Oecologia (2014) 174:121–130 DOI 10.1007/s00442-013-2751-3
POPULATION ECOLOGY - ORIGINAL RESEARCH
Linking individual diet variation and fecundity in an omnivorous marine consumer Blaine D. Griffen
Received: 4 March 2013 / Accepted: 14 August 2013 / Published online: 31 August 2013 © Springer-Verlag Berlin Heidelberg 2013
Abstract Individual diet variation is a common feature of populations. This variation may be particularly common in near-shore marine benthic habitats where omnivory is prevalent and prey availability is spatially variable. Accurately predicting population responses to anthropogenic change that is occurring rapidly in these systems requires a quantitative link between individual diet variation and fecundity. Here I develop this quantitative link for the European green crab Carcinus maenas, specifically focusing on variation in the relative amounts of plant and animal material included in the diet. I demonstrate both short- and long-term diet variation between crabs as well as large individual variation in fecundity. I then quantitatively link variation in diet and fecundity using a laboratory feeding experiment. Fecundity increased by approximately 5,200 eggs when daily consumption of animal tissue increased by 1 % of body weight, but was not influenced by the most commonly consumed algal species. Results presented here have important implications for understanding population dynamics in general, and also provide information necessary for accurately predicting population growth of this widespread invader. Keywords Bioenergetics · Individual specialization · Reproductive effort
Communicated by Pete Peterson. B. D. Griffen (*) Department of Biological Sciences and Marine Science Program, University of South Carolina, Columbia, SC 29208, USA e-mail:
[email protected]
Introduction Trait variation between individuals within a population is ubiquitous and can take many forms including variation in morphology, physiology, or behavior. This individual phenotypic variation can have important consequences for population and community dynamics. At the population level, individual variation is one of the primary determinants of population stability (Łomnicki 1988) and thus extinction risk (Connor and White 1999; Fox and Kendall 2002; Kendall and Fox 2002, 2003; Fox 2005; Melbourne and Hastings 2008; Vindenes et al. 2008). At the community level, individual variation can determine the strength, form, and importance of both direct and indirect ecological interactions (Bolnick et al. 2011; Griffen et al. 2012; Pruitt et al. 2012a, b). One common type of variation is individual resource or dietary specialization (Bolnick et al. 2003). Variation in diet within a population may occur because of endogenous factors, such as between-individual differences in search strategies (e.g., Elinger 1989; Austin et al. 2004), prey handling capabilities (e.g., Estes et al. 2003), or food preferences (e.g., Poor and Hill 2006), or because of exogenous factors, such as spatial variation in food availability (e.g., Elinger 1989). In addition to these natural sources of variation, diet variation is also commonly induced in populations as a result of anthropogenic environmental changes. These induced changes are especially prevalent in coastal areas that, due to their proximity to humans, are especially vulnerable to human-induced rapid environmental changes. Previous studies have demonstrated human-induced changes in diet of near-shore marine mammals (e.g., Hirons et al. 2001; McKinney et al. 2009), fish (e.g., Pihl 1994), sea birds (e.g., Sydeman et al. 2001), and crustaceans (e.g., Reichmuth et al. 2009). The result of these diet changes is
13
122
greater diet variation across populations when individuals differ in whether or how much they alter their diets. Further, some of these diet changes were implicated in concurrent changes in population dynamics (Hirons et al. 2001; Sydeman et al. 2001). With continued stress being placed on coastal ecosystems as a result of growing human populations in these areas, understanding the consequences of these diet shifts for marine animal populations is becoming increasingly important. Diet shifts and increases in diet variation may ultimately have important population-level consequences because of stoichiometric (Sterner and Elser 2002) or energetic (Kooijman 2010) links between diet and reproduction. Ultimately, scientists would like to predict how future changes to coastal environments will influence the species that live there. However, our ability to make these predictions is limited by, among other things, two deficiencies in our understanding of diet variation and in the link between diet and reproduction. First, while individual diet variation has been found in a wide range of phylogenetic groups (Bolnick et al. 2003), most research in this area has been conducted with fish, mammals, birds, reptiles, and insects (Araújo et al. 2011). We know much less about diet variation in non-insect invertebrates, yet these animals make up an important group of consumers in coastal waters (Davenport and Anderson 2007)—a group that now plays an enhanced role in top-down regulation of near-shore ecological communities with the loss of large predators from coastal ecosystems due to overfishing, local extinctions, and invasions (Steneck and Carlton 2001; Byrnes et al. 2007; Howarth et al. 2013). Second, while the link between diet and fecundity is often assumed or even demonstrated in a qualitative way (e.g., Anderson et al. 1982; Kennish 1997; Griffen et al. 2011), relatively few studies have quantitatively examined the link between diet