Journal of Paleolimnology

, Volume 49, Issue 2, pp 253–266 | Cite as

Cladoceran remains reveal presence of a keystone size-selective planktivore

Original paper

Abstract

We tested the use of cladoceran remains as a proxy for the presence and life history type of alewife (Alosa pseudoharengus) from pre-colonial times to present in a group of coastal lakes in southern New England. Alewife are a keystone predator that structure the zooplankton community through strong predation on large-bodied zooplankton species, which releases small zooplankton species, such as Bosmina spp., from competition and predation pressure. In southern New England there are lakes without alewife, lakes with anadromous alewife that only reside in lakes during the summer, and lakes with landlocked alewife that reside in lakes year-round. The entire zooplankton community of these lakes is structured differently based on the presence and type of alewife they contain. We examined differences in the morphology of Bosmina spp. from sediment core samples and contemporary zooplankton samples between lakes with different types of alewife. We found that there were significant differences in the morphology of Bosmina spp. between lakes with and without alewife. We also used discriminant analysis on the morphology of Bosmina spp. to classify lakes in terms of alewife presence and alewife type. We found that the morphology of Bosmina spp. can serve as a useful proxy for detecting the presence, but not the life history type of alewife from paleoecological and contemporary inferences.

Keywords

Cladoceran remains Bosmina spp. Size-selective predation Alewife 

Introduction

Understanding historical ecosystem structure and function is essential for predicting future ecosystem dynamics and restoring damaged ecosystems. Many pressing environmental problems, from overharvesting to habitat destruction, are the result of long-term anthropogenic disturbances (Jackson 2001; Jackson et al. 2001). To understand how systems functioned prior to major anthropogenic disturbances, ecologists have turned to the historical record to determine past presence and abundance of species, and how changes in species have altered ecosystem structure and function (Foster et al. 2003; Lotze and Milewski 2004).

In New England, USA anadromous fish restoration has proved especially challenging because declines in anadromous fishes predate the modern ecological record by hundreds of years (Limburg and Waldman 2009; Hall et al. 2011). Prior to Euro-American dam construction, lakes throughout New England with connections to the ocean and without natural barriers, such as high waterfalls or large beaver dams, supported abundant runs of anadromous fishes, such as alewife (Alosa pseudoharengus). Alewife and other anadromous species have been in a decline since colonial dams started to block access to freshwater spawning habitat (Limburg and Waldman 2009; Hall et al. 2011) and that decline has accelerated towards present (Atlantic States Marine Fisheries Council 2009). This recent acceleration in anadromous fish declines has prompted managers to list alewife as a federal species of concern (Atlantic States Marine Fisheries Council 2009) and has prompted renewed attention to anadromous fish restoration efforts throughout the region. However, because initial alewife declines predate the modern ecological record by hundreds of years, the written record alone cannot be used to determine the historical distribution of alewife and other anadromous fishes. In areas without reliable written records or anecdotal accounts, the paleoecological record is the one of the few tools available for determining if past anadromous fish populations were present.

Past anadromous fish populations can be detected based on the trophic and non-trophic effects that fishes have on ecosystem dynamics. Alewife and other fishes influence community structure through predation on smaller fishes and zooplankton (Brooks and Dodson 1965; Dodson 1974) and they recycle and transport nutrients (Vanni et al. 1997). They also affect community structure by mediating competition and predation in lower trophic levels through size-selective predation (Brooks and Dodson 1965; Dodson 1974) and their identity, presence, and abundance can alter lake species composition (Brooks and Dodson 1965; Rudstam et al. 1993; Post et al. 2008). Cladoceran zooplankton remains have been used to infer past food web structure and can reveal whether fishes or invertebrates dominate zooplanktivory (Kitchell and Kitchell 1980; Salo et al. 1989; Alexander and Hotchkiss 2010; Korosi et al. 2010). Fishes can also affect nutrient cycles by recycling limiting nutrients into bioavailable forms that primary producers can take up and use (Schindler 1992; Vanni et al. 1997; Vanni 2002) and by transporting nutrients between ecosystems (Kline et al. 1990; Naiman et al. 2002; Vanni 2002; Schindler et al. 2003; Walters et al. 2009; Flecker et al. 2010).

When strong size-selective planktivorous fishes, such as alewife, are the dominant zooplanktivores, they release Bosmina spp. and other small zooplankton species from predation (Dodson 1974; Carpenter et al. 1985; Post et al. 1995). Bosmina spp. and other cladoceran species exhibit phenotypic plasticity in their response to predation and grow energetically expensive defense morphology only when threatened by cues from predators (Kerfoot 1987; Dodson 1988; Dodson and Havel 1988). Size-selective planktivorous fishes reduce the threat of invertebrate predation on Bosmina spp. by preying upon invertebrate predators (Dodson 1974; Carpenter et al. 1985; Post et al. 1995). When size-selective planktivorous fishes dominate planktivory, Bosmina spp. and other small zooplankton species do not need to allocate as much energy to growing defensive morphological structures, such as longer mucrones and antennules (Post et al. 1995; Alexander and Hotchkiss 2010). As a result, the morphology of Bosmina spp. can serve as tool for reconstructing lake community structure and determining if fishes or invertebrates dominated planktivory (Kitchell and Kitchell 1980).

