ECOLOGY OF MENIDIA MENIDIA LARVAE IN TWO TEMPERATE ESTUARINE LITTORAL HABITATS

The Atlantic silverside (Menidia menidia) is one of the numerically most abundant fish species in estuaries along the East Coast of North America, but its ecology during the first two weeks post-hatch has not been described. Therefore ecological investigations into appropriate sampling methods, preferred habitats, feeding ecology and growth of these larvae will contribute valuable information to our knowledge base for this species. The upper reaches of two Rhode Island, USA, estuaries, with differing levels of anthropogenic inputs, were the study sites for this project. Previous studies have shown that during early spring M. menidia adults ripen for spawning by feeding exclusively on zooplankton. The zooplankton community in Upper Point Judith Pond (UPJP) is dominated by polychaete larvae, indicating a eutrophic environment, whereas the Upper Pettaquamscutt River (UPR) is dominated by crustaceans, indicating a relatively pristine environment. To assess and describe the habitat ecology of M. menidia larvae during their first two weeks of life in the littoral zone, four goals were set: (1) determine depth distribution of M. menidia larvae from both estuaries: (2) assess abundance and distribution of M. menidia larvae between estuaries; (3) compare feeding habits of the larvae in the two estuaries through gut content analysis; (4) compare growth of larvae in the two estuaries via age-length relationships based on otolith analysis. Of the four sampling devices used to collect larvae, the circular quadrat, which sampled the land-water interface, the aquarium net, which sampled water from 0.3 – 0.4 m depth, and the small plankton net, which sampled water from 0.4 – 0.5 m depth collected many larvae. A large plankton net, which sampled water > 1 m depth, did not. This indicates that M. menidia larvae can be found from the shoreline interface to 0.5 m depth. Analysis of collection data indicated a zero-inflated Poisson distribution, suggesting a patchy distribution of larvae in the field. Gut content of larvae between estuaries differed markedly, with 76.2% of the larval diet at UPR consisting of copepod eggs and 72.5% of that at UPJP consisting of copepod nauplii. The slopes of the age-length regressions of the larvae between estuaries were not significantly different, indicating that growth rates did not differ. These results provide new information on the feeding habits, growth, and distribution of M. menidia during its first two weeks of life in the field.


LIST OF TABLES
. Gut contents of Menidia menidia larvae for UPR (n = 51) and UPJP (n = 58). Each taxon is represented as a percent of the total gut contents for all larvae collected in each estuary. Results from the chi-square analysis show a significant difference in feeding habits of Menidia menidia larvae (p < 0.0001).…………………………………………………….…………28 Figure 4. Growth of Menidia menidia larvae collected from UPJP (open circles) and UPR (diamonds). Linear regressions represent the age-length relationship of Menidia menidia larvae from UPJP (dashed line), y = 0.66x + 2.98, and UPR (solid line), y = 0.65x + 3.06. In the linear equations Y = total length in millimeters and X = age in days. Results from the ANCOVA analysis show no significant differences in the slopes of the age-length relationship of larvae between estuaries (p = 0.8147). However, age is a significant indicator of the size of larvae (p < 0.0001 in structured habitat so that the risk of predation is reduced and food availability for larvae is high (Beck et al. 2001;Boesch and Turner 1984;Heck and Thoman 1981;Rooker and Holt 1997;Weinstein 1979). Such habitats include shallow, near-shore environments that may or may not have submerged aquatic vegetation.
The Atlantic silverside (Menidia menidia) is an estuarine species that occurs from Nova Scotia to Florida (Middaugh 1981) (Conover and Kynard 1984;Koltes 1984;Middaugh et al. 1981;Moore 1980). For example, Middaugh (1981), showed that M. menidia spawn during on a lunar cycle. In addition, Bengtson et al. (1987) investigated the relationship between maternal length and egg diameter. Research investigating the habitat ecology of M. menidia has been restricted to the juvenile and adult life stages (Barkman et al. 1981;Bengtson 1984). Adults spawn on grasses in the intertidal zone (Middaugh 1981). In Rhode Island, USA estuaries, spawning occurs between May and early July (Huber and Bengtson 1999).
In the upper reaches of two local estuaries, the Pettaquamscutt River (UPR) and Point Judith Pond (UPJP), the zooplankton communities are quite different in early spring, when adult M. menidia return from the winter in an emaciated condition and feed on zooplankton to ripen for spawning (Bengtson 1982;. The zooplankton community in UPR is dominated by crustaceans at this time, indicating a fairly pristine environment, and UPJP is dominated by polychaete larvae, indicating a eutrophied environment (Bengtson 1982). Given the propensity of marine larval fish to feed on copepods, an a priori assumption might be that M. menidia in UPR feed on higher quality prey than do those in UPJP. Volson (2012) (Lee 1980). The Pettaquamscutt River is a flooded river valley that is approximately 227 hectares (Gaines 1975).

