Aspects of Growth for the Northern Quahog, Mercenaria mercenaria

The growth of the northern quahog, Mercenaria mercenaria, has been mathematically modeled over both the nursery stage and the complete life so as to develop expectations of growth as a function of environmental suitability. Then, the growth of northern quahogs in an experimental nursery upweller was evaluated as a function of system operating parameters (flow rate and stocking density) and other environmental parameters to determine the limiting factors in this critical phase of shellfish aquaculture. The von Bertalanffy growth equation was used to predict increases in shell length (millimeters), weight (grams), and the relative growth rate(% increase per day) at various instantaneous growth coefficients (K). The relative growth rate (RGR) was also determined over a number of time intervals, including 1, 4, 7, 14, and 28 days. The age at which the maximum shell length and weight was reached varied with K. A higher K (0.30) resulted in rapid growth and an earlier asymptote, while a lower K (0.20 and 0.10) resulted in slower growth and a'later asymptote in the animal's maximum shell length and weight. The RGR averaged over an annual time interval (annual RGR), as predicted by von Bertalanffy, decreased rapidly as the northern quahog aged, approaching 0.5 % increase/day after age 2. Annual RGR at different K values was similar, indicating that RGR was insensitive to changes in K. During the first growing season (210 days in the northeast), the increase in shell length predicted by von Bertalanffy was linear with a slope determined by K, that is, a greater slope results in a higher K. A similar relationship was apparent with weight. The RGR,

This suggests that researchers should use data collected from northern quahogs in a size range similar to that being modeled when estimating biomass from length and abundance data. Predicted shell lengths and RGRs were compared to observed shell lengths and RGRs from a field experiment growing northern quahogs in an experimental-scale upweller (nursery stage). The northern quahogs grew at a K of 0.25 indicating favorable conditions for growth. Early in the experiment (between 70 and 100 days after spawning), the experimental RGR differed markedly from the predicted measure; however, after 100 days post spawning the experimental RGR was higher than expected and followed the general trend of decreasing RGR over time.
Northern quahog seed were grown from ~2 (longest axis) to~ 13 mm in an experimental-scale floating upweller from June 21 to August 19, 1999 in Point Judith Pond, Wakefield, Rhode Island. Flow rates and stocking densities were varied in order to produce a chlorophyll-a effective flow rate range of 360 to 1,500 µg per minute per liter of northern quahog volume (µg ·min-1 -r 1 ), and growth and environmental parameters were measured semiweekly. During the first two-week experiment (June 21 to July 7) an asymptotic relationship was observed between growth(% increase/day) and chlorophyll-a effective flow rate. A significant difference in growth was found between the treatments. The difference in the functional relationship between experiments 1 and 3 was possibly related to lower initial DO values, which reduced differential growth in experiment 3. In experiment 1, the low-biomass treatments grew faster than the high-biomass treatments. A significant difference in growth between treatments was also observed in experiment 3, although the asymptotic relationship was less pronounced. In experiment 3, the high-biomass replicates grew faster than the low-biomass replicates. Experiments 1 and 3 both experienced similar environmental conditions; however, experiment 1 encountered higher initial morning dissolved oxygen (DO) levels. In addition, the within experiment variability in experiment 3 was much less than the variability in experiment 1; therefore, accentuating growth differences in experiment 3. In both experiments 1 and 3, maximum growth occurred near treatment 2 in a range of chlorophyll-a effective flow rates of 550 to 650 .µg·min-1 · r 1 . In experiments 2 and 4, there were no significant differences in growth between treatments.
Growth appeared to be limited by environmental conditions. In order to eliminate the effect of food limitation on growth, the upper third of the replicates (fastest growing animals) were used to calculate the RGR during the two-month experiment. Growth was linearly correlated with morning-dissolved oxygen (R 2 = 0.42) and with chlorophyll-a (R 2 = 0.35). The critical DO threshold for growth in upwellers appears to be 5 ppm, below which growth is adversely affected. During this IV study, morning DO levels were less than 50 % saturated, indicating the potential for DO levels to be increased. Future research should investigate methods for elevating DO levels in upwellers.          Growth (% increase/day) as a function of chlorophyll-a effective flow rate for experiment 2 (July 7 to July 22, 1999 Figure 13.   and 100 days after spawning), the experimental RGR differed markedly from the predicted measure; however, after 100 days post spawning the experimental RGR was higher than predicted and followed the general trend of decreasing RGR over time.

INTRODUCTION
The success of a shellfish aquaculture operation depends on optimizing production, specifically, on maximizing growth and survival. Growth is defined as an increase in the size of an individual or the mean increase in the size of a population (Malouf and Bricelj, 1989). Growth is usually expressed as a change in shell length, weight, or volume. The particular method employed to quantify clam growth depends on the life stage of the animal, the application of the measurement, as well as the resources available.
The change in size (shell length, weight, or volume) per unit time is defined as the growth rate. The growth rate is usually expressed as an absolute growth rate, a relative growth rate, or an instantaneous (specific) growth rate (Ricker, 1975). The absolute growth rate describes an increase in size (shell length, weight, or volume) over a specific time interval, usually a month or year. The absolute growth rate does not account for differences in the initial size of the animal. Two northern quahogs may have the same absolute growth rate, but different initial sizes; therefore, the smaller northern quahog is growing more relative to its initial size than the larger northern quahog. To incorporate the effect of initial size on growth, the relative growth rate or specific growth rate are commonly employed. The relative growth rate measures the change in size of the animal relative to its initial size. The specific growth rate is a special case of the relative growth rate that uses a log or natural log transformation (Jobling, 1994).