We tested the use of cladoceran remains as a proxy with which to reconstruct the ecological history of alewife, a size-selective planktivore in coastal southern New England. Alewife function as a keystone species (Paine 1969; Power et al. 1996) by mediating competition and predation through strong size-selective predation on large zooplankton species (Brooks and Dodson 1965; Post et al. 2008). The best molecular genetic evidence suggests that several populations of alewife in southern Connecticut became genetically isolated from other anadromous alewife populations approximately 270–520 years ago, but these molecular estimates extend up to 5000 years before present (Palkovacs et al. 2008). Colonial dam construction, which began in the region approximately 350 years before present is suspected to have caused landlocking, but Native American fishing weir and beaver dam construction prior to Euro-American settlement may also have prevented populations of formerly anadromous alewife from accessing the ocean.

Connecticut’s landlocked alewife populations now live in lakes year-round and have evolved to be genetically, morphologically, and ecologically distinct from anadromous alewife (Palkovacs et al. 2008; Post et al. 2008). In lakes without any alewife, large zooplankton dominate the community year-round (Brooks and Dodson 1965; Post et al. 2008) while in lakes with populations of landlocked alewife large zooplankton are reduced to very low densities due to constant intense size-selective predation pressure from alewife (Brooks and Dodson 1965; Post et al. 2008). In lakes with anadromous alewife runs, large zooplankton species are common in the water column during the late fall to early spring when alewives are not present, but are typically eliminated from the water column by intense predation by young-of-the-year alewife by mid June (Post et al. 2008). Three distinct lake communities have formed in the region based on the presence and type of alewife: a community with large zooplankton species year-round in lakes without alewife, a community with a seasonal cycle in zooplankton species composition in lakes with anadromous alewife, and a community with small zooplankton species year-round when landlocked alewife are present (Brooks and Dodson 1965; Post et al. 2008).

In addition to the direct effects that the presence of and type of alewife (landlocked or anadromous) in a lake have on zooplankton community composition, we expect the presence and type of alewife to have an indirect effect on the defense morphology of small zooplankton, such as Bosmina spp. (Brooks and Dodson 1965; Post et al. 2008). Alewife prey intensely upon large zooplankton predators, such as Mesocylops edax, (Brooks and Dodson 1965; Post et al. 2008) which prey upon Bosmina spp., potentially releasing Bosmina spp. from predation pressure and the need to grow long mucrones (tail spines) for anti-predator defense. We expect Bosmina spp. to have consistently short mucrones in lakes with landlocked alewife, consistently long mucrones in lakes without alewife, and variable mucrones lengths in lakes with anadromous alewife. In lakes with anadromous alewife, we expect mucrones to be longest in the spring before adult alewife have arrived to spawn and young-of-year alewife have hatched and to be shortest in the fall after young-of-year alewife have removed most large invertebrate predators from the water column through their intense size-selective planktivory.

Here, we asked if the morphology Bosmina spp. from sediment cores and recent zooplankton hauls could be used as a proxy to discriminate between lakes with landlocked alewife and lakes anadromous alewife and between lakes with and without alewife. We conducted a regional historical study to determine when and where anadromous and landlocked alewife were present in four coastal lakes in southern New England. We used Bosmina spp. from recent zooplankton hauls from thirteen lakes where the presence and type of alewife is known to determine how the presence and type of alewife (anadromous or landlocked) affect the morphology of Bosmina spp. We then cored the lakes from our environmental history study and used the morphology of Bosmina spp. from sediment cores to test our estimates of past alewife type and presence. In doing this we have three goals: first, to determine if Bosmina spp. morphology can be used as a proxy for the presence and type of size-selective planktivorous fishes in a lake; second, to determine if coastal lakes in southern New England that now have populations of landlocked alewife, formerly supported runs of anadromous alewife prior to Euro-American dam construction; and third, to determine how ecological communities in the region’s lakes changed as alewife became landlocked.

Study sites

We conducted the paleoecological portion of our study in four kettle pond lakes in south central Connecticut: Bride Lake in East Lyme, CT (41º19′39″N–72º14′15″W), Linsley Pond in North Branford, CT (41º19′03″N–72º47′05″W), Rogers Lake in Old Lyme, CT (41º21′01′N–72º18′05″W) and Uncas Pond in Lyme, CT (41º22′27″N–72º18′48″W) (Fig. 1). We conducted the contemporary portion of our study in thirteen lakes in south central Connecticut, which are described in Post et al. (2008). All of the lakes we studied differ in their accessibility to alewife at present and in their histories of anadromous fish passage. The four lakes that we cored all lie within 15 km of the Long Island Sound and all of our study lakes are similar in terms of climate, surface temperature, (Post et al. 2008) and dissolved organic carbon levels (P. Raymond, personal communication). Most of these lakes are not undergoing major state changes and there are no significant differences in phosphorus levels, algal-macrophyte dominance, or other environmental factors in lakes based on different predation regimes (Post et al. 2008).
Fig. 1

Map of Fishways, Dams, and Study Lakes in South Central Connecticut. Study lakes are in black, asterisk are dams without fishways, and plus symbol are dams with fishways

Methods

Environmental history study

We used four kettle pond lakes in south central Connecticut, Bride Lake, Linsley Pond, Rogers Lake, and Uncas Pond, as case studies to understand the ecological history of the region. For each of these four lakes, we conducted an environmental history survey and created a timeline of the past presence and type (anadromous or landlocked) of alewife in the watershed. Bride Lake currently has a run of anadromous alewife, Rogers Lake and Uncas Pond currently have populations of landlocked alewife, and Linsley Pond currently does not have any alewife (Post et al. 2008). We searched state and municipal land records and maps for evidence of mill and dam construction in our four study watersheds from the mid-1600s to present to determine when dams would have first blocked anadromous fish passage and created populations of landlocked alewife. We also used published literature on the system (Brooks and Dodson 1965; Brugam 1978; Palkovacs et al. 2008; Post et al. 2008) to determine if populations of both anadromous and landlocked alewife were still extant and if not, when they became extinct. We used state stocking records (Connecticut Department of Environmental Protection 2011) and molecular genetic information (Palkovacs et al. 2008) to determine if extant populations of landlocked alewife in the region were recently stocked (within the past 50 years) or if they were older, independently derived populations that diverged from anadromous populations around the time of European settlement in the region 350 years before present.