Abundance and Distribution Sampling
To determine distribution patterns and densities, i.e., abundance per cubic The small plankton net filtered 0.32 m 3 of water. Finally, the large plankton net filtered 1.98 m 3 of water.
Larvae that were collected for laboratory analysis were euthanized using MS-222 mixed in seawater (90 g/mL), then preserved in either 95% ethanol (for otolith analysis) or 10% straight formalin mixed in seawater (for gut content analysis). Each larva collected in the field was measured to the nearest hundredth of a millimeter for total length (TL) using a dial caliper.
III. Laboratory Work

Gut Content Analysis
Foraging habits of M. menidia larvae were determined by gut content analysis of preserved M. menidia larvae collected from the field. In the laboratory, the gut was gently pulled apart and examined in a 50-mm Sedgwick-Rafter counting cell under a compound microscope. Each prey item was tallied and identified to the lowest possible taxon. From UPJP, a total of 58 guts were examined. From UPR, a total of 51 guts were examined. All larvae dissected, for both estuaries, were between 4.18 mm and 9.36 mm (TL). The number method was used to show food type as a percentage of the total gut contents of each larva (Zacharia and Abdurahiman 2004).
Each taxon was represented as a percent of the total gut contents for all the larvae dissected for each estuary.

Otolith Analysis
In the lab, one of the sagittal otoliths was extracted from each larva, placed on a microscope slide with one drop of immersion oil (Grade A from Cargille Laboratories), photographed using a light microscope camera (at 100X or 400X magnification), and the rings counted using methods of Barkman (1978). Otoliths were first examined using light microscopy (400 X magnification) to count daily growth rings to determine age (days). Measurements of diameters of the sagittae were taken as a proxy for growth. At a later time, a second reading was completed by the same observer from the photographs taken of the sagittal otoliths. Six pre-hatch rings were subtracted from the total number of daily rings on each sagittal otolith (Barkman, 1978). The sagittae did not require additional processing because the core was visible.
The relation of the number of daily rings (age) to larval length (TL) was determined for each estuary. The slopes of these linear relationships provided an estimate of growth (mm/day) of larvae in each of the estuaries. All measurements were in micrometers (µm) using the computer software program ImageJ ® (Abràmoff et al. 2004). Seventeen larvae were examined from UPJP and 19 larvae from UPR.

IV. Statistical Analysis
The relationship between age and length was determined for each estuary and the slopes of these regressions were analyzed with an analysis of covariance (ANCOVA). A chi-square analysis was applied to the gut content data to determine any significant difference in feeding habits of M. menidia larvae between estuaries.
Distribution and abundance data collected from the field were analyzed using a Zero-inflated Poisson model (White and Bennetts 1996). This type of generalized linear model assumes that the outcomes that have a zero value are due to two processes, collecting larvae vs. not collecting larvae. Not collecting larvae results in an outcome of zero. If larvae were collected, then the outcome becomes count data.
Where Y p is equal to the number of larvae collected, β � 01 and β � 0 are the intercepts, and β � quadrat, aquarium net, small plankton net, large plankton net, site, date, depth are the coefficients for each predictor. The predictors included in the part of the model that analyzes when a larva was not collected are H p which represents the quadrat, I p which represents the aquarium net, J p which represents the small plankton net, and K p which represents the large plankton net. The intercept was set at zero for this part of the analysis. The predictors included in the part of the model that analyzes when larvae were collected are A p which represents the sites (UPR and UPJP), B p which represents the duration of sampling (date), C p which represents the depth of the water, D p which represents the quadrat, E p which represents the aquarium net, F p which represents the small plankton net, and G p which represents the large plankton net. For this part of the analysis, the intercept could not be set at zero.