4
A number of growth models are utilized in the literature; however, the von Bertalanffy model is the standard for marine species (King, 1995) and has been successfully applied to the northern quahog (Jones et al., 1989). Traditionally, northern quahog growth studies have focused on aging northern quahogs in the natural environment based on sclerochronology or shell growth rings (Kennish and Loveland, 1980, Peterson et al., 1984, Jones et al., 1989, Rice et al. 1989, Arnold et al., 1991, Slattery et al., 1991. These studies have documented the effect of environmental conditions (Ansell, 1968, Jones et al., 1989, habitat (Peterson et al., 1984, Slattery et al., 1991, and fishing pressure (Rice et al. 1989) on the growth history of the animal.
In aquaculture research, sclerochronological measurements are not required for documenting growth because the age is based on time from spawning or settlement.
Growth of the northern quahog varies both temporally and spatially. Ansell (1968) documented growth of the northern quahog over its natural geographic range. Jones et al. (1989) compared Ansell's data to their own and found that northern quahogs from Narragansett Bay grow exceptionally fast during the first 2 years of life.
They also reported that growth varied widely thfoughout the bay. They found that the von Bertalanffy estimates of the maximum shell length (Lo) varied from 67 to 100 mm, the growth coefficient (K) varied from 0.16 to 0.30, and the time at zero length (to) varied from -0.05 to -0.81. The authors report that Mercenaria mercenaria have been known to live as long as 40 years, although only 10 % of the clams from Narragansett Bay lived longer than 30 years. Rice et al. (1989) also documented differences in growth throughout the bay. Northern quahogs sampled from Greenwich cove exhibited anL 00 of 87 mm and aK of0.09, while northern quahogs from the West Passage exhibited an Lro of 111 mm and a K of 0.10. There are a variety of factors that could account for these differences including fishing pressure (density), environmental conditions (temperature, salinity, oxygen, and food availability), sediment type, food concentration and quality, nutrient loading, and current speed.
The diversity of growth estimates used in shellfish aquaculture research is daunting.  described growth as an increase in shell length (mm) over the number of days from spawning, an increase in microns per day, as well as a percent weekly increase in packed volume.  defined growth in terms of increases in shell length and settled volume (biomass). Since the time intervals varied between measurements, the authors converted biomass increases into an equivalent monthly relative growth rate. Similarly, a number of studies have transformed biomass increases into daily, weekly, and monthly growth rates .
The variability present in the above experiments and in others makes it extremely difficult to compare growth between studies. In addition, because growth is being expressed during different time periods and over varying sizes, the validity of growth comparisons is questionable. Furthermore, investigators have not predicted or modeled growth under optimal conditions. Without an estimate or expectation of growth, researchers lack a baseline for comparison. In other words, observed differences in growth due to experimental treatments could be confounded by differences in the predictable growth of the animal depending on the measure used.
The purpose of this research was to: 6 I. Develop a set of expectations for northern quahog growth over both a time-line and the first season of growth based on units of shell length, weight, and volume; II. Investigate the sensitivity ofRGR to changes in the growth coefficient (K) and the time averaging period over which RGR was measured; and III. To finely apply the theoretical growth models to observations from the field and from a nursery upweller.
Growth of the northern quahog, Mercenaria mercenaria, was modeled over the course of the animal's life span and during the first growing season (nursery stage).
Comparisons were made with different growth coefficients (K) as well as during different time periods. Finally, predicted growth was compared to observed growth in an experimental-scale upweller.

7
The von Bertalanffy growth model was used to characterize growth of the northern quahog in the coastal waters of Rhode Island. The von Bertalanffy equation in terms of shell length is: where L 1 is the shell length at age t, Lw is the maximum shell length attained by the animal, K is the growth coefficient, and t 0 is the theoretical age (years) at zero length (King, 1995). Length in centimeters at time t (L 1 ) was converted to weight in grams at time t (Wi) by using: where a is a unit conversion factor and b is the volumetric expansion factor. The growth rate was defined as the daily relative growth rate (RGR) and as the specific growth rate (SGR). RGR was calculated as: RGR = ( Wfina1 -W;n;1;a1 J(r-1)

W;nitial
where Wfinat is the final weight (g) and W;nitial is the initial weight (g) and T (days) is the intervening time period (Ricker, 1975). The RGR is expressed as a% increase per day. The SGR was calculated as: 8 where LN is the natural log, W finat is the final weight (g), Winitial is the initial weight (g), and T (days) is the intervening time period. The SGR is also expressed as a %

increase per day
Growth over the lifetime: The increase in the length and weight of the northern quahog was characterized over the course of the bivalve's lifetime at varying growth coefficients (K) of 0.10, 0.20, and 0.30. The growth coefficients of 0.10, 0.20, and 0.30 were selected to encompass the range observed in Narragansett Bay. The maximum age of the northern quahog was assumed to be 40 years (Jones et al., 1989) and t 0 was assumed to be +0.10 years or 36 days. A value of +0.10 was chosen because the time from spawning to settlement can take anywhere from 3 to 5 weeks depending on water temperature (Rice, 1992). The L oo for the von Bertalanffy and the a and b weight coefficients for the length-weight relationship were based on samples collected in Narragansett Bay (Jones et al., 1989, Rice et al., 1989. The daily RGR and SGR averaged on an annual basis were calculated over the life span of the quahog.

