Variations in growth and reproduction of bay scallops (Argopecten irradians) (Lamark, 1819) from six subpopulations in the northeastern Gulf of Mexico.

ABSTRACT Bay scallops, Argopecten irradians, sustained a valuable commercial and recreational fishery in Florida during the middle part of the last century. In 1994, after decades of declining stocks, state managers closed this fishery to commercial harvest and severely limited the recreational harvest. In Florida, most bay scallops reproduce only once, generally during the fall at an age of roughly 12. The current 10-week recreational harvest season, July 1 through September 10, occurs at a time when somatic and reproductive tissues are changing rapidly in preparation for the fall spawning season. This study describes changes in tissue weights (reproductive, somatic, visceral) during the 2002 harvest season in scallops collected from six subpopulations along Florida's Gulf of Mexico coast. We observed tissue-specific patterns of weight change during the 7-month study and also noted regional differences. In three Panhandle sites (St. Joseph Bay, Lanark Village, and St. Marks), a shift from somatic growth to reproductive growth occurred later in the year and was more intense than in peninsular sites (Steinhatchee, Homosassa and Anclote). We also monitored recruitment of juvenile scallops at three of the sites from July of 2001 through June of 2003. There were protracted fall and winter peaks within a background of constant, low-level recruitment. The harvest limits allow each person to collect two gallons (7.55 L) of whole scallops or one pint (0.47 L) of adductor muscle meat each day. In June, just prior to the harvest season, the number of whole scallops that would be collected varied significantly between sites (range 55-203), as did the equivalent yield of adductor muscles (range 143-342). Muscle size peaked in August or September, depending on location. The allowable number of scallops collected within the volume-based limits had decreased (41-112 whole scallops or 84-116 shucked scallops) and between-site variability in the numerical harvest was lower.

KEY WORDS: Argopecten irradians, Florida, gonadal-somatic index, growth, harvest, recruitment, bay scallop 

INTRODUCTION

Bay scallops, Argopecten irradians (Lamark, 1819), were once a valuable commercial fisheries species for the state of Florida. During the 1950s, harvests commonly exceeded 100,000 pounds per year of adductor muscle (Murdock 1955) and peaked at 401,283 pounds in 1958 (Rosen 1959). There is also evidence for prehistoric sustenance harvest (Marelli & Arnold 2001) and a modern cultural heritage of recreational harvest. Stocks declined during the 1960s and 1970s, with a concomitant decrease in commercial harvest to levels between 10,000 and 20,000 pounds of adductor muscle per year by the 1980s (for example, see Joyce 1982). The abundance of scallops in many subpopulations declined to very low densities or disappeared completely. This decline was first documented in Pine Island Sound, then in Tampa Bay, the Anclote River Estuary and finally in the Homosassa River region. Similar declines have been observed in the western Florida Panhandle (Arnold et al. 2005).

In response to the decline of bay scallop abundance in Florida waters, harvest restrictions were instituted in 1985 and expanded in 1994. Management of the fishery targeted both commercial and recreational harvest. Commercial harvest was eliminated and recreational harvest was severely curtailed in 1994. All subpopulations south of the Suwannee River were closed to recreational harvest. The daily harvest limit of five gallons (18.88 L) of whole scallops per person was reduced to the current limit of two gallons (7.55 L) of whole scallops or one pint (0.47 L) of shucked adductor muscle per individual. The duration of the harvest season was reduced from 10 mo per year to 10 wk per year. Based on a survey program that was initiated in 1993, subpopulations are currently grouped into 3 classes: healthy (average density >25 scallops/600 [m.sup.2]), transitional (5-25 scallops/600 [m.sup.2]) or collapsed (<5 scallops/ 600 [m.sup.2]). In 1994, all populations south of the Suwannee could be classified as either collapsed or transitional. To accompany the management actions, active restoration was initiated in 1998 and continues to this day. The goal of the restoration program was not to produce scallops for direct harvest but to create high-density spawning aggregations that may then seed the natural populations. Hatchery-reared scallops are held in cages at densities up to 300 per 0.6 [m.sup.2] in Crystal Bay, Homosassa River estuary, the Anclote River estuary (hereafter Homosassa and Anclote), Tampa Bay and Sarasota Bay. Subsequent increases in scallop abundance in the Homosassa subpopulation (Arnold et al. 2005) have been extensive enough that the recreational fishery was reopened in 2002.

In Florida, recreational harvest occurs during summer to accommodate the most popular means of collection, snorkeling. Current harvest of bay scallops in Florida is concentrated on a limited number of traditional subpopulations. Although the harvest season occurs immediately prior to the spawning season, previous research indicates that harvest occurs when muscle yield per individual is maximal (Barber & Blake 1983). A similar management approach is applied in the northeastern United States (bay scallop harvest is timed to maximize the yield of adductor muscle; Rhodes 1991, Estabrooks 2003). However, those animals are allowed to pass through their first spawning season and are harvested at a time when adductor muscles have begun a period of new growth, when most animals in the population are roughly 15 mo of age.

