Separation of recently settled manila clams, Tapes philippinarum (A. Adams and Reeve, 1850), from three sediment types using sucrose density solution.

ABSTRACT Thorough study of the life history of important fishery and aquaculture shellfish species is essential for the most appropriate management and development of the resource. The early juvenile life stages of wild intertidal clam populations are difficult to study because of their existence in sediments of similar size and color to the animal. Here, we test the efficacy of methods for separation of recently settled Manila clams (Tapes philippinarum) from different types of intertidal sediments using density. Three aquaria were set up with sterilized sediments of each of the three sediment types--cobble/sand/shell, cobble/mud, mud/sand--resulting in nine aquaria total for the experiment. Each tank was filled with filtered, sterilized sea water; the water was heated to 20[degrees]C and aerated. Competent T. philippinarum larvae were added to each tank at a known density and were allowed to metamorphose and settle. Aquaria were then drained, and four sediment samples were taken from each aquarium for evaluation of number of juvenile clams present. Samples were wet-sieved to isolate the size fraction from 125 to 500 tam, and then this size faction was settled through high-density (1.9 g/mL) sucrose solution to separate stained clams from sediments. The average numbers of clams counted per sample for each sediment type were 57.9 (SD-37.1) for mud/sand. 60.0 (SD = 36.1) for cobble/sand/shell, and 53.1 (SD = 39.6) for cobble/mud. The average number of clams expected per sample based on the number of larvae added to each tank was 58.8 for all three sediment types. Statistical analysis revealed no significant difference between the mean number of clams per sample and the expected value of 58.8 for all three sediment types. Therefore, these results indicate that the methods used here can be used to extract rapidly recently settled Manila clams from all sediment types with a high degree of accuracy.

KEY WORDS: density separation, juvenile clams, sampling, sucrose, Types philippinarum

INTRODUCTION

The Manila clam (Tapes philippinarum. A. Adams and Reeve, 1850) is of major importance to both the wild fishery and aquaculture industry worldwide. In British Columbia, Canada, this species is non-native and is thought to have been introduced along with oyster seed from Japan in the 1930s (Bourne 1982). Since its introduction, T. philippinarum has become a very important species economically and is the basis of the current clam culture industry (Jones et al. 1993). The conditions in British Columbia are favorable for the Manila clam, and it has become well established throughout the southern coastline (Quayle 1974).

In the general bivalve life cycle, a pelagic larval stage is followed by metamorphosis, during which the swimming organ or velum is lost, and the bivalve transforms into the benthic or epibenthic juvenile form, depending on the species. The postmetamorphic juvenile stage of the Manila clam is found in the upper layers of the sediment; however, sampling juveniles front these sediments presents many problems because of the similar coloration and size of the clams and fine sediments. Consequently, very few studies have focused on this life stage. When sampling has been carried out and curly juveniles in the sediment are counted (Glock 1978, Jones 1974, Williams 1980), the work is extremely time-consuming and prone to error. Rarely, however, is reference made in published studies of bivalve settlement and early life-stage ecology to the accuracy of the sorting and counting methods involved. A simple, consistent, and effective method for separation of postsettlement bivalves from the sediments would allow for more studies to be carried out and more insight into the recruitment patterns of the Manila clam and other valuable clam species to be gained.

The density of juvenile postsettlement clams and cockles was estimated to be 1.036 1.076 g/mL in a study by de Montaudouin (1997) and 1.1 g/mL in a study by Jonsson et al. (1991), whereas minerals are more dense, typically with a specific weight of 2.5 g/mL and higher (Denny 1993). Density gradients of silica sols have been used in many studies to separate lighter meiofauna from higher density sediment fractions (Burgess 2001, Nichols 1979, Schwinghamer 1981). It has been shown that high-density sucrose solutions can also be used to separate meiofauna from muddy organic sediment (Heip et al. 1974). The technique explored herein involves wet seiving to isolate the size fraction of the sediment containing the bivalves, then allowing that size fraction to settle through a high-density sucrose solution (1.9 g/mL) to isolate the meifauna and allow for easier counting of the bivalves. This technique will be tested using three different sediment types to determine the efficacy of the method with various sediments.

The clam juveniles would be expected to float in a solution with a density of 1.9 g/mL. However, the high concentration of sucrose increases the osmotic pressure on the animal cells causing them to dehydrate, thereby increasing the density of the animals and causing them to sink (Bowen et al. 1972, Price et al. 1978). Although the juveniles become more dense and sink, they sink slowly in relation to the higher density mineral components of the sample and are found in the top layer of the sediment once it has settled out. This process of isopycnic sedimentation at one solute density to separate particles of different densities is called "rho spectrometry" by Price et al. (1977).