Growth during the first growing season:
To document growth of the northern quahog during the nursery stage (2 to 14 mm) shell length, weight, and RGR were modeled during the first growing season. In Rhode Island waters, the growing season is approximately seven months (210 days) and lasts from mid April to mid November (Ansell, 1968). The model was initiated at the time of spawning and to, the age that the clams had a size of 0 mm, was assumed to be 36 days to account for the time between spawning and settlement.
The increase in shell length, weight, RGR, and SGR were modeled during the first growing season (36 to 210 days) atKvalues of0.10, 0.20, and 0.30. The RGR during the first growing season was further investigated by varying the growth interval (1) between determinations. Specifically, growth intervals of 1, 4, 7, 14, and 28 days were used to calculate RGR ( Figure 5).
Application of the growth model to field and experimental data: The theoretical models for growth over the lifetime and during the nursery stage were applied to northern quahog growth data from Narragansett Bay and Point Judith Pond, Rhode Island. Weight-length data for northern quahogs in the size range of 60 to 140 mm were collected with a dredge in upper Narragansett Bay and analyzed for the a and b coefficients of the weight-length relationship. Data collected during the summer 1999 from an experimental-scale upweller in Point Judith Pond were used to define growth of the northern quahog during its first growing season. Specifically, the growth coefficient (K) was determined by non-linear regression methods (DeAlteris and Skrobe, unpublished); the time at length zero (t 0 ) was determined from a linear regression of length versus time after spawning; and the length-weight parameters (a and b) were calculated using linear regression of the log transformed length and weight data.
To elucidate differences between the observed growth rate of the northern quahog during the summer 1999 (Appleyard, 2000) and the predicted growth rate, 10 residuals between the predicted RGR and the experimental RGR were calculated. A linear regression of the residuals and morning-dissolved oxygen (DO) was performed to investigate the influence of environmental conditions on residual growth during the experiment. RGR residuals were also compared to morning temperature and chlorophyll-a concentration.

11
The parameters employed to model growth of the northern quahog in Rhode Island waters are included in Table 1 increases, the time required for the northern quahog to reach its maximum shell length and weight decreases. RGR over an annual time interval decreases from 3.3 % increase/day during the first year to 0.5 % increase/day during the juvenile years (age 1-2) and becomes negligible during the adult years (age 2+) ( Figure 2A). Although the same general trend was apparent for the SGR averaged over an annual time interval, the SGR during the juvenile years was about half the RGR ( Figure 2B).

Growth during the first growing season:
The parameters in Table 1 were also used to model growth of the northern quahog during the first growing season. The increase in shell length during the first seven-month (210 day) growing season was linear, as would be expected from the von Bertalanffy growth curve. The maximum shell length reached during the first growing season varied considerably with K. The northern quahogs reached a shell length of 27 mm with a K of 0.30, a shell length of 18 mm with a K of 0.20, and a shell length of 9 mm with a K of 0.10 ( Figure 3A). A similar relationship was apparent with weight during the first growing season. Specifically, a weight of 6.20 g, 2.20 g, and 0.30 g was reached during the first growing season with K values of 0.3, 0.2, and 0.1, respectively ( Figure 3B).
The difference in final shell length and weight observed at K values of 0.10, 0.20, and 0.30 were not apparent when growth was converted to a RGR ( Figure 4A).
Although there was a considerable change in the RGR during the first growing season, the northern quahogs displayed similar RGRs at varying K values. In particular, at 90 days after spawning, the northern quahogs grew at a rate of 11 % increase/day, while at 150 days after spawning they grew at 3 % increase/day. The relative growth rate was high early in the northern quahogs first growing season, but decreased substantially and leveled off after 180 days after spawning. The SGR during the first growing season mirrored the RGR ( Figure 4B).
The relative growth rate was also determined using a variety of growth intervals. There were no detectible differences between RGR determinations at a K of 0.10, 0.20, and 0.30 with varying time intervals ( Figures 4A, 4B, SA, SB, SC, and SD); however, the value of RGR at a specific time varied with the time interval used to 13 calculate RGR (Figure 6). At 90 days after spawning, the northern quahogs grew at 26 % increase/day during a 28-day growth interval, at 15 % increase/day during a 14-day growth interval, at 13 % increase/day during a 7-day growth interval, at 11 % increase/day during a 4-day growth interval, and at 10 % increase/day during a 1-day growth interval. In other words, the larger the growth interval that RGR was averaged over, the larger the RGR. This relationship becomes less obvious when the northern quahogs reach 120 days after spawning and disappears after the first growing season The results of this growth modeling exercise indicate that the growth rate of the northern quahog over its life span changes considerably (Figure lA and lB) as a function of K. In particular, early on growth of the northern quahog is nearly linear and reaches as asymptote as the animal ages. The age that growth reaches an asymptote depends on the particular growth coefficient (K) used. A higher K results in rapid growth and an earlier asymptote in the animal's maximum shell length and weight, while a lower K results in slower growth and a later asymptote; therefore, K is a useful indicator of environmental suitability for growth. RGR averaged over an annual time interval decreases rapidly as the northern quahog ages, approaching 0 % increase/day after age 2. RGR was relatively insensitive to changes in K. SGR closely followed the patterns observed for RGR; however, because of the log or LN transformation, SGR during the juvenile years was about half the RGR.

