Population biology of the Pacific geoduck clam, Panopea abrupta, in experimental plots, southern British Columbia, Canada.

ABSTRACT Information was presented on the density, recruitment to harvestable size, and growth of the Pacific geoduck clam, P. abrupta, in 0.25 ha experimental plots with controlled fishing in two areas of southern British Columbia over a 9-year period. In two plots that had high exploitation in the first 3 years, the geoduck population densities and the average age (by removing old individuals) were substantially reduced, but thereafter densities slowly increased due to geoduck recruitment. Recruitment was similar between plots after high and low exploitation rates, suggesting thai fishing and different geoduck densities had no long term effect on subsequent recruitment. After removing many old individuals from a geoduck population, the mean maximum asymptotic length ([L.sub.[infinity]]) of the you Bertalanffy growth model decreased, and the Brody growth function k values increased. In plots where there was low exploitation and geoduck densities remained similar over lime, [L.sub.[infinity]] values also remained similar, but k values varied with time suggesting other biologic or environmental factors could have influenced growth rates.

KEY WORDS: geoduck, Panopea abrupla, density, recruitment, growth

INTRODUCTION

The Pacific geoduck clam, Pare)pea abrupta (Conrad, 1849), is a large hiatellid bivalve found buried up to 1 m within mud and sand substrates. This geoduck species is distributed in coastal areas from the intertidal zone to at least 100 m (Jamison et al. 1984) from California to Alaska and to southern Japan (Coan et al. 2000). Commercial fisheries occur for P. abrupta in Washington State, British Columbia and Alaska (Campbell et al. 1998a, Bradbury & Tagart 2000, Hand & Bureau 2000) with annual landed values varying between 32 and 42 million Canadian dollars in British Columbia during 2000 to 2003. Adult geoduck generally have separate sexes and broadcast spawn usually during summer (Andersen 1971, Sloan & Robinson 1984, Campbell & Ming 2003). Planktonic larvae remain in the water column 16 to 47 days, settle as postlarvae on substrates, and subsequently burrow into suitable substrates (King 1986, Goodwin & Pease 1989). Geoduck juveniles and adults feed by filtering food particles (e.g., phytoplankton) from seawater (Goodwin & Pease 1989). Geoduck growth is variable, depending on a variety of environmental factors, but fastest during their first 10 y (Goodwin & Shaul 1984, Goodwin & Pease 1991, Harbo et al. 1983, Bureau et al. 2002, Noakes & Campbell 1992, Hoffmann el al. 2000). Geoduck are long-lived, reaching ages of up to 168 y (Bureau et al. 2002). Geodnck recruitment was considered to be variable but low (Breen & Shields 1983, Harbo et a1.1983), however recent studies of age frequencies have suggested strong recruitment events to adult geoduck populations throughout BC since 1988 (Bureau et al. 2002, Bureau et al. 2003). Natural mortality rate (M) estimates of adult geoduck in British Columbia range from 0.01-0.04 (Breen & Shields 1983, Harbo et al. 1983, Noakes 1992).

Little is known of the effects of fishing on the population biology of P. abrupta (e.g., Goodwin 1978, Campbell et al. 1998a). The objective of the study was to investigate the effects of fishing on geoduck population biology. The purpose of this study is to present information on the density, recruitment to harvestable size, and growth of P. abrupta in small experimental plots with controlled fishing in two areas of southern British Columbia over a 9-year period. Further analyses of geoduck recruitment and natural mortality of this study are presented in Zhang and Campbell (2004).

METHODS

Experimental Plots

The study areas were located at Ritchie Bay (northwest Meares Island on the west coast of Vancouver Island) and southwest Marina Island (Georgia Strait) of British Columbia. Ritchie Bay has been closed to commercial fishing and designated as a research reserve since 1980, when Fyfe (1984) started his geoduck study. The area around Marina Island was heavily harvested during 1978-1989 and was subsequently closed to commercial fishing for geoduck since 1990 (Campbell et al. 1996b). For this study, 4 experimental plots were established at Ritchie Bay [(1) 125[degrees]55.242'W, 49[degrees]13.502'N; (2) 125[degrees]55.179'W, 49[degrees]13.506'N; (3) 125[degrees]55.076'W, 49[degrees]13.488'N; (4) 125[degrees]55.046'W, 49[degrees]13.549'N] during 1991 and 3 plots established at Marina Island [(1) 125[degrees]03.734'W, 50[degrees]03.001'N; (2) 125[degrees]03.745'W, 50[degrees]02.959'N; (3) 125[degrees]03.826'W, 50[degrees]02.955'N] during 1992.

Each experimental plot comprising a 0.25 ha area (50 x 50 m), was delineated and subdivided into 10-m sections with lead lines to provide five 500 [m.sup.2] strips (50 x 10 m). Experimental plots 1, 2, and 3 at Ritchie Bay were slightly smaller (0.244 ha) than the other plots because 0.006 ha/plot were reserved for potential show factor plots (see later). The minimum-maximum depth (m) at datum for each experimental plot at Ritchie Bay was (1) 7.8-11.8; (2) 10.0-10.9; (3) 7.7-9.8: and (4) 10.7-11.6 and at Marina Island (1) 11.7 14.5; (2) 11.6-15.4: and (3) 12.2 16.8. The bottom substrates comprised of soft mud to fine sand and shell mixture at Ritchie Bay and fine sand with some small gravel at Marina Island.

Experimental Treatments

The chronology of events for each plot and year included an abundance (show factor and density) survey, biosample, and or commercial harvest (Table 1). Each plot was randomly chosen for a different treatment in terms of a depletion harvest regimen (Table 1, 2). The general harvest regimen in each plot for Ritchie Bay was (1) moderate, (2) none, (3) heavy, and (4) none, and for Marina Island was (1) none, (2) moderate, and (3) heavy. In an attempt to determine if seeding of hatchery-raised juvenile geoduck could supplement natural recruitment, Plot 1 and Plot 2 of Ritchie Bay, were artificially seeded at approximately 16 [m.sup.-2] with younger then 0.5-y-old hatchery-raised, juvenile geoduck of mean 11.3 mm shell length (SL) (min. 6 mm SL. max. 18.4 mm SL, n = 136) during September 1991. Plot 4 of Ritchie Bay and Plot I of Marina Island were considered the control treatment as the only removals made were for biosamples and were not seeded with hatchery-raised juvenile geoduck. No treatment was replicated with additional plots in the same study area because of the considerable logistical costs involved.

Densities

Geoduck density was estimated using survey methods described by Campbell et al. (1996 a, 1996 b). The survey method for each plot involved counting geoduck siphons in 5-[m.sup.2] (5 x 1 m) quadrats along ten 50-m one meter wide transects. Two transects were placed on either side of the four inner subdivision boundaries and one transect each was placed on the two inner sides of the study area border for a total of 10 transects per plot. Six of the eight inner transects for Plots 1, 2, and 3 in Ritchie Bay were shortened because they had 10 m reserved as potential show factor plots, where no geoduck was removed. Consequently, 10 [m.sup.2] for each of these 6 transects per Plot 1, 2, and 3 were not used in the density estimates for Ritchie Bay. The total area considered for geoduck density surveys and harvest was 2440 [m.sup.2] for Ritchie Bay Plots 1, 2, and 3 and 2500 [m.sup.2] for Ritchie Bay Plot 4 and all three plots of Marina Island. There were 88 quadrats surveyed for Ritchie Bay Plots 1, 2, and 3 and 100 quadrats surveyed for Ritchie Bay plot 4 and Marina Island Plots 1, 2, and 3 for each density survey.