Bay scallops have been described as quasi-semelparous; most individuals spawn only once, but a small portion survive for one or more additional spawns (Orensanz et al. 1991). The traditional view of this species suggests that local subpopulations will spawn synchronously, either during warming spring or cooling fall periods (Sastry 1963, 1970). Recent research suggests that this is not always the case in Florida subpopulations. Arnold et al. (1998) monitored recruitment of spat in four distinct subpopulations on Florida's west coast from early August through late February. They found that whereas distinct peaks occurred in each locale, some recruitment occurred throughout the study in three of the four subpopulations. In a separate study, Bologna (1998) collected recruits in 12 of 28 mo during 1994 through 1996 in St. Joseph Bay, Florida. Over the course of the study, recruits were collected in 8 of the 12 calendar months. Secondary recruitment periods have also observed in Long Island, New York (Tettlebach et al. 1999).

Tissue growth in bay scallops is most rapid during the first year of life. Somatic growth will often precede gonadal growth, with energy stored in the adductor muscle as glycogen and protein and in the digestive gland as lipid (Barber & Blake 1981). Growth of these two organs may slow or be negative once gametogenesis begins, at which point gonad growth is very rapid (Bricelj et al. 1987, Barber & Blake 1983). Postspawn Florida scallops can suffer high mortality that is correlated with depleted energy reserves (Barber & Blake 1983). In northeastern United States populations, a higher percentage of scallops survive for additional spawns (Bricelj 1992). The scallops that survive a spawning cycle will initiate new growth in the adductor muscle, and gonadalsomatic indices do not attain as high a level during subsequent spawns, although actual gonad weight continues to increase (Bricelj 1992).

In this study, we explore small-scale variations in tissue-specific size and growth during a single harvest season (2002) in six subpopulations of bay scallops along Florida's Gulf of Mexico coast. The results from monitoring juvenile scallop recruitment are used to refine the interpretation of tissue growth and reproduction in three of those six subpopulations. The changes in the amount of allowable harvest, in terms of the number of scallops harvested, that accompany the changes in scallop shell and tissue (in particular adductor muscle) volume are described.

MATERIALS AND METHODS

Adult Collections

Argopecten irradians were collected from six sites along the Florida Gulf coast between Port St. Joe and Tampa Bay (n = 1226, Fig. 1). Four sites (Anclote, Homosassa, Steinhatchee and St. Marks) are part of the Big Bend region of Florida, a zero-energy coastline. The coastal seagrass bed found here varies between 11 and 35 km wide, an area approximately 3,000 [km.sup.2], and is dominated by Thalassia testudinum, Syringodium filiforme and Halodule wrightii (Iverson & Bittaker 1986). The St. Joseph Bay and Lanark sites are also low energy and have similar seagrass communities but are protected by barrier islands.

[FIGURE 1 OMITTED]

The bay scallop abundance at four of the six sites is routinely monitored (Arnold et al. 2005). Anclote was the southernmost site. During 11 y of survey there (1994-2004), the average scallop density varied between 0.15 and 47.35/600 [m.sup.2]. During three of the years the density was considered healthy, during four years it was considered transitional, and during four years it was considered collapsed. It has not been open for harvest since 1993. Over the same period, the Homosassa population was considered healthy in 5 years, transitional in 3 years, and collapsed in 3 years. The study year, 2002, was the first year this site had been opened to harvest since 1993. Steinhatchee is the most stable subpopulation in the state; its density had been considered healthy in all years except 2004, when it was in classified as transitional. The St. Marks and Lanark subpopulation densities are not monitored. The St. Joseph Bay subpopulation was also generally healthy (7 of the 11 y), although it was classified as transitional in 2 years and collapsed in 2 years. All four monitored subpopulations were considered to be healthy during 2002.

Our goal was to collect samples of 30 adults from each study site at 3-week intervals between May and September 2002 (Table 1), with at least one sample set collected before and one sample set collected after the harvest season. However, we began collecting samples as early as April 24 at Anclote and Homosassa and May 16 to May 17 at Steinhathcee, Lanark and St. Joseph Bay (because collections were made opportunistically). The study was terminated as early as September 3 at Steinhatchee (because of logistic constraints) and as late as October 10 at the Homosassa site (Table 1). Live scallops were collected opportunistically by hand and held in coolers for transport to shore. Samples collected at the Anclote and Homosassa sites were placed directly into freezers at the Fish & Wildlife Research Institute (FWRI) in St. Petersburg. Samples collected at all other locations were initially frozen and then were shipped overnight (packaged with dry ice) to St. Petersburg. All samples were then held at -80[degrees]C until analysis.