Growth during the first growing season:
To predict growth during the nursery stage, growth of the northern quahog was investigated during the first growing season (210 days in Rhode Island). The increase in shell length and weight was linear ( Figure 3). The final shell length and weight reached during the first growing season varied considerably with K; however, the RGR at different K values was almost identical. Again, the measure of RGR was relatively insensitive to changes in K. The RGR did vary during the first growing 16 season. The RGR was extremely high early in the growing season (at 90 days after spawning the predicted RGR was 11 % increase/day), but then decreased substantially and leveled off (at 150 days after spawning the predicted RGR was less than 4 % increase/day).
The RGR was further investigated by altering the time period between RGR determinations. There were no observable differences between RGR at a K of 0 spawning there was a considerable difference between the RGR calculated over a 28day interval and the RGR calculated over a 4-day interval ( Figure 6).
Application of the growth model to experimental data: The a and b coefficients estimated for the weight-length relationship from the adults and nursery stage northern quahogs were different from each other and from other values published for Narragansett Bay northern quahogs (Figure 7). This suggests that researchers should always use data collected from locally available northern quahogs in a size range similar to that being modeled when estimating biomass from shell length and abundance data.
The relative growth rate changes considerably during the first growing season.
The growth rate depends on the size of the animal as well as the time period between measurements. Growth studies on upwellers have not taken into account these differences. In addition, the variability between growth measurements has prohibited meaningful comparisons between research studies.  converted their biomass increase to a monthly percent increase because the time interval between volume determinations varied. The investigator's failed to take into account changes in the growth rate at different time intervals. A study completed by  found a relationship between the daily growth rate (DGR) and flow ratio at varying northern quahog shell lengths. In particular, the authors found that the DGR increased as the size of the animal decreased. Based on the results of this modeling exercise, the RGR is expected to increase with decreasing size. The relationship developed by the authors correlates well with predicted growth in the natural environment.
During the experiment completed in the summer 1999 (Appleyard and DeAlteris, 2000) the northern quahogs grew at a K of 0.25 ( Figure 7B), indicating favorable conditions for growth. AK of 0.25 approaches the maximum K observed by Jones et al. (1989) in Narragansett Bay.
The predicted RGR was compared to the experimental RGR during the study.
Early in the experiment, between 70 and 100 days after spawning, the experimental RGR differs markedly from the predicted RGR; however, above 100 days after spawning the experimental RGR was higher than expected and followed the general trend of decreasing RGR over time ( Figure 10). To incorporate the influence of anticipated growth on the experiment, the residuals between the predicted and experimental RGR were determined. The residuals were then compared to key      18 The purpose of this study was to define the relationship between flow rate, stocking density, and growth in order to determine the flow rate and density that optimizes growth. Furthermore, this study was designed to investigate other significant environmental parameters influencing bivalve growth in an experimental-scale upweller system. Northern quahog, Mercenaria mercenaria, seed were grown from -2 (longest axis) to ~ 13 mm in an experimental-scale floating upweller from June 21 to August 19, 1999 in Point Judith Pond, Wakefield, Rhode Island. Flow rates and stocking densities were varied in order to produce a chlorophyll-a effective flow rate range of 360 to 1,500 µg ·min-1 T 1 , and growth and environmental parameters were measured semiweekly. During the first two-week experiment (June 21 to July 7) an asymptotic relationship was observed between growth (% increase/day) and chlorophyll-a effective flow rate. A significant difference in growth was found between the treatments. The difference in the functional relationship between experiments 1 and 3 was possibly related to lower DO values, which reduced differential growth in experiment 3. In experiment 1, the low-biomass treatments grew faster than the high-biomass treatments. A significant difference in growth between treatments was also observed in experiment 3, although the asymptotic relationship was less pronounced. In experiment 3, the high-biomass replicates grew faster than the low-biomass replicates. Experiments 1 and 3 both experienced similar environmental conditions; however, experiment 1 encountered higher morning dissolved oxygen (DO) levels. In addition, the within experiment variability in experiment 3 was much less than the variability in experiment 1; therefore, accentuating growth differences in experiment 3. In both experiments 1 and 3 maximum growth occurred near treatment 2 in a range of chlorophyll-a effective flow rates of 550 to 650 µg·min-1 · r 1 . In experiments 2 and 4 there were no significant differences in growth between treatments.
Growth appeared to be limited by low oxygen. In order to eliminate the effect of food limitation on growth, the upper third of the replicates (the fastest growing animals) were used to calculate the relative growth rate (RGR) during the two-month experiment. Growth was linearly correlated with morning-dissolved oxygen (R 2 = 0.42) and with chlorophyll-a (R 2 = 0.35). The critical DO threshold for growth in upwellers appears to be 5 ppm, below which growth is adversely affected. During this study, morning DO levels were less than 50 % saturated, indicating the potential for DO levels to be increased. Future research should investigate methods for elevating DO levels in upwellers.

