Figures
Abstract
Dispersal, the movement of an individual away from its natal or breeding ground, has been studied extensively in birds and mammals to understand the costs and benefits of movement behavior. Whether or not invertebrates disperse in response to such attributes as habitat quality or density of conspecifics remains uncertain, due in part to the difficulties in marking and recapturing invertebrates. In the upper Bay of Fundy, Canada, the intertidal amphipod Corophium volutator swims at night around the new or full moon. Furthermore, this species is regionally widespread across a large spatial scale with site-to-site variation in population structure. Such variation provides a backdrop against which biological determinants of dispersal can be investigated. We conducted a large-scale study at nine mudflats, and used swimmer density, sampled using stationary plankton nets, as a proxy for dispersing individuals. We also sampled mud residents using sediment cores over 3 sampling rounds (20–28 June, 10–17 July, 2–11 August 2010). Density of swimmers was most variable at the largest spatial scales, indicating important population-level variation. The smallest juveniles and large juveniles or small adults (particularly females) were consistently overrepresented as swimmers. Small juveniles swam at most times and locations, whereas swimming of young females decreased with increasing mud presence of young males, and swimming of large juveniles decreased with increasing mud presence of adults. Swimming in most stages increased with density of mud residents; however, proportionally less swimming occurred as total mud resident density increased. We suggest small juveniles move in search of C. volutator aggregations which possibly act as a proxy for better habitat. We also suggest large juveniles and small adults move if potential mates are limiting. Future studies can use sampling designs over large spatial scales with varying population structure to help understand the behavioral ecology of movement, and dispersal in invertebrate taxa.
Citation: Bringloe TT, Drolet D, Barbeau MA, Forbes MR, Gerwing TG (2013) Spatial Variation in Population Structure and Its Relation to Movement and the Potential for Dispersal in a Model Intertidal Invertebrate. PLoS ONE 8(7): e69091. https://doi.org/10.1371/journal.pone.0069091
Editor: Cédric Sueur, Institut Pluridisciplinaire Hubert Curien, France
Received: October 31, 2012; Accepted: June 12, 2013; Published: July 12, 2013
Copyright: © 2013 Bringloe et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The project was funded by a Strategic Project Grant from the Natural Science and Engineering Research Council of Canada (NSERC), a NSERC Discovery Grant to MAB, the Work-Study Program at UNB, and the Canada Summer Job Program. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exists.
Introduction
Movement patterns can have profound consequences on the structure and dynamics of populations and communities, and these consequences often depend on which particular stages in the life cycle of an organism are more mobile than others [1]. Studies, mostly on birds and mammals, identified two broad categories of movement: natal dispersal, the movement from the natal area to where breeding first occurs, and breeding dispersal, the movement between two successive breeding areas [1], [2]. Ultimately, dispersal of juveniles appears to be tuned to minimize inbreeding and kin competition and is proximately driven by intraspecific density [3], [4], [5], [6]. Ultimately, dispersal of adults appears to be tuned to maximize mating opportunities with the sex less invested in reproduction dispersing predominantly [3], [7]. Proximate factors influencing adult dispersal include availability of resources, density of the opposing sex, and interactions among these factors [8], [9]. Much progress has been made determining causes of dispersal in vertebrates and marine invertebrate larval stages (adults of marine invertebrate species are typically sessile, resulting in obligate larval dispersal [10], [11]), but similar information on dispersal for invertebrate taxa that are mobile throughout their life history is scarce. While it is known that certain stages of various invertebrate species move more than others (juveniles in marine amphipods [12], [13], [14], adults and juveniles in freshwater stream invertebrates [15], juveniles and sex-biased adults in land-based arthropods [16], [17], [18]), hypotheses regarding the ultimate factors causing these biases have garnered less attention. Examples, however, do include Beirinckx et al. [19], who observed female-biased dispersal in damselflies, and attributed this to maturation rate and foraging behavior in females. Caudill [20] reported female-biased dispersal in mayflies, attributed to the study species’ mating system where males emerge only to swarm for a single day around their natal pond. Albrecsten and Nachman [21] also reported female-biased dispersal in tephritid fly, attributed to pre-emptive competition for oviposition sites and a “sit-and-wait” strategy by receptive males. Baker [22] reported that the isopod Hemilepistus reaumuri disperses depending on its body size, with smaller males establishing new burrows more quickly in anticipation of not being able to compete with other males for female initiated burrows.
Knowledge of intraspecific interactions driving dispersal comes largely from long-term tracking studies, focusing on a single or a few populations [7]. These studies rely on temporal fluctuations in population structure and links to variation in dispersal. However, this approach is difficult to apply to invertebrates because individuals can be very small and numerous, rendering mark-recapture methods impractical in many cases (especially in marine environments). The inability to monitor individuals over time breaks the link between biological attributes of a population and dispersal (or proxies of dispersal) of individuals. Note that indirect measures of dispersal serve as suitable proxies where direct measures are not feasible. For example, continuity or isolation of populations is routinely approximated through genetic studies [23]. Measuring the moving portion of the population against a background of known population structure can also be used to identify more mobile life history stages [15].
One way to detect patterns of invertebrate dispersal is to characterize biological attributes of different populations over a large spatial extent and test if these attributes relate to the dispersal of certain life history stages. Typically, studying a given species in a large spatial context should yield variation in population dynamics, including differences in total densities and demographic structure. These differences among populations can in turn be tested for associations with the dispersal of individuals within populations. Such associations, if they exist, provide insight on possible explanations for dispersal, which can be used as working hypotheses in future studies that explicitly test these mechanisms. We used this approach to inform hypotheses about the ultimate causes of dispersal in a model marine invertebrate.
