B. formosae formosae and B. formosae habei could be interpreted at a specific rather than a subspecific level. In addition. B. areolata was sepa- rated from B. fonnosae formosae and B. formosae habei at the distances of 0.35 and 0.37. An UPGMA cluster phenogram is shown in Fig. 5. Only minor differentiation existed among the populations in each of the three species (D < 0.0002 to 0.0030). DISCUSSION Our results indicated the shoulder width differed among spe- cies: B. formosae habei > B. areolata > B. formosae formosae. respectively. The fixed allelic difference between B. formosae for- mosae and B. formosae habei was observed in three of the 12 examined loci. The Nei's genetic distance between B. formosae formosae and B. formosae habei was 0.25. These allozyme differ- ences are well above the specific level (Thorpe 1983; Richardson et al. 1986). Therefore, they should be recognized as two full species: B. formosae and B. habei. By using the allozyme electrophoretic technique, several gas- tropod species previously considered to be polytypic or subspecies are actually separate species: Oncomelania hupensis hupensis and O. hupensis quadrasi (with Nei's genetic distance [D] = 0.62) (Woodruff et al. 1988). Stramonita haemastoma canalictdata. and S. haemastoma floridana (with D = 0.28) (Liu et al. 1991). Nii- cella emarginata complex (with D = 0.16) (Palmer et al. 1990), Crepidula conve.xa complex (with D = 0.76) and C. plana com- plex (with D = 0.39) (Hoagland 1984). Although no simple re- lationship exists between genetic distance and taxonomic level, Thorpe (1983) found the Nei's genetic distances range from 0.19 to 2.59 for >95% of the congeneric invertebrates. Richardson et al. ( 1986) also suggested that fixed allelic difference can be diagnos- tic in separating species if the loci with fixed allelic differences are >209'f of the examined loci. In the present study, Nei's genetic distance (D = 0.25) and the level of fixed allelic differences (257^) all indicated that the difference between B. formosae for- mosae and B. formosae habei is on the specific level. According to the records of Altena and Gittenberger ( 1981 ), S. formosae habei are distributed on the northeast coast of Taiwan; however, we were unable to locate them. From a reliable record indicating the natural distribution of B. fonnosae formosae is on the southwest coast of Taiwan and that of B. fonnosae habei is on the southeast coa.st of China (Ke and Li 1991) (Ke and Li 1992). The occurrence of B. formosae habei in southern Taiwan was observed in 1987; since then it has been a very coinmon shellfish in fish markets. It has become very rare after 1996. with one or two individuals mixed in hundreds of B. areolata. Because of this unusual change in abundance, we suspected that B. formosae habei might have been introduced from China through fisheries opera- tions and is not a resident species in southern Taiwan. It is also speculated that range expansion could be caused by a temporal change in hydrographic condition. Introduction of an exotic species either by an accident or com- mercial purpose is quite common in Taiwan. For example, Perna viridis was believed to be imported by ship industry (Lai 1987). A South America freshwater apple snail, Ampullarius insularus was imported from Argentina for commercial use in 1980 (Chang 1985). Both species are now widespread in Taiwan. In addition, marine bivalve Mytilopsis sallei. freshwater snail Pila leopordvil- lensis. and land snails Achatina fnlica and Bradybaena similaris are also introduced species in Taiwan. The presence of B. formo- sae habei in Taiwan might be another case of an importation from southeast coast of China. The reproductive season of B. fonnosae formosae is known to be from October to January, with average egg diameter 0.55 mm (Chiu and Liu 1994); whereas, B. formosae habei spawns from June to September with average egg diameter 0.26 mm (Ke and Li 1991) (Ke and Li 1992). The colors of the female sperm-ingesting gland are also different, being dark brown and brown, in B. for- mosae formosae and B. fonnosae habei. respectively (personal observation). Although the breeding compatibility of these two subspecies is unknown, the differences in allozyme patterns indi- cated they are different and should be elevated to full species level; that is, B. formosae and B. habei. ACKNOWLEDGMENTS We thank Dr. H. Y. Chen for help with collecting ivory snails. We also thank Dr. C. C. Lu and S. K. Wu for their constructive comments on this manuscript. LITERATURE CITED Altena. C. O. V. R. & E. Gittenberger. 1981. The genus Buhykmia (Proso- branchia. Buccinidae). Zoologische Verhandelinfien 188:57 pp. Chang. W. C. 1985. The ecological studies on the Ampullaria snail (Cy- clophoracea: Ampullaridae). Bull. M(daail.. ROC. 11:43-52. Chiu. Y. W. & L. L. Liu. 1994. Copulation and egg-laying behaviors in the ivory shell. Babylonia formosue formosae (Neogastropoda: Buc- cinidae). Venus 53:49-55. Hedgecock, D. 1986. Is gene flow from pelagic larval dispersal important Separation of B. Formosae Formosae from B. Formosae Habei 977 in the adaptation and evolution of marine invertebrates'' Bull. Mm: Scl. 39:550-564. Hoagland. K. E. 1984. Use of molecular genetics to distinguish species of the gastropod genus Crepiduki (Prosobranchia: Calyptraeidae). Mala- cologia 25:607-628. Ke. C. H. & F. X. Li. 1991. Histology of gonad and reproductive cycle of Babylonia formosae. J. Oceanog. Taiwan Strait 10:213-220 (in Chi- nese). Ke, C. H. & F. X. Li. 1992. Ultrastructural studies on spermatogenesis and sperm morphology of Babylonia formosae (Sowerby) (Gastropoda). Acta Zool. Sinica. China 38:233-238 (in Chinese). Lai, K. Y. 1987. Marine Gastropods of Taiwan (2). Published by Taiwan Museum, Taipei, ROC. pp. 78-79 (in Chinese). Lan, T. C. 1990. Savon,- Mollusca. Published by Taiwan Museum. Taipei, ROC, pp. 31-32 (in Chinese). Liu, L. L., D. W. Foltz & W. B. Stickle. 1991. Genetic population structure of the southern oyster drill Stramonita { = Thais) haemostomu. Mar. Biol. 111:71-79. Murphy, R. W., J. W. Sites, D. G. Buth & C. H. Haulier. 1990. Proteins. L isozyme electrophoresis, pp. 45-126. In: D. M. Hillis and C. Moritz (eds.). Molecular Systematics. Sinauer, Sunderland, Massachusetts. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Palmer. A. R., S. D. Gayron & D. S. Woodruff. 1990. Reproductive, mor- phological, and genetic evidence for two cryptic species of northeastern Pacific Nucella. Veliger 33:325-338. Richardson, B. J., P. R. Baverstock & M. Adams. 1986. Allozyme elec- trophoresis: a handbook for animal systematics and population studies. Academic Press, San Diego. 410 pp. SAS Institute. Inc. 1985, SAS user's guide: statistics, version 5. SAS Institute, Inc., Cary, North Carolina. Sneath. P. H. A. & R. R. Sokal. 1973. Numerical taxonomy — the prin- ciples and practice of numerical classification. W. H. Freeman & Co., San Francisco. Swofford, D, L. & R, K. Selander, 1989. Biosys-1: a Fortran program for the analysis of allelic variation in population genetics and biochemical systematics. University of Illinois Press, Champaign. Illinois. Thorpe, J. P. 1983. Enzyme variation, genetic distance, and evolutionary divergence in relation to levels of taxonomic separation, pp. 131-152. In: G. S. Oxford and D. Rollinson (eds.). Protein polymorphism: Ad- aptative and Taxonomic Significance. Academic Press. San Diego. Woodniff, D. S., K. C. Staub, E. S. Upatham, V. Viyanant & H. C. Yuan. 1988. Genetic variation in Oncomelania hupensis: Schistosoma japoni- ciim transmitting snails in China and the Philippines are distinct spe- cies. Malacologia 29:347-361. Journal of Slwllfish Research. Vol, 17. No. 4, 474-98^, 1998. CHARACTERIZATION OF THE DIGESTIVE TRACT OF GREENLIP ABALONE, HALIOTIS LAEVIGATA DONOVAN. I. MORPHOLOGY AND HISTOLOGY JAMES O. HARRIS, CHRISTOPHER M. BURKE, AND GREG B. MAGUIRE School of AijiuiciiltiiiT University of Tasmania Launceston. Tasmania. Australia. 7250 ABSTRACT Australasian abalone such as the greenlip abalone, Haliulis laevigata, prefer a diet ol red algae (Rhodophyta); whereas, abalone from elsewhere more commonly prefer brown algae (Phaeophyta). Because of this feeding preference, the structure of the digestive tract of A/, laevigata was investigated using histological and scanning electron microscopy (SEM) techniques. The digestive tract of both starved and fed adult H. laevigata revealed the presence of ciliated, mucus, and secretory cells throughout the digestive tract. The esophagus contained secretory, ciliated, and large mucous cells, with fragmentation spherules also present. The crop extended from the esophagus to the stomach. It was surrounded by thin muscularis and consisted mainly of secretory cells, although some phagocytes were present. The stomach possessed mainly secretory cells, although some ciliated cells, mucous cells, and phagocytes were present. The style sac differed from the stomach, having more ciliated cells. In mtestinal regions I to III, the epithelium was shorter than in previous regions. Few cilia were present on the ridges, although many were observed in the gutters. Intestinal regions IV to V contained more mucous cells than intestine III. and more bactena were observed associated with the fecal string than in other regions. The low incidence of bacterial association with the gut epithelium was attributed to the occurrence and number of mucous cells, common throughout the digestive tract. Spherical bodies present in the lumen are believed to be fragmentation spherules involved in waste removal and enzyme release. Starved abalone contained fewer mucous cells in the esophagus, had less pronounced staining reactions in the stomach, contained large amounts of granular inclusions in the style sac. and had fewer phagocytes in the intestines. KEY WORDS: Abalone. Haliotis laevigata, histology, digestive tract, starvation INTRODUCTION Abalone are herbivorous archaeogastropods whose diet con- sists mainly of macroalgae, although diatoms and some detritus, including sand, are also ingested (Campbell 1965, Garland et al. 1985). Australasian abalone prefer red algae (Rhodophyta) and will consume brown algae (Phaeophyta) only when preferred spe- cies are less common (Poore 1972, .Shepherd 1973, Shepherd and Steinberg 1992). Brown algae are digested much more slowly (Foale and Day 1992). Abalone from elsewhere more commonly prefer brown algae (Shepherd and Steinberg 1992), suggesting possible differences in digestive strategy. Several enzymatic stud- ies of abalone from the northern hemisphere are available, but few data are available for Australasian species (Clark and Jowett 1978). Feeding preference differences for Australasian species have been attributed to algal toughness (McShane et al. 1994) or phenolic content (Shepherd and Steinberg 1992), although Shep- herd and Steinberg ( 1992) considered that feeding preferences are primarily attributable to the selective nature of environments on available algae. In such primitive gastropods as abalone, digestion begins with extracellular digestion followed by phagocytosis, or cellular up- take of particles, in the digestive gland and ingestion by mobile amoebocytes in some cases (Owen 1966). Intracellular digestion occurs within the duct cells of the digestive gland (Fretter and Graham 1962). Secretory cells occur throughout the digestive tract of Haliotis spp. They have been documented in the buccal cavity, crop, stomach, cecum, digestive gland, style sac, and intestine by Crofts (1929). Other locations for secretory cells include the buc- cal pouches (Fretter and Graham 1962) and esophagus (Beve- lander 1988). A wide variety of digestive enzymes have been identified from the gut of abalone with the carbohydrates fucoidan, carboxymeth- ylcellulose (CMC), and algin being among the most common sub- strates used for detecting enzyme activity (Duffas and Duffas 1968, Elyakova et al. 1981, Yamaguchi et al. 1989, Boyen et al, 1990). Protease, alginase, and amylase activity also appear in the crop fluid o'i Haliotis rufescens Swainson (McLean 1970). The nature of the digestive tract of Haliotis spp. was examined by Crofts (1929) (drawings for Haliotis tiiberculata Linnaeus). Campbell (1965) (drawings for Haliotis cracheroclii Leach), and Bevelander ( 1988) (a photographic study for H. tiiberculata). Ex- amination of the greenlip abalone digestive tract, by histology and scanning electron microscopy (SEM), strengthens our knowledge of normal abalone gut structure, particularly for species that prefer red algae. It also complements current gut physiology research (Hanis et al. 1998a). In addition, it facilitates the detection of dietary or toxicant-induced alterations in structure (Harris et al. 1998b). Contributions to host animals from digestive tract bacteria can come from either resident or transient populations. Alimentary tracts offer many habitats conducive to microbial activities, such as fermentation of complex organic molecules, the products of which can be used by the host. Intestinal surfaces are often colonized by bacteria, which then make up the autochthonous flora of the host (Savage 198.^), and which can contribute to the nutrition of the host (McBee 1971). Transient bacteria are ingested with, or as, food and encounter both physical or biological events during pas- sage that protect the resident populations from displacement (Or- pin and Anderson 1988). Scanning electron microscopy has been used on the oysters Crassostrea virginica Gmelin and Crassostrea gigas Thunberg to examine bacterial associations (Tall and Nau- man 1981. Garland et al. 1982a) and has revealed physical attach- ment by resident microbes to internal digestive surfaces of other invertebrates (Harris 1993. Jolly et al. 1993). In this study, the gut structure of greenlip abalone is charac- terized using histological and/or SEM techniques for fed and starved individuals. Because the digestive gland is the most com- 979 980 Harris et al. prehensively studied organ involved in digestion in abalone (Campbell 1965. McLean 1970. Bevelander 1988) and is known to be free of bacteria (Erasmus et al. 1997). it was not considered in this study. Our emphasis was on epithelial function, because this can influence bacterial associations with the gut wall. It comple- ments a study of the gut microenvironment in this species (Harris et al. submitted). MATERIALS AND METHODS Adult specimens of H. laevigata were maintained in a recircu- lating system consisting of 6 x 20 L plastic containers and a biofilter for up to 3 weeks prior to histological sampling. Greenlip abalone were collected from several locations in northern Tasma- nia, 40° to 4I°50'S, 146°50' to I48°50'E, (Petal Point, Foster Islands, Port Sorell, and Flinders Island) and consisted of adults of 135 to 185-mm length. Macroalgae were collected in southern Tasmania, 42°50' to 43°50'S, 147°50' to 148°E (Blackman Bay and Port Arthur). Macroalgae of the genera Polysiphonia sp.. Ulva sp., and other epiphytic algae associated with the macroalgae Am- phibohis spp. were collected by divers and used as food. Epiphytic algae of seagrasses are known to comprise up to 859f of the green- lip abalone diet in wild conditions (Shepherd 1973). The algae were added to the maintenance tank and left for 10 to 14 days. The abalone fed rapidly when Polysiplumia sp. were added, although they were "messy"" feeders. To remove algal debris, tanks were siphoned every second day. A diatom film, which was grazed by the abalone, developed within the tank during the study period. Abalone were removed from the maintenance tank by either a commercial abalone iron, a warm water siphon or a flat plastic spatula with grease. Several abalone were dissected for prelimi- nary investigation of the digestive tract. For histological examina- tion, separate groups of fed and starved animals were used. No macroalgae were added for 10 days to tanks containing starved abalone, allowing sufficient time for physiological changes to oc- cur within the abalone (Carefoot et al. 1993). A scalpel was used to cut the foot as close as possible to the shell without disrupting the mantle and visceral mass. The mantle surrounding the mantle cavity was removed to facilitate access to the digestive tract, taking care not to puncture the rectum (intestine V). Two parallel incisions were made in the integument covering the digestive tract (Fig. 1 ), joined by an additional two longitudinal cuts in the integument to complete a rectangular section. This layer was peeled away from the digestive tract with scalpel, tweezers, and a blunt probe. The intestine (sections 11, III. IV, and V) were teased with a blunt probe from the connective tissue and mem- branes, and excised with scissors and scalpel. Because the oe- sophagus lies under the intestines, it was carefully separated by teasing with a blunt probe. The digestive gland and gonad were removed by scraping them from the surface of the crop and stom- ach. Short lengths of gut (about 10 mm) were removed from the esophagus, the crop, the stomach wall, the style sac, and intestinal regions III and IV and placed into either phosphate-buffered for- malin, Zenker's fluid, or Bouin"s fluid during daylight hours. Samples were fixed for 24 hours at room temperature ( 15 to 18°C) then dehydrated through a graded ethanol series to xylene in a Tissue-Tek II tissue processor. Dehydrated tissue samples were embedded in paraffin resin on a Shandon Histocentre 2 and sec- tioned on a Microm HM 340 microtome at 4 p,m. Sections were oven dried overnight at 37"C. Routine Harris" Haematoxylin and Figure L Schematic diagram of greenlip abalone showing locations for incisions into integument (a.b — lateral cut site; c,d — longitudinal cut site). Eosin (H & E) staining in a Shandon Linistain GLX automatic tissue stainer was carried out on all tissues processed. Five abalone. all of which had been maintained in recirculating aquaria and fed with red, filamentous algae Polysiphonia sp. were used for SEM. Abalone