I. Field Abundance and Distribution
Average density and the number of larvae collected by each device are shown in Table 2. Only one M. menidia larva was collected with the large plankton net. The quadrat, aquarium net, and small plankton net all collected more larvae compared to the large plankton net. This indicates that M. menidia larvae are generally not in waters greater than one meter depth, but can be found in waters less than 0.5 m depth in the littoral zone. Further, the quadrat, the aquarium net, and small plankton net can all be used to collect M. menidia larvae from the littoral zone of estuaries.
Distribution and abundance of M. menidia larvae from the field are represented as the frequency of occurrence of the number of larvae collected ( Figure 2). The distribution of the data follows a Poisson curve with a high frequency of zeros ( Figure   2). This suggests that M. menidia larvae have a patchy distribution in the littoral zone.
It also shows how often and how many M. menidia larvae were collected per tow ( Figure 2). For field collections made before June 14 th , up to 69 larvae were collected per tow (Figure 2A). For field collections made on and after June 14 th , up to fifteen larvae were collected per tow ( Figure 2B).
The Zero-inflated Poisson analysis shows which predictors in this study influenced the number of larvae collected, as well as which predictors influenced the zeros in the data (Tables 3, 4). Before June 14 th , the presence of larvae in the littoral zone in both estuaries correlated with date, depth, and all sampling devices (Table 3).
By the middle of June, fewer larvae were collected from the littoral zone in UPR and UPJP compared to the beginning of this sampling season (β � date = -0.10, χ 2 = 46.01, p < 0.0001, Table 3). The number of larvae collected increased with depth in the littoral zone (β � depth = 1.34, χ 2 = 35.16, p < 0.0001, Table 3 For field collections made on and after June 14 th , depth, the quadrat, and the aquarium net influenced the number of larvae collected from the littoral zone in UPR and UPJP (Table 4). The number of larvae collected increased with depth in the littoral zone (β � depth = 5.28, χ 2 = 12.96, p < 0.05, Table 4). The quadrat and aquarium net did not consistently collect larvae in statistically significant amounts (β � quadrat = -247.90, χ 2 = 4.03, p < 0.05, β � aquarium net = -59.65, χ 2 = 4.68, p < 0.05, Table 4), as indicated by the negative coefficients. This suggests that larvae were most likely in waters between 0.4 and 0.5 meters depth in the littoral zone. No predictors significantly influenced the presence of zeros for this data.
One of the goals of this project was to determine if the density of M. menidia larvae differed between the estuaries. The results of the Zero-inflated Poisson analysis for site indicated that UPR had a higher density of M. menidia larvae compared to UPJP, for the entire sampling period.

III. Otolith Analysis
Results from the ANCOVA show a significant relationship between length of larvae and age for fish from both estuaries (p < 0.0001, Figure 4). Based on the age-