Survey dates for geoduck densities at Ritchie Bay were July 25 to August 14 1991; May 13-27, 1992; May 17 19, 1993; May 17-19, 1995; May 31 to June 1, 1996; May 24 25, 1997; July 9-10, 2000; and at Marina Island were July 15, 1992; July 8-9, 1993; June 6, 1994; June 4-5, 1996; September 18-19, 1997; October 5, 1999; September 7-8, 2001 (Table 1).

Show Factor

Because all geoduck may not show their siphons at the substrate surface at any one time, "show" plots were used to determine the percentage of exposed geoduck (i.e., showing their siphons each day). In Ritchie Bay, only the three show plots in Plot 2 were used throughout this study. In the Marina Island study area, the three show plots were placed 5-10 m outside the borders of the three experimental plots. Each 20 [m.sup.2] show plot (i.e., 2 x 10 m, subdivided into four 1 x 5 m quadrats) was marked with lead line providing a total of 12 quadrats in the three show plots of each study area. Each geoduck showing in the show plots was marked beside the neck with a "flag" (stainless steel wire) in the substrate. The proportion of geoduck showing (the show factor) on any given day was determined by dividing the number showing on that day by the cumulative total number of geoduck flagged during the period that the show plots were monitored each year. Monitoring of the show plots started several days before the density surveys were conducted in the experimental plots. Sometimes not all the show plots were monitored. On a few occasions, less than three show plots were surveyed, and the available data were used for the analyses.

For most plots, both the shows and density were surveyed on the same day. However, for a few days some Ritchie Bay plots did not have show factor and density surveys on the same day, in which case the show factor used was interpolated (averaged) between dates. One exception to a reliable interpolation occurred for the geoduck density estimate at Ritchie Bay Plot I which was surveyed during August 7-14, 1991 when the show factor had been measured on July 26 and 31 and August 1 and 15, 1991 and estimated to be 0.96. This show factor value was clearly incorrect and too high because era severe storm, with 115.6 mm of rain and south east winds of up to 43 km/hour, occurred on August 7, 1991 (Environment Canada web site for Tofino Airport [125[degrees]24'W, 49[degrees]4'N] http://www.climate.weatheroffice.ec.gc.ca/climateData/ canada_e.html]). The diver surveying for geoduck density on the August 15, 1991 also recorded that geoduck were not showing well due to a "winter like storm" the previous week. Campbell et al. (1996a) found that show factor decreased to about 0.65 following a strong storm near Comox during June 1993. All measured show factors at Ritchie Bay were higher than 0.80. Show factor data from the Marina study area on July 8, 1993 also provided a reduced show factor of about 0.66, due to increased wind speeds (hourly weather and wind speeds and direction were examined for Campbell River Airport [125[degrees]16'W, 49[degrees]57'N; see website: http:// www.climate.weatheroffice.ec.gc.ca/climateData/canada_e.html] a location near to Marina Island, during the survey periods and about 5 days prior to the survey). Consequently a show factor value of 0.66 was considered reasonable to be used in the density calculations for Ritchie Bay Plot 1 in 1991.

Removals

Geoduck were removed from each plot for two reasons: (1) to deplete the population through commercial harvest, and (2) to collect a biosample for aging. Removal of geoduck from each plot was always conducted soon after the density surveys had been completed. An experienced fisherman used the usual commercial harvest methods (Campbell et al. 1998a) to systematically harvest each of the five strips per plot right up to the rope borders on each day fished. This ensured, as much as possible (except for environmental influences such as tidal current and visibility variation), that all geoduck had an equal chance of being harvested in the plot on each day during the study. A target number for each biosample was a minimum of approximately 500 geoduck per plot (i.e., 100 per strip). Biosamples with 50 less than the 500 geoduck target were probably due to the divers not being able to find the extra individuals at low geoduck densities in the plots. When more than 1000 geoduck were removed per plot, the first approximately 500600 geoduck were used as a biosample. Geoduck within the potential three 20 [m.sup.2] show factor plots in each of the Ritchie Bay Plots 1, 2, and 3, however, were not removed, making the actual area harvested 2440 [m.sup.2] per plot. The total number of geoduck removed per plot per year is shown in Table 2.

Biosamples

For each geoduck, the shell length (SL) was measured as the straight-line distance between the anterior and posterior margins of the shell to the nearest mm with veneer calipers. The age of each geoduck was estimated using the acetate peel method of Shaul and Goodwin (1982). One valve per geoduck was sectioned through the hinge plate, the cut surface polished, etched with a 1% hydrochloric acid solution for approximately 1 minute, washed with distilled water, dried, and an acetate peel made by applying first acetone, then an acetate sheet on the hinge surface. Annual growth rings imprinted on the acetate peel were counted on a digitizing table after x40 magnification using a Neo-Promar projector. Shaul and Goodwin (1982) and Noakes and Campbell (1992) provide further details in shell preparation and age validation procedures. Although most individuals had both their SL and age measured, there were some that had only the SL or only the age measured; these latter individuals were included in the analysis only where appropriate.

To determine the weight of geoduck necks by age at recruitment, sub samples of between 45 and 115 geoduck, with as wide a size range as possible, were obtained from each biosample. For each geoduck, the neck was cut at the base from the body, cut longitudinally along the siphons to remove excess water and weighed in grams.

Data Analyses

Density

The mean show-adjusted density, D (number [m.sup.2]), for each plot was calculated as

(1) D = 1/n [n.summation over i=1] [d.sub.i]/1/m [m.summation over j=1] [S.sub.j]

where [d.sub.i] is the density of geoduck (i.e., number of geoduck [m.sup.-2] or number of geoduck per 1 x 5 m quadrat divided by 5) in the ith quadrat, [S.sub.j] is a proportion of geoduck showing in the jth show quadrat, n is the total number of quadrats sampled to estimate geoduck density, and m is the total number of show quadrats (i.e., twelve 1 x 5 m quadrats) used for estimating show factor in each study area. Hereafter density in this study refers to show-adjusted density.

Distribution-free unbiased bootstrap techniques (Efton & Tibshirani 1993) with S-plus 12001) computer programs were used to estimate 95% confidence intervals (CI) for the mean show-adjusted density. Eq. 1 was used to obtain a new mean show-adjusted density by randomly resampling, with replacement, m new show factors from the original tit show factors and n new density counts from the original n density counts. This procedure was repeated 1000 times. To produce the (1-2[alpha]) CI of show-adjusted density, the [alpha] percentile (2.5%) and the 1-[alpha], percentile (97.5%) of the bias-corrected and accelerated (BCa) distribution of the 1000 show-adjusted densities were used as the lower and upper CI bound, respectively (Efron & Tibshirani 1993).

Exploitation

Exploitation was calculated as R/D, where R is the removed density (number of geoduck [m.sup.-2] calculated as total number of geoduck divided by the plot area harvested), and D is the geoduck density (mean, 95% CI) surveyed just prior to removal.

Age Frequencies

The age frequency data were standardized for density to allow comparison of the number of geoduck [m.sup.-2] belonging to a particular age group between plots and years. The equation used was:

(2) [D.sub.a] = [p.sub.a] x D

where [D.sub.a] is the estimated mean density of geoduck of age a, [p.sub.a] is the proportion of geoduck of age a (mnnber of individuals of age a divided by total number of known age) in the bio-sample, and D is the mean show-adjusted density.