Shell and Tissue Analysis

Scallops were thawed for at least 1 hour before dissection. Many shells were covered in epibionts such as oysters, jingle shells and slippers snails. These fouling organisms were removed. The soft tissues of the scallop were separated into three components: gonad, adductor muscle, and all other tissue (hereafter termed viscera). Each of the tissues was placed into tared aluminum trays and the wet weights were measured to the nearest 0.1 mg. The volume of the adductor muscle was determined by submerging the muscle in a known volume of water in a graduated cylinder and observing the increase in volume to the nearest 0.2 mL. All components were then dried to constant weight in an oven at 60[degrees]C to 65[degrees]C for 20-24 h and the dry weights of each of the three tissues measured. Only mature specimens with a gonad dry weight <0.001 g were included in statistical comparisons.

After all tissue was removed, the valves of each scallop were rearticulated and the height, width and length were measured to the nearest 0.1 mm using digital calipers. To estimate the volume of a collection of whole scallops, rearticulated shells were taped back together and the entire sample from each sampling date and location was placed inside a 5,000-mL graduated beaker. The dry volume of the entire sample was estimated to the nearest 100 mL.

This process was repeated three times and an average volume was calculated to estimate the error variance associated with the packing of the shells.

The uneven sampling effort resulted in a matrix of sample dates and locations that was not amenable to a full factorial design. Initial analysis of the data was conducted using an analysis of covariance (ANCOVA), with the variable of interest as the dependent factor (shell height, adductor muscle dry weight, muscle volume, or gonadal-somatic index [GSI]), site as a categorical factor, and Julian date as a covariate. We calculated GSI as the ratio of reproductive tissue to all other tissue (muscle and viscera). When the ANCOVA model showed that the parameter of interest varied significantly either temporally or between sites, further analysis was conducted. We used either Dunn multiple comparison test or Tukey multiple comparison test for uneven sample sizes to identify differences between collection dates within each site or between sites within each month. We calculated growth rate for each subpopulation by least squares linear regression, using shell height as the dependent variable and Julian calendar day as the independent variable. In each of the six regressions the residuals were normally distributed and homoscedastic. The slope of the resulting equation can be interpreted as the growth rate in mm per day.

Recruit Monitoring

As part of an ongoing study (Arnold et al. 1998, 2005) recruit collectors were deployed in 3 of our study sites: Anclote, Homosassa and St. Joseph Bay. Each recruit collector consisted of a 4-mm mesh, 1/2-bushel citrus bag containing a 30 x 48-cm panel of 4-mm mesh polypropylene (Ambrose et al. 1992). The collector was attached 0.25 m above a cinder block anchor with 1/2-inch polypropylene rope. A small plastic donut float was attached to the opposite end of the trap, and a crab-trap float was attached to the distal end of the rope, each helping to keep the trap in a roughly vertical orientation off the bottom.

Each trap was placed in seagrass beds at depths of 0.5-1.5 m of water at low tide. Twenty-four collectors were deployed at each site, composed of 2 sets of 12, paired stations. At each station, one trap from each pair was pulled on alternate, 3-week target schedules, so that each trap soaked for roughly 6 weeks (median = 43 days). Actual soak times varied between 36 and 84 days. Each trap was retrieved and placed in a labeled garbage bag without other preservation and was examined at a later date for scallop recruits. All surfaces (inside and outside of the citrus bag, and both sides of the polypropylene mesh) were examined with the naked eye, and all spat were removed and counted. Recruitment rate was calculated as the number of recruits divided by the number of days the trap had soaked. Mean recruitment rates were calculated weekly for each station by averaging the recruitment rate of all collectors at each location. Results are presented for traps retrieved during the period of July 2001 to June 2003.

RESULTS

Initial analyses using ANCOVA showed that each of the 4 variables of interest (shell height, adductor muscle dry weight, adductor muscle volume and the ratio of gonadal-somatic index) varied both between sites and temporally (Table 2). In each case the interaction term was also significant, indicating that the differences between the six sites were not consistent during the 6 months of the study. Further analyses were focused on elucidating where these differences existed.

The overall trend was for shell height to increase at each site during the course of the study but at rates that varied between sites. The growth rate ranged from 0.075 mm [day.sup.-1] at Lanark Village to 0.208 mm [day.sup.-1] at St. Marks (Table 3). Immediately prior to the July 1st opening of the recreational harvest season, mean shell heights at Anclote and Lanark were significantly larger than they were at the other tour sites. During the middle of the harvest season (late July and early August) mean shell height had increased slightly and was uniform across the sample region, with the exception of one sample period (August 9-29), when the scallops at St. Joseph Bay were smaller than scallops at all other sites (Table 3). At the conclusion of the harvest season, shell heights at most sites had continued to increase slightly and were similar at most sites. The exceptions were the scallops at St. Marks, which had grown significantly larger than those at all other sites.