were removed from the shell, and the digestive tract was exposed. To preserve the contents, whole gut sections were removed aseptically. Where necessary, regions of the gut were teased from the integument with a blunt probe and/or forceps. Samples were taken from the esophagus, crop, stomach, style sac, and intestines III and IV. Samples were immediately transferred into 2.5% glutaraldehyde fixative in 0.2 m cacodylate buffer, pH 7.2 containing the major salts present in sea water: 1.6% NaCl, 0.6% MgCU, and CaCU (Garland et al. 1982b). Samples were fixed for 24 h at 4°C then rinsed through three O.IM caco- dylate buffers of decreasing osmolality (Lewis et al. 1985). At this stage, the samples were trimmed and cut open to expose internal surfaces. The samples were then dehydrated through a graded ethanol series (Hodson and Burke 1994). Absolute ethanol and acetone were prepared by storing the commercial grade chemicals over anhydrous copper sulphate (in dialysis tubing). Dehydrated samples were immediately transferred to acetone prior to being critical point dried in liquid CO,. Samples were dried using a Balzers CPD 020 Critical Point Dryer (CPD) apparatus. The samples were mounted onto aluminium SEM stubs with carbon paint and/or double-sided tape. Samples were then kept in a plastic desiccator and stored over CaCL at a vacuum pressure of 25 to 30 psi, to prevent rehydration of the dried samples (Garland et al. 1982b), and sputter coated with gold (Balzers coater) within 24 h. Gut sections were viewed with a Phillips 505 SEM at operating Gastrointestinal Structure of Greenlip Abalone 981 TABLE 1. Cellular types distributed nithin the digestive tract of greenlip abalone. H. laevigata. Gut Region External Color Cell Types Observed Esophagus Pale Mucus, ciliated, and secretory Crop Blue-gray Secretory, phagocytes Stomach Green-gray Secretory, ciliated, mucus, phagocytes, amoebocytes, muscularis Stvie sac Green-gray Secretory, ciliated Intestine 11 Light brown Intestine III Gray-black Ciliated, phagocytes, secretory Intestine TV Brown Mucus, secretory, ciliated Intestine V White voltages between 15 to 20 kV, using llford FP4 120 lllm for micrographs. RESULTS The epithelium of the greenlip abalone digestive tract varied in shape, cellular types, composition, and staining reaction (Table 1 ). The efficacy of fixative type had a marked influence on cellular appearance between and within different regions of the gut. Star- vation caused very minor effects to the epithelia of the esophagus, style sac, and intestines, with little effects noted elsewhere (Table 2), SEM of digestive tract epithelium revealed few bacterial cells, with most in the intestines. What is apparent from this investiga- tion is the widespread occurrence throughout the digestive tract of spherical bodies, 5. to 8-|xm in diameter, corresponding to the fragmentation spherules described by Morton (1953). believed to be involved in waste removal and/or delivery of enzymes to the lumen. Although their size and spherical nature suggested the pos- sibility of these spherules being bacterial, visible evidence of spherules being shed into lumen from the crop (see Fig. 7) and style sac (see Fig. 13) indicates otherwise. A general plan of the digestive tract of H. laevigata was developed from the morpho- logical observations (Fig. 2). Esophagus The esophagus of the greenlip abalone is oriented posterior to the cephalic region. The entrances to the esophageal pouches lie immediately posterior to the salivary glands. The right esophageal pouch is twisted over the esophagus, following the rest of the digestive tract posteriorly, to the region where intestine 111 begins. TABLE 2. Changes in the digesti>e epithelium of starved greenlip abalone, H. laevigata, as indicated by histology. Tissue Starved Esophagus Crop Stomach Style sac Intestine II Intestine III Intestine IV Intestine V Fewer mucous cells on ridges No difference Less pronounced staining Large amounts of granular inclusions Fewer phagocytes No difference More intense staining No difference Figure 2. General layout of greenlip abalone digestive tract, showing (left) digestive tract in situ, and (right) schematic diagram of the di- gestive tract, in dorsal viev* (a = cephalic region; b = right buccal pouch: c = esophagus; d = crop; e = stomach; f = stomach cecum; g = style sac; b = intestine I; i = intestine II; j = intestine III; k = intestine I\ : I = intestine V). The esophagus and crop are equivalent to the postesophageal regions I and II as described in Campbell (1965). The left esophageal pouch is smaller and is located underneath the right esophageal pouch. The midesophagus extends to the posterior end of the esophageal pouches, where it then becomes paler and continues to the crop. This paler section corresponds to the post- esophagus region 1 of H. cracherodii (Campbell 1965). but is referred to as the esophagus in this study. No muscularis was observed in this region. The oesophagus contains mucous cells, secretory cells, and ciliated cells (Fig. 3). Large mucous cells occur on the ridges of the esophagus, among secretory cells containing granules. Granular inclusions occurred in the distal cell tips of both fed and starved animals. Ciliated cells occurred on both the ridges and gutters. Comparison with starved abalone revealed tnany more mucous cells on the ridges of the esophageal cells of fed animals. Within the esophagus, fragmentation spherules were evident among mucus within the lumen (Fig. 4). Neither the gut wall nor ingested food showed bacterial association. Food particles were associated with the mucus (Fig. 5). Figure 3. Esophageal epithelium from fed abalone in histological sec- tion; tissue was fixed in Zenker's fluid; magnification 400 x (a = granu- lar inclusions; b = ciliated epithelium; c = mucous cell). 982 Harris et al. "\. Figure 4. Epithelial surface of esophagus using SEM; Bar = 10 uni (a = fragmentation spherules: b = mucous; c = cilia I. Crop The posterior oesophagus I expands into the crop, which is distinguished by its deep blue-gray color. The crop extends to the most posterior point of the right foot muscle, then narrows and twists 180° into the stomach. Two large folds extend from the crop into the stomach, forming a valve that controls entry of material into the stomach. Thin muscularis surrounded the crop. The epi- thelium of the crop contained mainly columnar secretory cells, although phagocytes were also present (Fig. 6). Observed within the crop were cells in various stages of constriction and fragmen- tation, demonstrating the stages involved in release of fragmenta- tion spherules (Fig. 7). including bulging through the mucus. The cells" surface also had a striated border. Granular inclusions were prevalent toward the distal cell tips in both fed and starved ani- mals. Nuclei were mainly located in the basal half of the cells. Heavily ciliated regions contained both debris and fragmentation spherules (Fig. 8). In the crop, greater cell definition was found with Zenker's fluid than other fixatives. An isolated helical bac- terial cell was observed in this region. Stomach The stomach of H. laevigata follows the general pattern for prosobranch gastropods as described by Morton (1953). A gastric Figure 6. Crop epithelium of slar\ed ahaione in histological section; tissue Has tl.ved in Zenker's fluid: niagniflcalion 4(10 x (a = nucleus of mucus cell: b = non-nucleated fragmentation spherules; c = muscula- ris). .shield is present on the anterior, right side wall. A furrowed ciliary sorting area takes up the floor of the stomach. The stomach cham- ber narrows anteriorly into the heavily ciliated style sac. in which is located the protostyle. The style sac continues to narrow into the intestine. The ducts known to lead to the digestive gland in other species of Haliotis were difficult to locate in H. laevigata, al- though these ducts are presumably present. Muscularis was also found below the epithelial cells. Cells coinprising the stomach epithelium included ciliated cells, mucous cells, secretory cells, and phagocytes (Fig. 9). Secretory cells constituted the majority of the epithelium, with mucous cells occurring consistently. Phago- cytes, although infrequently observed, were present both in basal and distal regions of the cells. Nuclei were located within the basal third of the cells. Amebocytes occurred in the stomach epithelia and were distinguishable from phagocytes by their irregular mor- phology. Some cilia were also visible under the spherules (Fig. 10). Bulging of the secretory cells through the epithelium and cavities in the mucus occurred in the style sac (Fig. 1 1). similar to the observations of the crop. Secretory cells involved in fragmen- tation appear uniform in view from the epithelial surface, but side views of fragmented cells indicate more variety in shape. The club-shaped tips of the cells were observed protruding from the Figure 5. Esophageal epithelium, showing mucopolysaccharide matrix and algal fragments using SEM. Bar = 50 \im (a = algal fragments in mucus). Figure 7. Epithelium of crop wall, showing secretory cells in various stages of constriction and fragmentation using SEM; Bar = 10 ]xm (a = constricting cell; b = fragmented cell; c = cilia). Gastrointestinal Structure of Greenlip Abalone 983 Figure S. Crop epithelium, sliowinj; a renioii where cilia are located. Fragmentation spherules are collected together in a furrow of the epithelium using SEM. Bar = 10 pm (a = fragmentation spherule; b = algal debris; c = cilia). mucus. The style sac has an evenly ciliated, striated epithelium (Fig. 12). although epithelia toward the stomach showed fewer cilia and more secretory cells. Bouin's fluid produced more defi- nition in cellular structure than other fixatives, although formalin- fixed tissue enabled phagocytes to be distinguished more easily. Interestingly, formalin-fixed stomach tissue retained a large layer of mucus that also showed separation from the epithelial surface. In star\ed animals, tissue anterior to the stomach showed large amounts of granular inclusions within the distal third of the cells and in the striated. borders. Nuclei of starved tissue cells also showed a less pronounced staining reaction (Figs. 13). Intestines I to III Intestine I continued from the narrowed style sac, crossed the oesophagus dorsally, into the first of the vertically oriented 180° intestinal twists. The second 180° twist and a change in exterior color to light brown/beige marked the beginning of intestine II, which extended to about the midpoint of the foot. Here the intes- tine abruptly changed exterior color to almost black, indicating the start of intestine III. No muscularis was observed in this region. Figure 10. Epithelial surface of the left stomach wall using SEM. Note the occurrence of secretory cells involved in fragmentation as the dom- inant cell types. Bar = 10 pm (a = fragmentation spherules; b = cilia). The epithelium was most similar to that of the crop. Cells within intestine III were more cuboidal than in intestines I and II. Few ciliated cells were present on the ridges (Fig. 14), though more occurred in the gutters, and were observed supporting fragmenta- tion spherules within the intestine III. Phagocytes were widespread among the cells, although mainly confined to the distal cell tips. Formalin-fixed tissues revealed more mucous cells in ridge regions than other fixatives, as well as phagocytes located in the distal parts of the cells. Few phagocytes were evident within the starved tissue samples (Fig. 15), although several phagocytes occurred in the fed tissue samples. SEM revealed large areas of the epithelium covered by fragmentation spherules, and a layer of mucus. Where this layer of mucus was removed from the intestinal surface, the fragmentation spherules were apparent, indicating mostly secre- tory cells as the dominant cell type (Fig. 16). This removal of mucus may have been an artifact of the glutaraldehyde fixation process, but it shows the arrangement of the cells beneath the mucous coat. Epithelial cells can be seen in cross section where tissue was trimmed after fixation (Fig. 17). These cells were over- laid with mucus, some debris, and fragmentation spherules. On epithelial surfaces, more fragmentation spherules were found, along with rod-shaped bacteria (Fig. 18). Figure 9. Stomach epithelium iif led al)al .05). 992 Harris et al. TABLE 3. Genera, location, and hydrolytic activity ol' bacterial isolates from the digestive tract of H. laevigata. Polymer Degrading No. Activity" Site Bacterial Groups Isolates CMC Agar Starch Esophagus Enterobacteriaceae 6 4 3 ■) Cytophaga 4 4 4 Altcnmionas 1 1 1 Crop Enterobacteriaceae 6 4 3 Aerococcus 2 ■y Stomach Enterobacteriaceae Neisseriaceae 6 1 3 1 I 3 1 Style sac Enterobacteriaceae Neisseriaceae Alleroiuonas 6 1 1 4 I 2 Intestine I] Enterobacteriaceae Alferomonas Listeria 6 1 I 5 1 1 3 1 Intestine IV Enterobacteriaceae 4 -) 1 Cytopluiga 3 3 3 Acinetobacter I 1 1 Aerococcus I Total 51 34 19 13 ■' Number of isolates showing positive hydrolytic activity. DISCUSSION Within the abalone gut. dissolved oxygen levels were below the detection level of this OME and the gut should, therefore, be regarded as anoxic, or at least niicroaerophilic. Low dissolved oxygen levels, similar to those found within the abalone digestive tract, have been reported in other invertebrates. Plante and Jumars (1992) found that even within the digestive tracts of deposit- feeders known to have consumed oxygenated sediment, oxygen levels were similar to animals known to have consumed anoxic sediments. From this. Plante and Jumars ( 1992) proposed that the low dissolved oxygen levels were attributable to biological or chemical processes in the foregut that quickly consumed added oxygen, with the gut contents effectively acting as an oxygen sink. The low oxygen tension and weakly acidic conditions within the digestive tract of H. laevigata provide a selective environment. Prieur et al. ( 1990) reviewed the microbiology of bivalve digestive tracts and noted a higher proportion of fermentative bacteria than in the surrounding seawater. Most of the bacteria isolated from the guts of aquatic invertebrates have been facultative aerobes, al- though obligate aerobes and anaerobes have been reported (Harris 1993). The metabolism of facultative aerobes quickly depletes the available oxygen, thereby creating conditions favorable for anaero- bic fermentation. However, anoxia is an insufficient variable with which to define microbial activity, and fermentation in particular, in an environment such as the abalone gut. because an anoxic environment tnay still have oxidizing conditions. Combined Eh and dissolved oxygen measurements provide a better understand- ing of the inicrobial environment (Plante and Jumars 1992). The pH profile along the greenlip abalone digestive tract is similar to other gastropod and bivalve mollusks. The lowest values recorded are in the stomach of Patella sp. (.'i.S.'i) (Hyinan 1967), Crepidula sp. (6.U0) (Hyman 1967). Buccinum sp. (5.6) (Hytnan 1967). Ostrea ediilis sp. (6.02) (Mathers 1974). and in the style sac oi Mya sp. (4.4) (Owen 1966). although few authors have reported pH levels in the crop of mollusks. The lower pH in the crop and stomach reduces the viscosity of mucus, allowing the gut contents to mix readily. Raising pH increases the viscosity of the mucus in the intestine, helping to consolidate the loosely bound mucus string into cohesive pellets (Morton 1968). Crop contents in H. cra- cherodii are considerably less viscid than in other regions of the gut (Campbell 1965). The only direct measurement of pH within the digestive tract of abalone was described by Gomez-Pinchetti and Garcia-Reina ( 1993). They measured the pH of digestive gland homogenates from Haliati.s coccinea ca)niriensis. and recorded values between 5.5-6.0. This suggests that the crop is the most acidic region in haliotids in general and specifically in H. laevi- gata. This organ is believed to act as a food storage and digestion organ, because both recognizable food pieces up to 3-cm long and unrecognizable food have been found (Campbell 1965). Esopha- geal valves restrict the tnovement of larger food particles from the crop into either the stomach or the stomach cecum (Crofts 1929). Natural seawater has pH values varying between 7.5 to 8.5 (Austin 1988). The decrea.se in pH within the abalone gut portrays an environment that differs from relatively stable, alkaline seawa- ter. The acidic abalone gut provides an environment that would select against organisms unable to tolerate acid pH. The genus Vibrio, for example, is tolerant of mildly alkaline conditions and is generally grown on media of pH 8.6 (Baumann and Schubert 1984); whereas, other marine bacteria such as Alcaligenes are commonly isolated in neutral pH (Kersters and De Ley 1984). The enzyme activity peaks found in other abalone also illustrate the pH changes found throughout the gut of H. laevigata. Diges- tive activity in the esophagus of H. rufescens is highest at pH levels between that of seawater and 6.8 (McLean 1970). Peak alginase activity in H. ntfescens and H. corrugata occurs from pH 7.4 to 7.6 (Nakada and Sweeney 1967). However, the lower pH levels found in the crop of H. laevigata are still within the range of pH that enables efficient amylase and protease activity in H. rufescens (McLean 1970). even though different enzymes with different pH optima are likely to be present in H. laevigata. The stomach functions to collect food and secretions from the salivary glands, cecum, and the digestive glands (Crofts 1929), so the pH within the stomach should also be a mixture of these influences and the secretions of the stomach. Observations from H. laevigata indicate that the stomach has similar pH to that of the crop and style sac. In the digestive diverticula of Haliotiis sp. ( =Haliotis). the maximum activity of enzymes such as fucoidan- ase occurs at pH = 5.4 (Thanassi and Nakada 1967). a