INTRODUCTION
Over the past decade the use of upwellers as bivalve nursery units has increased dramatically in North America . A number of studies have explored the relationships between flow rate, stocking density, and growth in upwellers . The majority ofresearch on upwellers has focused on the northern quahog, Mercenaria mercenaria, because of its significant aquaculture potential. In particular, the northern quahog grows well at high densities, has adapted to a variety of geographic sites along the northeast coast, and has a lucrative market. flow rate, the data is represented by an asymptotic relationship; in particular, as the chlorophyll-a effective flow rate increases, growth increases steeply and then levels off with increasing chlorophyll-a effective flow rates (Figure 1 ). Efficiency in this upweller system refers to economically optimizing both upweller space (density) and pumping capacity (flow). Theoretically, growth will be optimized at some percentage of the maximum growth rate; as indicated in Figure 1, 80 to 90% of the maximum growth rate equates to a chlorophyll-a effective flow rate range of 470 to 700 µg ·min-1 k -1 . g .  concluded that food supply was the primary limitation in their upweller system. Their data suggests that to obtain unlimited growth, northern quahog seed needed to remove approximately 150 µg·min-1 -kg-1 . The investigators deduce that northern quahog growth was reduced if more than 20 % of the ambient chlorophyll-a concentration (µg/l) was removed as water passed by the bivalves.
Consequently, to supply the necessary ration of 150 µg·min-1 ·kg-1 without exceeding 20 % removal, food must be supplied to the bivalves at a rate of 750 µg·min-1 ·kg-1 •  confirmed that ambient chlorophyll-a concentrations were reduced by ~20% through an initial silo of northern quahogs at similar stocking densities. However, they found that after water passed through an initial group of northern quahogs it could then support an additional equivalent biomass of northern quahogs at the same growth rate. They conclude that to achieve maximum growth of northern quahogs in an upweller it is necessary to pass more water through the animals than can actually be filtered; therefore, the low rate of chlorophyll-a removal reported by  may reflect a large amount of unused water through the system.
The authors hypothesize that this surplus water may be a physical requirement of the system where minimum flow rates are required to create uniform flows through the seedbed, remove waste products, maintain water quality, and maintain a minimum concentration of chlorophyll-a.  found that food was the primary limitation in their upweller system, while  concluded that flow rate was the primary limitation.  also speculate on the importance of environmental conditions, specifically water quality, but they fail to characterize these parameters in their system. Growth and survival of the northern quahog is clearly influenced by the surrounding environment. Northern quahog adults and juveniles can survive in water temperatures from 1 to 33 °C, but grow optimally at 23 °C . Northern quahogs can tolerate salinities between 10 and 35 %0 ) for short periods, but prefer to inhibit waters greater than 20 %0 . Northern quahogs have been known to endure oxygen concentrations below 1 mg 0 2 11 ) for more than three weeks; however, growth is significantly reduced and an oxygen debt is incurred when oxygen concentrations fall below 5 mg 02/l Dewitt, 1983, Hamwi, 1969).
Although there is a general disagreement as to the limiting parameter for growth in upwellers, in the literature growth is clearly related to both system operating parameters (flow rate and stocking density) and environmental conditions at the site (temperature). study were to: I. Develop a relationship between flow rate, stocking density, and growth so as to determine the chlorophyll-a effective flow rate that optimizes growth; and II. Determine the most significant limiting parameter for bivalve growth in Growth of northern quahog seed was studied over an 8-week period in an experimental-scale floating upweller system located in a nutrient-rich estuary. At the beginning of the experiment the ambient chlorophyll-a concentration (µg/l) was measured at the site, and flow rates and stocking densities were adjusted to achieve three nominal ranges of chlorophyll-a effective flow rates, including a low (-350 µg ·min-1 T 1 ), medium (-600 µg·min-1 ·r 1 ), and high range (-1,200 µg ·min-1 T 1 ). Each combination of effective flow rate (µg/min) and northern quahog biomass (1) or chlorophyll-a effective flow rate (µg·min-1 ·r 1 ) represents a treatment, as shown in Table 1. A sample calculation of initial chlorophyll-a effective flow rates for the first growth interval (June 21 to June 24, 1999) is illustrated in Table 2. The average chlorophyll-a concentration during the time period was 11.70 ± 2.06 µg/l (S.E.) and the flow rates were set at 4 l/min, 6 l/min, and 8 l/min resulting in three effective flow rates of38.36 µg/min, 57.84 µg/min, and 77.12 µg/min. The northern quahog seed were initially stocked at a biomass of 0.055 1 (density of 0.3 l/cm 2 ) and 0.1091 (density of 0.6 l/cm 2 ) resulting in the desired range of chlorophyll-a effective flow rates. The experiment was a two (density) by three (effective flow rate) factorial design with six treatments of chlorophyll-a effective flow rates. Each treatment was replicated in triplicate resulting in a total of 18 observations (silos).

Site Location
The Flow through each silo was manipulated with the ball valve and was measured volumetrically with a graduated cylinder and a stopwatch.
Northern quahog seed (300,000 at 0.6 mm) were purchased from Bluepoints Company, Inc., West Sayville, New York. The seed were held in the upweller until they reached > 2 mm (longest axis).

Data Collection
At the beginning of each experiment the seed were pulled from the unit, sieved, and randomly distributed throughout the 18 replicates at a biomass of 0.055 1 (wet volume) and 0.109 1. In addition, the valve length of a random sample (n = 75) of seed was measured to the nearest 0.01 mm with vernier calipers. Five sub-samples of northern quahogs were also taken to develop a relationship between wet volume (1) and wet weight (kg). Each experiment was terminated when the biomass in the slowest growing replicate doubled. This occurred approximately every two weeks during the summer. At the termination of each two-week experiment, the valve length of a random sample (n = 25) of northern quahogs from each replicate was determined.
Four two-week experiments were completed during the summer 1999.
The change in volume of each silo was measured semiweekly resulting in 3 to 4 day growth intervals. Semiweekly flow rates to each silo were also measured in the morning or late at night to minimize wave activity. Care was taken to ensure that the upweller unit was not altered during measurements and flows were adjusted accordingly.
Chlorophyll-a (Chl-a), particulate organic matter (POM), temperature, salinity, and dissolved oxygen (DO) were measured semiweekly from an empty silo. With the start of the second two-week experiment (July 7) all environmental parameters were taken in the morning, midday, and evening to quantify daily fluctuations at the site.
Discrete chlorophyll-a samples (n = 3) were taken with a syringe. Samples were prefiltered with a 150 µm Nytex screen to remove particulates that bivalves are unable to filter (Defossez and Hawkins, 1997). Samples (10 ml) were forced through a 25 mm diameter Whatman GF IF filter contained in a 25 mm Swinnex filter holder. The procedure for chlorophyll-a analysis is slightly modified from the standard procedure outlined in . When measuring the change in volume of northern quahog seed, each silo and screen were cleaned with freshwater. Once a week the remainder of the upweller manifold was cleaned by a diver to ensure consistent flow through the system.
Approximately 10 days into each of the experiments a feeding rate experiment Was completed to quantify the amount of ambient chlorophyll-a removed by the 54 northern quahogs, expressed as % clearance. Samples (n = 2) of 50 ml were taken from the empty silo as well as from each replicate. Sampling was conducted in the early afternoon of a clear day and was completed within a 10-minute window.
Samples were vacuum filtered in the laboratory through a 25 mm Whatman GF IF filter within 2 hours of sampling. Chlorophyll-a analysis was completed as previously outlined.