Adductor-muscle dry weight and volume both changed in similar ways during the course of the study, exhibiting two general trends. At St. Joseph Bay, adductor muscle volume increased throughout the season. At the other locations, muscle volume peaked in August then declined in September and October (Fig. 2). Adductor-muscle dry weight peaked in August and September at St. Marks and St. Joseph Bay. At the other four sample locations, August peaks were followed by declines in September and October (Fig. 3).

[FIGURES 2-3 OMITTED]

Gonad dry weight also changed over the course of the study, again having two general trends. At the peninsular locations (Anclote, Homosassa and Steinhatchee) gonad dry-weight initially increased slowly then leveled off in August and September. In Panhandle sites (St. Marks, Lanark and St. Joseph Bay), gonad dry weight also initially increased slowly but then increased rapidly in September to a much greater extent than at the three peninsular sites (Fig. 4).

[FIGURE 4 OMITTED]

Changes in gonadal-somatic index (GSI) followed two distinct trends. GSI at all sites was low when the study was initiated; mean values ranged from 0.02-0.05 during April through June (Fig. 5). At the three peninsular sites, significant increases in GSI were observed in August and September, but the mean GSI never exceeded 0.12. At the three Panhandle sites, the increases in GSI that occurred in September were more pronounced. Mean GSI in September increased to values of 0.20 in St. Marks, 0.14 in Lanark and 0.19 in St. Joseph Bay.

[FIGURE 5 OMITTED]

Three trends in the number of scallops estimated to be harvested by someone collecting their recreational limit were observed. First, the number of scallops that can be legally harvested generally declines as the season progresses, regardless of which measure is used (Fig. 6). This trend was not entirely linear, and an increase in the number of scallops harvested using the one-pint limit did occur late in the season because muscle size began to decline. Second, more scallops could be legally harvested by someone shucking their scallops at sea and collecting one pint of meat than by someone collecting two gallons of whole scallops and shucking them on shore. This pattern is especially true early in the season, but the disparity declines as the season progresses. Finally, during the harvest season Lanark had the largest scallops with the largest adductor muscles, so a fisher harvesting their legal limit from this area would take fewer scallops than at any other area.

[FIGURE 6 OMITTED]

If the harvest season were to open in May, a legal harvest would range from 65-215 whole scallops and between 213 and 429 shucked scallops (Fig. 6). Under current regulations, when the season opens on July 1, legal harvest would range from 55-203 whole scallops or 143-342 shucked scallops. When scallops at most sites achieved peak muscle size during early August (midway through the harvest season), legal harvests ranged from 41-112 whole scallops or 84-116 shucked scallops. Thus, there is a disparity between sites in the number of scallops that could be legally harvested, but this disparity declines as the season progresses.

In each year, all three monitored subpopulations had recruitment that occurred with peaks but also with protracted periods of very low levels of recruitment (Fig. 7). At Anclote, spat settled over a protracted period in both years. Recruitment peaked from late October 2001 through January 2002, but some recruits were collected in every sample period from August 2001 through May 2003. Modest peaks also occurred in August and during October through December 2002. At the Homosassa site, spat recruited to collectors from August through June, with peaks in September 2001 through January 2002 and again in August through November of 2002. In St. Joseph Bay, we observed recruitment throughout the study. A large peak in recruitment began in October 2001 and continued through January 2002. A more modest peak in recruitment was observed in late October and continuing through November in 2002. One clear trend is that Homosassa traps (maximum rate of 0.38 spat per bag per day) collected scallops at a rate roughly one order of magnitude less than Anclote traps did (maximum rate of 6.3 spat per bag per day), and Anclote traps collected scallops at a rate roughly one order of magnitude less than traps in St. Joseph Bay did (maximum rate of 48 per bag per day).

[FIGURE 7 OMITTED]

DISCUSSION

In the six subpopulations of bay scallops that we sampled from Florida waters, we observed higher than expected variability between sites and predictable but still variable temporal changes. In each subpopulation, we documented growth and development towards a maximal spawning condition in early fall, as predicted by previous studies (Barber & Blake 1983). All six subpopulations showed a positive trend in shell height and gonad weight, and all of the subpopulations experienced a relative increase in mean muscle weight and volume during the initial portion of the study. However, there were regional differences in the amount of energy devoted to somatic growth versus gonad growth and reproduction. The scallops from the Florida Panhandle sites had more pronounced increases in gonadal-somatic index in September than did scallops from the peninsular sites.