value of pH similar to that found in the crop, stomach, and style sac of H. laevigata. In some invertebrates, pH and redox conditions are sometimes at unusual levels that favor association between the host and spe- cific microbial communities (Hatris 1993). The selective nature of these changes in pH imparted on the microbial communities will favor those microbes best adapted to the conditions. The different pH readings and the microaerophilic environment found in the abalone digestive tract, therefore, provide several niches for mi- crobes to exploit and grow. ■ Most of the bacteria found within the abalotie gut were able to degrade starch. CMC. or agar. The ability of several different isolates to degrade both agar and CMC indicates that these bacteria were capable of growth on two of the more common substrates available in this environment. Althouiih the isolation media were Digestive Physiology of the Greenlip Abalone 993 as close as possible to the conditions within the gut environment. the use of selective media can sometimes fail to detect some bac- teria capable of hydrolytic activity (Harris 1993). Therefore, it is likely that there are some bacterial types present that were not isolated through the enrichment process. The larger diversity of bacteria isolated from the esophagus of the abalone. with a de- crease in species in the crop and stomach, is directly related to the selective conditions of the gut environment. Dissolved oxygen and pH are two variables likely to influence microbial growth strongly in the abalone. Bacterial genera found within the digestive tracts of bivalve molluscs include: Achromobacier, Flavohacterium/Cytophaga. Pseudomonas, Vibrio. Corynebacterium, Arthiobacter. Escheri- chia. Neisseria, Streptococcus. Micrococcus. Moraxetla. Acineto- bacter. and Aeromonas spp. (Prieur et al. 1990). Juvenile blacklip abalone. Haliotis rubra, have been shown to consume bacteria with coralline algae (Garland et al. 1985). These bacteria were predominantly Moraxella. although Pseudomonas, Vibrio. Altero- monas. and smaller numbers of Flavobacterimn/Cytophaga and Aeromonas spp. were also present. Bacterial isolates obtained from the South African abalone. H. midae. showed an ability to use a range of complex polysaccharides (Erasmus et al. 1997). In terms of hydrolytic capabilities, the types of bacteria found within the abalone gut are similar to those found in the sea hare (Gastropoda). Aplysia Juliana. (Vitalis et al. 1988). However, not all the bacteria ingested by abalone may be able to exploit the gut environment. From our study, it seems that the marine bacteria capable of growth at reduced pH are different in composition to those isolated by other authors at higher pH (Sawabe et al. 1995, Erasmus et al. 1997). Consequently, the reports of other authors may have re- vealed bacterial populations that are present, but not necessarily capable of contributing to the digestive ability of the host in a typical gut pH regime. It may be that the bacteria reported in this study differ from those reported elsewhere by being capable of digesting algae within the gut environment, from the wider variety of bacteria ingested by the abalone. Vibrio spp. have been recorded as predominant microorganisms in several marine invertebrate digestive tracts (Unkles 1977, So- chard et al. 1979. Hartis 1993). including abalone (Erasmus et al. 1997, Sawabe et al. 1995). It may seem surprising that so few isolates of Vibrio spp. were obtained from the greenlip abalone. However. Vibrio spp. are usually isolated on alkaline media (Bau- mann and Schubert 1984), suggesting growth is reduced or pre- vented in acidic conditions. The microbial isolates from the most acidic region, the crop, were almost entirely from the family En- terobacteriaceae, suggesting that these bacteria are well adapted to the acidic gut environment. The Enterobacteriaceae are rarely re- corded from the marine environment or from the guts of inverte- brates (Harris 1993). The occurtence of the Enterobacteriaceae in H. laevigata may represent a normal bacterial flora that specifi- cally developed within the gut and adapted to the reduced pH and microaerophilic environment. Their presence throughout the di- gestive tract suggests that these bacteria may be indigenous. Wild greenlip abalone are obligate drift algae consumers, able to consume many different types of algae (Shepherd 1973). Inges- tion of diatoms, detritus, bacteria, and sand also occurs as a result of the mode of feeding (Campbell 1965). The diet fed to the abalone during this study was limited in diversity as compared to that of abalone in the wild. The restrictions this would place on microbial growth may be subtle, although some decrease in normal microbial species diversity could be expected. By restricting avail- able food types, this may also reduce bacterial diversity. Digestive tract analysis of bacterial biota in other animals maintained in laboratory systems supports this theory, because the selective pres- sures imposed by the artificial environment influence the normal bacterial flora occurring in the gut of aquatic invertebrates (So- chard et al. 1979). Seaweeds are known to have epiphytic colonies of diatoms, yeasts, and bacteria (Austin 1988), some of which have known algal cell-degrading abilities, such as Cytophaga spp. Mechanical breakdown of algae by the radula would release cellular contents previously unavailable to epiphytic or free-living bacteria, and this would be expected to slunulate microbial growth. However, few bacteria were seen to be associated with the gut surface of the greenlip abalone. This may be caused by the action of cilia and the presence of mucous and secretory cells (Harris et al. 1998). and the results from this study that suggest that variation in pH from the external environment may also be a factor. We obtained 44 isolates of bacteria from the gut of H. laevigata. These bacteria were capable of degrading algal polysaccharides at levels of pH and dissolved oxygen similar to gut values. Therefore, bacteria may contribute to H. laevigata nutrition. Because the bacteria do not seem to be strongly associated with the gut epithelium, then bac- terial digestive activity is likely to be restricted to the lumen (Har- ris et al. 1998). Some output of feces still occurs several days after feeding has ceased (Wee et al. 1992). allowing the possibility for sustained bacterial activity within the intestines. The less intimate association of bacteria with bivalves as compared to terrestrial animals (Kueh and Chan 1985) also seems apparent in the greenlip abalone. Kueh and Chan (1985) suggested that, for oysters, the gut flora are mainly derived from the external environment and a more indigenous population of bacteria dominate the lower digestive tract because of selective pressures and multiplication. This situ- ation seems analogous to that of the greenlip abalone. CONCLUSION The presence of bacteria within the digestive tract of the green- lip abalone, and their ability to break down algal carbohydrates at pH levels found within the gut, suggests that bacteria are capable of contributing to the nutrition of their host, although the amount remains in question. The lack of physical association of these bacteria with gut epithelium suggests a different digestive strategy to terrestrial herbivores. If bacteria contribute significantly to host nutrition, they are more likely to contribute through activity within the gut lumen. The selective pressures of the gut environment give rise to bacterial populations that are different in composition to those reported from the external marine environment. ACKNOWLEDGMENTS This work was supported by the School of Aquaculture. Uni- versity of Tasmania at Launceston. The authors to thank Mr. James Mason of Funieaux Aquaculture for providing the abalone. Also, the authors thank Mr. Mark Heather, Mr. Brad Adams, and Mr. Nick Savva of Tasmanian Tiger Abalone for providing much of the algae fed to the abalone. We also thank Dr. Judith Handlinger for critical assessment of this manuscript. Present address for G.B.M.; Fisheries Western Australia. Research Division, P.O. Box 20, North Beach, WA. 6020. Australia. 994 Harris et al. LITERATURE CITED Austin. B. 1988. Marine microbiology. Cambridge University Press, Mel- bourne. 222 pp. Baumann, P. & R. H. W. Schubert. 1984. Vibrionaceae. pp. ."; 16-550. //;. N. R. Krieg and J. G. Holt (eds.). Bergey's Manual of Systematic Bac- teriology. Williams & Wilkins. Sydney. Campbell. J. L. 1965. The structure and function of the alimentary canal of the blacklip abalone. Haliotis cracherodii Leach. Trans. Am. Microsc. Soc. 84:376-395. Craigie. J. S. 1990. Cell walls, pp. 221-258. //;.■ M. Cole and R, G. Sheath (eds.). Biology of the Red Algae, Cambridge University Press, Mel- bourne. Crofts, D. R. 1929. Haliotis. Liverpool Marine Biological Committee Memoirs No. 29. pp. 50-81. Crosby, N. D. & R. G. B. Reid. 1971. Relationships between food, phy- logeny, and cellulose digestion in the Bivalvia. Can. J. Zool. 49:617- 622. Erasmus, J. H., P. A. Cook & V. E. Coyne. 1997. The role of bacteria in the digestion of seaweed by the abalone Halioris niidae. Aquaculture 155: 377-386. Galli, D. R. & A. C. Giese. 1959. Carbohydrate digestion in a herbivorous snail, Tegula fimebralis. J. Exp. Zool. 140:415-440. Garland, C, D., S. L. Cooke, J. F. Grant & T. A. McMeekin. 1985. Inges- tion of the bacteria on the cuticle of crustose (nonarticulated) coralline algae by postlarval and juvenile abalone {Haliotis ruber Leach) from Tasmania waters. J. Exp. Mar. Biol. Ecol. 91:137-149. Gomez-Pinchetti, J. L. & G. Garci'a-Reina. 1993. Enzymes from marine phycophages that degrade cell walls of seaweeds. Mar. Biol. 1 16:553- 558. Hahn, K. O. 1989. Biotic and abiotic factors affecting the culture of aba- lone. pp. 1 13-134. In: K. O. Hahn (ed.). Handbook of Culture of Aba- lone and other Marine Gastropods. CRC Press, Boca Raton, ¥h. Harris, J. M. 1993. The presence, nature, and role of gut microflora in aquatic invertebrates: a synthesis. Microb. Ecol. 25:195-231. Harris, J. O., C. M. Burke & G. B. Maguire. (1998). Characterization of the digestive tract of greenlip abalone. Haliotis laevigata Donovan. I. mor- phology and histology. / Shellfish Res. 17:1007-1016. Hyman, L. H. 1967. The invertebrates, vol. VI, Mollusca I. McGraw-Hill. Sydney, pp. 169-+01. Kersters, K. & J. De Ley. 1984. Genus Alcaligenes. pp. 361-373. In: N. R. Krieg and J. G. Holt (eds.), Bergey"s Manual of Systematic Bacteriol- ogy. Williams & Wilkins, Sydney. Kreger, D. R. 1962. Cell walls, pp. 315-336. //;.■ R. A. Lewin (ed). Physi- ology and Biochemistry of Algae. Academic Press, London. Kueh, C. S. W. & K. -Y. Chan. 1985. Bacteria in bivalve shellfish with special reference to the oysters. J. App. Bacteriol. 59:41—17. Lewis, T. E., C. D. Garland & T. A. McMeekin. 1992. Manual of hygiene for shellfish hatcheries. Department of Agricultural Science, University of Tasmania, Hobart. 45 pp. Mathers, N. F. 1974. Digestion and pH variation in two species of oysters. Proc. Malacol. Soc. (London) 41:37^0. McBee, R, H. 1971. Significance of intestinal microflora in herbivory. Ann. Rev. Ecol. Syst. 2:165-176. McLean, N. 1970. Digestion in Haliotis riifescens Swainson (Gastropoda: Prosobranchia). J. E.xp. Zool. 173:303-318. McCandless, E. L. 1981. Polysaccharides of the seaweeds, pp. 559-588. In: C. S. Lobban and M. J. Wynne (eds.). The Biology of Seaweeds. Blackwell Scientific, Melbourne. Morton, B. 1978. Feeding and digestion in shipworms. Oceanogr. Mar. Biol. Ann. Rev. 16:107-144. Morton. J. E. 1968. Molluscs. Hutchinson University Library, London, pp. 94-116. Musgrove, R. J. 1988. Digestive ability of the freshwater crayfish Para- nephrops zealandicus (White) (Parastacidae) and the role of microbial enzymes. Freshw. Biol. 20:305-314. Nakada, H. I. & P. C. Sweeny. 1967. Alginic acid degradation by elimi- nases from abalone hepatopancreas. J. Biol. Cheni. 242:845-851. Owen, G. 1966. Digestion, pp. 53-96. In: K. M. Wilbur & C. M. Yonge (eds.). Physiology of Mollusca, vol. 11. Academic Press, London. Plante. C. & P. Jumars. 1992. The microbial environment of marine de- posit-feeder guts characterized via microelectrodes. Microb. Ecol. 23: 257-277. Prieur, D.. G. Mevel. J. L. Nicolas. A. Plusqucllec & M. Vigneulle. 1990. Interactions between bivalve molluscs and bacteria in the marine en- vironment. Oceanogr. Mar. Biol. Ann. Rev. 28:277-352. Prim. P. & J. M. Lawrence. 1975. Utilization of marine plants and their constituents by bacteria isolated from the guts of echinoids (Echino- dermata). Mar. Biol. 33:167-173. Revsbech. N. P. & B. B. Jorgensen. 1986. Microelectrodes: their use in microbial ecology, pp. 293-352. In: K. C. Marshall (ed.). Advances in Microbial Ecology, vol. 9. Plenum Publishing, New York. Sawabe, T., Y. Oda, Y. Shiomi & Y. Ezura. 1995. Alginate degradation by bacteria isolated from the gut of sea urchins and abalones. Microb. Ecol. 30:193-202. Shepherd. S. A. 1973. Studies on southern Australian abalone (Genus Hali- otis). I. Ecology of five sympatric species. Aiist. J. Mar. Freshw. Res. 24:217-257. Sochard. M. R.. D. F. Wilson. B. Austin. & R. R. Colwell. 1979. Bactena associated with the surface and gut of marine copepods. App. Envir. Microbiol. 37:750-759. Sokal. R. R. & J. F. Rohlf 1995. Biometry. The principles and practice of statistics in biological research. W. H. Freeman. New York. 887 pp. Tall, B. D. & R. K. Nauman. 1981. Scanning electron microscopy o( Cris- lispira species in Chesapeake Bay oysters. App. Envir. Microbiol. 42: 336-343. Thanassi, N. M. & H. I. Nakada. 1967. Enzymic degradation of fucoidan by enzymes from the hepatopancreas of abalone. Haliotus species. Arch. Biochern. Biophys. 118:172-177. Underwood. A. J. 1981. Techniques of analysis of variance in experimental marine biology and ecology. Oceanogr. Mar. Biol. Ann. Rev. 19:513- 605. Unkles. S. E. 1977. Bacterial flora of the sea urchin Echinus esculentiis. App. Envir. Microbiol. 34:347-350. Vitalis, T. Z., M. J. Spence & T. H. Carefoot. 1988. The possible role of gut bacteria in nutrition and growth of the sea hare Aplysia. Veliger 30:333-341. Wee, K. L., G. B. Maguire & S. M. Hindrum. 1992. Methodology for digestibility studies with abalone. I. Preliminary studies on the feeding and defecatory behavior of blacklip abalone, Haliotis rubra, fed natural and artificial diets, pp. 192-196. In: G. L. Allan and W. Dall (eds.). Proceedings of the Aquaculture Nutrition Workshop. Salamander Bay, 15-17 April 1991, NSW Fisheries, Brackish Water Fish Culture Re- search Station, Salamander Bay. Australia. Zar. J. H. 1996. Biostatistical Analyses. Prentice-Hall of Australia. Pty. Ltd.. Sydney, pp. 65-91. Journal of Shcllfnli Rcn'orch. Veil. 17, Nii. 4. 445-1002. IWS. BACTERIAL COLONIZATION OF A FORMULATED ABALONE DIET DURING EXTENDED IMMERSION ANDREW BISSETT.' CHRIS BURKE,' " * GRAEME A. DUNSTAN,' ' GREG B. MAGUIRE' ^ 'School of Aqiiacitlture University of Tasnuiiiia Launcestoii, Tasmania. 7250. Australia 'The Cooperative Research Centre for Aqitacidtitre P.O. Box 123 Broadway Sydney. 2007. Australia ^CSIRO Division of Marine Research Hohart. Tasmania. 7001. Australia ABSTRACT The characteristics of the microbiota of a formulated abalone (Haliotis laevigata) diet were studied by scanning electron microscopy (SEM) and standard bacterial culture and taxonomic techniques. Microbes colonizing the diet (ABCHOW) were enumer- ated by SEM and partly identified after immersion of the diet in seawater for 0, 2, and 4 days with and without abalone. The fatty acid composition of the diet was also analyzed, after similar treatments, for biomass estimates and bacterial biomarker identification. Bacterial numbers on unimmersed diet and diet immersed in sterile seawater for 2 and 4 days were negligible. Bacteria proliferated after 2 days immersion in seawater with abalone ( 1.2 x 10' cells/mnr ) and without abalone (5.7 x lO'' cells/mnr) (p < .05). Numbers continued to rise between 2 and 4 days for diet immersed without abalone (6.5 x 10"* cells/mm"). However, a decrease in bacterial numbers was observed between 2 and 4 days immersion in seawater with abalone (7.7 x lO"* cells/mm" after 4 days), and this was accompanied by an increase in ciliate numbers (from to 10" ciliates/mm"). Ten distinct taxonomic groups of bacteria were identified from the diet after immersion; Cylophaga spp. was the most abundant group. Chemotaxonomic analysis, including fatty acid profiling, failed to provide microbial biomass estimates or bacterial biomarkers. The majority of the microbes were found to have the capacity to degrade a protein and a lipid source within the diet, but not two carbohydrate sources, including the binder. Bacteria were found to affect the physical form of the diet, but it is unlikely that they affected its macronutritional value to any great extent. INTRODUCTION Worldwicje decline in abalone fisheries has accelerated the de- velopment of abalone mariculture (Coote et al. 1996). One of the major constraints to the industry is the provision of an economi- cally viable, nutritionally suitable, formulated diet (Fleming et al. 1996). Given the high cost and logistical difficulties of supplying natural seaweeds to abalone (Coote et al. 1996), the industry pref- erence is for a cost-effective formulated diet. The development of such a diet depends not only on an understanding of the nutritional requirements of the abalone and their digestive processes, but also on the diet's performance in the culture system. Most formulated diets for fish are consumed rapidly, but, large industries now exist for slow-feeding invertebrates such as marine shrimp and marine gastropods. Abalone graze on food slowly, so it is not uncommon in cominercial situations for diet to be im- mersed for up to 4 days before it is consumed. Given the rapid leaching of water-soluble nutrients from formulated diets (Goldb- latt et al. 1979) and the ample time for microbial colonization, costly feeding strategies, based on frequent input, may be adopted. However, growth trials have suggested the opposite may actually be the case: faster growth rates were observed with less frequent feeding and cleaning intervals and therefore extended immersion (Maguire et al. 1996). When considering the effects of microbes on the nutritive value of formulated diets, several possibilities exist. They may have detrimental effects: by consuming the nutrients meant for the aba- lone, or decreasing the stability of the diet by facilitating its physi- cal breakdown and exacerbating water quality problems (Shigueno *Present address: Fisheries Western Australia, Research Division, PO Box 20, North Beach, WA 6020. Australia. 