Model Species
The burrow-dwelling amphipod Corophium volutator is a common species of intertidal mudflat ecosystems on the North American and European coasts of the North Atlantic, reaching densities of more than 60,000 individuals m–2 [24], [25]. North American populations typically produce two apparent generations each summer, one in late May and one in July that overwinters [24], [26]; both generations overlap throughout the summer [27]. C. volutator populations are typically female biased (2∶1 or 3∶1), and scarcity of males might limit reproductive output [28], [29]. Mated females brood eggs that hatch into juveniles that are released into the surrounding mud [30]. C. volutator is also a prominent swimmer; individuals of all sizes leave burrows during periods of immersion and drift in the water column [27], [31]. Details of swimming have mostly been studied in Europe [31], [32], [33], [34] and to our knowledge have only been investigated at one mudflat in North America [27], [35], [36]. The abundance of swimming C. volutator usually displays seasonal, lunar and diel rhythmicity with peaks occurring at night during periods of immersion around the new or full moon, between June and August [31], [33], [34], [35]. Several studies reported the over representation of small juveniles among swimmers [31], [34]. A detailed study examining differences between mud residents and swimmers was conducted by Drolet and Barbeau [27] at one site, who reported important temporal variation in swimmers that occasionally resulted in increased proportions of small juveniles and males swimming, while ovigerous females tended to swim less.
Swimming by C. volutator can serve as a proxy for dispersal. C. volutator actively controls its entry and time spent in the water column by swimming, or beating its pleopods [37], [38]. Swimming is also combined with passive movement driven by water currents, with individuals entering the water column during periods of peak velocity in tidal currents [32], [39]. Individuals swimming as high as 4 m into the water column may be transported large distances, up to 14.4 km in a single swimming event [39]. As a consequence, the genetic structure of C. volutator populations shows high connectivity among mudflats [40]. Given the high degree of tidal entrainment faced by individuals entering the water column, we worked under the assumption that the characteristics of individuals found in the water column at a particular site are representative of individuals that will travel large distances. Dispersal herein is therefore defined operationally by swimming density or propensity to swim (i.e., standardized swimming activity) of amphipods, and concerns movement away from respective populations.
C. volutator is an ideal model invertebrate to study biological drivers of dispersal. Although population dynamics vary little within a mudflat [27], they vary substantially among mudflats [24], providing the backdrop for investigating attributes of populations that associate with dispersal. As well, many Corophiid amphipods (possibly including C. volutator) are successful invasive species [41], highlighting the importance of dispersal to the biology of such amphipods. Here, we expand on previous population studies by considering the demographics of the dispersing subset in greater detail. By studying multiple mudflats simultaneously throughout the upper Bay of Fundy (spanning an approximate area of 10,000 km2), we observed dispersal under a broad range of population states (density, size structure and sex structure). Our first objective was to confirm that most variation in swimming occurs at the level of site (a.k.a. mudflat) to ensure we had appropriate variability to assess differences in swimmers over different population states. Our second objective was to identify the main life history stages of dispersal by determining which sizes or sexes of individuals were swimming relative to non-swimming individuals. Our third main objective was to use the among-site variation in swimming individuals to identify potential population factors that may drive dispersal. For this objective, we assessed if dispersal of different life history stages varied in relation to overall density (or biomass) of mud residents. We also assessed whether the swimming propensity of certain stages associated with certain characteristics of the mud resident population. More specifically, we examined (i) whether standardized swimming activity (density of a swimming stage standardized by its density in the mud) of juveniles correlated with the density of resident (mud) adults; (ii) whether standardized swimming activity of males correlated with resident density of females of different reproductive stages, and (iii) whether standardized swimming activity of non-ovigerous females correlated with resident density of males. Observing resident and swimming individuals at several mudflats spanning the upper Bay of Fundy provided a robust ground to better describe the dispersal behavior of C. volutator, which in turn allowed us to provide insights on possible mechanisms driving this behavior. By informing hypotheses regarding the dispersal of C. volutator, we hope to illustrate possible avenues for mechanistic models that would further our understanding of dispersal in invertebrate taxa.
Methods
Sampling of Swimmers
No specific permits were required for the described field studies, as we sampled non-protected marine invertebrates (amphipods, annelids, gastropods, bivalves, nematodes) in non-privately-owned or non-protected locations (intertidal mudflats located below the low high tide line, according to Canadian guidelines).
To evaluate spatial and temporal patterns in density and population structure of swimming C. volutator, we sampled the water column with stationary plankton nets at 9 mudflats during summer 2010. Site locations spanned throughout the upper Bay of Fundy in the Chignecto Bay (Marys Point, Daniels Flats, Grand Anse, Pecks Cove, Minudie) and Minas Basin regions (Noel Bay, Moose Cove, Avonport, Starrs Point; Figure 1), where tidal amplitude is amongst the highest in the world (up to 16 m). The sites selected represent many of the major mudflats in the upper Bay of Fundy, ranging in size from 1.4 to 17 km2 (6.0±4.6 km2, mean±SD, n = 9 mudflats).Three plankton nets were deployed at each site in a line parallel to the low water line with 5 m between adjacent nets (Figure 2). The nets were secured at the top of metal posts, 1 m above the substratum, and were free to rotate with the water current direction. We placed the nets at a fixed elevation of 7.5 m above 0 chart datum at 8 mudflats (which is a distance of 50–300 m from shore); elevation was measured with surveying equipment. The nets at Starrs Point were placed at an elevation of 11.7 m (and 20 m from shore); this mudflat has a very shallow incline, and hardly dips below 10.5 m in elevation 500 m onto the mudflat. These nets were, however, immersed each sampling round; at 7.5 m elevation, the nets were submerged for most of the high tide cycle (∼4 h). A plankton net consisted of a metal ring (20 cm diameter) at the mouth of the net, and fabric (180 µm mesh size, which captures all sizes of C. volutator [42]) extending 60 cm to the narrow end of the net, with a plastic funnel (with an 8 cm diameter opening) glued at the narrow end of the net (Figure 2) [35].
A Maritime provinces of Canada, showing the location of the Bay of Fundy; the rectangles represent Chignecto Bay and the Minas Basin. B and C detail the two main branches of the upper Bay of Fundy. Major mudflats are shown in light grey and the sites where sampling was conducted are identified.