DISCUSSION
This is the first report of the field ecology of M. menidia larvae, even though laboratory studies on this larval species have been conducted for decades (e.g., Austin et al. 1975;Middaugh and Lempesis 1976;Morgan and Prince 1977;Deacutis 1978;Bengtson 1985;Lankford et al. 2001). Field collections from the littoral zone of UPR and UPJP during the summer of 2012 showed that this larval fish can be collected at the shoreline interface to waters 0.5 m deep, can be collected with a variety of sampling devices, and displays a patchy distribution. The two estuaries sampled differed with regard to abundances of M. menidia larvae and the prey consumed by those larvae, but the larvae grew at the same rates regardless of those differences.
The quadrat, aquarium net, and small plankton net collected more larvae compared to the large plankton net with the aquarium net collecting the most M.
Menidia menidia larvae collected from the field followed a zero-inflated Poisson model, suggesting that these larvae are not distributed evenly in the littoral zone, but have a patchy distribution. Many fishes display a patchy distribution because of ocean processes (Pepin et al. 2003) and social behavior (Maynou et al. 2006), such as spawning, predator pressures, and feeding. Understanding why marine larvae have a patchy distribution has been a focus of many studies. Hewitt (1981) proposed that larvae display a patchy distribution because it benefits schooling, a behavior displayed in the juveniles and adult life stages. Shaw (1960Shaw ( , 1961 showed that M. menidia begin to school around 11 -12 mm SL; later research by Shaw and Sachs (1967) showed that optomotor responses, proposed to be involved with schooling, are present in newly hatched M. menidia. Since schooling is widely thought to reduce predation pressure, the development of such behavior likely allows M. menidia larvae to enter water that they occupy as newly hatched larvae. Future research should include collecting spatial data to determine where these patches of larvae are in the littoral zone and the size of each group.
Results from the gut content data show that M. menidia larvae in UPR consume mostly copepod eggs, whereas those in UPJP consume mostly copepod nauplii. Volson (2012) found a high abundance of calanoid copepods in UPR and a varying zooplankton community in UPJP, where from early spring (April to early May) to late spring (June), the dominant zooplankton present switches from polychaete larvae to copepods. Most of the sampling for this study took place in late spring, when the polychaete larvae had already settled and were not available to, or not preferred by, the larvae. The exact species of copepod that the eggs came from, for this study, was not determined.
These findings agree with previous research on M. menidia. In the Pataguanset Estuary in Connecticut, Cadigan and Fell (1985) found that one of the most commonly occurring food items in the guts of adult M. menidia were copepods. Further, research by Gilmurray and Daborn (1981) showed that small-sized M. menidia consumed smaller zooplankton species, such as copepods. In UPR and UPJP it may be that size class of M. menidia is also an important factor in prey selection. Fernandez-Diaz et al.
(1994) showed that mouth size of larval Sparus aurata correlates to the size of prey it consumes. For this study, the prey items found in the gut contents of larval M. menidia may be due to the available zooplankton in the estuary and mouth size.
It is necessary to stress the importance of future monitoring of the zooplankton communities in both of these estuaries, because of impacts due to anthropogenic influences. Increased nutrient levels in the water can change the community composition of estuaries (Pinckney et al. 1998), including UPJP and UPR. UPJP has a high abundance of polychaete larvae (Bengtson 1982), and is more eutrophic than UPR (Table 1) Few to no M. menidia adults and juveniles were collected in seine hauls that spring. It was determined that the jellyfish were consuming the zooplankton community in UPJP. The mechanisms that regulate this phenomenon are not well understood, but eutrophication has been suggested as one factor (Purcell 2012).
It has long been known that otoliths can be used as a proxy for fish growth.
The significant age-length relationship for M. menidia larvae in the current study has been previously shown in work by Barkman (1978).
Between estuaries, there was no significant difference in the age-length relationship of M. menidia larvae. According to the regression coefficients in the agelength equations, the larvae grow at 0.65 -0.66 mm/d. Barkman et al. (1981)  Temperature influences the growth of fish. In particular, for M. menidia, temperature is one factor that determines sex and size during the larval stage (Conover and Kynard 1981). Research by Conover and Kynard (1981) showed that warmer temperatures, 17 o -25 o C, produce more male fish compared to cooler temperatures, 11 o -19 o C, which produced more females. In addition, the male fish produced in those warmer temperatures tended to be smaller in size compared to female M. menidia (Conover and Kynard 1981). Water temperatures in UPJP are cooler than the water temperatures in UPR, even during the summer months (Volson 2012) when sampling occurred for this study (Table 1). Despite a larger length-at-hatch for M. menidia in UPJP, the cooler water temperature in this estuary may have resulted in a slower growth rate for the larvae. As a result, larval growth was not greater in UPJP compared to UPR.
Nutritional quality of the prey items might also influence the growth of M. menidia larvae. Previous research has shown that zooplankton from UPJP are more lipid-rich compared to zooplankton from UPR (Volson 2012). On the other hand, copepod eggs, with more yolk, might be expected to have a higher lipid content than do copepod nauplii. In any case, this difference in gut content between the estuaries did not influence larval growth during the time samples were collected for this study.
Furthermore, the number of copepod nauplii and copepod eggs consumed by M.
menidia larvae on an hourly or daily timescale in both estuaries is not known. As a result, any further remarks concerning the effect of the nutritional quality of prey on M. menidia larvae cannot be made.
The "growth-mortality" hypothesis states that larger fish have higher survivability than smaller fish (Anderson 1988). However, Gleason and Bengtson (1996)      For both graphs, red triangles represent the probability estimates following the Zero-inflated Poisson model. The blue dots, in each graph, represent the observed relative frequencies of the total number of larvae collected from the field. Both graphs also show how many Menidia menidia larvae were collected from the field in one tow. A.