The 95% CI for density of age a were calculated using a bootstrap algorithm (Efron & Tibshirani 1993, S-plus 2001). Eq. 1 and 2 were used to obtain a new estimate by randomly resampling, with replacement, m new show factors from the original m show factors, n new density counts from the original n density counts, and k new age estimates from the original k age estimates. The m new show factors and the n new density count were used to estimate D, and the [p.sub.a] was estimated by counting the number of age a in the k new age estimates. This procedure was repeated 1000 times. To produce the (1-2[alpha]) CI of show-adjusted density at age, the [alpha] percentile (2.5%) and the 1-[alpha] percentile (97.5%) of the bias-corrected and accelerated (BCa) distribution of the 1000 show-adjusted densities were used as the lower and upper CI bound, respectively (Efron & Tibshirani 1993).

Pair-wise comparisons to test for differences in age frequencies between plots and years were conducted using the Mann-Whitney test.

The figures for age frequencies adjusted to density were truncated to 100 y for graphical purposes because there were few geoduck found older than 100 y in both study areas.

Recruitment

Recruitment was defined as the density of geoduck growing into the population that were fully vulnerable to fishing. We assumed recruitment to occur when all of the necks of geoduck of the youngest age cohort were large enough to be clearly' seen by divers when showing at or above the substrate. The threshold criterion used for recruited geoduck neck size at the youngest age was where the lower 95% CI of the mean neck wet weight was equal to 100 g or higher. Recruitment for geoduck from the Ritchie Bay area was 6 y. (neck weight mean = 148.2, lower-upper 95% CI = 130.5-165.9, n = 34), and from the Marina Island area was 7 y (neck weight mean = 124.2, lower-upper 95% CI = 104.2 144.2, n = 22). The densities of age 6 and 7-y geoduck and 95% CI for each plot were calculated as described in the previous section above. Because there were no replicates for each plot-treatment, for comparisons between plots and years, we considered overlapping 95% CI to suggest similarities in mean densities at age, and nonoverlapping 95% CI to suggest a difference in mean densities at age.

Growth

The relationship between shell length and age was estimated with the von Bertalanffy (1938) growth model

(3) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where t is age in years, [L.sub.t] is shell length (mm) at age t, [L[infinity]] is the mean asymptotic length (mm), k is the Brody growth coefficient determining rate of increase or decrease in length increments, [t.sub.0] is a phase-variable, suggesting the hypothetical age at which the organism would have been at zero length, and [epsilon] N(0,[[sigma].sup.2]) is a normal variate. The parameters [L.sub.[infinity]], k, [t.sub.0], and [epsilon] were estimated using maxinmm likelihood methods (Bain & Engelhardt 1991) in the statistics program of S-Plus (2001). Only undalnaged geoduck shells were used for the growth analyses. Initial analyses indicated that [t.sub.0] values were either far too positively or negatively large in the biosamples that had few young geoduck. For consistency, we followed the same approach of Bureau et al. (2002) by setting to as zero for all samples causing the fitted curves through the origin and re-estimating [L.sub.[infinity]] k, and [epsilon].

The 95% CI for the parameters [L.sub.[infinity]], k, and [epsilon] from each biosample were calculated using a bootstrap algorithm as follows. Pairs of age length data for individual geoduck were randomly resampled, with replacement, from the original data up to the biosample size n, and then Eq. 3 (with [t.sub.0] = 0) was used to obtain esiimates of the parameters. This procedure was repeated 1000 times. To produce the (1-2[alpha]) CI for each of, [l.sub.[infinity]], k, and [epsilon]; the [alpha] percentile (2.5%) and the 1-[alpha] percentile (97.5%) of the bias-corrected and accelerated (BCa) distribution of the 1000 values of each parameter were used as the lower and upper CI bound, respectively (Efron & Tibshirani 1993).

Pair-wise comparisons for hypothesis testing of simultaneous differences in parameters [L.sub.[infinity]] and k between all biosamples were conducted using the likelihood ratio test outlined by Kimura (1980) to determine if geoduck growth rates were different between plots and years.

Relation Between Density, Recruitment, and k

Regression analyses (using the least squares method) were used to determine if there were significant relationships between the Brody growth coefficient k and mean overall density or recruitment density, for each plot and year at Ritchie Bay.

RESULTS

Density

For Ritchie Bay, geoduck density remained highest over time in unharvested Plot 2 and lowest in the highly exploited Plot 3 (Fig. 1A). The removals over 3 y of practically all geoduck in Plot 3 caused geoduck densities to decline significantly by 1993 and to remain low, although there was a small density increase by 2000 (Fig. 1A). Although there was 0.62 exploitation rate of the geoduck in Plot 1 during 1991 (Table 2), with a consequent decline in geoduck density in 1992, densities recovered by 1995 (Fig. 1A). Geoduck densities in Plot 4 fluctuated between about 2.68 to 3.85 [m.sup.-2] (Fig. 1A, Table 2).

Geoduck densities in the Marina Island experimental plots were considerably lower than those for Ritchie Bay (except for Plot 3 after 1992) (Fig. 1, Table 2). The removal of substantial numbers of geoduck reduced densities in both Marina Island Plots 2 and 3 (Fig. 1B, Table 2). Although Plot 3 had the highest and lowest density of the Marina Island plots in 1992 and 1996. respectively, Plot 3 densities recovered slightly to similar densities of control Plot 1 by 2001 (Fig. 1B. Table 2).

Exploitation rates were high especially in Plot 3 of Ritchie Bay and Marina Island, and in plots that had biosamples taken in already low density plots (Table 2). Mean estimates of exploitation values greater than 1.0 (i.e., Ritchie Bay Plot 3 in 1992, Marina Island Plot 2 in 1992) clearly were not possible and reflect the uncertainty of the density estimates of the survey method, but were possible within the lower 95% CI exploitation values (Table 2).

Show factors were generally recorded higher than 0.8 for most surveys in both study areas, except for Marina Island during 1993 to 1994 (Fig. 2) and for Ritchie Bay Plot 1 in August 1991 due to inclement weather (see also show factor in methods section). Strong winds, that could have affected geoduck to produce low show factors, were recorded at 30 and 28 km per hour from a northwest direction during July 5 6, 1993 respectively, and at 30 km per hour from a south east direction during June 4, 1994 near Marina Island. There were southeast winds of up to 43 km/hour recorded on August 7, 1991, near Ritchie Bay.

Age Frequencies

Following heavy harvesting there was a significant decrease in median ages (Table 3) and clear reduction in the densities of old geoduck in Ritchie Bay Plot 1 (Fig. 3A, 3B, 3C) between 1991 and 1995, Plot 3 (Fig. 3 F, 3G) between 1992 and 2000, and Marina Island Plot 3 (Fig. 4C, 4D) between 1992 and 2001. There was no significant change in geoduck median age for Ritchie Bay control Plot 4 between 1995 and 2000 (Table 3, Fig. 3H, 3I).

Recruitment

For Ritchie Bay, geoduck recruitment density at age 6 y, was generally higher for all plots in 2000 than in previous years (Table 4, Fig. 3). Recruitment within plots between years, was similar for Plot I between 1991 and 1995 but higher in 2000; was higher for Plot 2 in 2000 than 1995; higher for Plot 3 in 2000 than 1992; and similar for control Plot 4 in 1995 and 2000 (Table 4, Fig. 3). Geoduck recruitment density between Plots 1,2, and 4 in 1995 was similar, and recruitment for all four plots in 2000 was similar.