MATERIALS AND METHODS

Three different sediment types were obtained from intertidal beaches north of Nanaimo, British Columbia. Nine tanks were prepared with each of the three different sediment types (treatments) randomly assigned to three tanks per sediment type. The sediment types were cobble/mud, cobble/sand/shell and mud/sand according to their properties, and a subsample (approximately 20 [cm.sup.3]) of each sediment type was analyzed for grain size components by wet sieving using methods adapted from Komar (1998). Organic material was not removed prior to wet sieving. Briefly, the sediment was wet-sieved, dried at 60[degrees]C (Buyd & Tucker 1992) for 30 min, and weighed.

Sediments to be used in the experiment were autoclaved prior to placement in each aquarium. Each aquarium was filled to a depth of 5 cm (surface area of each lank was 800 [cm.sup.2]) with the assigned sediment type. and then filled with 15 L seawater (filtered to 1 [micro]m and sterilized with UV radiation). All aquaria were aerated and warmed to 20[degrees]C before addition of competent, hatchery-reared Manila clam larvae. A total of 2400 clams were added to each of the nine tanks.

Tanks were maintained at 20[degrees]C and clams were fed a combination of 50:50 Chaetoceros meullerii and Isochrysis spp (Tahitian strain) at a combined concentration of 20,000 30,000 cells/mL (Jones et al. 1993). Clams were left in the tanks for 11 days to settle and metamorphose. The tanks were then drained to simulate low-tide sampling conditions, and core samples of the sediment were taken. Four sediment cores were taken from each of the nine tanks (36 samples total). A small corer made of PVC pipe (5 cm internal diameter) was inserted I cm into the sediment, and a thin metal lid was slid under the pipe to prevent the sediment from falling out. Samples were placed in a plastic sample bag, labeled, and frozen for later counting. Freezing was chosen as a method of sample preservation to ensure consistency between laboratory methods and field methods that have been used for a parallel study.

For enumeration, samples were thawed then placed in 0.01% phloxine B dye fur at least 20 min (Williams 1978). Samples were then washed through a series of sieves; the fraction of sediment from 125 to 500 [micro]m was placed into a high-density sucrose solution (1.9 g/mL) in 30-mL test tubes. The test tubes were inverted to mix the sediments, to avoid particle-particle interactions (Price et al. 1977), then left to settle out by gravity (minimum of 25 min). The top layer of sediment was pipetted off the surface of the settled sediments in the test tube. The amount of sediment pipetted off the top was approximately 1-2 mL, which is less than 10% of the original core sample volume. The number of clams in each sample was counted under a dissecting microscope at 60x magnification. Statistical analysis of variance was calculated using JMP statistical software.

RESULTS

The grain size for each sediment type is shown below in Table 1. The cobble/mud sediment contained the largest fraction of the >2000-[micro]m size category with 82% by weight. The cobble/sand/ shell also contained mostly >2000-[micro]m-sized components; however, it should be noted that in this case, over half of this size category was composed of broken shell whereas the >2000-[micro]m fraction in the cobble/mud was entirely small cobble. The mud/ sand sediment was mostly composed of medium sand (70% dry weight); with coarse sand and fine sand making up another 18% dry weight. The components >2000 [micro]m were only broken shell in the mud/sand sediment.

Two thousand four hundred clams were added to each tank, and each tank had a sediment surface area of 800 [cm.sup.2]: therefore, the expected average density of clams in each tank was 3 clams/[cm.sup.2]. The surface area of each core was 19.7 [cm.sup.2], so the expected average number of clams per core was 58.8. The means and 95% confidence limits fur number of clams recovered from coring the three sediment types is shown in Figure 1. Comparison of the mean numbers of clams per core among the three sediment types was done to test if all means were equal regardless of sediment type. The mean values for clams per core were 60.0 for cobble/sand/ shell, 53.1 for cobble/mud, and 57.9 for mud/sand. Based on analysis of variance, there was no significant difference in the number of juvenile clams recovered from each of the different sediment types and the expected value 58.8 (n = 3, probability ranged from 0.63 to 0.94). Nor was there a significant difference between treatments (n = 3, probability = 0.90). Individual t-tests were also run to determine if the recovery of clams from coring of any of the experimental units (nine tanks x three treatments) departed significantly from the expected value based on larval seeding density. There was no statistically significant difference between the observed and expected values for any of the experimental units (n = 4, probability ranged from 0.30 to 0.75).