Changes in the value of a ratio can occur when either of its components change but can also remain the same when both components change similarly. In this study, increases in GSI might be expected at the peninsular sites, because scallops at all three had decreases in muscle weight and volume in the fall. However, gonad weight also declined (indicative of a spawning event), so the ratio changed little. These observations do not support the expectation that energy reserves were transferred from muscle to gonad. In two of the three Panhandle sites (Lanark and St. Marks), scallop muscle weight declined, whereas gonad weight increased, resulting in significant increases in GSI, evidence that would support the expectation of a transfer of energy from muscle to gonad. These two subpopulations are most similar to scallops observed by Sastry (1966) and Barber and Blake (1985) in which energy reserves in muscle and digestive gland were transferred into reproductive tissue. In St. Joseph Bay, both muscle and gonad weights increased, whereas GSI increased greatly. Thus, a contrast between scallops in the two regions suggests that in 2002, those at peninsular sites had limited energy available for an early fall spawn, whereas those at Panhandle sites had relatively more energy for a late fall spawn, as seems to be evidenced by the actual size of the gonad or the GSI.

Studies of recruitment can supply evidence supporting the assumption that peaks in gonad weight and GSI are correlated with a successful spawning event. Settlement of scallops on spat collectors occurred contemporaneously with peaks in GSI (this study). At Anclote, spat settled over a protracted period (August 2002 through January 2003). The initiation of this recruitment was concurrent with the weak peak in GSI. Similarly, at the Homosassa site, a weak peak in GSI was observed in August, and again, spat recruited to collectors in August through December. In St. Joseph Bay, before we discontinued this study, we observed a peak in GSI in September. A large peak in recruitment began in November 2002 and continued through January of 2003, suggesting that the large energetic reserves we observed translated into a large recruitment event in St. Joseph Bay.

Bologna (1998) suggested that the St. Joseph Bay subpopulation spawned year-round but that the peak of the spawning season varied from year to year. In that study, gonadal-somatic condition peaked during the winter, but recruits were collected in nearly every month. Barber & Blake (1981, 1983) found that the scallops in the Anclote estuary devoted most of their assimilated energy to somatic growth in spring. Histological analysis indicated that oogenesis was occurring without large increases in gonad size during May. Muscle, mantle and digestive tissue weights peaked in August. Further tissue growth was limited to increased gonad size in the early fall, including transfers of stored energy from muscle and viscera to the gonad. Cytoplasmic growth and vitellogenesis were associated with both increases in gonad weight (or body component index) and plateaus then declines in muscle weight (Barber & Blake 1983). The 3-year study found that the initiation of this process varied between the beginning of June and the middle of July but always peaked in September. Their observations compare well with ours.

The variability observed in shell size showed no easily discernable regional pattern. The two sites at which shell sizes were statistically larger than the others at the beginning of the study (Lanark and Anclote) are interspersed spatially among the other four sites. By September, all mean shell heights were around 60 mm at all of the sites. The simplest explanation would be that at or near that size, scallops along Florida's gulf coast begin to use available energy for reproduction. Evidence from recruitment studies supports this claim. At the Anclote and Homosassa sites, recruitment began in August and continued through the fall, whereas very little recruitment was observed in St. Joseph Bay (whose mean shell height was smallest in 2002) until January 2003. Alternatively, St. Joseph Bay animals may have been devoting energy to spawning in the spring of 2002, as suggested by observed recruitment throughout the spring, whereas the Anclote and Homosassa subpopulations did not produce a large number of spring recruits. The St. Joseph Bay animals may have been in a refractory state during summer of 2002 in preparation for a second major spawning event.

One possible explanation for the discrepancies in sizes between sites is that the subpopulations with smaller mean sizes in May had recruited more recently. The adults used in this study would have recruited sometime during fall 2001 or winter 2002. Evidence from recruitment in fall of 2001 and winter and spring of 2002 does not support the hypothesis for earlier recruitment at Anclote and Homosassa relative to St. Joseph Bay. At the Homosassa site, minor peaks in recruitment occurred in September through October and in December, but some recruits were collected in all months from August 2001 through May 2002. Both the St. Joseph Bay and Anclote subpopulations recruited in fairly distinct peaks during late fall 2001 and early winter 2002, but if anything, the peak recruitment in St. Joseph Bay was earlier, around November 1st, than at Anclote, which peaked closer to December 1st. The overlapping recruitment periods indicate that differences in settlement date were not responsible for the differences in size during May of 2002.