Data Analysis:
The chlorophyll-a effective flow rate (µg·min-'·r') for each replicate was calculated as the product of the average chlorophyll-a concentration (µg/l) during the period and the flow rate (l/min) to the replicate all divided by the average biomass (1) of the replicate during the same period. This study characterized growth as the relative growth rate (RGR) and was calculated as: where Chl-a (ambient) is the Chl-a concentration of incoming seawater and Chl-a (outflow) is the Chl-a concentration of outgoing seawater from each replicate.

55
To elucidate differences in growth between treatments the total RGR (% volume increase) was divided by the largest time period available, the length of each experiment. Since the RGR (% increase/day) measures the change in volume over each two-week experiment, the average chlorophyll-a concentration and average treatment biomass during the time period was used to calculate treatment chlorophylla effective flow rates. Prior to ANOV A analysis, the RGR (% increase/day) was arcsine transformed . Within each experiment, one-way ANOV As were performed for each two-week experiment with the average RGR (% increase/day) as the dependent variable and treatment as the independent variable.
Differences between treatment means were elucidated with the Tukey Honestly chlorophyll-a, and dissolved oxygen to determine which independent variables were significant in determining growth. Dissolved oxygen concentrations were converted to percent saturation based on temperature and salinity measurements . A step-wise linear regression analysis was also performed to elucidate the most significant parameter(s) for predicting growth in the experiment.
There were no observed mortalities during the course of the two-month experiment. Calculated mortalities were extremely variable in experiments 1 and 2 because counts per ml were not replicated. In addition, counts were not made before and after sieving. In experiments 3 and 4, counts per ml were replicated (n = 3) and counts were made before and after sieving. Mortality was calculated to be 111 ± 3 % (S.E.) and 99 ± 2 % (S.E.) respectively.
The average chlorophyll-a concentration ~as 16.42 ± 2.25 µg/l (S .E.) and the average treatment biomass ranged from 165.8 to 85.5 ml. The chlorophyll-a effective flow rate ranged from 420 to 1,445 µg·min-1 .r 1 roughly correlating with a RGR from 4.76 to 9.32 % increase/day. As the chlorophyll-a effective flow rate increased, the RGR increased until -650 µg·min-1 .r 1 at which point growth leveled off ( Figure 4).
Growth, as measured by RGR, was subjected to a one-way ANOV A with six levels of treatment. This was found to be statistically significant (E (5, 11) = 5.48, p < 0.05).
The strength of the relationship was 0.57 as indexed by the standard omega-squared (co 2 ). The Tukey HSD test indicated that the mean RGR for treatment 1 CM =5.54) was significantly lower than the means for treatment 4 CM =8.14), 5 CM =8.41), and 6 CM =9.26). To investigate the effect of effective flow rate and biomass on growth, a two-way ANOV A was performed with three levels of seston flux and two levels of .36, p < O. 05) were statistically significant. The strength of the relationship ( (J) 2 ) was 0.21 and 0.31, respectively. The interaction between effective flow rate and biomass was found to be ordinal; therefore, the main effects were examined by the Tukey HSD test. The Tukey HSD test indicated that the low-biomass treatments CM= 8.60) grew faster than the high-biomass treatments CM= 6.99).
The average chlorophyll-a concentration was 11.83 ± 1.16 µg/l (S.E.) and the average treatment biomass ranged from 205.6 ml to 91.6 ml. The chlorophyll-a effective flow rate ranged from 231 to 977 µg·min-1 T 1 roughly correlating with a RGR from 9.58 to 12.73 % increase/day ( Figure 5). RGR was consistently high within the chlorophyll-a effective flow rate range specified. RGR was subjected to a one-way ANOV A and there was no statistical difference between treatments CE (5, 11) = 1.48, p > 0.05).
The average chlorophyll-a concentration was 18.55 ± 2.12 µg/l (S.E.) and the average treatment biomass ranged from 184.8 to 84.6 ml. The chlorophyll-a effective flow rate ranged from 411to1,720 µg·min-1 .r 1 roughly corresponding to a RGR from 7.79 to 10.09 % increase/day ( Figure 6). The RGR increased slightly with an increase in the chlorophyll-a effective flow rate until -610 µg·min-1 .r 1 , at which point growth decreased and leveled off. RGR was subjected to a one-way ANOVA and was found to be statistically significant CE (5, 11) = 7.13, p < 0.05). The strength of the relationship was 0.64 as indexed by the o}. The Tukey HSD test indicated that the mean RGR for treatment 2 (M = 9.76) was significantly higher than the means for treatment 4 CM= 7.97), 5 CM= 8.62), and 6 CM= 8.38). In addition, the mean RGR for treatment 4 CM= 7.97) was significantly lower than the mean for treatment 3 (M ==9.20). A two-way ANOVA found both effective flow rate CE (2, 11) = 5.99, p < 0.05) and biomass CE (1, 11) = 22.33, p < 0.05) were statistically significant. The strength of the relationship (ro 2 ) was 0.21 and 0.45, respectively. The interaction between effective flow rate and biomass was found to be ordinal; therefore, the main effects were examined by the Tukey HSD test. The Tukey HSD test indicated that the high-biomass treatments (M = 9.25) grew faster than the low-biomass treatments CM= 8.32). The Tukey HSD test also found that the replicates with an effective flow rate of 111.3 µg/min CM = 9 .19) grew faster than the replicates with an effective flow rate of 74.2 µg/min CM= 8.87).
The fourth experiment began on August 5 and ended on August 19, 1999 (14 days) and the northern quahogs grew from 9.37 ± 0.12 mm (S.E.) to 11.47 ± 0.08 mm (S.E.). The average chlorophyll-a concentration was 17.91 ± 3.17 µg/l (S.E.) and the average treatment biomass ranged from 147.4 to 73.4 ml. The chlorophyll-a effective flow rate ranged from 491 to 1,905 µg·min· 1 .r 1 roughly corresponding to a RGR from 4.98 to 5.96 % increase/day (Figure 7). RGR was consistently low within the chlorophyll-a effective flow rate range specified. RGR was subjected to a one-way ANOVA and there was no statistical difference between treatments CE (5, 11) = 0.76, p > 0.05).