For Marina Island, geoduck recruitment density at age 7 y, was slightly higher for Plot 3 in 2001 than 1992 (Table 4, Fig. 4). Plots 2 and 3 had higher geoduck recruitment than Plot I in 2001 (Table 4, Fig. 4).

Geoduck recruitment at Ritchie Bay was well above that fin Marina Island in the last l0 y (Fig. 3, 4, Table 4). The highest density for each plot at a particular age and year: for Ritchie Bay Plot 1 was 0.60 at 4 y in 1995, 0.31 at 8 y in 2000; Plot 2 was 0.33 at 4 y in 1995, 0.51 at 8 y in 2000: Plot 3 was 0.08 at 4 y during 1992, 0.16 at 6 y in 2000; Plot 4 was 0.18 at 3 y in 1995, 0.15 at 7 y in 2000 (Fig. 3); and for Marina Island Plot I was 0.014 at 8 y in 2001; Plot 2 was 0.029 at 8 y in 2001; Plot 3 was 0.038 at 9 y in 1992, and 0.042 at 8 y in 2001 (Fig. 4).

The influence of artificial seeding on the density of the 1991 cohorts, 4 and 9 y later, was suggested by the higher geoduck densities in Plot 1 (4.28 and 1.90 times) and Plot 2 (2.36 and 2.40 times) compared with the geoduck densities in unseeded Plot 4 during 1995 and 2000, respectively (Table 5; Fig. 3).

Growth

The results on the geoduck growth curves (Table 6, Fig. 5) and the Kimura pair-wise comparisons (significant differences at P < 0.05) (Table 7) were grouped as follows.

Within Each Plot Between Years

The Brody growth coefficients, or k values, for geoduck in Ritchie Bay Plots 1, 2, and 4 were significantly higher in 1995 than for the same plots in 1991 and 2000. In contrast, the k value for Ritchie Bay Plot 3 was significantly higher in 2000 than in 1992. The k value for Marina Island Plot 3 was significantly higher in 2001 than 1992.

The [L.sub.[infinity]] values were significantly higher, prior to the removals, for most Ritchie Bay plots. (i.e., Plot 1 higher in 1991 than 1995 Plot 1 in 1995 and 2000 were not significantly different). Plot 2 and 4 larger in 1995 than 2000, Plot 3 larger in 1992 than for 2000. and Marina Island Plot 3 larger for 1992 than for 2001. In contrast, the [L.sub.[infinity]] values for Ritchie Bay Plot 4 were not significantly different between 1995 and 2000.

Between Plots Within Similar Years

The k value(s) for geoduck in Ritchie Bay Plot 1 in 1991 was not significantly different to Plot 3 in 1992: for 1995 Plots I, 2 and 4 were not significantly different; for 2000 Plots 1, 2. and 4 were not significantly different but all three were significantly lower than that for Plot 3. The k values were not significantly different for Marina Island Plots 1,2, and 3 in 200i. The k value was higher for Marina Island Plot 3 in 1992 than for Ritchie Bay Plot I in 1991 and Plot 3 in 1992.

The [L.sub.[infinity]] values were significantly different between most Ritchie Bay plots during 1995 and during 2000, except Plot 2 and 4 were not significantly different during 2000. The [L.sub.[infinity]] values for 2001 at Marina Island Plots 2 and 3 were not significantly different, but both were different from Plot 1. The [L.sub.[infinity]] value for Marina Island Plot 3 in 1992 was not significantly different for Ritchie Bay Plot 1 in 1991, but significantly different for Plot 3 in 1992.

Between Plots and Years

The k values were significantly different when comparing between most plots and years, except for Ritchie Bay Plots 1, 2, and 4 in 1995, which were similar to Plot 3 in 2000.

The [L.sub.[infinity]] values were not significantly different between Ritchie Bay Plot 1 in 1991 with Plot 2 and 4 in 1995, with Plots 2 and 4 in 2000, and Plot 4 in 1995 with Plot 2 in 2000.

Both k and [L.sub.[infinity]] geoduck growth values were significantly higher for the combined Ritchie Bay Plots 1, 2, 3, and 4 data in 2000 compared with the combined Marina Island Plots 1, 2, and 3 data in 2001.

Relation Between Density, Recruitment, and k

No significant (P > 0.05) relationships between recruitment and density, k and recruitment, or k and density could be detected when using data from all plots and years in the Ritchie Bay or the Marina Island study areas.

DISCUSSION

This study provides, for the first time, information on P. abrupta population biology from a 9-year study of experimental plots in two different areas of southern BC. Heavy exploitation for the first 3 years in Plot 3 of both Ritchie Bay and Marina Island substantially reduced geoduck population densities, reduced the average age (by removing old individuals), but thereafter densities slowly increased due to recruitment. A similar trend was observed for geoduck that were fished heavily only for the first year, but the initial recovery was more rapid in geoduck populations for Plot 1 of Ritchie Bay than for Plot 2 of Marina Island. Geoduck densities in the control plots fluctuated over time, but CIs often overlapped between years. Factors causing geoduck age densities to fluctuate could have included removals for biosamples, fishery induced mortality (harvesting methods causing incidental injuries to the remaining unfished geoduck), geoduck natural mortality (e.g., Zhang & Campbell 2004), recruitment and growth rates influenced by environmental conditions. The fluctuation was also likely related to the accuracy of the density estimates. Although experienced commercial divers were used for all density estimates, occasional poor visibility, due to plankton, drift algae, silt, or strong current at Ritchie Bay could have influenced the results.

Geoduck densities surveyed in the experimental plots were generally within the density range reported from surveys in similar regions: west coast of Vancouver Island (0.29-3.34 [m.sup.-2]) (Hand & Dovey 1999. Hand & Bureau 2000) and Inside waters including Georgia Strait (0.18 0.70 [m.sup.-2]) (Hand & Bureau 2000). The high density values of 5-7 [m.sup.-2] found in Plot 2, and generally from Ritchie Bay by Fyfe (1984) (4.00 7.69 [m.sup.-2]), were unusual for southern BC and only matched and surpassed by areas surveyed in northern BC with densities as high as 10.27 [m.sup.-2] found in the Moore Islands (Hand & Bureau 2000). A reason for the high geoduck density at Ritchie Bay was probably because of the area being closed to commercial fishing and designated as a research reserve since 1980 (Fyfe 1984). In contrast, the area around Marina Island was heavily harvested during 1978 1989 and closed to commercial fishing since 1990 (Campbell et al. 1996b).