A second possible explanation for the difference in size between subpopulations is that those differences are genetically based. Marelli et al. (1997) and Marelli & Arnold (2001) studied morphological and genetic differences between Florida's Gulf of Mexico bay scallop subpopulations. They suggested that the morphological differences between local subpopulations were minimal and that all of the subpopulations studied were part of the local subspecies. In those studies, neighboring subpopulations were often more dissimilar than more distant populations and little or no geographic pattern was apparent within the Florida subpopulations. We conducted size-adjusted principal components analysis on morphological data following Marelli et al. (1997) and found results similar to theirs (unpublished data). There was no readily apparent difference in the shape of any of the scallop subpopulations' shells nor was there a discernable geographic trend (two of the closest subpopulations, St. Marks and Lanark, were the most different). Patterns of gene flow between subpopulations, as influenced by the complex nearshore hydrography of the Florida Shelf (Yang & Weisberg 1999), may affect spatial and temporal patterns of genetic variation. It seems unlikely that any of the differences observed in tissue weights can be explained by genetic variation: rather, small-scale variation in the physical or biotic environments probably resulted in the observed differences in 2002.

Bricelj et al. (1987) summarized the changes in spawning patterns of bay scallops from 3 latitudinally distinct populations: Massachusetts, New York and Florida. The general conclusions were that northern populations showed more discrete peaks in spawning activity, whereas southern populations had more protracted spawning seasons. In the northern populations, periods of somatic growth occurred in spring, prior to spawning, and again in summer and early fall. In the northern populations, a considerable portion of the population will store sufficient energy reserves to overwinter and spawn a second time. These animals will rapidly resume growth of their adductor muscles during the fall. In the current study, we observed two distinct patterns of growth. In the peninsular subpopulations, changes in muscle and viscera accounted for most of the early-summer growth. Spawning began in late summer, reminiscent of the pattern previously observed in scallops from Massachusetts (Sastry 1970), albeit more protracted in Florida's scallops. Scallops from Florida's Panhandle subpopulations continued to grow until at least the conclusion of our study, and reproduction likely peaked late in the fall, reminiscent of the spawning patterns of populations from the Anclote estuary (Blake & Barber 1983) or North Carolina (Sastry 1970).

The recreational harvest season in Florida begins during a period when scallops are undergoing rapid growth, but the initiation of this period of rapid growth varies between sites. Thus, at the beginning of the season, there is high between-site variability in shell size and adductor muscle mass. The growth is most notable in changes in soft tissues, adductor muscles and gonads. As a result, at some locations, many more scallops could legally be harvested than at others. This finding is especially true if the scallops are shucked at sea. Between July 1st and August 1st, the number of legally harvested shucked scallops declines by about one third because of increases in adductor muscle volume, a process observed at all sites. This decline is less pronounced for whole scallops.

A short delay in the harvest season would result in fewer scallops being taken for harvest, possibly allowing more scallops to survive until fall spawning. A delay in the harvest season would also allow a more simple understanding of harvest, because most areas would have a similar number of scallops being harvested by each person collecting their limit. However, natural mortality rates are poorly understood in Florida's scallop subpopulations, and a delayed harvest may not result in a large net gain in the size of the reproductive population. Estimates of instantaneous daily natural mortality in Florida between 0.01 (Greenawalt 2002) and 0.014 (Bologna 1998) indicate that there would be roughly 25% fewer scallops present on opening day if it were moved from July 1st to August 1st. The effect of delaying the fishing season would then depend on whether fishers continued to collect scallops at the same rate l month later (potentially very deleterious) or if fishing mortality, as a percentage of the population, remained steady (little net difference).

Another consideration in changes of the harvest season is that Florida's scallops begin to lose muscle mass in September, as energy is transferred to the gonad (this study; Barber & Blake, 1985). A delay in opening the harvest season, assuming the 10-wk duration was maintained, might actually cause an increased harvest of those scallops that had managed to survive until the fall spawning season. Smaller adductor meats would lead to increased harvest if a person collected the one-pint limit of meat, but the larger shells present in the fall would result in smaller yield for a person collecting a limit of two gallons of whole scallops. A reduction in the duration of the harvest season would probably be unappealing to the tourism industry and recreational fishers unless considerable gains in stability of the population were realized. An analysis of the tradeoffs between biological gain and potential economic losses resulting from changing the dates of the harvest season should be made before any management actions are proposed.

ACKNOWLEDGMENTS

This study would not have been possible without the assistance of the many people who provided samples: Melanie Parker, Sarah Peters, Jaime Greenawalt, Troy Tuckey, Dan Marelli, Tammy Summers and Robert Turpin. April Granholm, Kim Parthemore, and Chris Matterson provided assistance in the laboratory. Sarah Waiters, Heather Wolfe, Jim Quinn, Judy Leiby and two anonymous reviewers provided valuable comments on earlier drafts of this document. This research was supported by revenues collected under the Florida Saltwater Fishing License.

LITERATURE CITED

Ambrose, W. G., Jr., C. H. Peterson, H. C. Summerson & J. Lin. 1992. Experimental tests of factors affecting recruitment of bay scallops (Argopecten irradians) to spat collectors. Aquaculture 108:67-86.