variation and fecundity in coastal species (but see Cruz-Rivera and Hay 2000a, b, 2001 for examples with amphipods). Here I examine the quantitative link between diet and fecundity for an omnivorous consumer that is now found in coastal habitats around the globe. Native to Europe, the green crab Carcinus maenas is a prolific invader that has now established numerous invasive populations worldwide (Carlton and Cohen 2003). One such population occurs on the northeast coast of North America. C. maenas has been present in this area since the early 1800s (Say 1817) and throughout that time has most commonly consumed bivalve prey such as clams (Mya arenaria) and mussels (Mytilus edulis), and only secondarily consumed other food items, including algae (Elner 1981). However, shipping traffic resulted in the introduction, through ballast water, of the Asian shore crab to the same shores in the early 1990s (Williams and McDermott 1990). As a result of interactions with this new invader, C. maenas shifts its diet
13
Oecologia (2014) 174:121–130
to include fewer animal prey and more algae (Griffen et al. 2008). With the spread of the new invader, C. maenas abundances have declined considerably (Lohrer and Whitlatch 2002), potentially due in part to reproductive failure as a result of this diet shift (Griffen et al. 2011). Understanding this decline or the declines of species in other systems as the result of human-induced changes may be facilitated by a clear, quantitative link between dietary intake and fecundity. The goal of this paper is to examine the quantitative link between the proportion of algal material vs. animal material in the diet of the green crab C. maenas and the resulting fecundity. As noted above, C. maenas has previously been shown to shift to an algal diet in the presence of the newer invasive Asian shore crab (Griffen et al. 2008). However, while this diet shift occurs on average, there appears to be considerable individual variation in diet, whether because some individuals alter their diets more than others or because of spatial variation in the availability of animal prey. The goal here is therefore to examine how the relative change in herbivory influences fecundity across individuals. I do this using several approaches. First, I use natural diets from field-sampled crabs to demonstrate that individual diet variation is common over both the short and long term. I also demonstrate abundant variation in fecundity. Next, I experimentally manipulate crab diets to examine how this variation influences reproductive effort and how it may influence future reproductive effort via changes in individual growth. Finally, I scale relative reproductive effort to a quantitative measure of fecundity using information on egg mass and crab reproductive physiology.
Materials and methods Crabs used in this study were collected intertidally by hand from Odiorne State Park and Fort Stark State Park, two beaches on the New Hampshire coast separated by approximately 1 km. Both are moderately exposed sites characterized by intertidal boulders overlying a substrate of shell, sand, and bedrock, and each supports a similar ecological community. Variation in short‑term diets I quantified variation in short-term individual diets using gut content analysis. I had two goals: first, to determine the relative amount of within-individual and betweenindividual diet variation; second, to determine the extent that individuals mixed their diets on a continual basis vs. consuming a single type of food for a single meal. I collected 53 adult female C. maenas (carapace width >30 mm) in the summer of 2006 at dawn on receding tides. Each
123
Oecologia (2014) 174:121–130
crab was dissected and the gut contents were spread over a grid of 100 points. The gut content item nearest to each point was then identified under a dissecting microscope in order to determine the percent contribution of different food types to the diet. I determined the total amount of variation present in the population (the total niche width; TNW) that was attributable to differences in gut contents between individuals (between-individual component; BIC) and to variation within a single individual’s gut contents (within-individual component; WIC) following Roughgarden (1972) and using the freeware program IndSpec4 (Bolnick et al. 2002). I used BIC/TNW to determine the proportion of total diet variation attributable to variation between individuals. I then used nonparametric bootstrap resampling (n = 1,000) to determine whether this observed variation between individuals (BIC/TNW) differed significantly from expected values if each consumer had sampled randomly from the population’s diet.
could be expected due to variation in egg size (i.e., if one individual produces generally larger eggs than another). Transitioning from reproductive effort to fecundity In the experiment described in the next section, I quantified reproductive effort as the mass of reproductive material (ovary + eggs). Ultimately, I wanted to predict the number of eggs that would be produced by vitellogenic ovaries. I did this by comparing the energy content of eggs from gravid crabs and ovaries from vitellogenic crabs, reasoning that if these two tissues had the same energy content per unit mass, then vitellogenic ovary mass could be directly translated into number of eggs simply using the mass of a single egg. I took a subsample (~0.1 g) from eggs or ovaries and determined the energy content of each subsample using a Parr 6725 semi-micro oxygen bomb calorimeter. I then used an unpaired t-test to compare the energy content of these two tissues.