1975, Moriarty 1986). The diet may also act as a reservoir for pathogenic microbes (Moriarty 1986, Muir and Sutton 1994). Con- versely, bacteria may be beneficial by forming extracellular par- ticulate matter after uptake of dissolved nutrients (Pearl 1978). which may then become available to the abalone. Dietary constitu- ents may be broken down and consumed by bacteria, which may in turn be consumed by the abalone. Both Garland et al. ( 1985) and Harris (1993) state that digestive enzymes may be supplied to abalone by bacteria. Finally, bacteria may produce certain feed attractants or micronutrients that make the diet more palatable or nutritionally adequate (Sakata 1987, McShane et al. 1994). The aim of this study was to quantify the microbial population colonizing a formulated abalone diet after periods of extended immersion and to isolate and partially identify members of this microbial population in order to examine their exoenzyme activity in relation to diet constituents. This information was sought using three distinct methods: scanning electron microscopy (SEM) enu- meration, traditional biochemical methods, and chemotaxonomic techniques. Results from this study could allow more informed decisions regarding feed strategies and diet formulation. MATERIALS AND METHODS Diet The formulated diet was a proprietary formulation (ABCHOW) produced with a pasta maker and subsequently dried as a biscuit (14 X 9 X 1 mm). ABCHOW was supplied by the South Australian Research and Development Institute (SARDI), Adelaide, Australia and was derived from diet 9 as used by Fleming et al. ( 1996). Diet was stored in a domestic freezer and added to treatment tanks manually. 995 996 BiSSETT ET AL. Abalone Culture Systems Treatment tanks comprised 70-L. round, aerated, center- draining, tlow-through, fiberglass aquaria supplied with approxi- mately 1.5 L/min of sand-filtered (40 to 50 (xm) sea water. Each treatment utilized three replicate tanks. Tanks were continously shaded in a black plastic (100% shade) enclosure to prevent ex- cessive benthic diatom colonization. The entire system was housed within a translucent fiberglass building. Light intensity was < 0.03 microEinsteins/m". Treatment tanks using abalone contained a bio- mass of approximately 100 x 1 g abalone/tank. Trials were con- ducted at ambient conditions (typically about 14°C, pH 8.2 and salinity 34 to 35 ppt) at a commercial abalone farm. Marine Shell- fish Hatcheries P/L, Bicheno, Tasmania, Australia. Tanks were cleaned by rapid draining before the sample diet was added. Treatments Seven treatments were employed in this study: 1 . Two days immersion in the presence of abalone 2. Two days immersion, abalone absent 3. Four days immersion in the presence of abalone 4. Four days immersion, abalone absent 5. Zero days immersion 6. Two days immersion in sterile seawater 7. Four days immersion in sterile seawater For treatments 1 to 4. samples were collected for both SEM enumeration and bacterial isolation directly from the aquaria. For treatments 6 and 7. 200 mL of coarse-filtered, aged seawater was added to 500 mL glass beakers and autocla\ed for 15 min at 121°C, 15 psi. Diet was added aseptically, and the beakers were then incubated under similar conditions to those experienced by other treatments, except they were static. Bacterial Isolation and Identification For isolation of bacteria, five pieces of the feed were collected from each replicate tank or beaker. Samples were collected using methylated-spirit disinfected filter forceps and placed in 10 mL of sterile saline. They were then hand homogenized and a further four, tenfold serial dilutions performed. Samples from the five dilutions were subsequently inoculated onto general carbon source plates and restricted carbon source plates and incubated at room temperature for 1 to 14 days. Restricted carbon source plates were used in the study to isolate bacteria utilizing nutrients specific to the ABCHOW diet. The restricted carbon source plates comprised 20 g agar, 10 g carbon source, and 500 mL each of distilled and filtered seawater. The nutrient source for each plate consisted of one of the principal ingredients of the ABCHOW diet: two carbo- hydrates (semolina), including the binder (sodium alginate), one high protein meal (casein), and one lipid (fish oil) source. Follow- ing incubation, plates were inspected daily for growth. Discrete colonies were removed and transferred to separate abalone-feed- nutrient-agar plates (identical to those above, substituting homog- enized ABCHOW diet for the individual nutrient sources) for pu- rification. Subculturing continued until pure cultures were ob- tained, and these were maintained on OrdaPs medium (Atlas 1993). Identification tests performed on the pure cultures were: Gram reaction, cellular and colonial morphology on Ordal's medium after 5 days incubation; glucose utilization (OF) test for metabolic type using a modification of the method of BariDW and Feltham (1993), whereby 500 niL of filtered seawater was substituted for the equivalent amount of distilled water; Craigie tube motility test (similarly modified from Barrow and Feltham 1993); oxidase and catalase (Barrow and Feltham 1993); anaerobic growth (Oxoid, Anaerogen system) on Ordal's medium; sensitivity to the antibi- otic 0/I29 (Oxoid). and growth on the restricted carbon sources specific to the ABCHOW diet. SEM Enumeration For each replicate tank or beaker, three whole pieces of feed were used. These were fixed in 2.5% glutaraldehyde in 0.2M cac- odylate buffer (pH 7.4) containing the major marine salts (Garland et al. 1982) immediately upon removal from experimental tanks. Samples were thus fixed for 2 h at room temperature, rinsed for 20 min in 0.1 m cacodylate buffer (Hodson and Burke 1994) and dehydrated through a graded ethanol (EtOH) series (Hodson and Burke 1994). The dehydration series was suspended at 70% EtOH until critical point drying was possible. Dehydrated samples were transferred to acetone and critical point dried using liquid CO,. Samples were dried using a Baltec CPD 030 or a Polaron E3000 CPD. Following drying, samples were mounted onto aluminum SEM stubs with conductive carbon paint as the adhesive. Samples were then sputter coated with gold (Balzers type coater) twice, to improve sample stability (Rosowski et al. 1981 ). as soon as practicable after drying. Samples were then stored in a vacuum desiccator (25-30 psi) over CaCI, (Garland et al. 1982). Samples were viewed and photographed under a Philips 505 SEM at a voltage of 15 kV. Random fields on the sample surface were chosen at low mag- nification (Lewis et al. 1985). then examined at a magnification of 2500 X. giving a viewing field of 1322 \x.m' (although it should be noted that the viewing field was not flat). The surface was focused and aligned at 90° to the viewing plane, and all organisms within the viewing field (including those that intersected the top and left sides) were counted. For each replicate, at least 10 full fields were counted. Representative micrographs were taken of each replicate. Data were transformed with a square root transformation (Sokal and Rohlf 1987) prior to statistical analysis, with a one-way analysis of variance ( ANOVA), to meet assumptions of normality and homogeneity of variance. For all tests, a significance level of p < .05 was adopted. Data for each immersion period were ana- lyzed separately, and environment (abalone present or absent or sterile seawater) was considered as a fixed factor. Pairs of means were compared using Fishers LSD (Sokal and Rohlf 1987). Fatty Acid Analysis The method of Bligh and Dyer ( 1959), as modified by Dunstan et al. (1995). was used for extraction. Fatty acid methyl ester (FAME) samples were analyzed with a Hewlett-Packard 5890 gas chromatograph (GC) that was equipped with a flame ionisation detector (FID). FAME samples were injected using an air-cooled on-column injection into a polar BPX-70 fused silica column (50 m X 0.32 mm ID). High-purity H, was the carrier gas. The GC oven temperature was initially held at 45°C for 2 min after injec- tion and then increased at 30°C/min to I20°C and at 3°C/min to 240°C, and was then held constant for 10 min. The retention index on both polar and nonpolar columns was used to identify fatty acids. Fatty acid identifications were verified with a Hewlett- Packard 5970B GC/MS system. Bacterial Colonization of Immersed ABCHOW 997 Figure I. A split view of the surface of diet immersed for 4 days without abalone. The left of the micrograph shows the interface be- tween areas of confluent growth and very little growth (arrows indi- cate boundary). The right demonstrates the density of cells within the area of growth. (Bar = 0.5 mm and refers to the left of the micrograph). RESULTS Enumeration At low magnification, tlie surface of eacli sample was observed to be typically undulating with irregular depressions. The inegu- larity of the diet's topography was accentuated with immersion tiine. At high magnification (2,500 X) it was possible to distin- guish individual bacterial cells, despite the surface corrugations and often dense mucilage. Cells colonized different areas of the diet to differing degrees; lipid globules (identified visually under EM) were less densely colonized than the rest of the diet matrix (Fig. 6). Areas of confluent growth, adjacent to areas of no growth, were also evident (Fig. 1 ). These areas of confluent growth seemed S3 ^ « s Immersion T (D avs) m e Figure 2. Mean bacterial numbers on ,\BCH()\V diet at (I, 2, and 4 days immersion with abalone, without abalone, and in sterile seawater. Means for the same immersion time that share a common superscript are not significantly different (p > .05). Vertical lines represent stan- dard errors of the means (n = 3). Figure i. \ typical random counting area from unimmersed diet (2500 X, Bar = 10 fiml. No bacterial cells were present. to be the result of bacteria spreading from initial sites that were favorable for growth. SEM examination showed negligible bacte- ria colonizing treatments 5 to 7, but a substantial proliferation in bacterial numbers for treatments 1 to 4 (Figs. 2^). One-way ANOVAs, based on all treatments except 5. con- firmed a significant treatment effect (p < .001) on bacterial abun- dance for 2- and 4-day data. Comparisons of means for the same immersion time showed that bacterial numbers for all treatment levels (n = 3) (abalone absent = 5.7 x 10"* cells/mm", abalone present = 1.2 x 10' cells/mm", sterile = 1.5 x 10^ cells/mnr) were significantly different after 2 days immersion (p < .05). There was no significant difference between the bacterial densities for nonsterile treatments after 4 days, with (mean = 7.7 x 10"* cells/ mm-) or without (mean = 6.5 x lO"* cells/mm^) abalone. but mean counts from diet immersed in sterile seawater (mean = 7.0 x lO' cells/mm") were significantly different (p < .05) to other treat- ments after this immersion time. SEM also revealed the presence of ciliates (Fig. 5) on the diet subjected to treatment 3. The average number of ciliates present on treatment 4 diet was 0.8 ciliates/counling field. The ciliates were approximately 30 x 20 |jLm in size, were all of the same morpho- type and were observed on all replicate samples from treatment 3, but no other samples. Figure 4. A typical random counting area from diet immersed for 4 days, without abalone. (2500 X, Bar = 10 ^m) showing massive bacte- rial colonization. 998 BiSSETT ET AL. Figure 5. A micrograph of an unidentifled ciliate (193000 X, bar = 10 fim), typically observed on treatment 4. Identification A large number of bacteria (108) were isolated directly from agars containing ingredients of the ABCHOW diet. No bacteria were isolated from diet immersed in sterile seawater, because no colonies had formed on plates from these treatments after 14 days. Bacteria were identified and placed into one of 10 taxonomic groups (Table 1 1. Most groups were observed in all treatments; bacteria falling into the group Cytophaga being the most numer- ous. The identification scheme used was adopted from Cropp and Garland ( 1988). All isolates were Gram-negative rods, catalase {+) and grew aerobically on Ordal's medium. Isolates were initially separated on their ability to produce acid from glucose (OF test) and on their ability to grow anaerobically on Ordal's medium. Three distinct groups resulted (Table 2). The first group were able to produce acid from glucose aerobically, but were incapable of anaerobic growth on Ordal's medium. The second group did not produce acid from glucose aerobically or anaerobically. hence they were oxidative, but some were capable of anaerobic growth on Ordal's medium. The final group were fermentative, producing acid from glucose both aerobically and anaerobically and were capable of anaerobic growth on Ordal's medium. All bacteria isolated demonstrated the ability to degrade a wide range of carbon sources (Table 3). Results were similar for both TABLE 1. Bacterial types isolated and identified from treatments 1 to 5, showing the total numbers of isolates in each taxonomic group. Bacterial Type Number of Isolates Treatments Cytophaga 27 Mesophilohacter or Monuella 19 Vibrionaceae 17 Alcaligeiies or P.seiidomoiias 12 Enterobacteriaceae 13 Acinetobacter 8 Moraxella or Paracoccus 5 Aeromunas 4 Flavobacterium or Phenylobaclenitm 2 AUeronumas 1 Total 108 1 to 5 1 to 5 1 to 5 1 to 5 2 to 5 2 to 5 1,2,3,5 1,2,3 3 &4 4 Figure 6. Lipid globules in the diet from 2 days immersion, with aba- lone (2500 X, Bar = 10 fim), showing typical sparse colonization of the lipids. carbohydrates. All isolates were able to use protein meal as a nutrient source. At least one isolate from each treatment was able to degrade either the carbohydrate or the lipid sources. At least one isolate from each taxonomic group m treatment 1 was able to utilize all nutrient sources. However, members of the families Vibrionaceae and Enterobacteriaceae demonstrated poor utiliza- tion of the binder. No Pseiidomonas or Acinetobacter spp. was observed to utilize the carbohydrate source. All other groups de- tennined to have oxidative metabolism were generally poor users of carbohydrate energy sources. Lipase activity was expressed by a large proportion of isolates of most taxonomic groups. Fatty Acid Analysis Analysis of the ABCHOW diet's fatty acid content after 0. 2. and 4 days immersion (with and without abalone) showed an in- crease in total fatty acids from 3.26 g/100 g dry wt at day immersion, to 4.55 g/lOOg dry wt at 4 days immersion without abalone (Table 4). Analysis of straight and branched-chain fatty acids in the ABCHOW diet revealed no change in relative com- position after extended immersion (Table 5. 