Based on our environmental history survey, we classified past and present lake states in terms of alewife presence and lake type (Table 1). We treated lakes with existing or past populations of anadromous or landlocked alewife as presence. We treated lakes without existing or past populations of anadromous or landlocked alewife as no alewife. We treated lakes with existing or past connections to the ocean and without natural (high waterfalls) or human-created (dams) barriers as anadromous prior to estimated date of dam construction. We treated lakes as landlocked if they had both past connections to the ocean and dams or if they had natural barriers to the ocean and were recently stocked (last 50-60 years) with landlocked alewife.
Table 1

Lake histories

Lake

Time period

Alewife state

Bride Lakea

Post-glaciation to present

Anadromous alewife

Linsley Pondb

Posts-glaciation to 1700s

Anadromous alewife

Linsley Pondb,c

1700s to 1960s

Landlocked alewife

Linsley Ponda,b

1960s to present

No alewife

Rogers Lakec

Post-glaciation to 1670s

Anadromous alewife

Rogers Lakea,c

1670s to present

Landlocked alewife

Uncas Pondd

Post-glaciation to 1960s

No alewife

Uncas Pondd,e

1960s to present

Landlocked alewife

aPost et al. (2008)

bBrugam (1978)

cBrooks and Dodson (1965)

dPalkovacs et al. (2008)

eCT DEP 2011

Coring and dating

We cored four glacial kettle pond lakes in south central Connecticut: Bride Lake, Linsley Pond, Rogers Lake, and Uncas Pond. We used a Livingston improved square rod piston corer to core Bride Lake in August 2009 and Rogers Lake, Uncas Pond, and Linsley Pond in June 2010. In Bride Lake, we obtained a 284-cm core in 8.1 m of water. In Linsley Pond, we obtained a 332-cm core 13.7 m of water. In Rogers Lake, we obtained a 289-cm core in 10.5 m of water. In Uncas Pond, we obtained a 276-cm core in 10.9 m of water. We extruded the cores into split PVC pipes, wrapped them in plastic wrap and kept them refrigerated after being transported back to the lab.

We dated Bride Lake, Rogers Lake, and Uncas Pond with 210Pb based on gamma ray detection in the Yale Environmental Chemistry Lab, New Haven, CT, USA. Prior to dating, samples were dried for over 12 h at 70 °C and then ground into a homogenous powder. We calculated total and excess 210Pb activity in our cores following the methods of Cutshall et al. (1983) and Benoit and Rozan (2001). We dated nine sections of core from Bride Lake and Rogers Lake and ten sections of core from Uncas Pond. We allowed all samples to equilibrate for ≥14 days prior to gamma detector counting. We evaluated radionuclide self-absorption and efficiency corrections with 137Cs, 226Ra, and 210Pb standards following the methods of Benoit and Rozan (2001). For each core, we checked our 210Pb curves against 137Cs and 226Ra curves to determine if our supported 210Pb activities were suitable to for calculating dates.

We used the constant rate of sedimentation (CRS) model to date cores with 210Pb following the methods of Appleby and Oldfield (1978), assuming constant atmospheric 210Pb deposition over time. We chose the CRS model because our lakes were all disturbed and 210Pb activities showed non-exponential decay with depth. For depths beyond the range of 210Pb and 137Cs dating in Bride Lake and Uncas Pond, we used sedimentation rates from the deepest portions of core dated with 210Pb and assumed that sedimentation rates beyond the 210Pb dating range remained relatively constant over time.

To date our core from Linsley Pond, we used sedimentation rates and 210Pb dates from a previous study of sedimentation rates in the lake (Benoit and Rozan 2001). We cored the exact same buoy-marked deep spot in the Linsley Pond basin as Benoit and Rozan (2001). We matched up our core with the top of theirs by assuming the sediment water interface at the top of our core to be present and then using the sedimentation rate at top of their core to determine where their core matched up with ours. In calculating where their core matched up with ours, we assumed that sedimentation rates in the 14 years since Benoit and Rozan cored Linsley Pond in 1996 were approximately the same as those that they found at the top of their core.

We also obtained sedimentation rates from below 150 cm in Linsley Pond and below 60 cm in Rogers Lake based on 14C dates determined at the National Ocean Sciences Accelerator Mass Spectrometry Facility, Woods Hole, MA, USA dates from the two lakes. We assumed that pre-Euro-American sedimentation rates remained relatively constant over time.

Bosmina morphology

We analyzed the morphology of Bosmina spp. by lake type and alewife presence because the type of lake and presence of alewife have major impacts on zooplankton community structure in our lakes (Brooks and Dodson 1965; Post et al. 2008), which may also influence Bosmina spp. morphology. We first examined the morphology of contemporary Bosmina spp. individuals to determine if patterns in Bosmina spp. morphology reflect overall patterns in lake community structure (Post et al. 2008). We arbitrarily selected Bosmina spp. from contemporary zooplankton samples collected in 2004 and 2005 from a set of thirteen lakes across south-central Connecticut, which differ in their accessibility to alewife and histories of anadromous fish passage, and are described in Post et al. (2008). Three of these lakes, Bride Lake, Dodge Pond and Gorton Pond, have annual runs of anadromous alewife. Five of these lakes, Avery Pond, Pattagansett Lake, Quonnipaug Lake, Rogers Lake, and Uncas Lake, have populations of resident landlocked alewife. Five of these lakes, Black Pond, Gardner Lake, Green Falls Lake, Hayward Lake, and Wyassup Lake, have neither anadromous alewife runs nor landlocked alewife populations. To minimize the possibility of detecting differences in Bosmina spp. morphology due to seasonality alone we examined samples from throughout the season (late March to early November) per lake.