Geoduck percentage shows were high (>80%) in both study areas for most of the summers studied. Low show factors (<80%) recorded in the Marina study area during 1993 1994 were preceded by strong winds of 28 30 km per hour a few days prior to the density surveys. High winds probably caused the geoduck to retract their siphons via shifting sand or debris on the substrate surface at the study sites where depths were fairly shallow (<17 m). Variation in the direction, duration, and/or intensity of wind may have influenced the level of geoduck show factor in different locations, depending on topography and depth of the area. Low salinity caused by precipitation and rainwater runoff from nearby land due to storms may also be other factors affecting geoduck shows. Show factor estimates could be important in estimating density, especially when density surveys are conducted after a storm or during fall-winter months. The percent of geoduck in a bed "'showing" their siphons at or above the substrate surface has been found in different studies to vary under different environmental and physiologic conditions. A number of authors (Goodwin 1973, Goodwin 1977, Cox & Charman 1980, Turner & Cox 1981, Fyfe 1984) have reported that the highest seasonal percentage of shows (>80%) was during the summer (April to September) when feeding and reproduction occurred and the lowest (<50%) was during the winter (November to February) when many geoduck were inactive. Shows were highest when local water currents were not excessive and there was no mechanical disturbance of the bottom (Goodwin 1977). Campbell et al. (1996a) documented the percentage shows was reduced from about 80% to 65% after a storm in a wave-exposed area, but returned to about 80% within 2-3 wk of the storm event.

The recruit age for geoduck was lower for Ritchie Bay (6 y) than for Marina island (7 y) because geoduck growth rates were highest in Ritchie Bay of the two study areas. This suggests that growth rates, physical or physiologic characteristics and age estimation for geoduck from different locations need to be examined prior to defining an age for a particular type of recruitment, (e.g., neck size for fishery purposes or for age at maturity, Campbell & Ming 2003).

Recruitment was variable temporally and at both small and large spatial scales. Recruitment was generally higher for the Ritchie Bay plots during 2000 and for Marina Island Plot 3 during 2001 than the previous years. Recruitment was considerably higher for geoduck in Ritchie Bay during 2000 than 101 Marina Island during 2001. Recruitment was similar for Ritchie Bay plots with high or low geoduck densities during 2000. Spatial and temporal variability in geoduck recruitment have been reported previously (Breen & Shields 1983, Harbo et al. 1983, Fyfe 1984, Goodwin & Shaul 1984, Sloan & Robinson 1984, Breen et al. 1991, Noakes & Campbell 1992, Hand & Dovey 1999, Orensanz et al. 2000, Bureau et al. 2002, Bureau et al. 2003). Recruitment may be influenced by various factors (e.g., size of breeding brood stock providing progeny, predation on the larval and early juvenile stages, environmental variables such as local currents, and/or general temperature trends affecting larval and early juvenile growth and survival). The timing of reproduction may also affect larval settlement and juvenile recruitment through food availability and growth rates. Gribben et al. (2004) found that differences in the timing of spawning between two P. zelalTdica populations could be attributed to latitudinal temperature gradients. Bureau et al. (2002, 2003) found that age frequency data of geoduck sampled throughout BC during the 1993 to 2002 period indicated strong geoduck recruitment trends, which might be linked to a general warming trend or El Nino events.

Local geoduck recruitment also could be affected by the physical disturbances of local fishing events. The study of Goodwin and Shaul (1984), which lasted approximately 20 mo and sampled >1 year-old geoduck with a venture dredge, suggested that fishing could have an adverse effect on recruitment in the short term. Our 9-y study examined the long term side effects of fishing on recruitment, since young geoduck (<4 y) could not be detected until the survivors recruited into 6-7 y age groups and became visible to divers several years following the fishing event. In this study, we did not detect any long-term effect of heavy fishing or low geoduck densities on subsequent recruitment. However, Zhang and Campbell (2004) found from retrospective age structure modeling of these populations that recruitment was generally lower after fishing geoduck in Plot 3 compared with the other three plots from Ritchie Bay suggesting that fishing, low geoduck densities or juvenile mortality could have influenced recruitment in Plot 3, at least in the first 1-3 y after fishing. Assuming there was sufficient geoduck larval supply from other areas, we concluded (this study and Zhang & Campbell 2004) that a heavy fishing event could reduce local recruitment in the short term, but probably not influence recruitment in the long term (e.g., >3 y). However, other environmental or biologic factors could mask the effects of fishing on recruitment as was observed in the Marina Island study area. In contrast to the Ritchie Bay observations, at Marina Island, the control unfished Plot 1 had poor recruitment compared with the other two fished plots.

Artificial seeding of Ritchie Bay Plots 1 and 2 provided a slight increase above the natural recruitment. Much (96% to 99%) of the large number of seed released (estimated to be a mean 16.0 [m.sup.-2]) into these two plots probably did not survive. Without providing adequate protection from predators (e.g., fish, crabs, and starfish) (Goodwin & Pease 1989) artificial geoduck seeding was considered not economically feasible for enhancement of natural geoduck populations (Beattie 1992, Beattie & Blake 1999, Clapp 2000).

Growth characteristics of geoduck front different plots and years varied considerably. Although no simple broad statistical relationship could be detected between geoduck density and the growth parameters when all plots and years for each study area were combined some interesting observations on geoduck growth could be made. Following removal of many geoduck in a plot, subsequent growth characteristics indicated a lowering of the apparent mean max asymptotic length ([L.sub.[infinity]]) and an increase in the Brody growth coefficient (k) of the von Bertalanffy growth model. Removal of old individuals from a geoduck population such as in Ritchie Bay Plots 1 and 3 and Marina Island Plot 3 (Fig. 1, 3, and Table 3) probably influenced the [L.sub.[infinity]] to decline and the Brody growth function k values to increase (Fig. 5 and Table 6). Where there was no major exploitation and geoduck densities remained similar over time such as in Ritchie Bay Plots 2 and 4, [L.sub.[infinity]] values remained similar, but k values varied with time. Reasons for the differences in P. abrupta growth rates between areas and plots could be generally attributed to various environmental and biologic factors associated with different habitats (e.g., depth, substrate type, temperature, exposure to water surge activity, pollution, food availability, geoduck density, and or genetic characteristics) (Breen & Shields 1983, Harbo et al. 1983, Goodwin & Shaul 1984, Goodwin & Pease 1991, Noakes & Campbell 1992, Hoffmann et al. 2000, Bureau et al. 2002).

TABLE 1.

Annual schedule of abundance survey (A), bio-sampling (B) and or
commercial harvest removal (C), and hatchery-reared geoduck juvenile
seeding (E) in each experimental plot of the Ritchie Bay and Marina
Island study area.

                                                  Marina Island
         Ritchie Bay Experimental Plot          Experimental Plot
           Number and Harvest Regime        Number and Harvest Regime

Year       1         2        3       4        1        2        3
        Moderate    None    Heavy    None   Moderate   None    Heavy

1991   A, B, C, E   A, E   A, C      A
1992   A            A      A, B, C   A        A        A, C   A, B, C
1993   A            A      A, C      A        A        A      A
1994                                          A        A      A
1995   A, B         A, B   A         A, B
1996   A            A      A         A        A        A      A
1997   A            A      A         A        A        A      A
1998
1999                                          A        A      A
2000   A, B         A, B   A, B      A, B
2001                                          A, B     A, B   A, B

The chronological sequence of events for each plot and year is shown in
alphabetical order.

TABLE 2.

Summary of pre-harvest geoduck densities, number and density of
geoduck removed, through harvest and or biosample removals, and
exploitation rate, by experimental plot at the Ritchie Bay and Marina
Island study areas.