6). DISCUSSION Enumeration Microorganisms are known to colonize many natural and arti- ficial marine surfaces, hence it is not surprising that this study TABLE 2. Metabolic types of bacteria isolated from treatments 1 to 5 as determined by glucose utilization (OF test! and anaerobic growth Oxidative Oxidative Total Obligate Facultativelv Number Treatment .Aerobe ■Anaerobic Fermentative of Isolates I 1 3 4 5 Total 13(9) 10 23 13(2) 7 20 16(2) 6 22 20(3) 4 25 13(2) 5 18 75(18) 32 108 Figures in parentheses indicate numbers of organisms capable of anaerobic growth. Bacterial Colonization of Immersed ABCHOW 999 TABLE 3. Bacterial types isolated frum treatments 1 to 5, showing total number of Isolates in each group and the number of isolates capable of degrading specific nutrient sources found in the ABCHOW diet. Number Binder Lipid of (Sodium Carbohydrate Protein (Fish Treatment Bacterial Group Isolates Alginate) iSemolinal (Casein) Oil) Aeromonas spp. 2 1 1 -) 2 Akaligenes or Pseiidomoiun spp. 3 2 1 3 3 Cytophaga spp. 4 2 1 4 4 Mesophilohacler spp. 4 4 2 4 4 Moraxella or Paracoccus sp. 1 1 1 1 1 Pseudomonas sp. 1 1 1 1 1 Vibrio spp. 4 4 2 4 4 Vibrionaceae 4 1 4 4 2 Acinetobacter sp. 1 1 1 1 2 Aeromonas spp. 1 1 1 1 2 Akaligenes or Pseudomonas sp. 1 1 1 2 Cytophaga spp. 7 1 4 7 7 2 Enterobacteriaceae 4 1 4 4 2 Mesophilobacter or Moraxella spp. 4 3 2 4 4 2 Moraxella or Paracoccus sp. 1 1 1 1 2 Vibrionaceae 3 3 3 3 3 Acinetobacter sp. 3 3 3 3 Aeromonas spp. 1 1 1 1 3 Cytophaga spp. 9 2 3 9 9 3 Enterobacteriaceae T 1 2 3 Flavobaclerium or Phenylobacterium sp. 1 1 1 1 3 Mesophilobacter or Moraxella spp. 1 1 1 1 3 Moraxella or Paracoccus sp. 1 1 3 Pseudomonas sp. 1 1 1 1 3 Vibrionaceae 3 3 3 4 Acinetobacter sp. 2 2 2 4 Alcaligenes or Pseudomonas spp. 4 3 3 4 Alteromonas sp. 1 1 1 4 Cytophaga spp. 4 1 4 4 4 Enterobacteriaceae 4 2 3 2 4 Flavobacterium or Phenylobacterium sp. 1 1 1 I 4 Flavobacterium sp. 1 1 1 1 4 Mesophilobacter or Moraxella spp. 7 1 6 6 4 Pseudomonas sp. 1 1 1 4 Vibrionaceae 1 I 1 5 Acinetobacter sp. 2 2 1 5 Alcaligenes or Pseudomonas spp. 2 1 2 2 5 Cytophaga spp. 6 3 6 6 5 Enterobacteriaceae 3 3 3 5 Mesophilobacter or Moraxella spp. 2 1 1 2 2 5 Vibrio sp. 1 1 I 1 demonstrates microbial colonization of the surface of the AB- CHOW diet after 2 to 4 days immersion. Microorganisms were clearly visible on the surface of the diet, under SEM examination, despite the fragility of the diet. The diet's postimmersion fragility was also expected, especially given dry matter loss of 2'&'7c after 48 h immersion (Maguire 1996). This fragility and dry matter loss are noteworthy for two reasons, the first being the relationship be- tween dry matter loss and available surface area. Dry matter loss does not necessarily imply a decline in the surface area. The loss of food particles and the expansion of food particle size following hydration may actually increase the surface area available for mi- crobial colonization. This point should be considered when inter- preting bacterial density figures. Second, the fragility of the diet postimmersion presented a problem in handling for SEM exami- nation. It may be of benefit to stop the dehydration process at 70% EtOH and store the samples until ready for SEM examination. Samples that were viewed immediately after the dehydration and critical point drying processes were completed, appeared more stable under SEM. Bacterial numbers observed ranged from lO" to 10'' cells/mm". It is not surprising to find low bacterial numbers on samples from treatments 5 to 7. The diet contains very little moisture (approxi- mately 6%) (Maguire unpublished data) and was stored in a freezer before application to the culture environment. Such conditions will 1000 BiSSETT ET AL. TABLE 4. Total fatty acids (g/100 g \vt dry \vt) for the ABCHOW diet after various periods of immersion, with and without abalone (n = 3). Immersion Abalone Total Fativ Treatment Time (d) Present ( +)/Absent (-) Acids 5 - 3.26 1 2 + 3.86 2 2 - 3.98 3 4 + 4.5.'; 4 4 - 4.44 cause many bacteria to become donnant. if not render them unvi- able. The small increase in bacterial numbers over 4 days immer- sion in sterile seawater demonstrates the very low initial numbers present. The bacterial densities observed from treatments 1 to 4 are similar to those reported in the literature. Lewis et al. (1985) re- ported bacterial numbers in the order lO"* cells/mm' on crustose coralline algae, a preferred natural settlement substrate of juvenile abalone. The microbial community existing on natural seaweeds is well established and stable, although seasonal fluctuations do oc- cur (Lewis et al. 1985). Bacteria examined in this study have had only 2 to 4 days to colonize the substrate and proliferate. The fact that the microbial biomass is higher on the formulated diet after 4 days than on the abalone's natural diet indicates the suitability of the diet as a substrate for bacteria. The difference between bacterial numbers observed in treat- ments I and 2 (p < .05) (Fig. 2) is a result of the presence of abalone. Bacteria are known to be associated with abalone, both externally and internally (Prieur et al. 1990. Harris 1993). The association of bacteria with abalone places more bacteria in con- tact with the diet than if the abalone were absent. Abalone may also transfer bacteria between diet pieces. Free-floating bacteria are reliant upon chance contact with the diet before they can locate a suitable substrate (via chemotaxis). Although movement of mo- tile bacteria is relatively fast (Schlegel 1993). their movement is TABLE 5. Distribution and total branched-chain fatty acids from the ABCHOW diet after various periods of immersion, with and without abalone. Immersion Time (dl 2 2 4 4 Branched-Chain With Without With Without Fatty Acids Abalone Abalone Abalone Abalone il4:0 0.0 0.0 0.0 0.0 0.0 il5:0 0.2 0.2 0.2 0.2 0.2 a 15:0 0.1 0.1 0.1 0.1 0.1 il6:0 0.1 0.1 0.1 0.1 0.1 il7:0 0.4 0.4 0.4 0.4 0.4 al7:0 0.0 0.0 0.0 0.0 0.0 il7:l 0.2 0.2 0.1 0.1 0.2 il8:0 0.5 0.5 0.5 0.5 0.5 il8:l 0.1 0.1 0.1 0.1 0.1 brl9:l 0.8 0.7 0.7 0.7 0.8 Total 2.4 2.3 2.2 2.2 2.4 Figures represent branched-chain fatty acids as a percentage proportion of total fatty acids (n = 3). effectively confined to very small areas and does not play any major role in the distribution of bacteria over large areas. How- ever, the water column itself is also an important microbial source, as is indicated by the difference (p < .05) in bacterial numbers between treatments 4 and 7. This suggests that bacteria that come into contact with the diet are able to move, via chemotaxis. to a suitable substrate and proliferate. Bacterial numbers were not significantly different between treatinents 3 and 4 (p > .05 ). This apparent loss of treatment effect was brought about by an increase in bacterial numbers in treatment 4 and a decrease in treatment 3 (Fig. 2). It is unlikely that the nutrient content of the diet would be exhausted in such a small sampling time, so it was expected that bacterial numbers would continue to increase over this period. A factor that may have contributed to the decline in bacterial numbers is the presence of protozoan ciliates (6.2 x lO'/mm") in treatment 3. Ciliates are known to graze heavily on bacteria; indeed, it has been shown that ciliates can clear approximately 10^ bacteria/ciliate/h (Iriberri et al. 1994. Solic and Krustulovic 1994). This being the case, the ciliates observed on treatment 3 could be capable of consuming 10^ cells/ mm-/h. This figure equates to the whole standing crop of bacteria and thus, may explain how bacterial numbers declined in treatment 3. The abalone themselves may also have been ingesting bacteria from the diet surface or disturbing surface films. Identification One of the aims of this study was to categorize, to a degree that allowed an assessment of metabolic activity and capacity, the bac- terial microflora colonizing the ABCHOW diet. The majority of the bacteria found on the diet demonstrated an ability to degrade a range of nutrient sources presented to them. All isolates demonstrated protease activity (Table 3). Although proteins are among the most expensive of the diet's components, their degradation may not be deleterious. Fleming et al. (1996) suggested that protein partially digested by bacteria may be more efficiently digested by abalone. It should be noted, however, that faster protein decomposition generally leads to water quality prob- lems. The binder was poorly utilized by many of the bacterial isolates (Table 3). This is surprising given that many of the binders used in formulated diets, for example alginate and cellulose, are readily available in the marine environment. If binder is not readily ac- cessible to the bacteria, it may be able to perform its task of holding water-soluble nutrients longer. Lipase activity was demonstrated by most of the bacterial iso- lates (Table 3). It is interesting to note, however, that SEM ex- amination of the diet revealed that lipids were not very heavily colonized in relation to other areas (Fig. 6). This indicates that although most of the isolates were able to utilize lipids, they may not have been the preferred nutrient source. The lack of bacterial colonization of lipids seen under SEM examination may be attrib- uted to the hydrophobic/hydrophilic interactions that must be over- come by the bacteria at the lipid/water interface. Because many of the oxidative organisms were unable to pro- duce acid from glucose, it is reasonable to suggest that they are likely to be poor users of complex carbohydrates, which are ini- tially hydrolyzed to glucose. No isolates identified as Pseudomo- nas or Acinetobacter were able to use the carbohydrate source supplied. Abalone consume a natural diet that is high (40-50%) in carbohydrates and possess many enzymes capable of carbohydrate hydrolysis (Fleming et al. 1996). As a result, many formulated Bacterial Colonization of Immersed ABCHOW 1001 TABLE 6. Straight-chain fatty acids from the ABCHOW diet after various immersidn times with and without abalone. Immersion Time (d) Straight-Chain 2 Fatty Acids With Abalone Saturated 12:0 0.1 0.1 14:0 4.1 4.2 15:0 0.4 0.4 16:0 17.4 17.9 17:0 0.2 0.2 18:0 3.2 3.2 Total 25.4 26.0 Monoenoic 16:l(n-9) 0.2 0.1 16:l(n-7) 4.4 4.5 18:l(n-9) 10.6 10.6 18:l(n-7) 1.8 1.8 18:l(n-5) 0.2 0.2 Total 16.2 17.2 Without Abalone With Abalone Without Abalone 0.1 4.1 0.4 IS.O 0.2 3.3 26.1 0.1 4.5 10.9 2.0 0.2 17.7 0.1 4.2 0.4 18.1 0.2 3.3 26.3 0.2 4.6 11.0 2.0 0.2 17.8 0.1 4.2 0.4 18.2 0.3 3.3 26.5 0.2 4.8 11.0 2.1 0.2 18.3 Figures represent straight-chain fatty acids as a percentage proportion of total fatty acids (n = 3). abalone diets compiise up to 60% carbohydrate. The bacteria as- sociated with surface of abalone would also be expected to per- form well on carbohydrates, but this does not seem to be the ca,se. It is unlikely, then, that the predominantly oxidative isolates in this studs would have a great effect on carbohydrate availability dunng 4 days immersion. Fatty Acid Analysis The quantitative increase in total fatty acids is thought to have resulted from a decrease through leaching in amounts of other food components. Many lipids are not water soluble, so a rise in lipid content and total fatty acids (g/lOO g dry wl). as water-soluble nutrients are lost, is expected. Many Gram-negative bacteria, including: Pseudomonas, Al- teromonas, Moraxella, Cytophaga, and Flavobactehwn (Kaneda 1991), all of which have been isolated from the ABCHOW diet, contain greater than 20'7f branched-chain fatty acids. It has been noted, however, that branched-chain fatty acids are more common in Gram-positive bacteria; whereas, straight-chain fatty acids are more common in Gram-negative bacteria (Kaneda 1991 ). Bacterial growth on the diet should, then, be indicated by an increa.se in these fatty acids. No such change was demonstrated in the present study. SEM analysis has demonstrated clearly that bacterial coloniza- tion of the diet occurs after immersion. Chemotaxonomic tech- niques have previously been utilized for ecological studies inves- tigating relatively low-nutrient environments, so very small shifts in the fatty acid spectrum have been observable. However, the chemotaxonomic method chosen failed to demonstrate an obvious increase in microbial biomass. The most likely explanation for this is that it is inadequate for use on samples of high original lipid content. Table 5 indicates that the ABCHOW diet was relatively high in initial fatty acid levels. This high lipid content masks the presence of bacterial lipids, which make up only a very small percentage of total lipids. ACKNOWLEDGMENTS Thanks are due to the University of Tasmania and the Coop- erative Research Centre for Aquaculture for funding and facilities. Marine Shellfish Hatcheries Ltd., for use of their research facility, Stephen Hodson for SEM advice and comments, and Stephen Hin- drum and Deon Johns for technical assistance. LITERATURE CITED Atlas, R. M. 1993. Handbook of microbiological media. CRC Press. Lon- don, pp. 222. Barrow, G. I. & R. K. A. Feltham. 1993. Cowan and Steele's manual for the identification of medical bacteria. Cambridge University Press, Cambndge, UK. pp. 189-238. Bligh, E. G. & W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biocliem. Physiol. 37:911-917. Coote, T. A., P. W. Hone, R. Kenyon & G. B. Maguire. 1996. The effect of combinations of dietary calcium and phosphorus on the growth of juvenile Haliolis laevigata. Aquaculture 145:267-279. Cropp. C. M. & C. D. Garland. 1988. A scheme for the identification of marine bacteria. Ausl. Microhini 9:27-34. Dunstan, G. A.. H. J. Bailie. S. M. Barrett & J. K. Volkman. 1995. Effect of diet on the lipid composition of wild and cultured abalone. Aqua- culture 140:115-127. Fleming, A. E., R. J. Van Bameveld & P. W. Hone. 1996. The develop- ment of formulated diets for abalone: a review and future directions. Aquaculture 140:5-53. Garland, C. D., G. V. Nash & T. A. McMeekin. 1982. The preservation of mucus and surface associated microorganisms using acrolein vapor fixation. J. Microscopy 128:307-312. Garland, C. D., S. L. Cooke, J. F. Grant & T. A. McMeekin. 1985. Inges- tion of the bacteria on and the cuticle of crustose (nonarticulated) coralline algae by post larval and juvenile abalone (Haliotis ruber Leach) from Tasmanian waters. / E.xper. Mar. Biol. Ecol. 91:137-149. Goldblatt. M. J.. D. E. Conklin & W. D. Brown. 1979. Nutrient leaching from pelleted rations. Proceedings of the World Symposium on Finfish Nutrition and Fishfeed Technology. Hamburg. Germany. 1978. vol. 11. pp. 117-129. 1002 BiSSETT ET AL. Harris. J. 1993. The presence, nature, and role of gut microflora in aquatic invertebrates: a synthesis. Microh. Ecol. 25:195-231. Hodson. S. L. & C. Burke. 1994. Microfouling of salmon cage netting: a preliminary investigation. Biofoiilmg 8:3-105. Iriberri. J., B. Ayo, E. Santamaria. 1. Barcina, & J. Egea. 1994. Influence of bacterial density and water temperature on the grazing activity of two freshwater ciliates. Freshw. Biol. 33:223-231. Kaneda. T. 1991. Iso- and anteiso-fatty acids in bacteria: biosynthesis, function, and taxonomic significance. Microbiolog. Rev. 55:288-302. Lewis, T. E., C. D. Garland & T. A. McMeekin. 1985. The bacterial biota on crustose (nonarticulated) coralline algae from Tasmanian waters. Microh. Ecol. 11:221-230. Maguire, G. B. 1996. Effects of extended feed immersion, feeding fre- quency, and feed rate on growth of juvenile green-lip abalone Haliotis laevigata and on composition and microbial colonisation of a formu- lated diet. pp. 89-91. In: P. Hone and A. Fleming (eds.). Proceedings First and Second Annual Abalone Aquaculture Workshops, South Aus- tralian Research and Development Institute. Adelaide. SA. Maguire, G. B., S. Hindrum, D. R., Johns. G. A. Dunstan & M. Cropp. 1996. Effects of tank drainage frequency on growth of juvenile greenlip abalone Haliotis laevigata, pp. 74-83. In: P. Hone (ed.). Proceedings of the Third FRDC/CRC Abalone Aquaculture Workshop. August 1996. Port Lincoln. South Australia. South Australian Research and Devel- opment Institute, Adelaide, SA. McShane, P. E., H. K. Gorfine & I. A. Knuckey. 1994. Factors influencing food selection in the abalone Haliotis nihra (Mollusca: Gastropoda). J. Exper. Mar. Biol. Ecol. 176:15-26. Moriarty, D. J. W. 1986. Bacterial productivity in ponds used for culture of penaeid prawns. M/crofe Ecol. 12:259-269. Muir, P. R. & D. C. Sutton. 1994. Bacterial degradation of microencapsu- lated foods in larval culture. J. World Aquacul. Sac. 25:371-377. Pearl. H. 1978. Microbial organic carbon recovery in aquatic ecosystems. Liimiol. Oceanog. 23:927-935. Prieur, D., G. Mevel, J. L. Nicolas, A. Plusquellec & M. Vigneulle. 1990. Interactions between bivalve molluscs and bacteria in the marine en- vironment. Oceanog. Mar. Biol. .Ann. Rev. 28:277-352. Rosowski, J. R., K. D. Hoagland, S. C. Roemer & K. W. Lee. 1981. Im- proving the image of complex and delicate biological surfaces. 5a;ii- nmg 4:181-187. Sakata, K. 1987. Feeding attractants and stimulants for marine gastropods, pp. 115-129. In: Scheuer, P.J. (ed,). Bioorganic Marine Chemistry, vol. 3. Schlegel, H. G. 1993. General microbiology, 7th ed. Cambridge University Press, Cambridge UK. pp. 606. Shigueno. K. 1975. Shrimp culture in Japan. Association for International Technical Promotion, Tokyo. 153 pp. Sokal, R. R. & F. J. Rohlf 1987. Introduction to biostatistics (2nd ed.). W. H. Freeman. New York. pp. 363. Solic, M. & N. Krustulovic. 1994. Role of predation in controlling bacterial and heterotrophic dinoflagellate standing stocks in the coastal Adriatic Sea: seasonal patterns. Mar. Ecol. Prog. Ser. 114:219-235. Jounud of SlwUfisb Research. Vol. 17. No. 4. KJO.VIOOS. 1998. ABUNDANCE, RECRUITMENT, AND MORTALITY OF PACIFIC LITTLENECK CLAMS PROTOTHACA STAMINEA AT CHUGACHIK ISLAND, ALASKA WILLIAM R. BECHTOL AND RICHARD L. GUSTAFSON Alaska Department of Fish and Game Division of Commercial Fisheries Homer. Alaska 99603 ABSTRACT From 1992 to 1996, the Alaska Department ot Fish and Game annually surveyed the Pacific littleneck clam Pruloiluica shiminea (Conrad, 1857) population at Chugachik Island, Alaska. Estimates based on randomly placed survey quadrats indicated the population declined from 7,2 million clams in 1992 to 3.3 million clams in 1995 and increased to 5.5 million clams in 1996. Survey hiomass estimates similarly declined from 136,887 kg in 1992 to 65.852 kg in 1995, and then increased to 1 15,495 kg in 1996. Annual harvest hiomass. mean weight-at-age, and survey abundance and age composition data from 1992 to 1996 were also used as inputs in an age-structured model of the Chugachik clam population. The age model used independent selectivity curves to estimate age-specific recruitment to the fishery and the survey. Age of 505^ selectivity to the fishery ranged from 7 to 11 years, depending upon model emphasis of survey age composition data. This agreed well with size-at-age data, indicating few clams recruited to legal harvest size of 38 mm prior to age 5 and only 50% had recruited by age 7. Length-based estimates of annual recruitment to legal size ranged from 6 to \2% of the population, averaging 10%. Greater age model emphasis on survey age data generally increased both estimated survival and the age when 50% of a cohort recruited into the fishery, and decreased the age when 50% of a cohort recruited into the survey. Age-model estimates of population abundance also varied with weighting applied to survey age composition data. Population abun- dance trends from the model agreed well with survey trends from 1992 to 1995. although the model usually exceeded survey estimates. Model estimates were 15 to 25% less than survey estimates for the 1996 population, probably because of a lag in model response to abundance trend changes. KEY WORDS: Littleneck clams, Protothacu suimmeu, age-model. Alaska INTRODUCTION Hardshell clams have long been an important component of the recreational and commercial fisheries in Cook Inlet. Alaska. The commercial fishery dates to the 1950s, when butter clams Saxido- mus giganleiis (Deshayes, 1839) were sold in canned and fresh markets. Sales of canned clams contaminated with paralytic shell- fish poisoning (PSP) from southeast Alaska subsequently led to a market collapse in the late 1950s. The commercial hardshell clam fishery in Cook Inlet re-emerged, targeting Pacific littleneck clams Protothaca staminea (Conrad, 1857) in 1986 after the Alaska De- partment of Environmental Conservation (DEC) certified the Chugachik Island beach in Kachemak Bay, Alaska, for commer- cial harvesting. Commercial harvests during 1986 to 1991 aver- aged 7.107 kg of clams annually from Chugachik Island (Fig. I). In addition, recreational diggers accounted for 77% of all hardshell clam harvests in southern Cook Inlet from 1986 to 1996. although beach-specific harvest information is unavailable (Scott Meyer, Alaska Department of Fish and Game. Homer. Alaska, unpub- lished data). The Alaska Department of Fish and Game (ADF&G) initiated annual surveys on Chugachik Island in 1992. because this area had the longest history of DEC certification for commercial clamming in Kachemak Bay. Average harvests increased to 14.876 kg annu- ally during 1992 to 1994 (Gustafson 1995). The commercial fish- ery was closed in 1995 and 1996 after ADF&G surveys indicated 3 consecutive years of abundance declines. Management strategies for the commercial hardshell clam fishery in Kachemak Bay now include: minimum legal size of 38 mm ( 1.5 inches); 1 April annual registration deadline; fishing district segregation into two groups that open on alternate years; quarterly harvest allocations with maximum quotas for specific beaches; closures in areas of high recreational use; and closures during 1 November to 15 March when ambient air temperatures are below freezing and during weekends from 15 May to 15 September. To evaluate commercial fishery impacts on Pacific littleneck clams at Chugachik Island, we used survey data in a length-based inodel to estimate age-specific recruitment to legal size. Our field surveys were designed to estimate legal clam abundance in the study area. Most clam studies wash the removed substrate through screens (Paul and Feder 1973). However, because our budget and survey time was limited, we did not wash the removed substrate through screens. As a result, as the survey design developed, we were able to sample more quadrats during a site visit, but with the recognition that our survey was biased toward larger clams be- cause of a lack of substrate screening. To back-calculate the sub- legal component of the population and estimate the true clam population abundance, recruitment, and mortality, we developed an age-structured model that accommodated survey bias through selectivity functions. MATERIALS AND METHODS Survey Data Abundance, biomass. and age composition of the hardshell clam population at Chugachik Island in Kachemak Bay. Alaska were estimated from surveys conducted in May during 1992 to 1996 (Gustafson 1995). The littleneck clam bed. defined to include habitat located between the -1.5 m mean low water level and the blue mussel Mytihis ediiliis (Linnaeus. 