Variation in long‑term diets I quantified variation in long-term individual diets using stable isotope ratios of nitrogen (δ15N) and carbon (δ13C), where δ15N provides a relative measure of trophic level and δ13C provides a measure of diet variation within trophic level (Post 2002). I collected 146 adult female C. maenas in spring of 2009. I used muscle tissue from a single walking leg from each crab for isotope analysis. δ15N and δ13C were measured simultaneously using an Isoprime mass spectrometer connected via continuous flow to a EuroVector Elemental Analyzer. Three internal standards were run approximately every 40 samples to calibrate the system and to compensate for potential drift over time (nitrogen— USGS40, N1 and N2; carbon—USGS40, USGS24, sucrose NIST8542). Variation in fecundity I quantified natural variation in fecundity across 45 eggbearing C. maenas. Eggs were removed from the pleopods of each female and each female was then dissected to ensure that they were post-vitellogenic (i.e., that all egg production was completed). Each egg mass was dried for 48 h at 70 °C and then weighed. Three subsamples of eggs were taken from each of 16 randomly selected crabs. Subsamples were weighed and then counted under a dissecting microscope, and the three values were averaged. I then used linear regression to determine the relationship between egg count and egg mass. The slope of this relationship indicates the average dry mass of a single egg. I used this average value, together with the clutch mass, to estimate the total number of eggs in each clutch. I also used the SE in egg mass to estimate the amount of variation in clutch size that
Experimental test of link between diet and reproductive effort I examined the influence of diet variation on reproductive investment by C. maenas in an experiment that capitalized on the ability to induce reproduction in crustaceans by removing the eyestalk which contains a gland that secretes gonad-inhibiting hormone (Adiyodi and Adiyodi 1970). Unilateral eyestalk ablation generally stimulates reproduction in adult C. maenas within 2–3 weeks (Bliss 1966; Griffen et al. 2011; and reviewed in Stowasser 2008). I collected 40 mature female crabs (33.6–48.9 mm carapace width) during May 2009, which, in the Gulf of Maine, is the height of the reproductive season for C. maenas (Berrill 1982). I selected only crabs with a green carapace, indicating that they had recently molted (Styrishave et al. 2004) and were thus not already engaged in vitellogenesis (Reid et al. 1994). Crabs were transported to the University of South Carolina where they were placed into individual chambers within a recirculating tank filled with seawater collected from the local South Carolina coast. The water temperature was maintained at 12–13 °C throughout the duration of the experiment, mimicking spring/summer water temperatures commonly found on the New Hampshire coast where the crabs were collected. Each chamber received a continual supply of water at a rate of ~3 L/h. Consumers commonly compensate for low-quality diets by increasing the amount consumed (Simpson and Simpson 1990). Thus, variation in diet quantity can compensate to some extent for variation in diet quality. I incorporated both types of variation in my experiment. I randomly assigned each crab one of 20 experimental diets that varied both the amount of food present (four food levels: 0.2, 0.4, 0.8, and
13
124
1.6 g every other day) and the proportion of that food that was animal tissue or algae (five levels: 1.0:0, 0.75:0.25, 0.5:0.5, 0.25:0.75, 0:1.0 animal tissue:algae). Two crabs were assigned to each of these experimental diets; however, each crab had its own unique consumption of plant and animal material due to individual variation in consumption over the 6-week experiment. Thus, even when different individuals have similar food resources available, differences in food selection leads to individual diet variation that may result in differential reproductive success. The statistical analyses described below appropriately accounted for this individual variation in dietary intake. Mussels (Mytilus edulis) are the most common animal food consumed by C. maenas at the site where these crabs were collected (Griffen et al. 2008). However, mussel tissue in preliminary experiments degraded quickly, causing unacceptably high nonconsumptive food loss. I therefore used fish muscle tissue (tilapia) as an animal food source in this experiment because tilapia did not disintegrate over the course of a single feeding period. However, it should be noted that tilapia is a higher quality food than mussel tissue, being higher in both percent nitrogen (tilapia = 12.5 % N and mussels = 9.2 % N, determined using Perkin Elmer 2400 CHNSO Elemental Analyzer, PE 2400) as well as in energy content (tilapia = 22.59 kJ/g and mussels = 19.71 kJ/g, determined using calorimetry), and may also differ from mussels in other ways (i.e., other nutrients, assimilability, etc.). Chondrus crispus is the most common algal food consumed by Carcinus maenas in the Gulf of Maine (Griffen et al. 2008) and was used as the algal food source in this experiment. C. crispus is a much lower quality food than