We tested for the past presence and type of alewife in each lake based on the morphology of Bosmina spp. (Bosmina longirostris and Eubosmina tubicen) remains from sediment core samples. For our analysis of Bosmina spp. morphology from sediment cores, we sampled sediment cores at three core sections per lake. We took sediment samples at core depths of 4, 150, 280 cm for Linsley Pond; depths of 12, 60, 237 cm for Rogers Lake; and depths of 5, 50 and 100 cm for Uncas Lake and Bride Lake. The 280 and 150 cm core samples from Linsley Pond correspond to periods when the lake would have been accessible to anadromous alewife while the 4 cm sample corresponds to a very recent period after landlocked alewife became extirpated from the lake. The 237 cm sample from Rogers Lake corresponds to a period when the lake would have been accessible to anadromous alewife while the 60 and 12 cm samples correspond to periods during which only landlocked alewife were present in the lake. The 100 and 50 cm core samples from Uncas Pond correspond to periods before alewife were stocked and the 5 cm sample corresponds to a more recent period after landlocked alewife had been stocked. The core samples from Bride Lake all represent periods when the lake was accessible to and had runs of anadromous alewife.

We followed the Korhola and Rautio (2001) procedure for preparing cladoceran remains by boiling sediment in KOH to break down organic material. We boiled approximately 2.5 mL of sediment per depth for 45 min at 70–80 °C. We then rinsed the sample through a 53-μm mesh sieve, first checking to ensure that material passing through the mesh did not contain any cladoceran remains. We stored the samples in a refrigerator in 15-mL Biorad tubes with several drops of 75 % ethanol as a preservative and several drops of safranin-glycerin solution as coloring. We pipetted 100 μL of sample onto each slide and covered the slide with a 24 × 50-mm coverslip. We made permanent slides by painting nail varnish around the edges of the coverslips.

For both sediment samples and contemporary samples, we measured the mucro length and the body length of 20 Bosmina spp. individuals per sample on a compound microscope at 20×. We used the distance from the base to the tip of the mucro as mucro length (Fig. 2). We used the distance from the posterior margin of the carapace to the base of the head as body length (Fig. 2) because the head and antennae had detached from the rest of the body making a measurement of total body length impossible in most individuals from sediment samples. We also included the ratio of mucro length to body length in our analyses in order to remove the potential influence of body size on mucro length.
Fig. 2

Morphological Measurements of Bosmina spp. (C.W. Twining)

We analyzed patterns in Bosmina spp. morphology from both contemporary and sediment core samples using analysis of variance and discriminant analysis. We performed ANOVA by lake type (no alewife, anadromous alewife, or landlocked alewife) on Bosmina spp. mucro length, body length, and the ratio of mucro length to body length from contemporary zooplankton samples and sediment core samples. To avoid pseudoreplication, we used lake means from contemporary zooplankton samples and lake-core depth means from sediment core samples. Because seasonal changes occur in the zooplankton community (Post et al. 2008) there may be seasonal changes in Bosmina spp. morphology, which could influence our results through time averaging. Therefore, we also performed ANOVA by lake type, season, and the interaction of lake type and season on Bosmina spp. mucro length, body length, and the ratio of mucro length to body length from contemporary zooplankton samples. For samples from sediment cores, we created estimates of past alewife presence and lake type based on our 14C and 210Pb dates and estimated dates of dam construction, alewife extirpation, and alewife stocking from our historical study of regional dam construction, and previous studies in these lakes (Brooks and Dodson 1965; Brugam 1978; Palkovacs et al. 2008; Post et al. 2008). To avoid pseudoreplication in our ANOVAs, we used lake-season means. In our ANOVAs we treated lake type, season, and the interaction of lake type and season as factors and performed Tukey post hoc tests on all analyses of variance. We performed analysis of variance in R (R version 2.10.1).

We performed discriminant analysis by lake type and alewife presence on the morphology of Bosmina spp. because there are major differences in zooplankton community structure (Brooks and Dodson 1965; Post et al. 2008), which may also influence the morphology of Bosmina spp. and may allow us to discriminate between lakes using the morphology of Bosmina spp. We performed discriminate analysis on Bosmina spp. mucro length, body length, and the ratio of mucro length to body length to classify lakes by lake type and alewife presence. We ran discriminant analysis on all of our data and on the top 25 % of data, which captured only the largest size class of Bosmina spp. individuals. We performed discriminant analysis in SPSS (SPSS version 16). In our discriminant analyses, we used within groups covariance matrices and assigned equal prior probabilities to all groups.

Results

Lake history and geochronology

Bride Lake has a current and historical run of anadromous alewife. There has never been a major anthropogenic barrier to fish passage on the outlet stream (Palkovacs et al. 2008; Post et al. 2008; Table 1) and all thus of our sediment core samples from Bride Lake represent periods when anadromous alewife were present in the lake (Fig. 3a, ESM Fig. 1a).
Fig. 3

Pb-210 Core date versus core depth with error in a Bride Lake, b Rogers Lake, and c Uncas Pond. Slopes used to calculate dates from core depth and R2 values are included parentheses

Linsley Pond had a population of landlocked alewife in 1950s (Brooks and Dodson 1965), prior to extensive state alewife stocking efforts in the 1960s (Connecticut Department of Energy and Environmental Protection 2011). By the mid-1970s alewife were extirpated from Linsley Pond entirely (Brugam 1978) and the lake does not contain any alewife today (Post et al. 2008; Table 1). The outlet stream from Linsley Pond was dammed by 1758, which is very close to the 270–520 year best estimate for evolutionary divergence time for other landlocked alewife populations across the region (Palkovacs et al. 2008). This suggests that Linsley Pond’s former population of landlocked alewife also likely evolved from a previous anadromous run. In Linsley Pond, 210Pb and 14C dating indicated that our sediment core samples from 150 to 280 cm represented periods when anadromous alewife were present in the lake while our sample from 4 cm represented a period when no alewife were present in the lake (Benoit and Rozan 2001).