1 758) bed. was estimated to encompass 61.254 m''. Substrate in the clam bed was a mi.xture of 1 to 8-cm coarse rock and muddy sand (Gustafson 1996). During tides exposing the beach to the -1.5 m mean low water level, substrate was removed by hand digging with a rake to a depth of 30 cm within randomly placed quadrats measuring 0.5 m x 0.5 m. The number of quadrats sampled ranged annually from 12 in 1992 1003 1004 Bechtol and Gustafson ISO 1211 1 — ^Hmtst ^ SurHEJ BOITBSS ■u 'Z -.- Effort ((Sggers) 0. n in .n I ifl .n .f . , ..20 Year Figure 1. Estimates of survey biomass and commercial fishery harvest and effort for Pacific littleneck clams at Chugachik Island, Alaska. 1986 to 1996. to 33 in 1995 (Table 1). During substrate removal and replace- ment, all observed clams were retained. Clams were transported to the laboratory' to obtain age, weight, and length data; weighl-at-age was estimated from samples collected in 1997 only. Clams were aged by counting concentric growth rings on the external surface of the clamshell. The use of these rings to indicate annual growth in littleneck clams has been validated through mark-recapture and size distribution studies (Houghton 1973, Paul and Feder 1973). Although cautioning against the use of growth lines as annuli for littleneck clams in a clean-sand habitat. Peterson and Ambrose (1985) found that specimens deposited a single growth line over 12 months in a muddy-sand environment, as is found at Chugachik Island. All age and size-at-age data used in the model were derived from survey samples. A simple random sample design was used to estimate litdeneck clam abundance at Chugachik Island. Abundance was calculated by multiplying the mean density of littleneck clams in sample quadrats (0.25 m") by the total area where littleneck clams were found. Standard variance estimates for simple random sampling were used to calculate the variance (Cochran 1977). The sampling fraction of quadrats was less than Kf at Chugachik Island. Finite population corrections were not included in the variance estimate of abundance, because finite population coiTCctions can generally be ignored if the sampling fraction does not exceed 5% (Cochran 1977). We estimated annual fishery recruitment using length-at-age survey data and assuming knife-edged recruitment at the minimum legal size for the fishery. For the length-based model we calculated Mainland 25 5 Km Figure 2. Study area showing sequential quadrat selection during the 1996 survey as an example of the simple random survey design at Chugachik Island. Alaska. recruitment as differences in mean percentage legal clams by age class. For age class <(. the mean proportion of legal clams p„ among survey years was calculated from the following: (1) where l„ ,, is the number of legal clams age a clams in year y. and n,, , is the abundance of age a clams in year y. Mean annual rate of recruitment to age class a was estimated as the difference between the proportions of legal clams in ages a and a - I using the fol- lowins: = /'„ -/'„-! ■ (2) TABLE 1. Estimates of annual sur\ey abundance and length-based recruitment for Pacific littleneck clams at Chugachik Island, 1992 to 1996. Sample Quadrats n Sample Densities Annual Abundance Annual Recruitment Year Mean (Clams/m") SD (Clams) 95% CI. (Clams) Percentage 1992 12 117.7 62.98 7.207.5(12 ±2.923.9.^0 858,548 11.9% 1993 16 89.8 52.68 -5.497.507 ±1.767.208 463.089 8.4% 1994 33 79.3 74.35 4.888.737 ±1.791.592 278.764 5.7% 1995 35 53.3 38.24 3.262.213 ±1,021.592 369.098 11.3% 1996 33 88.2 67.62 5.405.201 ±1.886.510 628.649 1 1 .6% Mean 26 85.7 5.252.232 519.6.W 9.9% LiTTLENECK ClAMS AT CHUGACHIK ISLAND 1005 The total year y recruitment R^ was the cumulative products of age class abundance and age-specific recruitment rates as in the fol- lowina: ^('\, X "<,.,)■ (3) Age-Structured Model Our primary objectives in using an age-structured model were to estimate natural mortality, fishery selectivity, survey selectivity, and annual population abundance for 1992 to 1996. Information supplied to the age model included survey data and commercial harvest data. Commercial harvest weights were obtained from ADF&G fish tickets during 1992 to 1994 when the littleneck clam fishery occurred. Age-structured models that incorporate hetero- geneous data have been reviewed by Hilboni and Walters ( 1992). Megrey (1989), and Quinn and Szarzi (1993). The Chugachik model incorporated auxiliary information, similar to age-based models developed by Deriso et al. ( 1985). In our conceptual model of the annual cycle of events affecting Pacific littleneck clams at Chugachik Island (Fig. 3). age increments occur at the end of winter to coincide with the approximate time of annulus formation. The population is then subjected to age-specific mortality through commercial fishing, followed by natural mortality prior to again incrementing to the next year class. The Chugachik Island resource also incurs unquantified. rec- reational and subsistence harvests. It is likely that strong year classes, once recruited to legal size, are subjected to greater non- commercial mortality than weak year classes. Although noncom- mercial mortality abundance varies annually, annual mortality rates may be moderately stable, because the abundance removals from strong year classes are greater than abundance removals from weak year classes. Thus, the proportion of clams dying from non- commercial sources is assumed to be stable among years and is treated as natural mortality. Natural mortality is described by a single exponential decay function for all years and cohorts. Our age-structured model used a reduction equation to describe annual survival. The number of age-rt clams in a cohort in the spring of year y after winter annulus formation was the following: ^„.i.,., = 5(N.,,, - C„,,) (4) where S is the annual survival rate, a model-estimated parameter, and C„ ,, is commercial fishery harvest. The population model as- sumes that clams from age 2 to 14 are present in the estimated population. Although age classes outside this range were observed in all Chugachik surveys, clams younger than age 2 have not consistently appeared in surveys, and cohorts older than age 14 are a minor component of the population. Age 2 recruitment in 1996 was calculated as the median of age 2 clam abundance estimates for the 1992 to 1995 population years. Through independent logistic functions describing fishery and survey selectivity, the model accommodated differences among age compositions of the underlying population, the field surveys, and the commercial fisheries. Relationships between clam age and fishery and survey selectivity were as.sumed to be constant among years. Annual age compositions in the commercial fishery were estimated by the age-structured model, because commercial har- vests were not sampled. Composition of the annual commercial harvest/, , was estimated from an age-specific selectivity function .V, , and model-estimated cohort abundance usina the followina: L.y = — • and X [^ vvas the survey estimate and A', ,„,„;_., the model estimate of abundance in year y. We used the natural log of clams numbers because a log-normal error structure is commonly asso- ciated with abundance data (Funk 1994). Model sensitivity was examined through several scenarios that varied the emphasis, or weighting, on available data sources. In some cases, model runs were rejected, because they yielded unre- alistic results, such as the age of 50% selectivity being greater than 15 or a population abundance estimate that was negative. RESULTS AND DISCUSSION Siiney and Length-Based Estimates Survey estimates showed population abundance declined from 7.2 million clams in 1992 to 3.3 million clams in 1995 before staging a moderate increase to 5.4 million clams in 1996 (Table 1, Fig. 4). Survey biomass similarly declined from 136.887 kg in 1992 to 65,852 kg in 1995, and then increased to 1 15,495 kg in 1996. Based on size-at-age of littleneck clams among all survey years, the length-based model indicated that few clams recruited to legal size prior to age 5 (Fig. 5). Age-specific clam recruitments were 0.4% for age 5, 5.8% for age 6. 44.3% for age 7, 39.1% for age 8. 9.2% for age 9, and 1.1% for age 10. Cumulative increases in the legal component of the surveyed population indicated age 7 to be the age of 50% recruitment to the commercial fisher)'. Ap- plication of age-specific mean recruitment to estimated annual age composition produced annual recruitment rates ranging from 5.7% to 11.9% and averaging 9.9% during 1992 to 1996 survey time series (Table 1 ). Annual recruitment as a percentage of the total abundance declined during 1992 to 1994. the years of the com- Figure 4. Comparison of Pacific littleneck clam abundance estimates from field surveys (thick line) and from an age-structured model with different weighting of the survey age composition data, Chugachik Island, .\laska, 1986 to 1996. LiTTLENECK CLAMS AT ChUGACHIK ISLAND 1007 100%- 80%. / / / '^ Selected 1 r ^- 0.1 • ••- 2 Percent l/h ' -o- 10 —4- 50 / /' / ' - «- 100 20% ■ / » b ' j ...» _yC^ O 1 ( ) 2 4 6 8 10 12 14 16 Clam Age (>ears) Figure 5. Fishery selectivity for Pacific littleneck clams calculated from a length-based model (thick line) and from an age-structured model applying different weights to survey age composition data. mercial fishery, and increased in 1995 and 1996 when the fishery was closed. Age-Structured Model Estimates An age-structured model was previously used to estimate sus- tained recreational fishery yield for Pacific razor clams Siliqiia panda (Dixon. 1788) in Cook Inlet (Quinn and Szarzi 1993). This model relied heavily on fecundity data and spawner-recruit rela- tionships. In contrast, our approach for Chugachik clams was more generic in dealing with known commercial harvests but unknown recreational removals to evaluate the underlying population abun- dance. Althouch model estimates of the Chugachik Island clam population varied with weighting applied to survey age composi- tion data, results agreed well with the 1992 to 1995 population decrease observed in surveys (Fig. 4). For most weighting options, model estimates of the population slightly exceeded survey esti- mates. This supports the assumption of systematic survey selec- tivity. Some studies both within and outside of the Cook Inlet area have attempted to reduce selectivity by passing the removed sub- strate through mesh screens to reduce the clam nondetection (Gustafson 1996). However, using screens also decreases the num- ber of sample quadrats that can be dug. Through the logistic func- tion, our model accommodates systematic survey selectivity that results from a greater sample rate but with slightly less scrutiny of quadrats. The primary exception to model estimates slightly ex- ceeding survey estimates was 1996. when all model runs suggested 12 T lO-: 2-. t1.0 — •— Surv e> Selectivity — »— Fishery Selectivity — »- Annual Survival ■+- -+- 10 20 30 Survey Age Composition Weighting 40 ■ 0.2 0.0 50 Figure 6. The effect of increased weighting of survey age composition data on the estimated annual survival and the ages of 50% selectivity in the commercial fishery and in abundance surveys for Pacific littleneck clams at Chugachik Island, Alaska. 1008 Bechtol and Gustafson the true population was 15 to ISVc. or up to 1.0 million clams, less than the survey estimate. Because the model tracks abundance throughout the life of a cohort from the age of recruitment to the survey or fishery, a long time series is typically required to follow changes in population trends reliably. Thus, the model may lag behind trend changes detected by the survey. The lack of model response to the 1996 increase could reflect either a response lag or errors in survey estimation. Interestingly, preliminary results from the 1997 survey showed a slight increase over the 1996 survey estimate. The age model estimated that about 50% of a cohort recruited into the fishery at ages ranging from 7 to 1 1 years, depending upon weighting applied to survey age composition data (Fig. 5). In general, increased weighting resulted in increases to estimated survival and age of 50*^ selectivity by the fishery, and decreased the age of 50'7f selectivity by the survey (Fig. 6). Annual survival generally ranged from 50 to 609c. The age of 509^ selectivity in the fishery exceeded that in the survey by I to 6 years, a result con- sistent with field observations. In summary, the age-structured model estimates agreed mod- erately well with length-based model estimates. The age model was found to be sensitive to starting parameters and some anoma- lies observed in model run results were probably attributable to inappropriate initial parameters. Fit of an age-structured model to available data may improve if: ( I ) age composition estimates of the commercial harvests were based on commercial fishery samples rather than model estimates; and (2) noncommercial har- vests could be separated from natural mortality. ACKNOWLEDGMENTS The comments of Steve Fried, Robert Wilbur, Tim Baker, Pat- rick Sullivan, and several anonymous reviewers helped clarify this manuscript. Individuals contributing to field data collection in- cluded Tom Sigurdsson. Greg Demers, Trish McNeill, Ted Otis, Daisy Morton, Henry Yuen, Phil Cowan, Mike Parish, Mike Hol- man, Tracy Gotthardt. Nicky Szarzi, and Scott Meyer. The ASA mode! from which this clam model was derived was originally developed by Fritz Funk as a spreadsheet application for Pacific herring. This manuscript is contribution PP-165 of the Alaska Department of Fish and Game. Commercial Fisheries Division, Juneau, Alaska. LITERATURE CITED Cochran. W. G. l')77. Sampling techniques. 3rd ed. John Wiley & Sons, New York. Deriso. R. B., T. J. Quinn, II & P. R. Neal. 1985. Catch-age analysis with auxiliary information. Can. J. Fish. Aqual. Sci. 42:8215-824. Funk, F. 1994. Forecast of the Pacific herring biomass in Prince William Sound, Alaska, 1993. Alaska Department of Fish and Game, Commer- cial Fisheries Management and Development Division. Regional Infor- mation Rep. 5J94-04. Juneau. Alaska. Gustafson. R. 1995. Kachemak Bay littleneck clam assessinents, 1990- 1994. Alaska Department of Fish and Game, Commercial Fisheries Management and Development Division, Regional Information Report 2A95-19, Anchorage. Alaska. Gustafson. R. 1996. Kachemak Bay littleneck clam assessments. 1995. Alaska Department of Fish and Game. Commercial Fisheries Manage- ment and Development Division, Regional Information Rep. 2A96-12. Anchorage. Alaska. Hilborn. R. & C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics, and uncertainty. Chapman & Hall. New York. Houghton. J. P. 1973. The intertidal ecology of Kiket Island. Washington, with emphasis on age and growth of Pioiorluica staininen and Saxido- niis giganteus (Lamellibranchia: Veneridae). Ph.D. thesis. University of Washington, Seattle, WA. Megrey. B. A. 1989. Review and comparison of age-structured stock as- sessment models from theoretical and applied points of view. Am. Fish. Soc. Syinp. 6:8^8. Paul. A. J. & H. M. Feder. 1973. Growth, recruitment, and distribution of the littleneck clam. Protothaca shiinineu. in Galena Bay. Pnnce Wil- liam Sound. Alaska. Fish. Bull 7\:6b5-6n. Peterson, C. H. & W. G. Ambrose, Jr. 1985. Potential habitat dependence in deposition rate of presumptive annual lines in shells of the bivalve Protothaca staminae. Lethaia 18:257-260. Quinn, T. J., n & N. J. Szarzi. 