[FIGURE 1 OMITTED]

DISCUSSION

Analysis of grain size components (Table 1) shows that there were large differences between the compositions of the three sediment types used in this experiment. This was important because we were testing the accuracy of the sampling methods for extraction and counting of postsettlement juveniles in different types of sediments. For example, the mud/sand sediment was composed of 70% medium sand, which means 70% of the entire sample was the same size class in which the bivalves are found. This has the potential to lead to difficulty and inaccuracy in extraction of the clams from this large sediment fraction. In this experiment, no difficulties were encountered: The numbers of clams counted per core did not differ in the three types of sediment. Furthermore, for all three sediment types, the number of clams counted in each sample did not differ from the value expected per sample based on the number of clams initially placed in tanks. This means that with this method, there was an estimated 100% recovery of bivalves front sediment regardless of sediment type.

Individual tanks within each sediment type were tested to ensure that mean numbers of clams counted per sample did not differ statistically from tree tank to the next. Large standard errors were seen in data from some individual tanks. These were overcome when all samples of each sediment type were analyzed together. This may be interpreted as a result of the patchy settlement of clam larvae (Williams 1980), especially in coarser and heterogeneous substrates, and implies that the sampling effort may need to be increased for such sediment types.

The recovery of 100% of the bivalves that were placed in tanks also suggests 100% survival from the time of introduction of the larvae to the time of recovery of postsettlement clams. Survival rates for T. philippinarum in a hatchery generally vary from 50 90% depending on larval quality (Utting & Spencer 1991). The recovery of an estimated 100% of added larvae (had the entire sediment surface been sampled) is probably due to the relatively short duration of the study. Some of the postsettlement clams, in fact, may have been nonviable or dead at the time of sampling, but freezing and subsequent staining would not distinguish recently deceased clams from live ones. In other circumstances, it might be expected that some mortality would occur prior In settlement, so a failure to account for 100% of the introduced larvae might not be attributable to the sorting techniques in other studies.

Separation and counting of live clams prior to freezing was not attempted, leaving some question about whether separation as described herein would be equally effective for live specimens or those preserved using other methods such as formalin fixation. Schwinghamer (1981) conducted tests on live separation of benthos from mud and sediments using centrifugation in sorbitol and Percoll and found it to be an effective separation method that allowed for proper identification and observation of sampled benthos.

In a study by Burgess (2001), density separation of meiofauna front sediment was carried out using Ludox. Sediment samples were mixed with Ludox, then centrifuged to separate the meifauna. Using this method, Burgess was able to recover 95.9% of the bivalves in the sample. Jonge and Bouwman (1977) also found the use of density separation of nematodes and copepods from sediment and detritus to be more effective and accurate than handsorting decantation methods. Both Burgeses (2001) and Jonge and Bouwman (1977) note that a potential shortcoming of the density separation method is that animals may attach to sediments and therefore sink with them. Postmetamorphic bivalve juveniles have the ability to attach to larger sediments using a byssus. In this study, the sediment fractions larger than 500 [micro]m were not examined to look for attached juveniles; however, recovery was estimated at 100% in the size fraction examined, so few if any clams were likely to have been found in larger fractions. It is possible that if there were any clams attached to sediments by byssal threads, the threads were released when the sediment samples were frozen.

The use of the sucrose-density separation method described here is effective for counting newly settled juvenile clams from sediment. The use of the high-density sucrose solution to isolate the lower density animals increases sampling efficiency by decreasing the time to sort through sediment. Decreasing the lime to sort increases sampling accuracy because less physical and psychologic variance is introduced (Price et al. 1977). This decrease in sorting time would be especially important for sediments like the mud/sand sediment used here, where sieving would result in retention of the majority of the sediments along with the bivalves and, thus, hand sorting and counting would be quite tedious and prone to error.

These results show that these methods can be used in the field with the confidence to count recently settled clams in sediment samples involving a variety of sediment types. This would be useful in the assessment of farming techniques and questions of industry sustainability and environmental protection. Juvenile density estimates are important in the understanding of settlement patterns and early mobility in this economically important species. More generally, such studies are crucial to the understanding of bivalve larval ecology, demographics, and environmental/climatic factors that control recruitment, as well as invasion rates of exotic bivalves.

TABLE 1.
Grain size components, percentage by dry weight, or each sediment
type.

                                          % Overall Weight

                                              Cobble/
                   Size Range                  Sand/
   Size Class      ([micro]m)    Cobble/Mud    Shell    Mud/Sand

Shell                (>2000)         --         43          4
Granule+             (>2000)         82         30         --
Very coarse sand   (1000-2000)        3          8          3
Coarse sand        (500-1000)         4          7         11
Medium sand         (125-500)         7          9         70
Fine sand           (75-125)          1          2          7
Silt                  (<75)           2          1          5

The size category >2000 [micro]m in contains both granule+ and
broken shell.

ACKNOWLEDGMENTS

We would like to acknowledge the generous donation of lab space and larvae by the Center for Shellfish Research at Malaspina University-College; particularly the help of Dr. Bill Pennell for his guidance and enthusiasm, Gord Edmondson for help with equipment preparation, and Jenny Dawson-Coates for help with algal rearing and feeding. Also, the authors thank Terri Sutherland for her helpful comments.