60
A one-way ANOV A was performed to compare the RGR between the twoweek experiments. This was found to be statistically significant Cf (3, 64) = 135.34, p < 0.05) with an o/ of 0.80. The Tukey HSD test indicated that there was a significant difference between all the mean RGRs, with growth highest in experiment 2 CM = ll.89) and decreasing in experiments 3 CM= 8.81), 1CM=7.71), and 4 CM= 5.57).

The effect of environmental characteristics on growth
The Temperature during the experiment varied from 21.4 to 27.3 °C (Figure 9).
Other than a brief drop in temperature in mid July due to a rainstorm, temperature was fairly consistent during the experiment. A linear regression analysis indicated that temperature was not significant in determining growth as indicated by the upper 1/3 and August 17, more than half of the percent clearance determinations were negative.

RGR CE
In other words, the amount of incoming chlorophyll-a was less than the amount leaving. In the remaining clearance rate experiments there was no detectable trend.
The one-way ANOV As found significant differences in growth between In the first experiment, growth(% increase/day) followed the relationship presented in Figure 1. Growth increased as the chlorophyll-a effective flow rate increased until~ 650 µg·min-1 · r 1 at which point growth reached an asymptote. The one-way ANOV A found a significant difference in growth between the treatments with a relatively strong relationship as indexed by the standard omega-squared (ro 2 = 0.57). Furthermore, treatment 1 grew significantly slower than treatments 4, 5, and 6.
Treatments 4, 5, and 6 represent the asymptote of the function where growth asymptotes regardless of an increase in the chlorophyll-a effective flow rate. In addition, treatments 4, 5, and 6 were those with a low initial stocking density of 0.3 llcm 2 . In order to further investigate the effect of effective flow rate and biomass on 64 growth a two-way ANOV A was performed. There were significant differences between growth with the levels of effective flow rate and biomass. In particular, the low-biomass replicates (treatments 4, 5, and 6) grew faster than the high-biomass replicates (treatments 1, 2, and 3).
In the third experiment, the one-way ANOV A also indicated a significant difference in growth between the treatments with an even stronger relationship ( o/ = 0.64). The functional relationship between growth and chlorophyll-a effective flow rate was different than that postulated in Figure 1. Growth increased slightly with increasing chlorophyll-a effective flow rate, but then decreased slightly, reaching an asymptote above 1,000 µg·min-1 • r 1 . This trend is supported by the Tukey HSD test, which indicated that treatment 2 grew significantly faster than treatments 4, 5, and 6.
In addition, treatment 3 grew significantly faster than treatment 4. Since treatment 3 and treatment 4 have nearly the same chlorophyll-a effective flow rate, a significant difference in growth indicates an effect of biomass on growth, with the higher biomass treatment growing faster that the lower biomass treatment. The two-way ANOV A also found a significant effect of biomass on growth with the high-biomass replicates growing faster that the low-biomass replicates. The two-way ANOV