1993. Determination of sustained yield in Alaska's recreational fisheries, pp. 61-84. In: Proceedings of the In- ternational Symposium on Management Strategies for Exploited Fish Populations. Alaska Sea Grant Rep. 93-02. University of Alaska, Fair- banks. Journal of Shellfish Research. Vol. 17. Nii. 4, 11)04-11)1.^, 1998. REPRODUCTIVE CYCLE OF THE GIANT REEF CLAM PERIGLYPTA MULTICOSTATA (SOWERBY, 1835) (PELECYPODA: VENERIDAE) AT ISLA ESPIRITU SANTO, BAJA CALIFORNIA SUR, MEXICO FEDERICO GARCIA-DOMINGUEZ, BERTHA PATRICIA CEBALLOS-VAZQUEZ, MARCIAL VILLALEJO-FUERTE. AND MARCIAL ARELLANO-MARTINEZ Ccnlro IiUcrdiscipUnario de Ciencias Marinas liisliliilo Polileciiico Nacional La Paz. BCS 23000. Mexico ABSTRACT The reproductive cycle of Perii;hpta muliieostata (S.). was studied at Isla Espiritu Santo, Gulf of California. Mexico, from October 1992 to December 1993. The reproductive activity was present throughout the study period, except in February 1993, nevertheless a distinct seasonality was ob.served with three distinct peaks of spawning activity. A clear relationship between spawning and temperature or photosynthetic pigments concentration was not observed. Spawning occurs all year, but at a lower rate in the months with the lowest water temperatures. KEY WORDS: Reproductive cycle, gametogenesis, bivalves, Veneridae. Perighpta INTRODUCTION The giant reef clam, Periglypta multicostata (Sowerby, 1835), is the heaviest if not the largest of the Panamic members of the family Veneridae. They inhabit the sand ainong rocks at extreme low tide from Gulf of California to Peru (Keen 1971 ). This species is a dominant cotnponent of the zone of the coral Pocillopora elegans in the rocky substrata communities of Isla Espiritu Santo and coexists with three bivalve species. Veiuricokiria isocardia. Megapitaria uurantiaca. and Chione tumens. In Baja California Sur, wild stocks of giant reef clams remain practically untouched and are considered as a potential fisheries resource; however, aquaculture is not recommended (Baqueiro 1989). No research has been done on the biology or life history of this species. Clams are gathered by free diving and are dug out with hands, knifes, or forks. The synchronization of reproductive activity in local popula- tions is very important for successful fertilization. Reproduction seems to be cyclic, with events coordinated on an annual cycle (Eversole 1989). Environmental factors may intluetice the timing of reproduction in clams. The most commonly cited are food avail- ability and temperature (Bayne and Newell 1983. MacDonald and Thompson 1985. Jaramillo et al, 1993). The water temperature and its variation with latitude is used by many authors to attempt to explain reproductive timing in bivalves (Newell et al. 1982. Lozada and Bustos 1984. Manzi et al, 1985, Maiachowski 1988, Hesselman et al. 1989. Garci'a-Domi'nguez et al, 1993), The food availability is used to attempt to explain spawning timing in bi- valves in the sense of that if the spawning coincides with the highest food availability, this enables the larvae to exploit the phytoplankton bloom (Jaramillo et al, 1993, Villalejo-Fuerte et al, 1996a), The lack of biological information for proper management has led to overexploitation and misuse of stocks. It is essential to know the life cycle of the target species, and documentation of the re- productive cycle in a fishery is one necessary step in determining when recruitment might occur. This study describes the reproduc- tive cycle and the spawning season of P. multicostata in relation to the temperature and food availability. MATERIAL AND METHODS Monthly, 20 to 25 specimens of a wild and unexploited popu- lation of P. tnidticoslaia were collected from October 1992 to December 1993 at Isla Espiritu Santo, Bahia de La Paz. Gulf of California. Mexico ( 1 10°24'27"W. 24''28'54"N) by a scuba diver at 3- to 6-m depth, A total of 310 organisms were captured. When the biological samples were collected, water temperature was re- corded. The photosynthetic pigment concentration (mg chloro- phyll/m') in Bahi'a de La Paz. Gulf of California was obtained from satellite-derived information (Trant et al. 1993). this was considered to be an estimation of the food availability for the clams. Mantle, adductor muscles, gills, labial palps, and siphons were removed, keeping only the visceral inass (gonad, liver, and gas- trointestinal tract) and the foot. These tissues were fixed in buff- ered lO^f formalin, A slice of tissue of each clam was obtained from the dorsal area of the visceral mass and embedded in paraffin. Sections 7- to 9-[j,m thick were stained with hematoxylin and eosin (Luna 1968). This method was adopted after verifying, in 25 speci- mens, that gonadal maturity was uniform in different parts of gonad. The reproductive process (either spemiatogenesis or oogenesis) of P. multicostata was categorized in five stages based solely on morphological observations, characterized by the structure of the gonad, presence, absence, and quantity and development of ga- metes (Table 1. Figs. 1. 2. 3). In all the stages of gonadal devel- opment, phagocytes were present in varying proportions. Individuals were sexed by microscopic examination of histo- logical slides. Sex ratios were analyzed with chi-square to test the significance of the deviation from the expected sex ratio of 1 : 1 , for the total sample. The indifferent stage clams \vere not considered. Mean oocyte diameters, and their standard deviation, of six females selected randomly per month were determined from his- tological sections using an eyepiece graticule calibrated with a stage micrometer. At least 100 oocytes sectioned through the nucleus (i.e,. near the maximum diameter) per individual were measured along the longest axis. Individuals with few measurable oocytes and extensive phagocytosis ( "spent" specimens) were not 1009 1010 Garci'a-Dominguez et al. TABLE 1. Developmental Stages of P. multicostata Gonads. Maturity Stage Female Male Indifferent Characterized by presence of acinis with total absence tissue is abundant. Developing Oocytes inside of follicles are conspicuous, young oocytes with pear shape growing attached to folicular walls. The area of connective tissue decreasing. Ripe Free large oocytes present in the lumen with maximum size, few oocytes with pear shape attached to folicular walls. Connective tissue absent. Partially spawned Follicles containing some oocytes and large spaces, while others were empty. Some connective tissue visible. Spent Few residual oocytes, being phagocytized by amebocvtes. No evidence of active oosenesis. of gametes. It is not possible to distinguish the sex. The connective A variable quantity of germinal cells and spermatozoa were present inside the follicles. The area of connective tissue decreasing. Follicles filled with spermatozoa. Other spermatogenic cells restricted to a thick layer on the folicular walls. Connective tissue absent. Follicles partially empty. A marked decrease in the number of spermatozoa filling the lumen. Some connective tissue visible. Follicles collapsed, amebocytes phagocytizing residual spermatozoa. No evidence of active spermatogenesis. considered, using the criteria of Gratit and Tyler ( 1983a) and Grant and Tyler (1983b). RESULTS Spawning activity was present throughout the study period, except in February 1993 (Fig. 4). Nevertheless, there seems to be a distinct seasonality in the reproductive cycle, because P. multi- costata has fluctuations in its reproductive intensity, showing three distinct peaks of spawning activity during the study period: Octo- ber to December 1992, July to September 1993, and November to December 1993. Giant reef clams in the indifferent stage were observed every month except October 1992. Developing clams were present over the year except in February, September, November, and December 1993. The highest frequency of ripe organisms was present in October 1992. March, June to September, and November 1993. Partially spawned individuals were observed all year, except in Vl'^'K Figure 1. Indifferent stage; scale bar = 5(1 nm. February. The highest frequency of partially spawned individuals was observed from June to September. The spent stage was re- corded throughout the year, except in April. Of 310 ciatns examined. 149 (48.1%) were females and 83 (26,8%) males. The remaining 78 (25,1%) were undifferentiated. The sex ratio of the sexed sample (1,8 F: 1 m, n = 232) differs significantly (p «,05) from the expected ratio of 1:1. During the year, the mean oocyte diameter was > 47 |jim, except in October, when it was 41 \x.m (Fig. 5). In February, all the clams were spent and indifferent, so there were no oocytes. The pattern observed in oocyte diameters was consistent with the his- tological observations, which suggests spawning throughout the year, with the exception of February 1993. The standard deviations were wide in all months, indicating the presence of both small and large oocytes characteristics of ripe and spawning stages. The water tetnperature during the study period varied from 22°C to 31°C. The highest values were in October 1992 and Au- gust 1993. and the lowest values were in February and March 1993 (Fig. 5). Photosynthelic pigment concentration (mg chlorophyll/m'') in Bahi'a de La Paz was greater in the colder months than in the wanner ones. The maxiiTiuin value was in January (4.87 mg chlo- rophyll/m''). and the minimum was in Septetnber (1.36 mg chlo- rophyll/in'') (Fig. 5). DISCUSSION The annual reproductive cycle of P. midticoslata at Isla Espi'ritu Santo showed seasonality, with a protracted period of reproduction indicated by the consistent presence of spawning activity through- out the study period, with the exception of February 1993. Other species of bivalves, abundant in this locality, such as Megapitaria iiiirantiaca (Garci'a-Dominguez et al. 1994) or Pinctada mazat- Uiiiica (Garcia-Dominguez et al. 1996), have no .sea.sonal repro- dtictive cycles, and their spawning activity is continuous. In other bivalves, such as Meicenaria spp.. the spawning is essentially continuous in lower latitudes, but there are still cycles (Hesselman et al. 1989). Among other venerid clams from other localities along the Mexican Pacific coast, several other species lack distinct seasonal Reprodlictive Cycle of Periglypta multicostata 1011 <£f^' ^^ ,-^^**^<* r Figure 2. Photomicrographs of gonadal stages of the fetnale giant reef clam, P. multicostata. (A) developing stage, (B) ripe stage, (C) partially spawned stage, and (D) spent stage; scale bar = 50 pm. Figure 3. Photomicrographs of gonadal stages of the male giant reef clam, P. multicostata. (Ai developing stage, (B) ripe stage, (C) partially spawned stage, and (D) spent stage; scale bar = 50 pm. 1012 Garci'a-Dominguez et al. 100 N D J F M A M J J 1992 1993 □ INDIFFERENT DEVELOPING RIPE I PART SPAWN □ SPENT Figure 4. Reproductive cycle of P. multicostata at Isla Espi'ritu Santo. Relative frequency of gonadal stages between October 1992 and De- cember 1993. Observations of males and females are combined. reproductive cycles; Megapitaria aurantiaca. M. sqiialida. Dosinia ponderosa (Baqueiro and Stuardo 1977). Chione unda- tella (Baqueiro and Masso 1988). and M. sqiialida (Villalejo- Fuerte et al. 1996b). Other bivalves, such as Meixenaria merce- naria. displayed a synchronized polymodal breeding pattern, al- though not every year. In some years, it is polymodal and in others, it is bimodal with continuous spawning (Hetfeman et al. 19891. The sex ratio of the sample differs significantly from the ex- pected ratio of 1:1, with females being dominant, which suggests females outnumbered males in the population. The same was ob- served for the pearl oyster P. mazatkinka in the same locality (Garcia-Dominguez et al. 1996). this condition may be related to the fact that P. mazatlanica is a protandric hermaphrodite (Sevilla 1969, Saucedo and Monteforte 1994. Garcia-Dominguez et al. 1996). However, in the case of P. multicostata. evidence of he- maphroditism was not observed. Also, the fact that within the population as a whole, the majority are females is considered typical of freshwater and brackish water bivalves (Morton 1985). Oocyte diameters reflect the gametogenic cycle, thus minimum diameters coincide with the developing stage, and maximum di- ameters coincide with the mature and partially spawned stages. This pattern is similar for other species such as Argopecten ciirii- laris. Glycynieris gigantea, and Laevicardium elatum (Villalejo- Fuerte and Ochoa-Baez 1993. Villalejo-Fuerte et al. 1995. Villa- lejo-Fuerte et al. 1996a). MacDonald and Thompson (1985) suggested that bivalve ga- mete production is strongly influenced by such environmental fac- tors as temperature and food availability set in a seasonal context. The reproductive cycle of Periglypta multicostata was not clearly related to the water temperature. The same has been observed for such other venerid clams as Megapitaria aurantiaca. M. squalida. and D. ponderosa from Bahi'a Zihuatanejo (Baqueiro and Stuardo 1977), M. aurantiaca from Isla Espi'ritu Santo (Garcia-Dominguez et al. 1994). and in other bivalves such as Pinctada mazatlanica (Garci'a-Dominguez et al. 1996) from Isla Espiritu Santo. The re- lation between the temperature and spawning of other bivalves of the Mexican Pacific coast has been well documented in several species; Modiolus capax (Garza-Aguirre and BUckle-Rami'rez 1989), Chione californiensis (Garcia-Dominguez et al. 1993). Gly- cymeris gigantea (Villalejo-Fuerte et al. 1995), and Laevicardium elatum (Villalejo-Fuerte et al. 1996a). Accordingly with Sastry (1970). although temperature affects reproduction, other environ- mental factors seem to interact with it in determining the pattern of annual gonad activity in a given geographical area. It is likely the variation in annual reproductive activity of a species will be the phenotypic response of a single genotype. Food availability has been related to the timing of reproduction in some bivalves (Sastry 1979, Bayne and Newell 1983, Mac- Donald and Thompson 1985, Jaramillo et al. 1993). For example, in Chlamys amandi. the spawning time seemed to be related to food availability (Jaramillo et al. 1993); whereas. Hinnites gigan- teus showed no correlation between food availability and spawn- ing (Malachowski 1988). The reproductive cycle of P. multicos- tata at Isla Espi'ritu Santo did not exhibit a clear relation with food availability, because the spawning extends all year, independent of food availability, expressed as photosynthetic pigment concentra- tion. In Pinctada mazatlanica from the same location, maximum spawning did not coincide with maximum food availability (Gar- cia-Dominguez et al. 1996). Although, spawning of P. multicostata did not seem to be re- lated to temperature or food availability, its spawning may be triggered by other factors, such as day length, a particular lunar E ^ tu \~ tu < Q tu I- > o o o 80 70 . 60 . 50 . 40 30 20 ■ ' ■ ■ o tu en < tr LU 26 Q. LU 24 I- 22 ^t5 ■ ■ ■ ■ J ' O N D J 1992 M A M J J 1993 O N D Figure 5. Mean oocyte diameters of P. multicostata (bars = standard deviation), water temperature, and photosynthetic pigment concentra- tion from Isla Espiritu Santo, BCS, Mexico. Reproductive Cycle of Periglypta multicostata 1013 phase, salinity fluctuations, tidal cycle, or a combination of se\ eral of these. Unfortunately, in this study did not consider these other environmental factors. ACKNOWLEDGMENTS Our gratitude to the Direccion de Estudios de Posgrado e In- vestigacion del Instituto Politecnico Nacional (IPN), who gave us the funds for this work, to Ciro Arista. Arturo Tripp Q.. and Jose Luis Castro O. for their help in collecting samples, and Dr. Ellis Glazier for his editorial help on the English manuscript. We ac- knowledge the fellowships of Comision de Operacion y Fomento de actividades Academicas del IPN to F. Garcia-Dominguez and M. Villalejo-Fuerte. F. Garci'a-Domi'nguez is a Ph.D. student from PICP of the Universidad de Colima. Mexico. LITERATURE CITED Baqueiro, E. 1989. Clam culture in Mexico: pa.st. present, and future, pp. 383-394. In: J. J. Manzi and M. Castagna (eds.). Clam Mariculture in North America. Elsevier Science. New York Baqueiro. E. & J. A. Masso. 1988. Variacione.s p^lblaclonale^ y reproduc- cion de dos poblaciones de Chione undalellu (Sovverby. 1835) bajo diferentes regimenes de pesca en la Bahia de La Paz, B.C.S.. Mexico. Cieiic. Pesq. but. Nai de la Pesca. Mexico 6:51-67. Baqueiro. E. & J. Stuardo. 1977. Observaciones sobre la biologia. ecologia y explotacion de Megapilaria auranliaca (Sow, 1835), A/, squalida (Sow, 1835) y Dosinia ponderosa (Gray, 1838) (Bivalvia: Veneridae) de la Bahi'a de Zihutanejo e Isla Ixtapa, Gro.. Mexico. An. Centra. Cienc. del Mar y Limnoi Univ. Nal. Anton. Mexico 4:161-208. Bayne. B. L. & R. C. Newell. 1983. Physiological energetics of marine molluscs, pp 491^98. In: A. S. M. Saleuddm and K. M, Wilbur (eds.). The Mollu.sca. vol. 4. Academic Press. San Diego. Eversole, A. G. 1989. Gametogenesis and spawning in North American clam populations: implications for culture, pp 75-108. In: Manzi. J. J. and M. Castagna (eds.). Clam Mariculture In North America. Elsevier Science, New York. Garci'a-Domfnguez, P.. G. Garcia-Melgar & P. Gonzalez-Ramirez. 1993. Reproductive cycle of the clam Chione californiensis ( Broderip. 1 835 ). in Bahia Magdalena. Baja California Sur. Mexico. Ciencias Marina.s 19:15-28. Garcia-Dominguez, P., S. A. Garcia-Gasca & J. L. Castro-Ortiz. 1994. Spawning cycle of the red clam Megapitaria auranliaca (Sowerby. 1831) (Venendae) at Isla Espi'ritu Santo. Mexico. / Shellfish Res. 13:417-423. Garcfa-Dominguez, P., B. P. Ceballos- Vazquez & A. Tripp-Quezada. 1996. Spawning cycle of the pearl oyster Pinclada niazailanica {Han- ley. 1836) (Pteriiidae) at Isla Espi'ritu Santo, Mexico. J. Shellfish Res. 15:297-303. Garza-Aguirre. M. C. & L. p. Biickle-Ramirez. 