The triplicate nets were aligned parallel to the low water line. The tidal height at this particular location is ∼5 m above the mudflat surface. Photographs by KR Richard.
Sampling was conducted three times during summer 2010 (25–28 June, 10–13 July and 8–11 August) for three consecutive nights bounding new or full moons. Plastic bottles (500 ml, each with a 2 cm x 4 cm window covered with 180 µm mesh to prevent pressure build-up) were attached to the funnel of the nets each evening before immersion of the sites, and samples were collected in the morning following emersion. The samples were rinsed in a 250 µm sieve [42] stored in 95% ethanol, and later processed using a dissecting microscope. The number of individuals, and size (body length from tip of rostrum to end of telson) and sex of adults (>4 mm in body length; sexing detailed by Schneider et al. [28], and Barbeau and Grecian [43]) were recorded. Intersex individuals (consisting of 2.0±0.5% [mean±SE, n = 241 samples] of swimming adults) were pooled with males, since they are functional males [44]. A Folsom plankton splitter was used to subsample (1/2 to 1/512 subsamples) when an original sample exceeded 200 individuals to have 100–200 individuals to process. The number of individuals per sample was converted to a density (ind. M–3) by dividing by the volume of water filtered through each net. The volumes of water were estimated using dissolution of plaster hemispheres [45] attached at the mouth of each net; the procedure was calibrated using a saltwater flume [27].
Sampling of Mud Residents
We used mud core samples collected at each of our sites to sample mud residents. At each site, we sampled along a transect extending the eulittoral zone of the mudflat (range of transect lengths: 700 to 1800 m). Core samples were collected at 12 predetermined stratified random locations along each transect, using a 7 cm diameter corer, which was pushed into the mud to the depth of the compact hypoxic layer (3–6 cm deep). The mud was sampled: 20–24 June, 14–17 July and 2–7 August; only two mudflats (one randomly chosen per region [Chignecto Bay and Minas Basin]) could be sampled per day (we had two sampling teams working concurrently). Samples were rinsed in a 250-µm sieve and stored in 95% ethanol. C. volutator were processed from the samples as described above, though mud samples were not subsampled.
Data Analysis
Our first objective was to verify that variation in swimming C. volutator occurs at large spatial scales (i.e. level of mudflats). We used ANOVAs with the random factors of Site (9 mudflats), Round (3 sampling rounds: June, July and August), and Night (3 sampling dates per round) nested in Round. The structure of the ANOVA is presented in Table 1, and appropriate denominators for F-ratios and Quasi F-ratios were determined [46], [47]. Variance component analysis was done to estimate the proportion of variation explained by the different sources of variation [47]. This analysis was used to examine several dependent variables: total density of swimmers (ind. M–3 of water), densities of different size classes (<1.5, 1.5–2.5, 2.5–4, 4–6 and >6 body length) and densities of adult males, non-ovigerous females, and ovigerous females. Density variables were transformed using log10(datum+1) to correct for heterogeneity of variance.
To identify dispersal stages in C. volutator, we compared the size distribution of swimmers (0.5 mm increments, from 1 to 11 mm in body length) to that of the mud resident using bootstrap randomization procedures. We assumed that the expected size distribution of swimmers is a random subsample of the mud residents. For each site and round, we generated 95% confidence intervals around the size distribution of mud residents (i.e., the expected distribution based on 12 pooled cores). We randomly selected individuals out of the mud resident size distribution 1000 times and removed the 25 smallest and largest bootstrap values for each size class to get the lower and upper confidence limits. For each iteration, the number of randomly-selected individuals was the actual number of individuals measured in the plankton net samples. The observed proportions for swimmers (means from 7–9 nets) were then compared to the expected proportions, and size classes having observed proportions lying outside of the confidence interval limits of the expected distributions were declared significantly different. This analysis was done only for site-round combinations with at least 100 swimming individuals and 100 mud residents. To compare the stage structure of swimming adults (i.e., observed values) to mud resident adults (i.e., expected values), we used G-tests [48] with five amphipod stages: small and large males (4–6 mm and >6 mm), small and large non-ovigerous females (4–6 and >6 mm), and ovigerous females (which ranged between 4.7 and 10.8 mm). We computed the percentages of adults in each stage for Site*Round combinations with >100 total mud residents (pooled over 12 mud cores) and >100 individual swimmers pooled over 3 nights (n = 9 nets). If a stage had 0 individuals, it was pooled with another appropriate stage; specifically, large males were pooled with small males, and large females were pooled with ovigerous females (given that small males mature into larger males and that ovigerous females were frequently the largest individuals).
To gain insight on potential biological drivers of dispersal in C. volutator, we tested the effect of mud resident density (or biomass) on densities of swimming individuals using Model II regressions on log-transformed data. Specifically, reduced major axis regression analysis was used to account for variation in the estimate of our independent variable (Site*Round-averaged densities [ind. M–2] or biomasses [mg m–2] of total mud residents; each estimate was calculated from 12 core samples) [48], [49], [50]. The dependent variables were Site*Round-averaged densities of swimmers (total density and densities according to size class [<1.5, 1.5–2.5, 2.5–4, 4–6, >6 mm body length] or adult stage [males, non-ovigerous females and ovigerous females]). We first tested the slope for each regression against the value of 0 using a standard Student’s t-test (to determine if there was a significant relationship). If we had a significant relationship (i.e., slope different from 0), then we used Clark’s T-test to determine if the slope was significantly different from 1. Determining if a significant slope is greater or smaller than 1 is relevant because this indicates the type of relationship between the two variables (Figure S1). If the slope (on a log-log scale) is smaller than 1, then we have an ascending but decelerating relationship, suggesting a decreasing proportion of swimming individuals as mud density increases. If the slope equals 1, then a constant proportion of individuals swim across mud densities. Finally, if the slope is greater than 1, then an increasing proportion of individuals swim with increasing density of residents (ascending accelerating relationship). This analysis was done with two measures of C. volutator abundance in the mud as the independent variable, density (number of individuals m–2, which reflects the amount of mudflat surface inhabited by C. volutator) and biomass (mg m–2, which reflects the total mass of C. volutator within a site). Biomass per sample was calculated from the length-weight relationship in Boates and Smith [51].