either mussels or tilapia, and contains only 0.4–1.4 % N (Chopin and Floch 1992) and 8.37 kJ/g (determined from calorimetry). I collected C. crispus from Odiorne State Park, New Hampshire at the same time that I collected experimental crabs. Each crab was fed a constant experimental diet every other day and uneaten food was removed after 24 h. Crabs generally shred their food upon consumption, and experimental chambers were designed to collect small pieces of uneaten food. These were collected by filtration and the dry weight of uneaten food was determined at each feeding cycle. In addition, food mass controls, consisting of 1 g of fish and algae submerged in mesh bags within the experimental recirculating tank, were used to account for nonconsumptive changes in food during each feeding period (Peterson and Renaud 1989). Crabs were maintained under these conditions for 6 weeks, after which I performed unilateral eyestalk ablation to stimulate reproduction. Crabs were then fed for an additional 2 weeks while vitellogenesis proceeded. At the conclusion of that time, I dissected each crab and removed the ovaries and eggs. These were dried together at 70 °C for 72 h. I then divided the ovary dry mass
13
Oecologia (2014) 174:121–130
by the dry mass of the remainder of the crab yielding a sizeindependent index of reproductive effort [i.e., the gonadosomatic index (GSI), following Kyomo (1988)]. Statistically, the data were analyzed as follows. I first examined how the amount and type of food offered influenced what was consumed. Specifically, I used separate linear models to examine how the average daily amount of fish and algae consumed varied with the total amount of food offered daily (g) and the proportion of that offered food that was animal tissue as opposed to plant material. Next, I used a linear model to determine how GSI varied with the amount of fish and algae consumed per unit biomass of crab. (While maximum potential consumption could be constrained by controlling the amount of food offered, actual consumption within these limits could not be precisely controlled due to differences in feeding habits or preferences of individual crabs. I capitalized on this fact by using a regression analysis with actual food consumption, rather than experimental food treatment, as the predictor variable. This simultaneously eliminated the need for replication of the large number of food treatments and allowed for differences in individual diet selection.) One crab died during the experiment and was excluded from analyses. Two other crabs were extreme outliers and therefore apparently did not respond to eyestalk ablation with reproduction. I repeated the analyses both with and without these crabs. Exclusion of these crabs did not qualitatively alter the results of this experiment and so results are only presented with these two crabs removed. Finally, the analyses described in the preceding section supported the conversion of ovary mass directly into the number of egg that would be produced. I made this calculation as follows:
No. Eggs =
(GSI − 0.02) × massbody massegg
(1)
Previous work indicates that the average mass of the ovary in nonreproductive crabs is 2 % of the body weight (Griffen et al. 2011). For this reason, 0.02 is subtracted from GSI in the above equation, reflecting the mass of a vitellogenic ovary that is not converted to eggs. I compared changes in the calculated number of eggs produced as a function of the weight-specific amount of fish and algae consumed by crabs using a linear model. Finally, fecundity in crabs is size dependent (Hines 1982)—for this reason I examined how diet may influence future reproductive effort via its effect on growth. Crabs grow in increments during molts, and no crabs molted during this experiment. So I examined changes in soft body tissue that may lead to differences in size changes at molting. I first determined the allometric relationship between carapace width and dry mass for all crabs in the experiment and then used the residual from this relationship as an estimate
125
Oecologia (2014) 174:121–130
Results
12 11
15N (‰)
13
14
of crab soft tissue growth—crabs with positive residuals would have grown more than average, whereas crabs with negative residuals would have had less than average growth. I then used a linear model to compare crab growth as a function of average daily fish and algal consumption.
Variation in long‑term diets Stable nitrogen and carbon isotopes indicated abundant variation in long-term diets as well (Fig. 2). Assuming the commonly used trophic fractionation rate of δ15N of 3.4 ‰
-20
-19
-18
-17
-16
-15
-14
13C (‰) Fig. 2 Variation in long-term diets of adult female C. maenas as illustrated by stable isotope ratios of nitrogen (δ15N) (‰; variation across trophic levels) and carbon (δ13C) (‰; variation within trophic levels)
(Post 2002), these data represent approximately one full trophic level of long-term diet variation. In addition, the broad range of δ13C values (Fig. 2) indicates considerable diet variation within trophic levels as well. Variation in fecundity
15
20
25
The mass of eggs increased linearly with the number of eggs present in the subsample (linear model, p