Rogers Lake currently has a population of landlocked alewife and population genetic studies suggest this population diverged from runs of anadromous alewife 270–522 years before present (Palkovacs et al. 2008; Table 1). The outlet stream from Rogers Lake was dammed by 1672 thereafter preventing anadromous fish passage. This date falls well within the 270–520 year best estimate for evolutionary divergence time for landlocked alewife across the region and suggests that the current population of landlocked alewife in the lake became landlocked when the first dam in the watershed went in. In Rogers Lake, 210Pb and 14C dating indicated that our sediment core sample from 237 cm represented a period when anadromous alewife were present in the lake while our samples from 12 to 60 cm represented periods when landlocked alewife were present in the lake (Fig. 3b, ESM Fig. 1b). We did not calculate sedimentation rates at the top of our Rogers Lake core because there appeared to be mixing within the top 20 cm of sediment.

Uncas Lake currently has a population of landlocked alewife, which the state stocked in the 1960s (Post et al. 2008; Connecticut Department of Environmental Protection 2011; Table 1). Population genetics suggest that the alewife in Uncas Pond were stocked from a population of landlocked alewife from Amos Lake in Preston, CT. The landlocked alewife population in Amos Lake diverged from anadromous ancestors approximately 270–520 years before present (Palkovacs et al. 2008). Prior to stocking, Uncas Pond would not have supported runs of alewife or other anadromous fishes because it lies above a steep natural waterfall. In Uncas Pond, 210Pb dating suggested that our sediment core samples from 50 to 100 cm are representative of periods well before the lake was stocked with alewife, while our sample from 5 cm represent the post-alewife stocking period (Fig. 3c, ESM Fig. 1c).

Bosmina morphology

ANOVA of Bosmina spp. morphology from contemporary zooplankton samples and sediment core samples by lake type suggested that there were significant differences in the morphology of Bosmina spp. between lakes with and without alewife. Lake type was significant for contemporary Bosmina spp. mucro length (F2,10 = 9.73, p < 0.01), body length (F2,10 = 4.95, p < 0.05), and the ratio of mucro length to body length (F2,10 = 9.39, p < 0.01) and sediment core Bosmina spp. mucro length (F2,9 = 5.04, p < 0.05), body length (F2,9 = 8.54, p < 0.01) and the ratio of mucro length to body length (F2,9 = 4.66 and p < 0.05). Post-hoc Tukey tests of our analysis of contemporary samples by lake type suggested that there were significant differences between lakes with landlocked or anadromous alewife and lakes without alewife for mucro length (p < 0.05 between anadromous and no alewife lakes and p < 0.01 between landlocked and no alewife lakes), marginally significant differences in body length (p < 0.10 between anadromous and no alewife lake and p < 0.10 between landlocked and no alewife lakes), and significant differences in the ratio of mucro length to body length (p < 0.05 between anadromous and no alewife lakes and p < 0.01 between landlocked and no alewife). Post-hoc Tukey tests of our analysis of sediment core samples by lake type indicated that there were significant differences between lake states without alewife and lake states with anadromous alewife for mucro length (p < 0.05 between anadromous and no alewife lake states), body length, (p < 0.01 between anadromous and no alewife lake states), and the ratio of mucro length to body length (p < 0.05 between anadromous and no alewife lake states), but that there were no significant differences between lake states without alewife and lake states with landlocked alewife or between lake states with anadromous alewife and lake states with landlocked alewife.

For mucro length from contemporary samples, we found significant effects of lake type (F2,26 = 14.11, p < 0.001) and the interaction of lake type and season (F4,26 = 3.76 and p < 0.05). For body length, we found significant effects of lake type (F2,26 = 6.62 and p < 0.01) and season (F2,26 = 13.97 and p < 0.001). For the ratio of mucro length to body length, we found significant effects of lake type (F2,26 = 15.06 and p < 0.001) and the interaction of lake type and season (F4,26 = 4.24 and p < 0.01). Mucro length and the ratio of mucro length to body length started out similar in all lakes at the beginning of the season, but mucro length in lakes without alewife diverged from lakes with anadromous alewife and landlocked alewife as the season progressed (Fig. 5a, c, Electronic Supplementary Material (ESM) 7, ESM 11). In lakes without alewife, mucro length and the ratio of mucro length to body length increased as the season progressed (Fig. 5a, c, ESM 7, ESM 11, ESM 13). In lakes with anadromous alewife, mucro length and the ratio of mucro length to body length started out slightly higher than in lakes with landlocked alewife, but the two lake types converged by the end of the season as mucro length and the ratio of mucro length to body length decreased in both (Fig. 5a, c, ESM 7, ESM 11, ESM 13). Body length decreased in all lake types as the season progressed, but the largest decrease was in lakes with anadromous alewife and lakes with landlocked alewife (Fig. 5b, ESM 7, ESM 12). Lakes with anadromous alewife and lakes with landlocked alewife started out different, but converged by the end of the season as body length decreased in both, while body length stayed high in lakes without alewife (Fig. 5, ESM 7, ESM 11, ESM 12, ESM 13).