                         Removed    Biosample       Removed
                          total       total         density
Plot        Year           no.         no.       [no.m.sup.-2]

         Ritchie Bay

1           1991          5580         624           2.29
            1995           532         532           0.22
            2000           508         508           0.21
2           1995           506         506           0.21
            2000           494         494           0.20
3           1991          2930         --            1.20
            1992          4024         489           1.65
            1993           183         --            0.08
            2000           448         448           0.18
4           1995           521         521           0.21
            2000           516         516           0.21

        Marina Island

1           2001           417         417           0.17
2           1992           878         --            0.35
            2001           291         291           0.12
3           1992          1356         556           0.54
            1993           334         --            0.13
            1994           163         --            0.07
            2001           330         330           0.13

                         Geoduck Density ([no.m.sup.-2])

                                      95% CI

Plot        Year             Mean     Lower    Upper

         Ritchie Bay

1           1991             3.67*    2.84     5.06
            1995             3.17     2.82     3.58
            2000             2.00     1.75     2.27
2           1995             7.35     6.87     7.86
            2000             5.72     5.35     6.09
3           1991             2.54     2.33     2.75
            1992             1.55     1.34     1.77
            1993             0.15     0.10     0.19
            2000             0.76     0.65     0.88
4           1995             3.85     3.54     4.15
            2000             2.68     2.45     2.91

        Marina Island

1           2001             0.17     0.13     0.22
2           1992             0.34     0.27     0.41
            2001             0.14     0.10     0.19
3           1992             0.71     0.61     0.83
            1993             0.33     0.23     0.53
            1994             0.12     0.07     0.17
            2001             0.18     0.14     0.22

                            Exploitation Rate

                                  95% CI

Plot        Year         Mean     Lower    Upper

         Ritchie Bay

1           1991         0.62     0.45     0.81
            1995         0.07     0.06     0.08
            2000         0.10     0.09     0.12
2           1995         0.03     0.03     0.03
            2000         0.04     0.03     0.04
3           1991         0.47     0.44     0.51
            1992         1.06     0.93     1.23
            1993         0.51     0.39     0.76
            2000         0.24     0.21     0.28
4           1995         0.05     0.05     0.06
            2000         0.08     0.07     0.08

        Marina Island

1           2001         0.96     0.75     1.33
2           1992         1.04     0.86     1.12
            2001         0.82     0.63     1.11
3           1992         0.76     0.66     0.89
            1993         0.40     0.25     0.59
            1994         0.56     0.39     0.88
            2001         0.73     0.60     0.92

Plots 1, 2, and 3 from Ritchie Bay were 0.244 ha/plot, the rest were
0.250 ha/plot. Removed total and density includes the harvest and
biosample removals. no, number; CI, confidence interval.

TABLE 3.

Summary statistics of unmodified age data from the Ritchie Bay
and Marina Island experimental plot biosample data.

                                   Age (years)

Year    Plot    Mean    Median           (Min-Max)     SE      n

Ritchie Bay

1991    1       34.1     34.0     a       (3-99)      0.87    522
1992    3       38.9     43.0     b       (3-111)     1.03    443
1995    1       24.0     16.0     c       (2-104)     0.94    512
1995    2       42.8     46.0     e       (1-123)     0.98    487
1995    4       34.4     35.0     a       (2-123)     1.07    495
2000    1       19.2      9.0     c       (3-90)      0.79    498
2000    2       37.1     41.0     b z     (3-115)     1.12    447
2000    3       12.9      7.0     f       (2-111)     0.76    443
2000    4       33.7     34.0     ay      (3-124)     1.14    507

Marina Island

1992    3       36.7     30.0     a z     (1-125)     1.16    510
2001    1       30.2     19.0     y       (3-112)     1.19    411
2001    2       16.1      9.0     x       (3-98)      1.09    289
2001    3       13.8      8.0     w       (3-116)     0.88    325

SE = standard error of mean, n = number of geoduck sampled. For each
study area, medians followed by the same letter indicated age
distributions were not significantly different (Mann-Whitney pair wise
test p > 0.05), medians followed by a different letter indicated age
distributions were significantly different (Mann-Whitney test,
P < 0.05).

TABLE 4.

Density estimate of geoduck recruitment to harvestable size in
experimental plots of Ritchie Bay and Marina Island.

                               Number of geoduck [m.sup.-2]
                 Age
Year    Plot    (year)    Mean     Lower 95% CI    Upper 95% CI

Ritchie Bay

1991    1         6       0.104       0.053           0.179
1992    3         6       0.021       0.006           0.039
1995    1         6       0.068       0.032           0.114
1995    2         6       0.051       0.000           0.096
1995    4         6       0.109       0.055           0.169
2000    1         6       0.177       0.122           0.231
2000    2         6       0.179       0.095           0.285
2000    3         6       0.165       0.127           0.203
2000    4         6       0.137       0.089           0.192

Marina Island

1992    3         7       0.028       0.017           0.043
2001    1         7       0.009       0.005           0.014
2001    2         7       0.020       0.013           0.029
2001    3         7       0.030       0.021           0.041

CI, confidence interval.

TABLE 5.

Density of geoduck from the sampled Ritchie Bay study plots to
show the possible influence on recruitment of 4-year-olds in 1995
and 9-year-olds in 2000 by artificial seeding in 1991.

                              Number of Geoduck [m.sup.-2]
                 Age
Year    Plot    (year)    Mean    Lower 95% CI    Upper 95% CI

1995    1 *       4       0.60        0.47            0.74
1995    2 *       4       0.33        0.21            0.48
1995    4         4       0.14        0.08            0.20
2000    1 *       9       0.19        0.14            0.25
2000    2 *       9       0.24        0.14            0.36
2000    3         9       0.07        0.05            0.10
2000    4         9       0.10        0.06            0.14

* Plots 1 and 2 were seeded with hatchery raised geoduck juveniles
([greater than or equal to] 0.6 year) during September 1991. Plots 3
and 4 were not artificially seeded. CI = confidence interval.

TABLE 6.

von Bertalanffy growth parameters for P. abrupta from experimental
plots at Ritchie Bay and Marina Island.

    Plot         Year       [L.sub.[infinity]]

 Ritchie Bay

      1          1991    150.5 bcx (146.3-155.7)
      1          1995      140.4 a (139.2-141.9)
      1          2000      139.9 a (137.5-141.6)
      2          1995      150.5 c (149.5-151.7)
      2          2000      148.5 b (147.2-149.8)
      3          1992        159.0 (151.1-164.8)
      3          2000        144.3 (142.2-146.4)
      4          1995      147.6 b (146.3-148.9)
      4          2000      148.0 b (146.8-149.3)
   1,2,3,4       2000        145.8 (145.0-146.8)

Marina Island

      1          2001        137.8 (136.5-139.2)
      2          2001      134.4 m (132.5-136.5)
      3          1992      149.6 x (144.7-155.9)
      3          2001      131.0 m (129.0-133.1)
    1,2,3        2001        136.1 (135.0-137.1)

    Plot                   k                     [sigma]

 Ritchie Bay

      1           0.234 a (0.204-0.269)    16.62 (14.48-19.54)
      1           0.475 b (0.432-0.521)    10.54 (9.93-11.49)
      1           0.319 c (0.294-0.352)    13.48 (12.59-14.59)
      2           0.405 d (0.372-0.437)    10.81 (10.06-11.60)
      2           0.304 c (0.282-0.335)    10.95 (10.22-11.94)
      3           0.236 a (0.217-0.266)    16.36 (13.98-20.61)
      3          0.413 db (0.382-0.461)     9.96 (9.18-10.94)
      4           0.370 d (0.339-0.412)    11.62 (10.76-12.75)
      4           0.306 c (0.285-0.342)    10.37 (9.74-11.19)
   1,2,3,4          0.328 (0.314-0.340)    12.14 (11.64-12.64)