To determine if the propensity to swim of certain stages of C. volutator is related to the structure of mud resident population, we used Pearson correlation analysis between standardized swimming activity for a stage of interest and density of a mud resident stage of interest. Because each correlation tests a different null hypothesis, our Type 1 error rate would not be inflated (which is a problem when repeatedly testing the same hypothesis). Standardized swimming activity was calculated by dividing the density of a given swimming stage by its respective density in the mud, thus providing a measurement of swimming activity relative to availability in the mud. All variables were calculated using the averages (n = 9 nets for swimmers or n = 12 mud cores for residents) for each Site*Round combination. Swimming stages investigated included small, medium and large juveniles (<1.5, 1.5–2.5 and 2.5–4 mm body length, respectively), total males, small and large males (4–6 and >6 mm, respectively), total females, and small and large non-ovigerous females (4–6 and >6 mm, respectively). The mud resident variables were total adult density for the questions about juvenile swimming, densities of different reproductive stages of females (i.e., total females, total non-ovigerous females, small and large non-ovigerous females, and ovigerous females) for questions about male swimming, and densities of males (i.e., total males, and small and large males) for questions about female swimming.
Results
Does the Density of Swimmers Vary Spatially?
Swimming occurred throughout the upper Bay of Fundy, although considerably less occurred in the Minas Basin than in Chignecto Bay (Figure 3). Spatial variation in the density of swimmers was most important at the large scale, with the factor site accounting for most of the variation in total density (Table 2). Similar patterns were observed when considering size and sex classes (Table S1).
Data are for sites in the upper Bay of Fundy during June, July and August sampling rounds: MP = Marys Point, DF = Daniels Flats, GA = Grand Anse, PC = Pecks Cove, MN = Minudie, NB = Noel Bay, MC = Moose Cove, AV = Avonport, SP = Starrs Point. n = 7–9 nets (2–3 replicate nets over each of 3 replicate nights. The inserts for June and July are magnified charts for the Minas Basin sites. The density for Noel Bay in August was 0.27±0.06 ind. M–3.
Does the Stage Structure of Swimmers Reflect that of Mud Residents?
The observed size distribution of swimmers was different from the expected size distribution (if swimmers are a random subsample of the residents) for all site and round combinations for which the randomization test could be performed (Figures 4, 5, 6). One of two distinct patterns was apparent for all site and round combinations. First, the smallest size class (1–1.5 mm) was often overrepresented in the plankton samples, at a few sites in June and July (Figures 4 and 5, respectively), and at all the sites in August (Figure 6). Second, one or several of the intermediate size classes (large juveniles and small adults) were overrepresented in the plankton samples, which generally occurred when the small juveniles were not overrepresented (11 out of 15 times). The other times, both small juveniles and large juveniles were overrepresented. The stage structure of adult swimmers differed significantly from that expected based on adult residents in all Site*Round combinations tested, except Minudie in July (Figure 7). In nearly all cases that could be tested and were significant, large adults (males and non-ovigerous females >6 mm) and ovigerous females swam less than expected. Where significant, small adult males (4–6 mm) often swam more than expected, and small non-ovigerous females almost always swam more than expected (with 5 exceptions for males, and 1 exception for females).
Distributions are for sites in the upper Bay of Fundy during June 2010. The expected proportions correspond to the observed proportions residing in the mud (pooling over 12 cores). Error bars are 95% confidence intervals obtained by subsampling (bootstrapping) the mud resident population. The observed proportions for swimmers are the mean, calculated from the 9 nets (3 replicate nets for each of 3 replicate nights). Symbols above bars in an observed distribution indicate 0.5-mm size classes which had a greater (+) or smaller (−) proportion swimming than expected (based on the confidence intervals of the expected distribution). The numbers in parentheses represent the number of individuals measured to obtain the size distributions.
Distributions are for sites in the upper Bay of Fundy during July 2010. The expected proportions correspond to the observed proportions residing in the mud (pooling over 12 cores). Error bars are 95% confidence intervals obtained by subsampling (bootstrapping) the mud resident population. The observed proportions for swimmers are the mean, calculated from the 9 nets (3 replicate nets for each of 3 replicate nights). Symbols above bars in an observed distribution indicate 0.5-mm size classes which had a greater (+) or smaller (−) proportion swimming than expected (based on the confidence intervals of the expected distribution). The numbers in parentheses represent the number of individuals measured to obtain the size distributions.
Distributions are for sites in the upper Bay of Fundy during August 2010. The expected proportions correspond to the observed proportions residing in the mud (pooling over 12 cores). Error bars are 95% confidence intervals obtained by subsampling (bootstrapping) the mud resident population. The observed proportions for swimmers are the mean, calculated from the 7–9 nets (2–3 replicate nets for each of 3 replicate nights). Symbols above bars in an observed distribution indicate 0.5-mm size classes which had a greater (+) or smaller (−) proportion swimming than expected (based on the confidence intervals of the expected distribution). The numbers in parentheses represent the number of individuals measured to obtain the size distributions.
Data are for sites in the upper Bay of Fundy, summer 2010. The G-test statistic (subscript = degrees of freedom) and associated p-value compare expected (mud resident) and observed (swimming) distributions. For expected values, data were pooled over 12 cores; for observed values, the 2–3 nets per night were pooled, and the 3 replicate nights averaged. The df in the G-test is <4, when one or more adult stages with 0 individuals needed to be pooled with another stage. Site-round combinations with less than 100 C. volutator adults in the mud were not analysed. Small males and small non-ovigerous females are 4–6 mm body length, and large males and large non-ovigerous females are >6 mm. Ovigerous females ranged from 4.7 to 10.8 mm.