Discriminant analysis of Bosmina spp. mucro length, body length, and the ratio of mucro length to body length from contemporary zooplankton samples and core samples indicated that there were meaningful differences in Bosmina spp. morphology between lakes with and without alewife, but that there were no differences in Bosmina spp. morphology between lakes with landlocked alewife and anadromous alewife. Bosmina spp. individuals from lakes without alewife had longer mucrones and longer body length than individuals from lakes with either landlocked or anadromous alewife (Fig. 4). Discriminant analysis between lakes based on both presence and type of alewife (no alewife, anadromous alewife and landlocked alewife) yielded poor classification results: classification predicted only 48.5 and 59.6 % of original grouped cases correctly for contemporary and sediment core samples respectively (Table 2). Discriminant analysis between lakes based on alewife presence alone yielded much improved classification results: classification predicted 76.9 and 80.4 % of original grouped cases correctly for contemporary and sediment core samples respectively (Table 3).
Fig. 4

Mean ± Standard Error a mucro length, b body length, and c ratio of mucro length to body length from recent tows taken from 13 lakes in Connecticut and d mucro length, e body length, and f ratio of mucro length to body length from sediment cores from Bride Lake, Linsley Pond, Rogers Lake, and Uncas Pond

Table 2

Discriminant analysis by lake type

Original group

Predicted group

No alewife (%)

Anadromous (%)

Landlocked (%)

a. Recent tows

 No alewife

59.0

17.2

23.8

 Anadromous

21.5

36.5

42.1

 Landlocked

16.6

32.1

51.3

b. Cores

 No alewife

55.0

11.7

33.3

 Anadromous

3.3

76.7

20.0

 Landlocked

35.0

45.0

30.0

Table 3

Discriminant analysis by alewife presence

Original group

Predicted group

No Alewife (%)

Alewife (%)

a. Recent tows

 No alewife

59.8

40.2

 Alewife

18.7

81.3

b. Cores

 No alewife

65.0

35.0

 Alewife

14.4

85.6

Discriminant analysis on the largest size class of Bosmina spp. individuals from contemporary and sediment core samples yielded slightly improved results for both alewife type and alewife presence. When only the top 25 % of Bosmina spp. morphological data was used there were more meaningful differences in Bosmina spp. morphology between lakes with and without alewife, but differences between lake states with landlocked and anadromous alewife were still not significant. Discriminant analysis of the top 25 % Bosmina spp. morphological data between lakes based on the presence and type of alewife yielded marginally improved classification results compared with the analysis with all data: classification predicted only 55.6 and 65.0 % of original grouped cases correctly for contemporary samples and sediment core samples respectively (ESM 8). Discriminant analysis of the top 25 % Bosmina spp. morphological data between lakes based on the alewife presence also yielded slightly improved classification results compared to the analysis with all data: classification predicted 87.2 and 81.7 % of original grouped cases correctly for contemporary samples and sediment core samples respectively (ESM 8).

Discussion

Previous researchers have used cladoceran remains to make inferences about past food web structure, such as the abundance of planktivorous fishes and invertebrates (Kitchell and Kitchell 1980; Salo et al. 1989; Alexander and Hotchkiss 2010; Korosi et al. 2010). Here, we used the remains of Bosmina spp., a small herbivorous cladoceran, from lake sediment cores to reconstruct the presence of a keystone size-selective planktivorous fish, the alewife. We asked if the morphology of Bosmina spp. could be used as a proxy to discriminate between lakes with and without a keystone size-selective planktivore as well as between lakes with different types of a keystone size-selective planktivore (anadromous alewife and landlocked alewife). We found that the morphology of Bosmina spp. could be used to discriminate between lakes with alewife and lakes without alewife, but that it did not provide the detail necessary to discriminate between lakes with different types of keystone size-selective planktivores. Our results suggest that the morphology of Bosmina spp. can serve as suitable proxy for detecting the presence of strong size-selective planktivorous fishes.

We expected to find differences in Bosmina spp. morphology between lakes with anadromous alewife and lakes with landlocked alewife based on the major differences in the year-round zooplankton communities of these lakes (Brooks and Dodson 1965; Post et al. 2008). However, our results suggested that Bosmina spp. morphology does not provide the fine scale detail necessary to discriminate between past and present lake communities based on the type of size-selective planktivore they contain. Although lake type (no alewife, anadromous alewife, or landlocked alewife) was significant, post hoc tests confirmed that only differences between no alewife lakes and landlocked or anadromous alewife lakes were significant (ESM 4, ESM 5). In addition, our discriminant analysis only correctly classified lakes by the type of alewife they contain about 50–70 % of the time (Table 2) even when only the largest size class was analyzed (ESM 9, ESM 10). This suggests that the morphology of Bosmina spp. is not a suitable means of discerning the type of alewife in lakes and that Bosmina spp. morphology does not provide a reliable proxy with which to detect the identity of strong size-selective planktivorous fishes in the past.