B.
Figure 3: Gut contents of Menidia menidia larvae for UPR (n = 51) and UPJP (n = 58). Each taxon is represented as a percent of the total gut contents for all larvae collected in each estuary. Results from the chi-square analysis show a significant difference in feeding habits of Menidia menidia larvae (p < 0.0001).  Figure 4: Growth of Menidia menidia larvae collected from UPJP (open circles) and UPR (diamonds). Linear regressions represent the age-length relationship of Menidia menidia larvae from UPJP (dashed line), y = 0.66x + 2.98, and UPR (solid line), y = 0.65x + 3.06. In the linear equations Y = total length in millimeters and X = age in days. Results from the ANCOVA analysis show no significant differences in the slopes of the age-length relationship of larvae between estuaries (p = 0.8147). However, age is a significant indicator of the size of larvae (p < 0.0001).

APPENDIX I. DEPTH PREFERENCE EXPERIMENT
A mock littoral zone was created in the laboratory to determine if different size classes of M. menidia larva prefer a certain depth when residing in the littoral zone.
Larval M. menidia used for the experiment were spawned from four gravid male and four gravid female adults; collected in the field. In the lab, the adults were stripspawned, their fertilized eggs incubated, and the larvae reared according to methods described by Barkman and Beck (1976). During incubation and for the duration of the experiment, room temperature was kept at a constant 22 o C with an 18 h light: 6 h dark cycle. After seven days of incubation, the newly hatched larvae were added to one of the three aquaria. To the first aquarium, 54 larvae were added. To the second aquarium, 34 larvae were added. And to the third aquarium, 40 larvae were added.
Each aquarium was 113.56 liters (76.2 cm X 30.4 cm X 31.7 cm) and lined with sand that was sloped 12 o . A piece of acrylic glass, 0.64 cm thick, was cut to the dimensions of the aquarium and sealed over the sand ( Figure 5). The main purpose of the sand was to act as a support for the acrylic glass. The purpose of creating a slope was to mock the natural slope of the littoral zone found along the shores of Point Judith Pond and the Pettaquamscutt River. Each aquarium was divided into three sections according to depth. The shallow area ranged from 1.27 to 3.80 cm deep, the middle area ranged from 3.80 cm to 7.62 cm deep, and the deep area ranged from 7.62 cm to 10.16 cm deep. To ensure that the larvae would not be disturbed by the entrance of the observer, a black tarp was hung in front of the aquaria. During the experiment, larvae were fed Artemia nauplii every other day ad libitum. The experiment ran for two weeks. Three observations per aquarium were made daily to determine the number of larvae in each depth zone.
The exact number of larvae in each depth was unknown through most of the experiment, therefore; the number of larvae in each depth zone could only be estimated. Initial populations in each aquarium were determined by tallying the number of dead larvae removed from each tank each day.
Although the larvae were too delicate for determination of total length at the beginning of the experiment, length-at-hatch of larvae from these estuaries is about 4.5 mm (Volson 2012).

Analysis & Findings
A replicated test of goodness of fit (heterogeneity test) was constructed to determine any significant differences in depth preferences by larvae. Results showed no significant differences in depth preference by the M. menidia larvae (G T = 11.8, p > 0.05).
Contrary to observations in other laboratory settings (Bengtson, pers. comm.), the larvae did not seem to prefer specific depths over time. It was postulated that the larvae would prefer the shallow zone or be found very close to the water's edge, on the day of hatch. As time progressed, and the larvae grew, we expected them to move into the deeper zones of the aquaria. However, there did not appear to be a pattern of preference for the duration of the experiment. µm mesh to collect samples in water that was slightly greater than 1 m deep. Due to the different dimensions of the devices, each could only be used between the depths described above.