How does Swimming (or Potential for Dispersal) Relate to Density or Biomass of Mud Residents?
Swimming density of most stages of C. volutator was positively related to the overall abundance of residents, measured both as density (number of individuals m–2) and as biomass (mg m–2), as indicated by a positive slope significantly different from 0 (Table 3, Figure 8). The relationship was not significant for large adults (>6 mm) and ovigerous females. Significant slopes were shallowest for adults and steepest for small juveniles (<1.5 mm). However, all significant slopes were significantly less than 1, indicating a swimming response of most stages that is density dependent, with a decelerating rate of increase (Figure S1).
Data are for various C. volutator size and sex classes, and are site-round combinations in the upper Bay of Fundy, summer 2010. For swimmer values, mean of 7–9 nets (2–3 replicate nets for each of the 3 replicate nights); for mud resident values, mean of 12 replicate cores. See Table 3 for results of statistical analyses.
Is the Propensity to Swim of Certain Stages Related to Other Stages of Mud Residents?
Standardized swimming activity of certain stages was significantly associated with characteristics of the mud resident population for two stages (Figure 9, Table S2): large juveniles and small non-ovigerous females (2.5–4 and 4–6 mm respectively, Figure 9C, F). Standardized swimming activity of large juveniles was negatively correlated to density of resident adults (Figure 9C), indicating that these juveniles are less likely to swim at higher adult densities in the mud. However, standardized swimming activity of smaller juveniles (<1.5, 1.5–2.5 mm) was not associated with density of resident adults (Figure 9A, B). Standardized swimming activity of males (whether it be all males, small males or large males) was not strongly associated with any resident female features (densities for all females, non-ovigerous females or ovigerous females; p>0.08; Figure 9D, E; Table S2). Standardized swimming activity of small non-ovigerous females (4–6 mm) was negatively correlated with small resident males (4–6 mm; Figure 9F). No other significant associations were found between swimming propensity of females and any other resident male features (p>0.09; Table S2).
Data are site-round combinations in the upper Bay of Fundy, summer 2010. Standardized swimming activity was calculated as density of swimmers [m–3] divided by the respective density of mud residents [m–2]. For swimmer values, mean of n = 7–9 nets; for mud resident values, mean of 12 replicate cores. Small juveniles <1.5 mm body length, medium juveniles = 1.5–2.5 mm, large juveniles = 2.5–4 mm, small males and small non-ovigerous females = 4–6 mm, large males and large non-ovigerous females >6 mm. Ovigerous females ranged between 4.7 and 10.8 mm. The correlation coefficient (r, subscript = degrees of freedom) and p-value for Pearson’s correlation test on the untransformed data are presented; for significant correlations, the line of best fit is plotted.
Discussion
The mobility, abundance, and broad diversity of life histories of invertebrates can provide invaluable systems to test and refine hypotheses about dispersal, yet methodologies to accurately study such small animals remain scarce. Here, we used large spatial scale variation in population structure of both mud residents and swimmers to provide insight on mechanisms of dispersal in a model invertebrate species.
Spatial Variation in Density of Swimmers
Studies of swimming behavior in a key mudflat species Corophium volutator led to a better understanding of its population distribution and maintenance [27], [36], [39], [52]. However, the adaptive significance of dispersal in this species remained poorly understood. Prior to our study, swimming of C. volutator was only investigated at one site in the Bay of Fundy (Pecks Cove); our first objective was to evaluate variability in swimming across mudflats. Our results clearly demonstrated that swimming occurs at all the sites sampled, which correspond to many of the major mudflats of the upper Bay of Fundy (Figure 3). Most of the variation in the density of swimmers occurred among mudflats, notably between the two basins (Chignecto Bay and Minas Basin). Population structure of individuals in the mud can differ greatly among mudflats [24], and so can that of swimmers among mudflats. Distinct population structure of swimmers, even for mudflats adjacent to one another, suggests that movement by individuals largely occurs within the boundaries of mudflats. However, high rates of gene flow suggest that dispersing individuals are reaching adjacent mudflats [40]; individuals found in the water column are most likely representative of those individuals dispersing large distances [39].
Dispersal Stages
Individuals of all stages can disperse, but marked differences in the degree of swimming occurred among stages. We identified two main stages for dispersal: the smallest juveniles are prominent dispersers, swimming at most times and places, and large juveniles/small adults (particularly females) disperse under certain conditions. Both stages are pre-reproductive (or starting to reproduce), and so correspond to juvenile dispersal [1]. In contrast, large adults, presumably individuals who have already produced offspring, dispersed relatively little, meaning that true adult dispersal likely does not occur in C. volutator. This apparent lack or low level of adult dispersal may be a consequence of the life history of C. volutator, which usually live through a single breeding season [24].
Factors Driving Dispersal of Small Juveniles
Juvenile dispersal occurs when individuals move away from their place of birth, but the evolutionary underpinnings of this behavior remain debated [3]. C. volutator exhibited juvenile dispersal as recently hatched individuals (Figure 6); Stevens et al. [14] also reported significant net movements of small (<2.2 mm) juvenile corophiid amphipods in a New Zealand estuary. Dispersal by the smallest juveniles in our study did not appear to be driven by intraspecific competition: we did not detect positive density dependence (which would be an ascending accelerating relationship; Figures 8 and S1, Table 3) and there were no associations between the propensity to swim of young juveniles and presence of adults (Figure 9A). One hypothesis suggested by the observed ascending but decelerating rate of dispersal with increasing mud density, is that C. volutator seeks high densities of individuals in the mud, particularly early in life (Figure 8, Table 3); gregarious behavior has been previously observed in C. volutator [53], [54], though movement from high to low density patches was reported from mesocosm experiments [55], [56], and field experiments [57, 58; see Drolet et al. (54) for possible explanations for differences in results]. If C. volutator exhibits gregarious behavior, it is possible that high mud density is a good proxy for habitat suitability, though what constitutes “good habitat” needs to be examined using measured environmental variables beyond population structure (e.g. abundance of diatom food [59], presence/absence of predators [60]). The tendency to aggregate may also benefit the individual by facilitating mate finding earlier in life. On the other hand, the lack of an association between the dispersal of the smallest juveniles and surrounding population structure may indicate that dispersal early in life, if not precipitated by environmental factors, is a “hard-wired” behavior; the possibilities of inbreeding or kin competition avoidance, though difficult to discern, could be investigated [61], [62]. These hypotheses require further testing.