The effect of environmental characteristics on growth
When environmental conditions were suitable for northern quahog growth, especially in the beginning of experiment 1, the effect of food limitation on growth was apparent. When environmental conditions were less than optimal, as in experiments 2 and 4, growth appears constant over a wide range of chlorophyll-a effective flow rates. In other words, growth was not controlled by food limitation, but some other factor. To quantify the effect of environmental conditions on growth, the upper 113 of replicates, the fastest growing northern quahogs, were used to determine growth. By eliminating the slowest 2/3 replicates, the effect of food limitation on growth was minimized; therefore, differences in growth were constrained by the environmental conditions at the time.  concluded that food limited growth in their experimentalscale upweller. Although there were signs of food limitation on growth in experiments 1 and 3, growth in experiments 2 and 4 were controlled by other factors.  determined that the flow rate limited growth in their upweller system. They surmise that flow through the upweller had to be above a critical threshold in order to create a uniform flow (distribute food evenly among the clam seed), maintain water quality, remove wastes, and provide a sufficient chlorophyll-a concentration to the northern quahogs. Although  were unable to quantify the effect of water quality on growth, they eluded to the importance of environmental conditions on growth.
Over the course of the two-month experiment, growth was positively correlated with morning DO and negatively correlated with chlorophyll-a. In late June and early July, the experimental site in Point Judith Pond experienced a pronounced algae bloom. The bloom was evident as an increase and peak in the chlorophyll-a concentration ( Figure 11 ). There was a clear relationship between chlorophyll-a and morning DO, specifically as the chlorophyll-a concentration increased, morning DO levels decreased (Figure 16). The decrease in morning DO was a result of a combination of algae decomposition and algae respiration. At night, the algae were constantly respiring, converting captured energy into simple sugars, an oxygen consuming and carbon dioxide producing process. The algae were also continually dying off and decomposing, again an oxygen consuming process. A second algae bloom in the upper pond was apparent in mid August. Again, the same relationship between chlorophyll-a and morning DO was apparent. In late July, the chlorophyll-a concentration decreased substantially and morning DO levels increased. This decrease in chlorophyll-a was most likely a result of zooplankton grazing described by . Alternatively, the decrease in chlorophyll-a could have been caused by a crash or die off of a particular species of algae. The cyclic pattern of algae in the upper pond could be further verified by quantifying the species of algae present as well as the amount of zooplankton at the study site.
Regardless of the specific controlling mechanisms, an increase in algae biomass caused a distinct decrease in morning DO(< 5 ppm) resulting in depressed clam growth.  determined that Mercenaria mercenaria were able to maintain a constant rate of respiration with decreasing oxygen levels until 5 ppm. The northern quahog is a classic oxygen regulator . As the oxygen concentration decreases, bivalves can increase their rate of oxygen consumption through two mechanisms: 1) increasing their pumping rate; or 2) increasing their percentage of oxygen utilization.  determined that the pumping rate of northern quahogs remained constant with decreasing oxygen concentrations; however, northern quahogs were able to regulate 0 2 consumption by increasing the percentage of oxygen utilized. When oxygen levels reached 5 ppm or below,  found that oxygen uptake in northern quahogs decreased continuously and an oxygen debt was incurred. Once conditions were favorable, the oxygen debt was rapidly repaid in a matter of hours and northern quahogs were able to function normally.
Although juvenile northern quahogs can survive in oxygen concentrations below 1 ppm for up to three weeks (Stanley and Dewitt, 1989), 5 ppm is the critical threshold for northern quahog growth. There have been a number of studies that have investigated the effect of low oxygen levels on survival and tolerance, yet none have investigated the effect of low oxygen levels on growth. Based on the work completed by , 5 ppm is the critical threshold for northern quahog growth. When oxygen concentrations fall below 5 ppm, the northern quahogs cannot maintain sufficient oxygen uptake and incur an oxygen debt. In essence, the northern quahogs shut down and stop growing until oxygen levels rise above this critical threshold.
The results of this study stress the importance of sufficient oxygen concentrations for northern quahog growth in upweller systems. A number of 69 methods could be used to ensure optimal oxygen levels in an upweller. The upweller could be moved to a site that experiences lower chlorophyll-a values and higher morning DO values, but food for the northern quahog would be compromised.
Alternatively, the oxygen concentration in the upweller could be increased. During periods oflow morning DO(< 4 ppm), the% saturation was below 60 ( Figure 15); therefore, during periods of low morning DO, oxygen concentrations have the potential of being increased. Future research should investigate the most cost effective and efficient method of increasing dissolved oxygen levels in this upweller as well as in the more traditional passive flow upwellers. With optimal DO levels, the effect of food limitation on growth can be further defined and replicated.
Clearance rate experiments  postulated that the growth rate of northern quahogs was consistently reduced when % clearance of ambient chlorophyll-a was above 20 %, regardless of the overall amount of chlorophyll-a removed. This result suggests a threshold feeding response.  found that scallop clearance rates were significantly reduced when chlorophyll concentrations fell below 12 % of ambient concentrations.  also found that approximately 17 % of the ambient chlorophyll-a concentration was removed after the first pass through a silo of seed northern quahogs. The water was then passed through a second series of upwellers and the percentage of chlorophyll-a removed ranged from 36 to 73 %. They concluded that factors other than food limited growth of seed northern quahogs in an upweller silo.  (1996) developed a more economical method for measuring food depletion in the field. They measured food depletion in an experimental flume with a continuous flow Turner Designs fluorometer. By allowing the fluorometer readings to stabilize for at least one minute they were able to reduce the variability associated with discrete sampling.

SUMMARY AND CONCLUSIONS
The hypothesized relationship between growth and chlorophyll-a effective flow rate was only apparent during the first two-week experiment (experiment 1 ).
Although there were significant differences in growth between treatments in the third two-week experiment (experiment 3), these differences were most likely the result of small within sample variability. For the remainder of the experiment, northern quahog growth was limited by environmental conditions. Specifically, the relative growth rate of the upper one-third of the replicates was positively correlated with morningdissolved oxygen (R 2 = 0.42) and negatively correlated with chlorophyll-a (R 2 = 0.35).
The critical dissolved oxygen threshold for northern quahog growth in the experimental-scale upweller appeared to be 5 ppm, below which growth was adversely affected. Future research should investigate the most effective method for elevating DO levels in commercial upwellers.