Arnold, W. S., D. C. Marelli, C. P. Bray & M. M. Harrison. 1998. Recruitment of bay scallops Argopecten irradians in Floridian Gulf of Mexico waters: scales of coherence. Mar. Ecol. Prog. Ser. 170:143-157.

Arnold, W. S., S. P. Geiger, M. Parker, S. C. Peters, J. Cobb, B. Pittinger & C. Beals. 2005. Annual report of the bay scallop project 2004. Florida Fish and Wildlife Conservation Commission.

Barber, B. J. & N. J. Blake. 1985. Intra-organ biochemical transformations associated with oogenesis in the bay scallop, Argopecten irradians concentricus (Say), as indicated by [sup.14]C incorporation. Biol. Bull. 168: 39-49.

Barber, B.J. & N. J. Blake. 1983. Growth and reproduction of the bay scallop, Argopecten irradians (Lamark) at its southern distributional limit. J. Exp. Mar. Biol. Ecol. 66:247-256.

Barber, B. J. & N. J. Blake. 1981. Energy storage and utilization in relation to gametogenesis in Argopecten irradians concentricus (Say). d. Exp. Mar. Biol. Ecol. 52:121-134.

Bologna. P. A. X. 1998. Growth, production, and reproduction in bay scallops Argopecten irradians concentricus (Say) from the northern Gulf of Mexico. J. Shellfish Res. 17:911-917.

Bricelj, V. M., J. Epp & R.E. Malouf. 1987. Intraspecific variation in reproductive and somatic growth cycles of bay scallops Argopecten irradians. Mar. Ecol. Prog. Ser. 36:123-137.

Bricelj, V. M. & M. K. Krause. 1992. Resource allocation and population genetics of the bay scallop, Argopecten irradians irradians: effect of age and allozyme heterozygosity on reproductive output. Mar. Biol. 113:253-261.

Estabrooks, S. L. 2003. A rapid test for the determination of the spawning status of the bay scallop, Argopecten irradians (Lamark, 1819). J. Shellfish Res. 22:389-391.

Greenawalt, J. M. 2002. Mortality estimates and distributional patterns of the southern bay scallop along the Gulf Coast of Florida. M.S. Thesis, Univ. of Florida, Gainesville, FL. 57 pp.

Iverson, R. L. & H. F. Bittaker. 1986. Seagrass distribution and abundance in eastern Gulf of Mexico coastal waters. Est. Coast. Shelf Sci. 22:577-602.

Joyce, A. J., Jr. 1982. Summary of Florida commercial landings, 1981. Report of the Florida Department of Natural Resources.

Marelli, D. C. & W. S. Arnold. 2001. Shell morphologies of bay scallops, Argopecten irradians, from extant and prehistoric populations from the Florida Gulf Coast: implications for the biology of past and present metapupulations. J. Archaeol. Sci. 28:577-586.

Marelli, D. C., W. G. Lyons, W. S. Arnold & M. K. Krause. 1997. Sub-specific status of Argopecten irradians concentricus (Say, 1822) and of the bay scallops of Florida. Nautilus 110:42-44.

Murdock, J. F. 1955. Investigation of the Lee County bay scallop fishery. Florida State Board of Conservation, unpublished report.

Orensanz, J. M., A. M. Parma & O. O. Iribane. 1991. Population dynamics and management of natural stocks. In: S. E. Shumway, editor. Scallops: biology, ecology and aquaculture. Amsterdam: Elsevier Science Publishers B.V. pp. 625-713.

Rhodes, E.W. 1991. Fisheries and aquaculture of the bay scallop, Argopecten irradians, in the eastern United States. In: S. E. Shumway, editor. Scallops: biology, ecology and aquaculture. Amsterdam: Elsevier Science Publishers B.V. pp. 913-924.

Rosen, A. 1959. Summary of Florida marine landings and an analysis of the catch and effort of certain species, 1958. Report to the Florida State Board of Conservation.

Sastry, A. N. 1963. Reproduction of the bay scallop, Aequipecten irradians Lamark. Influence of temperature on maturation and spawning. Biol. Bull. 125:146-153.

Sastry, A. N. 1966. Temperature effects in reproduction of the bay scallop, Aequipecten irradians Lamark. Biol. Bull. 119:118-134.

Sastry, A. N. 1970. Reproductive physiological variations in latitudinally separated populations of the bay scallop, Aequipecten irradians Lamark. Biol. Bull. 138:56-65.

Tettlebach, S. T., C.F. Smith. R. Smolowitz, K. Tetrault & S. Dumais. 1999. Evidence for fall spawning of northern bay scallops Argopecten irradians irradians (Lamark 1819) in New York. J. Shellfish Res. 18:47-58.