Factors Driving Dispersal of Larger Individuals
Sexually mature individuals typically disperse to increase probabilities of finding a mate [2], [7]. Dispersal by C. volutator seems in tune with ideas about reproductive opportunities. Increased dispersal was observed with the onset of adulthood by large juveniles and small adults, especially females, possibly driven by mate limitation. Similar to small juveniles, swimming of large juveniles and small adults was not positively density dependent (or exhibiting intraspecific competition; Figure 8, Table 3). However, intraspecific interactions appeared to exist between these individuals and other mud resident stages. The propensity to swim of large juveniles was negatively associated with the presence of resident adults (Figure 9C), meaning that large juveniles swam proportionally less as adult presence increased in the mud. Adult presence would be a good indicator of the presence of potential mates, and so large juveniles from areas with low adult presence might be dispersing to secure mating opportunities. Dispersal by small adults might also be a mechanism for mate finding. Small females dispersed more than expected (Figure 7) and the propensity to swim of small females was negatively associated with the presence of mud resident small males (Figure 9F). The tendency for females to disperse substantially could be a reflection of the skewed sex ratio present in C. volutator populations, where females typically outnumber males 2∶1 or 3∶1 [24], [26], and in extreme cases 20∶1 [27]. The lower relative abundance of males leads to a smaller proportion of females becoming ovigerous and to reduced brood size, and so females may compete for males [29]. Therefore, presence of males may be an important factor in driving dispersal at the onset of sexual maturity. Our results suggest that females disperse to maximize their chances of finding a mate in a limited pool of males. If this hypothesis is true, males may be maximizing their reproductive output and minimizing their own risks (mortality through dispersal) by simply allowing females to find them. Female-biased juvenile dispersal has been reported in the corophiid amphipod Paracorophium spp. [14] and female-biased adult dispersal has been reported in several insects [18], [19], [20].
Conclusions
We have demonstrated the importance of studies at large spatial scales, focused on population variation, when studying invertebrate dispersal. By expanding sampling of swimming C. volutator from one site to several major mudflats in the Bay of Fundy, we identified C. volutator’s main dispersal stages and offered possible explanations as to why C. volutator disperses. Small juveniles (<1.5 mm) disperse regardless of surrounding population structure; the implications or reasons for aggregating in the mud as young juveniles should be investigated. In addition, we found associations between the propensity to disperse in large juveniles (2.5–4 mm) and small females (4–6 mm) and the density of mud resident adults and males, respectively. Our findings suggest that C. volutator on the cusp of reproduction disperse in response to mating requirements, and that females often assume the role of dispersal given sex-ratios are skewed in their favor. The observations we reported required variation in resident population structure as a backdrop to our sampling of dispersing individuals, and future studies can follow up on these observations and test for behavioral mechanisms of invertebrate dispersal. Future studies could also adopt our sampling design and statistical approach to identify dispersal stages in other invertebrate species, which will in turn enhance understanding of dispersal over a wider range of taxa; the necessity of large spatial contexts should not be overlooked.
Supporting Information
Figure S1.
Linear and exponential (increasing at an accelerating rate and increasing at a decelerating rate) relationships in a) untransformed space and b) in log10 transformed space. The slopes for log-transformed lines are indicated. If the x-axis represents population density, the density dependent relationships can be thought of in a similar way to functional responses in predation theory [63].
https://doi.org/10.1371/journal.pone.0069091.s001
(TIF)
Table S1.
ANOVA results for densities (ind. m-3, log10[datum+1]-transformed) of swimming Corophium volutator in the upper Bay of Fundy, summer 2010. All sources of variation are random. The degrees of freedom are the same as in Table 2.
https://doi.org/10.1371/journal.pone.0069091.s002
(DOCX)
Table S2.
Results for correlation analyses between standardized swimming activity and mud resident variables of Corophium volutator. Significant results are bolded. Patchiness was calculated as variance/mean for the given variable. Body lengths are in mm.
https://doi.org/10.1371/journal.pone.0069091.s003
(DOCX)
Acknowledgments
We thank Steph MacNeil, Alyssa Allen Gerwing, Allie Mayberry, Cindy Ochieng, Sebastian Richard, Julianne Richard and Kenny Richard for assistance in the field and in processing samples; Jason Addison, Janice Lawrence and the mudflat ecology group at University of New Brunswick (UNB), Mount Allison University and Carleton University for useful discussions; and 2 anonymous reviewers for useful comments.
Author Contributions
Conceived and designed the experiments: DD MAB TTB. Performed the experiments: TTB DD TGG. Analyzed the data: TTB DD MAB. Contributed reagents/materials/analysis tools: MAB. Wrote the paper: TTB DD MRF MAB TGG.
References
- 1.
Clobert J, Danchin E, Dhondt AA, Nichols JD (2001) Dispersal. USA: Oxford University Press. 480 p.
- 2. Greenwood PJ, Harvey PH (1982) The natal and breeding dispersal of birds. Annu Rev Ecol Syst 13: 1–21.
- 3.
Perrin N, Goudet J (2001) Inbreeding, kinship, and the evolution of natal dispersal. In: Clobert J, Danchin E, Dhondt AA, Nichols JD, editors. Dispersal. USA: Oxford University Press. 123–142.
- 4. Strickland D (1991) Juvenile dispersal in gray jays: dominant brood member expels siblings from natal territory. Can J Zool 69: 2935–45.