Marina Island

      1           0.277 x  (0.259-0.300)    10.78 (9.95-11.61)
      2           0.262 x  (0.245-0.284)    11.57 (10.53-12.91)
      3              0.155 (0.136-0.172)    19.18 (17.21-22.03)
      3           0.266 x  (0.250-0.285)    10.82 (9.93-11.99)
    1,2,3           0.252  (0.244-0.263)    11.35 (10.84-12.03)

    Plot           n

 Ritchie Bay

      1            82
      1           314
      1           427
      2           397
      2           365
      3            58
      3           297
      4           365
      4           421
   1,2,3,4       1510

Marina Island

      1           348
      2           223
      3            90
      3           255
    1,2,3         826

Values in brackets are approximate 95% confidence intervals. The von
Bertalanffy parameter [t.sub.o] was set at zero. Mean
[L.sub.[infinity]] and k values followed by the same letter and column
are not significantly different (P > 0.05); whereas values not followed
by the same letter are significantly different (P < 0.05) using the
Kiruma method (see text and Table 6 for more details). Comparisons of
geoduck growth in plots between study areas were conducted only for
plots that were sampled in the same year or one year apart. [sigma]
is the mean square error of the model.

TABLE 7.

Results of the likelihood ratio tests using the Kimura (1980) method
to compare von Bertalanffy growth parameters of geoduck between
years and plots for Ritchie Bay and Marina Island.

             Comparisons                     P-values Hypotheses

 Plot      Year versus    Plot     Year      H1      H2       H3

     Ritchie Bay           Ritchie Bay

   1          1991          1      1995     0        0       0
   1          1991          1      2000     0        0       0
   1          1991          2      1995     0        1.0     0
   1          1991          2      2000     0        0.3     0
   1          1991          3      1992     0.88     0.02    0.03
   1          1991          3      2000     0        0       0
   1          1991          4      1995     0        0.13    0
   1          1991          4      2000     0        0.16    0
   1          1995          1      2000     0        0.65    0
   1          1995          2      1995     0.049    0       0
   1          1995          2      2000     0        0       0
   1          1995          3      1992     0        0       0
   1          1995          3      2000     0.06     0       0
   1          1995          4      1995     0        0       0
   1          1995          4      2000     0        0       0
   1          2000          2      1995     0        0       0
   1          2000          2      2000     0.44     0       0
   1          2000          3      1992     0        0       0
   1          2000          3      2000     0        0       0
   1          2000          4      1995     0.04     0       0
   1          2000          4      2000     0.44     0       0
   2          1995          2      2000     0        0.02    0
   2          1995          3      1992     0        0       0
   2          1995          3      2000     0.76     0       0
   2          1995          4      1995     0.26     0       0
   2          1995          4      2000     0        0       0
   2          2000          3      1992     0        0       0
   2          2000          3      2000     0        0       0
   2          2000          4      1995     0        0.31    0.03
   2          2000          4      2000     0.84     0.57    0.83
   3          1992          3      2000     0        0       0
   3          1992          4      1995     0        0       0
   3          1992          4      2000     0        0       0
   3          2000          4      1995     0.13     0.01    0.04
   3          2000          4      2000     0        0       0
   4          1995          4      2000     0        0.61    0.01

    Marina Island         Marina Island

   1          2001          2      2001     0.30     0.02    0
   1          2001          3      1992     0        0       0
   1          2001          3      2001     0.42     0       0
   2          2001          3      1992     0        0       0
   2          2001          3      2001     0.70     0.08    0.03
   3          1992          3      2001     0        0       0

     Ritchie Bay          Marina Island

   1          1991          3      1992     0        0.81    0
   3          1992          3      1992     0        0.02    0
1,2,3,4       2000        1,2,3    2001     0        0       0

Hypotheses: H1: Common k, i.e., [L.sub.[infinity]](1) [not equal to]
[L.sub.[infinity]](2) and k(1) = k(2); H2: Common [L.sub.[infinity]],
i.e., [L.sub.[infinity]](1) = [L.sub.[infinity]](2) and k(1)
[not equal to] k(2); H3: Common k and [L.sub.[infinity]], i.e.,
[L.sub.[infinity]](1) = [L.sub.[infinity]](2) and k(1) = k(2). The
null hypothesis rejected at [alpha] <0.05 level. P-values of 0 are
<0.005.

ACKNOWLEDGMENTS

The authors thank D. Brouwer, D. Bureau, W. Carolsfeld, B. Clapp, J. Clarke, L. Convey, R. Khan, C. Hand, R. Harbo, S. Heizer, L. Lee, J. McArthur, F. Merilees, J. Nash, J. Rogers, T. Scott, K. Southey, N. W. Surry, W. Thompson: and T. White for technical assistance: W. Hajas for statistical advice: many geoduck commercial fishermen: R. Antifave, M. Barcelonne, R. Barton, M. Boudreau, K. Brewer, P. Boulton B. Crawford, G. Hache, R. Hats, G. Hickie, J. Hume, T. Keith, R. Kowatski, D. Larsom J. M. LeGuerrier, M. Lanoie, G. MacDonald, G. McKay, K. Montgomery, D. Mousseau, M. Renshaw, S. Renshaw, A. Schofield, T. Somerville, G. Sorensen and N. Sorenson, S. Soucy, R. Sutherland, K. Tull, L. Tulloch, B. Wallace-Tarry: Underwater Harvesters Association, for assisting with the field experiments; Pelegos Tracking and CF McLean Trucking for transporting samples and shells: J. H. Beattie, Washington Department of Fish and Wildlife, for supplying the hatchery reared juvenile geoduck: C. Hand, W. Hajas and anonymous reviewers for helpful comments, which improved this manuscript.

LITERATURE CITED 

Andersen, A. M. 1971. Spawning, growth and spatial distribution of the geoduck clam. Panope generosa (Gould), in Hood Canal, Washington. Ph.D. thesis, University of Washington, Seattle. 128 pp.

Bain, L. J. & M. Engelhardt. 1991. Introduction to probability and mathematical statistics, 2nd ed. Belmont: Duxbury Press. 644 pp.

Beattie. J. H. 1992. Geoduck enhancement in Washington State. Bull. Aquac. Assoc. Can. 4:18 24.

Beattie. J. H. & B. Blake. 1999. Development of culture methods for the geoduck clam in the USA (Washington State) and Canada (British Columbia). World Aquac. Sep.:50 53.

von Bertalanffy, L. 1938. A quantitative theory of organic growth (inquiries on growth laws II). Human Bio. 10:181 213.

Bradbury, A. & J. V. Tagart. 2000. Modeling geoduck, Panopea abrupta (Conrad, 1849) population dynamics. II. Natural mortality and equilibrium yield. J. Shellfish Res. 19:63 70.

Breen, P. A. & T. L. Shields. 1983. Age and size structure in five populations of geoduc clams (Panope generosa) in British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 1169:62.

Breen, P. A., C. Gabriel & T. Tyson. 1991. Preliminary estimates of age, mortality, growth, and reproduction in the hiatellid clam Panopea zelandica in New Zealand. New Zealand J. Mar. Freshwater Res. 25: 231-237.

Bureau, D., W. Hajas, C. M. Hand & G. Dovey. 2003. Age, size structure and growth parameters of geoducks (Panopea abrupta, Conrad 1849) from seven locations in British Columbia sampled in 2001 and 2002. Can. Tech. Rep. Fish. Aquat. Sci. 2494:29.