Seasonality likely limited our ability to our discriminate between Bosmina spp. from lakes with anadromous alewife and lakes with landlocked alewife because the morphology of Bosmina spp. varies considerably across season in all of our lake types. We found significant differences in the morphology of Bosmina spp. by lake type, season, and the interaction of lake type and season (ESM 1, ESM 2, ESM 3, ESM 7). In the spring, the morphology of Bosmina spp. in lakes with anadromous alewife was most similar to the morphology of Bosmina spp. in lakes without alewives (Fig. 5, ESM 7, ESM 11, ESM 12, ESM 13). Before anadromous alewife arrive and their young hatch in lakes with anadromous alewife runs Bosmina spp. must defend themselves from large invertebrate predators in the same way as Bosmina spp. in lakes without alewife. By summer and fall the morphology of Bosmina spp. from lakes with anadromous alewife had become most similar to the morphology of Bosmina spp. from lakes with landlocked alewife (Fig. 5, ESM 7, ESM 11, ESM 12, ESM 13). After young-of-the-year anadromous alewife hatch and graze down the large invertebrate predators of Bosmina spp. in the summer and fall (Post et al. 2008), Bosmina spp. no longer need to invest as much of their energy in defensive morphology. Differences in lakes with and without alewife increased throughout the season (Fig. 5, ESM 7, ESM 11, ESM 12, ESM 13) because Bosmina spp. in lakes without alewife must invest in more defensive morphology as their invertebrate predators increase both in number and size.
Fig. 5

Mean ± Standard Error a mucro length, b body length, and c ratio of mucro length to body length from recent tows taken from 13 lakes across Connecticut

Our analyses of contemporary samples and sediment core samples are likely suffering from time averaging (Kowalewski et al. 1998; Krause et al. 2010) and are not picking up the significant seasonal shifts in the morphology Bosmina spp. that we observed when we grouped contemporary samples by season. Although we included contemporary zooplankton samples from throughout the year, most our samples were taken during the summer when both anadromous and landlocked alewife are present in the lakes we studied and have similar effects on the zooplankton community (Post et al. 2008) and on the morphology Bosmina spp. (Fig. 4, ESM 7, ESM 11, ESM 12, ESM 13). Sediment core samples represent a time integrated sample of zooplankton remains from throughout the entire year. In our sediment core samples, Bosmina spp. biomass and the contribution of Bosmina spp. remains to sediment cores likely peaks during the summer when the morphology of Bosmina spp. in lakes with landlocked alewife and lakes with anadromous alewife is most similar (Fig. 5, ESM 7, ESM 11, ESM 12, ESM 13). As a result Bosmina spp. from sediment core samples do not capture the seasonal differences in zooplankton communities in lakes with anadromous alewife and lakes with landlocked alewife, which are very different in the winter and early spring when only landlocked alewife are present in coastal lakes (Fig. 5, ESM 4, ESM 5, ESM 6).

While the results of this study suggest that Bosmina spp. morphology cannot be used to discriminate between lakes with different types of alewife, they do suggest that Bosmina spp. morphology can be used to discriminate between lakes with and without alewife. We expected to find substantial differences the morphology of Bosmina spp. from lakes with alewife and lakes without alewife based on the major differences between the communities of lakes with alewife and lakes without alewife (Brooks and Dodson 1965; Post et al. 2008). Our results show significant differences between lakes with alewife and lakes without alewife, and the discriminant analyses correctly classifying lakes by alewife presence in about 80 % of cases (Table 3). Our discriminant analyses of Bosmina spp. from contemporary zooplankton samples, for which alewife presence was known (Table 3; Fig. 4a–c), and from sediment core samples (Table 3; Fig. 4d–f), for which alewife presence was estimated based on our historical survey, were highly consistent. The consistency between our contemporary and paleoecological analyses suggests that the morphology of Bosmina spp. provides a reliable proxy with which to detect the presence of keystone size-selective planktivorous fishes on both contemporary and long-term time scales.

Our results provide evidence that alewife were historically native in lakes with past connections to the ocean, including lakes that currently contain populations of landlocked alewife. In the past there has been debate as to whether the current populations of landlocked alewife in the southern New England are native or introduced and if so when they were introduced to the region (Brugam 1978; Palkovacs et al. 2008). High classification success in discriminant analysis confirmed our estimates of past alewife presence from our historical study and suggested that prior to colonial Euro-American dam construction alewife were historically native in coastal lakes throughout the region. These results complement molecular genetic evidence, which suggested that many populations of landlocked alewife in the region were relict populations of past anadromous alewife runs that independently evolved a landlocked lifestyle when isolated from anadromous ancestors at least 300 years before present (Palkovacs et al. 2008). When there is a dearth of information on past fish community composition in the written historical record, the morphology of Bosmina spp. can provide a reasonably accurate proxy for past keystone size-selective fish presence.

Our results have important implications for anadromous fish restoration and management. The morphology of Bosmina spp. in the paleoecologial record can serve as a useful proxy to detect where alewife and other anadromous size-selective zooplanktivores, such as American shad, Alosa sapidissima, or blueback herring, Alosa aestivalis, were native historically, and when they became extirpated from lakes. Anadromous fish restoration is an expensive and often highly politicized process, especially when historically native fishes have not been present in systems for decades to centuries, as is often the case for alewife and other declining anadromous fish species in the Eastern United States. Paleoecological evidence demonstrating that alewife and other anadromous species were historically native to these systems may help overcome the fears of those who have concerns that alewife and other anadromous fishes are being restored to watersheds where they are non-native. Paleoecological evidence may also be helpful in determining when landlocked alewife populations are the natural consequence of evolution from anadromous ancestors.

Notes

Acknowledgments

We thank Christoph Geiss, Michael Oleskewicz, and Derek West for assistance with coring, Elizabeth Hatton for assistance with sample preparation, Suzanne Alonzo for use of equipment, and Helmut Ernstberger, Gaboury Benoit, Troy Hill, and Peter Raymond for assistance with dating cores. This work was funded by a Yale Environmental Summer Fellowship, a Richter Undergraduate Summer Research Fellowship, a Mellon Undergraduate Research Grant, a Yale Environmental Studies Program Grant, and a Yale School of Forestry and Environmental Studies Summer Globalization Internship and Research Grant to C.W.T. Comments provided by two anonymous reviewers and Isabel Larocque improved this paper.