- 5. Wheelwright NT, Mauck RA (1998) Philopatry, natal dispersal and inbreeding avoidance in an island population of savannah sparrows. Ecol 79: 755–767.
- 6. Costello CM, Creel SR, Kalinowski ST, Vu NV, Quigley HB (2008) Sex-based natal dispersal and inbreeding avoidance in American black bears as revealed by spatial genetic analyses. Mol Ecol 17: 4713–4723.
- 7. Greenwood PJ (1980) Mating systems, philopatry and dispersal in birds and mammals. Anim Behav 28: 1140–1162.
- 8.
Danchin E, Heg D, Doligez B (2001) Public information and breeding habitat selection. In: Clobert J, Danchin E, Dhondt AA, Nichols JD, editors. Dispersal. USA: Oxford University Press. 243–258.
- 9.
Lambin X, Aars J, Piertney SB (2001) Dispersal, intraspecific competition, kin competition and kin facilitation: a review of the empirical evidence. In: Clobert J, Danchin E, Dhondt AA, Nichols JD, editors. Dispersal. USA: Oxford University Press. 110–122.
- 10. Strathmann RR (1993) Hypothesis on the origins of marine larvae. Annu Rev Ecol Syst 24: 89–117.
- 11. Pechenik JA (1999) On the advantages and disadvantages of larval stages in benthic marine invertebrate life cycles. Mar Eco-Prog Ser 177: 269–297.
- 12. Munguia P, Mackie C, Levitan DR (2007) The influence of stage-dependent dispersal on the population dynamics of three amphipod species. Oecologia 153: 533–541.
- 13. Franz DR, Mohamed Y (1989) Short-distance dispersal in a fouling community amphipod crustacean, Jassa marmorata Holmes. J Exp Mar Biol Ecol 133: 1–14.
- 14. Stevens MI, Hogg ID, Pilditch CA (2005) Evidence for female-biased juvenile dispersal in corophiid amphipods from a New Zealand estuary. J Exp Mar Biol Ecol 331: 9–20.
- 15. Bilton DT, Freeland JR, Okamura B (2001) Dispersal in freshwater invertebrates. Annu Rev Ecol Syst 32: 159–81.
- 16. Lagisz M, Wolff K, Sanderson RA, Laskowski R (2010) Genetic population structure of the ground beetle, Pterostichus oblongopunctatus, inhabiting a fragmented and polluted landscape: Evidence for sex-biased dispersal. J Insect Biol 10(105): 1–20.
- 17. Johns RC, Quiring DT, Ostaff DP (2009) Sex biased intra-tree oviposition site selection and larval foraging behavior of a specialist herbivore. Entomol Exp Appl 132: 264–274.
- 18. Johannesen J, Wennmann JT, Lubin Y (2012) Dispersal behavior and colony structure in a colonial spider. Behav Ecol Sociobiol 66: 1387–1398.
- 19. Beirinckx K, Gossum HS, Lajeunesse MJ, Forbes MR (2006) Sex biases in dispersal and philopatry: insights from a meta-analysis based on capture-mark-recapture studies of damselflies. Oikos 113: 539–547.
- 20. Caudill CC (2003) Measuring dispersal in a metapopulation using stable isotope enrichment: high rates of sex-biased dispersal between patches in a mayfly population. Oikos 101: 624–630.
- 21. Albrectsen B, Nachman G (2001) Female-biased density-dependent dispersal of a tephritid fly in a fragmented habitat and its implication for population regulation. Oikos 94: 263–272.
- 22. Baker MB (2004) Sex biased state dependence in natal dispersal in desert isopods, Hemilepistus reaumuri. J Insect Behav 17: 579–598.
- 23. Broquet T, Petit EJ (2009) Molecular estimation of dispersal for ecology and population genetics. Annu Rev Ecol Evol Syst 40: 193–216.
- 24. Barbeau MA, Grecian LA, Arnold EE, Sheahan DC, Hamilton DJ (2009) Spatial and temporal variation in the population dynamics of the intertidal amphipod Corophium volutator in the upper Bay of Fundy, Canada. J Crustacean Biol 29: 491–506.
- 25. Backer AD, Van Coillie F, Montserrat F, Provoost P, Van Colen C, et al. (2011) Bioturbation effects of Corophium volutator: importance of density and behavioral activity. Estuar, Coast Shelf Sci 91: 306–313.
- 26. Peer DL, Linkletter LE, Hicklin PW (1986) Life history and reproductive biology of Corophium volutator (Crustacea: Amphipoda) and the influence of shorebird predation on population structure in Chignecto Bay, Bay of Fundy, Canada. Netherlands J Sea Res 20: 359–373.
- 27. Drolet D, Barbeau MA (2012) Population structure of resident, immigrant and swimming Corophium volutator (Amphipoda) on an intertidal mudflat in the Bay of Fundy, Canada. J Sea Res 70: 1–13.
- 28. Schneider SD, Boates JS, Forbes M. 1994. Sex ratios of Corophium volutator (Pallas) (Crustacea: Amphipoda) in Bay of Fundy populations. Can J Zool 72: 1915–1921.
- 29. Forbes MR, McCurdy DG, Lui K, Mautner SI, Boates JS (2005) Evidence for seasonal mate limitation in populations of an intertidal amphipod, Corophium volutator (Pallas). Behav Ecol Sociobiol 60: 87–95.
- 30. Fish JD, Mills A (1979) The reproductive biology of Corophium volutator and C. arenarium (Crustacea: Amphipoda). J Mar Biol Assoc UK 59: 355–368.
- 31. Hughes G (1988) Dispersal by benthic invertebrates: the in situ swimming behaviour of the amphipod Corophium volutator. J Mar Biol Assoc UK 68: 565–580.
- 32. Essink K, Kleef H, Visser W (1989) On the pelagic occurrence and dispersal of the benthic amphipod Corophium volutator. J Mar Biol Assoc UK 69: 11–16.