Yang, H. & R. H. Weisberg. 1999. Response of the West Florida Shelf circulation to climatological wind stress forcing. J. Geophys. Res. 104: 5301-5320.

STEPHEN P. GEIGER, * JANESSA COBB AND WILLIAM S. ARNOLD

Florida Fish & Wildlife Research Institute, 100 8th Ave. SE, St. Petersburg, Florida 33701

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

TABLE 1.
Summary of collection dates and sample size of bay scallops
(Argopecten irradians) during 2002. The six collection
locations are listed in the first column, and total sample
sizes, n, for each location are shown in the last column.
Individual collection dates and sample sizes for each
date (given in parentheses) are listed by month.

Location         April       May         June         July

Anclote          4/24 (28)   5/08 (19)                7/01 (29)
                             5/30 (29)                7/16 (30)

Homosassa        4/24 (30)   5/08 (34)   6/21 (23)    7/16 (30)
                             5/30 (30)

Steinhatchee                 5/17 (29)   6/20 (30)    7/10 (30)
                                                      7/17 (30)
                                                      7/31 (30)

St. Marks                                6/25 (33)    7/15 (30)

Lanark                       5/16 (30)   6/05 (30)    7/09 (30)
                                                      7/31 (30)

St. Joseph Bay               5/17 (05)   6/12 (29)    7/17 (32)
                             5/26 (28)
                             5/31 (10)

Location         August      September   October      Total

Anclote          8/09 (33)   9/17 (30)   10/08 (32)   260
                 8/29 (30)

Homosassa        8/12 (30)   9/17 (30)   10/10 (30)   267
                 8/29 (30)

Steinhatchee     8/13 (30)   9/03 (30)                209

St. Marks        8/05 (30)   9/20 (30)                156
                 8/27 (33)

Lanark           8/22 (30)   9/12 (30)                180

St. Joseph Bay   8/15 (30)   9/19 (30)                154

TABLE 2.
Analysis of covariance results from analyses of site-specific and
temporal variation in four measures of growth and reproduction in bay
scallops (Argopecten irradians) during 2002: shell height, adductor
muscle dry weight, muscle volume, and the ratio of reproductive to
somatic tissue. For each variable, there was one degree of freedom
for each of the model and date components, five degrees of freedom for
both the by-site comparisons and interaction terms, and 1,214 degrees
of freedom for the error term. The mean square and F value for
ANCOVA results for each regression term are given. All
P values were <0.01.

                            Error      Intercept       Site

Shell             MS        23.26      25177.70       539.65
height            F                     1082.60       23.20

Adductor          MS        0.12         30.39         1.03
dry weight        F                     250.78         8.47

Muscle            MS        1.34        531.23        14.00
volume            F                     392.62        10.40

Gonad:            MS       0.0011       1.0101        0.0989
muscle            F                     892.88        87.41
ratio

                                        Site x
                                         Date
                            Date      Interaction

Shell             MS      25779.99      538.68
height            F        1108.49       23.16

Adductor          MS        30.71        1.03
dry weight        F        253.49        8.51

Muscle            MS       536.62        14.06
volume            F        398.62        10.44

Gonad:            MS       1.0147       0.0990
muscle            F        897.01        87.5
ratio

TABLE 3.

Observed growth rates in bay scallops (Argopecten irradians). The
growth rate (GR) of the shell height (mm [day.sup.-1]) for each
subpopulation, and coefficient of determination ([r.sup.2]) for
the least squares linear regression are given. Each regression
equation was significant at P [much less than] 0.001. Sample sizes
are the same as the total sample size in Table 1. Mean shell
heights are given for the period immediately prior to harvest
season (6/5 to 6/26; except Anclote, where samples were collected
on the first day of the harvest season, 7/1), during mid-season
(8/4 to 8/29), and after the harvest season had ended (9/12 to 9/20;
except Steinhatchee, where the last collection occured on 9/3, seven
days prior to the end of the harvest season). Similar groups, as
determined by Tukey multiple comparison test for unequal sample
sizes, are indicated by the same letter.

                  GR      [r.sub.2]   Pre-season

Anclote         0.090      0.4676      54.3 (a)
Homosassa       0.087      0.4897      45.8 (b)
Steinhatchee    0.132      0.5023      43.8 (b)
St. Marks       0.207      0.5663      43.1 (b)
Lanark
  Village       0.075      0.4056      54.1 (a)
St. Joseph
  Bay           0.135      0.6301      45.7 (b)

                 Mid-       Post-
                season     season

Anclote        60.1 (a)   59.0 (b)
Homosassa      57.3 (a)   58.9 (b)
Steinhatchee   57.7 (a)   55.9 (b)
St. Marks      56.2 (a)   63.3 (a)
Lanark
  Village      59.6 (a)   60.7 (b)
St. Joseph
  Bay          54.5 (b)   57.8 (b)