1989. Ciclo reproductivo del mejillon Modiolus capax (Conrad. 1837) (Bivalvia. Mytilidae. Ani- somyana) en la Bahi'a de Los Angeles. Baja California. Mexico. An. Inst. Cienc. del Mar y Limnoi. Univ. Nal. Auldn. Mexico 16:157-160. Grant, A. & P. A. Tyler. 1983a. The analysis of data in studies of inver- tebrate reproduction. I. introduction and statistical analysis of gonad indices and maturity indices. Int. J. Invert. Reprod. 6:259-269. Grant. A. & P. A. Tyler. 1983b. The analysis of data in studies of inver- tebrate reproduction. II. the analysis of oocyte size/frequency data, and comparison of different types of data. Int. J. Invert. Reprod. 6:259-269. Heffeman. P. B.. R. L. Walker & J. L. Carr. 1989. Gametogenic cycles of three bivalves in Wassaw Sound. Georgia. I. Merccnaria mercenaria (Linnaeus, 1758). / Shellfish Res. 8:51-60. Hesselman. DM.. B.J. Barber & N.J. Blake. 1989. The reproductive cycle of adult hard clams, Mercenaria spp. in the Indian River Lagoon, Florida. J. Shellfish Res. 8:43^9. Jaramillo. R., J. Winter, J. Valencia & A. Rivera. 1993. Gametogenic cycle of the chiloe Scallop (Chlamys amandi). J. Shellfish Res. 12:59-64. Keen, A. M. 1971. Sea Shells of tropical west America. Stanford Univer- sity Press. Stanford. CA. Lozada. E. & H. Bustos. 1984. Madurez sexual y fecundidad de Venus antiqua antiqua King y Broderip 1835 en la Bahia Ancud (Mollusca: Bivalvia: Veneridae). Rev. Biol. Mar. Valparaiso 20:91-112. Luna. L. G. (ed.). 1968. Manual of histologic staining methods of the Armed Forces Institute of Pathology (3rd ed.). McGraw-Hill. New York. MacDonald. B. A. & R. J. Thompson. 1985. InOuence of temperature and food availability on the ecological energetics of the giant scallop Pla- copecten magellanicus. II. reproductive output and total production. Mar. Ecol. Prog. Ser. 25:295-303. Malachowski, M. 1988. The reproductive cycle of the rock scallop Hinnites giganteus (Grey) in Humboldt Bay. California. / Shellfish Res. 7:341- 348. Manzi. J. J.. M. Y. Bobo & V. G. Burrell. 1985. Gametogenesis in a popu- lation of the hard clam. Mercenaria mercenaria (Linnaeus), in North Santee Bay. South Carolina. Veliger 28:186-194. Morton, B. 1985. The reproductive strategy of the mangrove bivalve Polr- mesoda (getoina) erosa (Bivalvia: Corbiculoidea) in Hong Kong. Ma- lacolog. Rev. 18:83-89. Newell, R. I.. T. J. Hilbish. R. K. Koehn & C. J. Newell. 1982. Temporal variation in the reproductive cycle of Mytilus edulis (Bivalvia: Mytil- idae) from localities on the East Coast of the United States. Biol. Bull. 162:229-310. Sastry, A. N. 1970. Reproductive physiological variation in latitudinally separated populations of the bay scallop, Aequipecten irradians La- marck. Biol. Bidl. 138:56-^5. Sastry. A.N. 1979. Pelecypoda (excluding Ostreidae). pp 131-192. In: A. C. Giese & J, S. Pearse (eds. l. Reproduction of Marine Inverte- brates. Academic Press. San Diego. Saucedo. P. & M. Monteforte. 1994. Breeding cycle of pearl oysters Pinclada mazatlanica and Pleria sterna in Bahia de La Paz. South Baja California. Mexico (Abstract: Pearls '94). / Shellfish Res. 13:348-349. Sevilla. M. L. 1969. Contribucion al conocimiento de la madreperla Pinctada mazatlanica (Hanley. 1945). Rev. Soc. Mex. Hist. Nat. 30: 223-262. Trant. A. V.. E. Smith. J. Hyon, R. Evans, O. Brown & G. Pedman. 1993. Satellite-derived multichannel sea surface temperature and phytoplank- ton pigment concentration data: a CD-ROM set containing monthly mean distribution for the global oceans (user's manual). Jet Propulsion Laboratory DAAC, Pasadena, CA. 32 pp. Villalejo-Puerte, M. & R. I. Ochoa-Baez. 1993. The reproductive cycle of the scallop Argopecten circularis (Sowerby, 1835) in relation to tem- perature and photoperiod. in Bahi'a Concepcion. B.C.S.. Mexico. Cien- cias Marinas 19:181-202. Villalejo-Puerte. M.. p. Garcia-Dominguez & R. I. Ochoa-Baez. 1995. Re- productive cycle of Glycxmeris gigantea (Reeve. 1843) (Bi\alvia: Gly- cymendidae) in Bahia Concepcion. Baja California Sur. Mexico. Ve- /iger 38:126-132. Villalejo-Fuerte. M., B. P. Ceballos-Vazquez & F. Garcia-Dominguez. 1996a. Reproductive cycle of Laevicardium elatmn (Sowerby. 1833) (Bivalvia: Cardiidae) in Bahi'a Concepcion, Baja California Sur, Mexico. ,/. Shellfish Res. 15:741-745. Villalejo-Fuerte. M., G. Garci'a-Melgar, R. I. Ochoa & A. Garci'a-Gasca. 1996b. Ciclo reproductivo de Megapitaria squalida (Sowerby, 1835) (Bivalvia: Veneridae) en Bahia Concepcion, Baja California Sur, Mexico. Boletin Cientifico (Santa Fe de Bogota) 4:29-39. Journal of Shellfish Research. Vol. 17, No. 4. 1015-1036. 1998. THE EVOLUTION OF A MUNICIPAL QUAHOG (HARDCLAM), MERCENARIA MERCENARIA MANAGEMENT PROGRAM, A 20- YEAR HISTORY: 1975-1995 SANDRA L. MACFARLANE Town of Orleans ConseiTcition Department Orleans. MA 02653 ABSTRACT Local municipal control of shellfisheries has been in existence in Massachusetts since 1942. The shellfish management programs of Orleans, MA. a town located at the elbow of Cape Cod. have evolved from transplants of native stock to use of hatchery-raised seed. For the period 1975 to 1995, the town utilized several forms of nursery culture in three separate estuaries including bottom culture, raft culture, a municipal hatchery, a land-based upweller system, tidal upweller. and floating trays. Financial constraints, as well as political and social perceptions determined the extent of the program at any given time. Management decisions were based primarily on survival of the seed rather than such factors as fast growth. The most successful method was a land-based upweller system with which we raised 1 million seed per year at 95% survival that were transplanted throughout the town. Survival in the field was directly related to water temperature at time of planting, which was most successful when water temperature was about 45=F(7=Cl. KEY WORDS: Quahog. hard clam seed. Mercenaria mer xenana, si hellfish management, nursery culture, aquaculture INTRODUCTION Efforts to observe growth of northern quahogs. also known as hard clams (Mercenaria mercenaria Linne) or to increase the natu- ral production have been attempted since the early part of this century. Belding (19121 described both bottom culture and off- bottom methods used in four separate locations in Massachusetts. Haskin ( 1952). Carriker ( 1959), and Caniker ( 1961 ) added data to our understanding of environmental aspects of water and sediment that increase quahog growth, survival, and abundance. Larval culture began with Wells, who cultured five molluscan species through metamorphosis and patented his methods in 1933 (Manzi and Castagna 1989). Loosanoff and his colleagues at the Bureau of Commercial Fisheries Laboratory in Milford. Connecti- cut (now National Marine Fisheries Service Laboratory) are cred- ited with numerous developments in rearing of bivalve moUusks from spawning through the juvenile stage (Loosanoff and Davis 1950), (Loosanoff and Davis 1951). and (Loosanoff and Davis 1963). Once larvae and juveniles were readily available, knowl- edge of quahog culture expanded enormously (Judson et al. 1977. Manzi and Castagna 1989. Rice, 1992). As a direct result of the culture efforts, entrepreneurs developed commercial hatcheries and field grow-out businesses. Towns in Massachusetts that manage their own shellfish resources pur- chased seed from hatcheries. Municipal shellfish programs, such as those on Cape Cod and Martha's Vineyard. Massachusetts, utilized this source of hatchery seed to develop their propagation schemes. George Souza. Shellfish Constable of Falmouth. MA was the first shellfish officer to take advantage of these seed. Working with staff biologists from the Massachusetts Division of Marine Fisheries, in 1972, he developed an off-bottom culture system for raising small seed to be planted in the wild. By 1977. eight Cape Cod towns: Bourne. Barnstable. Chatham. Dennis. Eastham. Or- leans. Wellfleet. and Yarmouth and the Martha's Vineyard Shell- fish Group followed his example. Information was shared among the towns through workshops sponsored by the Division of Ma- rine Fisheries. The focus of this paper is to describe the munici- pal propagation efforts that took place in Orleans from 1975 to 1997. Municipal Management Massachusetts is one of the few states where municipal man- agement of shellfish resources is the norm. Towns are encouraged to promote and protect those resources under broad guidelines set by the Commonwealth. Towns appoint a shellfish constable and staff, promulgate and enforce local regulations, issue harvest li- censes, determine harvesting techniques, determine areas suitable for harvest, conduct experiments in propagation, and issue licenses for private aquaculture. The state sets size limits, determines the species to be regulated, administers all aspects of contaminated areas, sets maxitnum fees for shellfish leases, and surveys potential lease sites for suitability and compliance with state law. Belding (1912) examined many aspects of quahog life history and concluded that, without culturing this animal, the fishery would collapse. Over half a century later, in 1975. the town of Orleans pur- chased 10.000 seed quahogs from Coastal Zone Resources, a hatchery in North Carolina to begin propagation experiments. Our primary objectives were to determine: 1. if hatchery-raised seed would survive transplant into a pro- tected environment; 2. if seed would survive and grow under varying environmen- tal conditions; 3. limiting factors of seed growth, including density; 4. if seed quahogs would survive in areas devoid of native stock: and 5. predators and estimate loss. Initial field trials were somewhat successful, and we continued and expanded the program throughout the next 14 years. During this period, we developed a number of secondary objectives. These were: 6. to monitor transplants and determine survival; 7. to find economical methods for growing seed; 8. to produce as many quahogs as funds would allow; and 9. to increase quahog stocks in areas of Pleasant Bay that had been marginally productive for 20 or more years. The program is presented chronologically. This emphasizes the transition as we built on success, learned from mistakes, and ex- perimented with alternatives. Each phase of the program was gov- 1015 1016 Macfarlane emed by logic, financial constraints, and management principles. Much of the discussion is based on the observations of the author, only some of which were quantified. The program is divided into eight separate sections as follows: 1. bottom culture; 2. raft (off-bottom) culture; 3. hatchery; 4. upweller; 5. transplants to the natural environment; 6. changes in direction; 7. private aquaculture; and 8. budget constraints. Methods for evaluation varied according to the specific propa- gation culture used at the time. Volumetric counts were made for each shipment of seed delivered. Square foot samples were dug with bottom culture, and survival and growth were estimated; samples of seed (appro.ximately 100 randomly gathered) from the rafts were measured for growth, and the entire harvest was volu- metrically counted to arrive at survival percentages; and volumet- ric counts were made from the upweller technique. Lack of tech- nical staff and a desire to transplant as soon as possible from harvest limited our ability for more detailed statistical analysis. Seed were measured in metric units; construction details are dealt with in standard units. Study Area Orleans is located at the "elbow" of Cape Cod, MA (Figs. 1 and 2). It has three separate embayments within its boundaries; Cape Cod Bay, Nauset/Town Cove, and Pleasant Bay. Although different from one another, all support, or historically have sup- ported, natural populations of quahogs. The Orleans municipal jurisdiction in Cape Cod Bay extends six miles (9.7 km) from shore and includes approximately \.5 miles (2.4 km) of intertidal sand flats that are nonproductive for quahogs because of harsh environmental conditions, including heavy ice buildup and shifting sands. The depth offshore ranges frotn to 25-1- feet (7.5 ml. Abundant beds of quahogs have his- torically been found in much of the deeper waters of the bay and are still harvested commercially by a few medium-sized (35 ft, 10.5 m) draggers. The Nauset estuary (2.333 acre) is a very productive area (Ro- man et al. 1989), shared by the towns of Orleans and Eastham and protected from the Atlantic Ocean by a migrating barrier beach. Approximately 1,150 acres are in Orleans, but Orleans and East- ham share a reciprocal fishing agreement. Quahogs are found along the edge of the Town Cove, in eelgrass (Zoslera marina), sand/silt/mud combinations, and along the steep gradient that leads to deeper water (6-18 ft; 2-6 m), as well as many parts of Nauset Harbor. They are harvested by hand scratchers or bull rakes. Pleasant Bay. a larger estuary, is shared with Chatham. Har- wich, and Brewster (no reciprocal fishing agreements); 3.500 acres are within the boundary of Orleans. A migrating barrier beach extending approximately 12 miles (19 km) to Chatham Inlet pro- tects the bay from the Atlantic Ocean. Of the three estuaries. Pleasant Bay has had the least stable quahog population (based on landings). This may result from the barrier beach dynamics. A large set of seed (less than 2" legal size. [50 mm] in longest diameter) was discovered in Pleasant Bay in the late 1950s. Gates (1964) conducted a survey of Big Bay. making 33 samples in 27 acres, and found up to 1 80/0.3 m" with an average density of about 50/0.3 m". He estimated a standing crop of 60.7 million animals worth $1,026 million at that time, providing a steady supply of shellfish and employment. They were well known in the market- place for their long shelf life and could be easily identified as originating from Pleasant Bay. because their growth rings were almost indistinguishable from one another. The population lasted until the early 1970s, at which time, quahogs became rare in the bay. with no recurring set. Figure I. Locus map of area, Cod. Cape Cod is the easternmost land mass of Massachusetts. The Town of Orleans is situated at the "elbow" of Cape A Municipal Quahog (Mercenaria mercenaia) Management Program 1017 Figure 2. Map of Orleans, Cape Cod Bay, Nauset/Town Cove and Pleasant Bay. BOTTOM CULTURE We used bottom culture of seed for 3 years at a total o'i 20 locations (Fig. 3). Three different enclosure designs were deployed in all three estuaries. 1975 We chose 10 separate locations, with varying sediment, current, and other environmental conditions, to plant the 10,000 seed qua- hogs, approximately 8-mm in size. Frames, 3' x 6' were con- structed of 1" X 2" '"strapping" on end to which a 3/16" mesh netting was stapled. Each intertidal area was raked to loosen the substrate and remove any visible predators, and the seed was broadcast within the raked area (Fig. 4). The frame was placed over the seed with the edges buried and was secured by stakes at the corners and attached to the frames. Results Table 1 provides details of the experimental planting for 1975. We considered survival above 90% to be excellent, 75 to 907r very good, 50 to 75% good, 25 to 50% poor, and less than 25% very poor. With the exception of the frame at Namequoit Point, which was lost in a storm, survival in the summer was excellent. Quahogs grew from 8 mm in July to 1 1 to 18 mm in October, depending, in part, on location. The seed at Snow Shore exhibited the least growth. Native stock in the general vicinity also had slow growth (and unusually thick shells, often associated with slow growth). Survival after the winter was excellent in two locations, very good at two, and disappointing in five locations where the survival was poor or very poor, despite an unusually mild winter. Most of the stock at Snow Shore and Meetinghouse River and half from Asa's Landing and the Yacht Club had died. The frame at Skaket had disappeared. We observed no correlation between sediment type or specific estuary and loss. In the spring at Asa's Landing, we observed live animals interspersed with empty shells in the top 25 mm of substrate. Below them, at around 33 mm, patches of black empty shells were found in black sulfurous smelling sand, adjacent to live seed in nonblack sediment. This initial observation was seen throughout the years in relation to the phenomenon known as "winter kill." 1018 Macfarlane ORLEANS SEED BOTTOM CULTURE QUAHAUGS Figure 3. Bottom culture experiments were conducted in 20 separate locations in all three estuaries over a 3-year period. 1976 In 1 976, we expanded the program and purchased 280,000 seed from two sources: 150,000 from Coastal Zone Resources and 130,000 from ARC (Aquacultural Resources Corp.). a local hatch- ery. Nearly 150.000 were used for intertidal bottom culture. A portion. 30,000, were large enough to plant directly into the natural environment, and the remaining 100.000 were grown in floating rafts (see next section). The bottom culture enclosures were larger. 9' x 6' constructed of 1" X 6" wooden boards. Each box was divided longitudinally for strength. We added netting to the bottom to enhance recovery, because a few quahogs had been found outside the frames in 1975. We also built smaller boxes, 4' x 3', installed at seven additional sites with 5.000 ARC seed at each site. Two boxes were installed side by side at three locations: Doane Way, Yacht Club, and Lonnie's Pond. One box was planted with ARC stock and the other with CZR stock in equal numbers to see if there was a variation in growth or survival depending on the source. Results Table 2 provides details of the plantings for 1976. 1. The ARC stock was smaller at the beginning of the season (6-8 mm) and delivered 2 weeks later than the CZR stock (8-10 mm): 25,0(W seed from .ARC was delivered in mid- July (8-10 mm). 2. At Meetinghouse River and Quanset Pond, seed grew faster than those at other sites. 6 to 8