- 33. Hughes G, Horsfall M (1990) Differences in the swimming behaviour of the amphipod Corophium volutator from different populations. J Mar Biol Assoc UK 70: 143–148.
- 34. Lawrie M, Raffaelli G (1998) In situ swimming behaviour of amphipod Corophium volutator (Pallas). J Exp Mar Biol Ecol 224: 237–251.
- 35. Drolet D, Barbeau MA (2009) Diel and semi-lunar cycles in the swimming activity of the intertidal, benthic amphipod Corophium volutator in the upper Bay of Fundy. J Crustacean Biol 29: 51–56.
- 36. Drolet D, Barbeau MA (2011) Movement patterns drive within-mudflat distribution of an intertidal amphipod. Mar Ecol Prog Ser 431: 255–265.
- 37. Ford RB, Paterson DM (2001) Behaviour of Corophium volutator in still versus flowing water. Estuar Coast Shelf Sci 52: 357–362.
- 38. Meadows PS, Reid A (1966) The behaviour of Corophium volutator (Pallas). J Zool 150: 387–399.
- 39. Drolet D, Bringloe TT, Coffin MRS, Barbeau MA, Hamilton DJ (2012) Potential for between-mudflat movement and metapopulation dynamics in an intertidal burrowing amphipod. Mar Ecol Prog Ser 449: 197–209.
- 40. Einfeldt AL, Addison JA (2013) Hydrology influences population genetic structure and connectivity of the intertidal amphipod Corophium volutator in the northwest Atlantic. Mar Biol 160(4): 1015–1027.
- 41. Buckley P, Dussart G, Trigwell JA (2004) Invasion and expansion of Corophiidae (Amphipoda) in the Stour estuary (Kent, UK). Crustaceana 77(4): 425–433.
- 42. Crewe TL, Hamilton DJ, Diamond AW (2001) Effects of mesh size on sieved samples of Corophium volutator. Estuar Coast Shelf Sci 53: 151–154.
- 43. Barbeau MA, Grecian LA (2003) Occurrence of intersexuality in the amphipod Corophium volutator (Pallas) in the upper Bay of Fundy, Canada. Crustaceana 76: 665–679.
- 44. McCurdy DG, Forbes MR, Logan SP, Lancaster D, Mautner SI (2004) The functional significance of intersexes in the amphipod Corophium volutator. J Crustacean Biol 24: 261–265.
- 45. Komatsu T, Kawai H (1992) Measurements of time-averaged intensity of water motion with plaster balls. J Oceanogr 48: 353–365.
- 46.
Winer BJ, Brown DR, Michels KM (1991) Statistical Principles in Experimental Design, 3rd ed. McGraw Hill. 1057 p.
- 47.
Underwood AJ (1997) Experiments in Ecology: Their Logical Design and Interpretation Using Analysis of Variance. England: Cambridge University Press. 524 p.
- 48.
Sokal RR, Rohlf JF (1981) Biometry. San Fransisco: Freeman. 859 p.
- 49. McArdle BH (1988) The structural relationship: regression in biology. Can J Zool 66: 2329–2339.
- 50.
Keough MJ, Quinn GP (2002) Experimental Design and Data Analysis for Biologists. England: Cambridge University Press. 556 p.
- 51. Boates JS, Smith PC (1979) Length-weight relationships, energy content and the effects of predation on Corophium volutator (Pallas) (Crustacea: Amphipoda). Proc NS Inst Sci 29: 489–499.
- 52. Drolet D, Barbeau MA (2009) Differential emigration causes aggregation of the amphipod Corophium volutator (Pallas) in tide pools on mudflats of the upper Bay of Fundy, Canada. J Exp Mar Biol Ecol 370: 41–47.
- 53. Meadows PS (1964) Experiments on substrate selection by Corophium volutator (Pallas): depth selection and population density. J Exp Biol 41: 677–687.
- 54.
Drolet D, Coffin MRS, Barbeau MA, Hamilton DJ (2013) Influence of intra- and interspecific interactions on short-term movement of the amphipod Corophium volutator in varying environmental conditions. Estuar Coast DOI10.1007/s12237-013-9605-4.
- 55. Bulling MT, Solan M, Dyson KE, Hernandez-Milian G, Luque P, et al. (2008) Species effects on ecosystem processes are modified by faunal responses to habitat composition. Oecologia 158: 511–520.
- 56. Godbold JA, Bulling MT, Solan M (2011) Habitat structure mediates biodiversity effects on ecosystem properties. Proc R Soc B 278: 2510–2518.
- 57. Wilson WH (1989) Predation and the mediation of intraspecific competition in an infaunal community in the Bay of Fundy. J Exp Mar Biol Ecol 132: 221–245.
- 58. Jensen KT, Kristensen LD (1990) A field experiment on competition between Corophium volutator (Pallas) and Corophium arenarium crawford (Crustacea, Amphipoda)-effects on survival, reproduction and recruitment. J Exp Mar Biol Ecol 137: 1–24.
- 59. Gerdol V, Hughes RG (1994) Feeding behavior and diet of Corophium volutator in an estuary in southeastern England. Mar Ecol Prog Ser 114: 103–108.
- 60. Coffin MRS, Barbeau MA, Hamilton DJ, Drolet D (2012) Effect of the mud snail Ilyanassa obsoleta on vital rates of the intertidal amphipod Corophium volutator. J Exp Mar Biol Ecol 418: 12–23.
- 61. De Luca PA, Cocroft RB (2008) The effects of age and relatedness on mating patterns in thornbug treehoppers: inbreeding avoidance or inbreeding tolerance? Behav Ecol Sociobiol 62: 1869–1875.
- 62. Bretman A, Rodriguez-Munoz R, Walling C, Slate J, Tregenza T (2011) Fine-scale population structure, inbreeding risk and avoidance in a wild insect population. Mol Ecol 20: 3045–3055.
- 63.
Taylor RJ (1984) Predation. Chapman & Hall, New York.