Bureau, D., W. Hajas, N. W. Surry, C. M. Hand, G. Dovey & A. Campbell. 2002. Age, size structure and growth parameters of geoducks (Panopea abrupta. Conrad 1849) from 34 locations in British Columbia sampled between 1993 and 2000. Can. Tech. Rep. Fish. Aquat. Sci. 2413:84.

Campbell A., R. Harbo & S. Heizer. 1996a. A survey of geoduck population density near Sandy Island, Comox, 1993. Can. Tech. Rep. Fish. Aquat. Sci. 2089:132-156.

Campbell A., R. Harbo & S. Heizer. 1996b. A survey of geoduck population density at Marina Island, 1992. Can. Tech. Rep. Fish. Aquat. Sci. 2089:157-203.

Campbell A., R. M. Harbo & C. M. Hand. 1998a. Harvesting and distribution of Pacific geoduck clams, Panopea abrupta, in British Columbia. Can. Spec. Publ. Fish. Aquat. Sci. 125:349-358.

Campbell. A., C. M. Hand, C. Paltiel, K. N. Rajwani & C. J. Schwartz. 1998b. Evaluation of some survey methods for geoducks. Can. Tech. Rep. Fish. Aquat. Sci. 2221:5-42.

Campbell, A. & M. D. Ming. 2003. Maturity and growth of the Pacific geoduck clam, Panopea abrupta, in southern British Columbia. J. Shellfish Res. 22:85-90.

Clapp, B. 2000. Underwater harvester's association geoduck enhancement program. J. Shellfish Res. 19:620.

Coan, E. V., P. H. Scott & F. R. Bernard. 2000. Bivalve seashells of western North America. Santa Barbara: Museum of Natural History Publication. 566 pp.

Cox, R. K. & E. M. Charman. 1980. A survey of abundance and distribution (1977) of the geoduck clam (Panope generosa) in Queen Charlotte, Johnstone and Georgia Straits, British Columbia. Marine Res. Branch. Fish. Management Rep. 16:122.

Efron, B. & R. J. Tibshirani. 1993. An introduction to the Bootstrap. New York: Chapman & Hall. 436 pp.

Fyfe, D. A. 1984. The effect of conspecific association on growth and dispersion of the geoduck clam, Panope generosa. M.Sc. thesis, Simon Fraser Univ. 110 pp.

Goodwin, C. L. 1973. Subtidal geoducks of Puget Sound, Washington. Washington Dept. Fish. Tech. Rep. 13:64.

Goodwin, C. L. 1977. The effects of season on visual and photographic assessment of subtidal geoduck clam (Panope generosa Gould) population. The Veliger 20:155-158.

Goodwin, C. L. 1978. Some effects of subtidal geoduck (Panope generosa) harvest on a small experimental plot in Hood Canal, Washington. Washington Dept. Fish. Prog. Rep. 66:21.

Goodwin, C. L. & B. Pease. 1989. Species profiles: life histories and environmental requirements of coastal fish and invertebrates (Pacific Northwest)-Pacific geoduck clam. (11.120). US Army Corps of Engineers. US. Wildl. Serv. Biol. Rep. 82-84:15.

Goodwin, C. L. & B. C. Pease. 1991. Geoduck, Panopea abrupta (Conrad, 1849), size, density, and quality as related to various environmental parameters in Puget Sound, Washington. J. Shellfish Res. 10:65-77.

Goodwin, C. L. & W. Shaul. 1984. Age, recruitment and growth of the geoduck clam (Panope generosa, Gould) in Puget Sound Washington. Wash. Dept. Fish. Prog. Rep. 215:30.

Gribben, P. E., J. Helson & A. G. Jeffs. 2004. Reproductive cycle of the New Zealand geoduck, Panopea zelandica, in two north island populations. The Veliger 47:53-65.

Hand, C. M. & D. Bureau. 2000. Quota options for the geoduck (Panopea abrupta) fishery in British Columbia for 2001 and 2002. Ottawa: Can Stock Assessment Secretariat Rex. Doc. 2000/163. Available at (http://www.dfo-mpo.gc.ca/csas/Csas/English/Publications/ Research_Doc_e.htm).

Hand, C. M. & G. Dovey. 1999. A survey of geoduck populations in the Elbow Bank and Yellow Bank area of Clayquot Sound, west Vancouver Island in 1994 and 1995. Can. Manus. Rep. Fish. Aquat. Sci. 2479: 33.

Harbo, R. M., B. E. Adkins, P. A. Breen & K. L. Hobbs. 1983. Age and size in market samples of geoduc clams (Panope generosa). Can. Manus. Rep. Fish. Aquat. Sci. 1714:77.

Hoffmann, A., A. Bradbury & C. L. Goodwin. 200l). Modeling geoduck, Panopea abrupta (Conrad, 1849) population dynamics. I. Growth. J. Shellfish Res. 19:57-62.

Jamison, D., R. Heggen & J. Lukes. 1984. Underwater video in a regional benthos survey. Honolulu: Proceedings Pacific Congress on Marine Technology, Mar. Tech. Soc., April 24 to 27, 1984.

Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776.

King, J. J. 1986. Juvenile feeding ontogony of the geoduck, Panope abrupta (Bivalvia: Saxicavacea), and comparative ontogeny and evolution of feeding in bivalves. M.Sc. thesis, Univ. Victoria. 281 pp.

Noakes, D. J. 1992. On growth and mortality of geoduck clams (Panope abrupta) (or how fast do all "ducs go to heaven"?). Can. Manus. Rep. Fish. Aquat. Sci 2169:22-34.

Noakes, D. J. & A. Campbell. 1992. Use of geoduck clams to indicate changes in the marine environment of Ladysmith Harbour, British Columbia. Environmetrics 3:81-97.

Orensanz, J. M., R. Hilborn & A. M. Parma. 2000. Harvesting Methuselah's clams--is the geoduck fishery sustainable, or just apparently so? Can. Stock Assessment Secretariat Res. Doc. 2000/175. (http:// www.dfo-mpn.gc.ca/csas/Csas/English/Publications/Research_ Doc_e.htm).

Shaul, W. & C. L. Goodwin. 1982. Geoduck (Panope generosa: Bivalvia) age as determined by internal growth lines in the shell. Can. J. Fish. Aquat. Sci. 39:632-636.

Sloan, N. A. & S. M. C. Robinson. 1984. Age and gonad development in the geoduck clam Panope abrupta (Conrad) from southern British Columbia, Canada. J. Shellfish Res. 4:131-137.

S-plus. 2001. Version 6.0. Seattle, WA: Insightful Corp.

Turner. K. C. & R. K. Cox. 1981. Seasonal reproductive cycle and show factor variation of the geoduck clam, Panope generosa (Gould), in British Columbia. J. Shellfish Res. 1:125.

Zhang, Z. & A. Campbell. 2004. Natural mortality and recruitment rates of the pacific geoduck clam, Panopea abrupta, in experimental plots. J. Shellfish Res 23:675-682.

A. CAMPBELL, (1) C. W. YEUNG, (1) G. DOVEY (2) AND Z. ZHANG (1)

(1) Pacific Biological Station, Nanaimo, BC, Canada V9T 6N7; (2) West 123[degrees] Resource Consulting Inc., PO Box 781, Ladysmith, BC, Canada V9G IA6