LITERATURE CITED

Bourne, N. 1982. Distribution, reproduction and growth of Manila clam, Tapes philippinarum (Adams and Reeve), in British Columbia. J. Shellfish Res. 2:47-54.

Bowen. R. A., J. M. St. Onge, J. B. Colton & C. A. Price. 1972. Density-gradient centrifugation as an aid to sorting planktonic organisms. 1. Gradient materials. Mar. Biol. 14:242-247.

Boyd, C. E. & C. S. Tucker. 1992. Water quality and pond soil analysis for aquaculture. Alabama Agricultural Experiment Station, Auburn University. Auburn, AL. 183 pp.

Burgess, R. 2001. An improved protocol for separating meiofauna from sediments using colloidal silica sols. Mar. Ecol. Pro. Ser. 214:161-165.

de Montaudouin, X. 1997. Potential of bivalves' secondary settlement differs with species: a comparison between cockle (Cerastoderma edule) and clam (Ruditapes philippinarum) juvenile resuspension. Mar. Biol. 128:639-648.

Denny, M. W. 1993. Air and water. The biology and physics of life's media. Princeton: Princeton University Press. 342 pp.

Glock, J. W. 1978. Growth, recovery and movement of Manila clams, Venerupis japonica, planted under protective devices and on open beaches at Squaxin Island, Washington. M.S. Thesis, University of Washington, Seattle, WA.

Heip, C., N. Stool & W. Hautekiet. 1974. A rapid method of extracting meiobenthic nematodes and copepods from mud and detritus. Mar. Biol. 28:79-81.

Jones, C. R. 1974. Initial mortality and growth of hatchery-reared Manila clams, Venerupis japonica, planted in Puget Sound, Washington beaches. M.S. Thesis, University of Washington, Seattle, WA.

Jones, G. G., C. L. Sanford & B. L. Jones. 1993. Manila clam hatchery and nursery methods. Innovative Aquaculture Products Ltd. and Science Council of British Columbia, Victoria, BC. 73 pp.

de Jonge, V. N. & L. A. Bouwman. 1977. A simple density separation technique for quantitative isolation of meiobenthos using the colloidal silica Ludox[R]. Mar. Biol. 42:143-148.

Jonsson, R., C. Andre & M. Lindegarth. 1991. Swimming behaviour of marine bivalve larvae in a flume boundary-layer flow: evidence for near bottom confinement. Mar. Ecol. Pro. Ser. 79:67-76.

Komar, P. D. 1998. Beach processes and sedimentation, 2nd ed. Englewood Cliffs, NJ: Prentice Hall. 544 pp.

Nichols, J. A. 1979. A simple floatation technique for separating meibenthic nematodes from fine-grained sediments. Trans. Amer. Micros. Soc. 98(1): 127-130.

Price. C. A., J. M. St. Onge-Burns, J. B. Coulton & J. E. Joyce. 1977. Automatic sorting of zooplankton by isopycnic sedimentation in gradients of silica: performance of "rho spectrometer." Mar. Biol. 42:225-231.

Price, C. A., E. M. Reardon & R. R. L. Guillard. 1978. Collection of dinoflagellates and other marine microalgae by centrifugation in density gradients of a modified silica sol. Limnol. Oceanogr. 23:548-553.

Quayle, D. B. 1974. The intertidal bivalves of British Columbia, Victoria, BC. British Columbia Provincial Museum. Handbook No. 17. 104 pp.

Schwinghamer, P. 1981. Extraction of living meiofauna from marine sediments by cetrifugation in a silica sol-sorbitol mixture. Can. J. Fish. Aqua. Sci. 38:476-478.

Utting, S. D. & B. E. Spencer. 1991. The hatchery culture of bivalve mollusc larvae and juveniles. Lab. Leafl. No. 68., Ministry of Agriculture Food and Fisheries Fisheries Research Lowenstoft. 31 pp.

Williams, J. G. 1978. The influence of adults on the settlement, growth, and survival of spat in the commercially important clam, Tapes japonica Deshayes. Ph.D. Thesis, University of Washington, Seattle, WA.

Williams, J. G. 1980. Growth and survival in newly settled spat of the Manila clam, Tapes japonica. Fish. Bull. 77:891-900.

DAPHNE M. MUNROE, (1) * DOUG BRIGHT (2) AND SCOTT MCKINLEY (1) (1) Center for Aquaculture and the Environment, Faculty of Agricultural Sciences, University of British Columbia, West Vancouver, BC. Canada, V7V 1N6; (2) UMA Engineering Ltd., Victoria, BC, Canada V8T 2W1

* Corresponding author. E-mail: [e-mail]