Supplementary material

10933_2012_9672_MOESM1_ESM.docx (41 kb)
Supplementary material 1 (DOCX 41 kb)

References

  1. Alexander ML, Hotchkiss SC (2010) Bosmina remains in lake sediment as indicators of zooplankton community composition. J Paleolimnol 43:51–59CrossRefGoogle Scholar
  2. Appleby P, Oldfield F (1978) The calculation of lead-210 dates assuming a constant rate of supply of unsupported lead-210 to the sediment. Catena 5:1–8CrossRefGoogle Scholar
  3. Atlantic States Marine Fisheries Commission (2009) Amendment 2 to the Interstate Fishery Management Plan for shad and river herring. Atlantic States Marine Fisheries Commission, Washington, DCGoogle Scholar
  4. Benoit G, Rozan TF (2001) Pb-210 and Cs-137 dating methods in lakes: a retrospective study. J Paleolimnol 25:455–465CrossRefGoogle Scholar
  5. Brooks JL, Dodson SI (1965) Predation, body size and composition of plankton. Science 150:28–35CrossRefGoogle Scholar
  6. Brugam RB (1978) Human disturbance and historical development of Linsley Pond. Ecology 59:19–36CrossRefGoogle Scholar
  7. Carpenter SR, Kitchell JF, Hodgson JR (1985) Cascading trophic interactions and lake productivity. Bioscience 35:634–639CrossRefGoogle Scholar
  8. Connecticut Department of Environmental Protection Fish Stocking Records 1860–2011Google Scholar
  9. Cutshall NH, Larsen IL, Olsen CR (1983) Direct analysis of lead-210 in sediment samples: self-absorption corrections. Nucl Instrum Methods 206:309–312CrossRefGoogle Scholar
  10. Dodson SI (1974) Zooplankton competition and predation—experimental test of size-efficiency hypothesis. Ecology 55:605–613CrossRefGoogle Scholar
  11. Dodson SI (1988) Cyclomorphosis in Daphnia-galatea-mendotaw birge and Daphnia retrocurva forbes as a predator-induced response. Freshw Biol 19:109–114CrossRefGoogle Scholar
  12. Dodson SI, Havel JE (1988) Indirect prey effects—some morphological and life-history responses of Daphnia-pulex exposed to Notonecta-undulata. Limnol Oceanogr 33:1274–1285CrossRefGoogle Scholar
  13. Flecker AS, McIntyre PB, Moore JW, Anderson JT, Taylor BW, Hall RO Jr (2010) Migratory fishes as material and process subsidies in riverine ecosystems. Am Fish Soc Symp 73:559–592Google Scholar
  14. Foster D, Swanson F, Aber J, Burke I, Brokaw N, Tilman D, Knapp A (2003) The importance of land-use legacies to ecology and conservation. Bioscience 53:77–88CrossRefGoogle Scholar
  15. Hall CJ, Jordaan A, Frisk MG (2011) The historic influence of dams on diadromous fish habitat with a focus on river herring and hydrologic longitudinal connectivity. Landsc Ecol 26:95–107CrossRefGoogle Scholar
  16. Jackson JBC (2001) What was natural in the coastal oceans? Proc Natl Acad Sci U S A 98:5411–5418CrossRefGoogle Scholar
  17. Jackson JBC, Kirby MX, Berger WH, Bjorndal KA, Botsford LW, Bourque BJ, Bradbury RH, Cooke R, Erlandson J, Estes JA, Hughes TP, Kidwell S, Lange CB, Lenihan HS, Pandolfi JM, Peterson CH, Steneck RS, Tegner MJ, Warner RR (2001) Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629–638CrossRefGoogle Scholar
  18. Kerfoot WC (1987) Translocation experiments—Bosmina responses to copepod predation. Ecology 68:596–610CrossRefGoogle Scholar
  19. Kitchell JA, Kitchell JF (1980) Size-selective predation, light transmission, and oxygen stratification—evidence from the recent sediments of manipulated lakes. Limnol Oceanogr 25:389–402CrossRefGoogle Scholar
  20. Kline TC, Goering JJ, Mathisen OA, Poe PH, Parker PL (1990) Recyling of elements transported upstream by runs of Pacific salmon: delta N-15 and delta C-13 evidence in Sashin Creek, Southeastern Alaska. Can J Fish Aquat Sci 47:136–144CrossRefGoogle Scholar
  21. Korhola A, Rautio MR (2001) Cladocera and other branchiopod crustaceans. In: Smol JP, Birks HJB, Last WM (eds) Tracking environmental change using lake sediments. Kluwer Academic Publishers, DordrechtGoogle Scholar
  22. Korosi JB, Paterson AM, DeSellas AM, Smol JP (2010) A comparison of pre-industrial and present-day changes in Bosmina and Daphnia size structure from soft-water Ontario lakes. Can J Fish Aquat Sci 67:754–762CrossRefGoogle Scholar
  23. Kowalewski M, Goodfriend GA, Flessa KW (1998) High-resolution estimates of temporal mixing within shell beds: the evils and virtues of time-averaging. Paleobiology 24:287–304Google Scholar
  24. Krause RA, Barbour SL, Kowalewski M, Kaufman DS, Romanek CS, Simoes MG, Wehmiller JF (2010) Quantitative comparisons and models of time-averaging in bivalve and brachiopod shell accumulations. Paleobiology 36:428–452CrossRefGoogle Scholar
  25. Limburg KE, Waldman JR (2009) Dramatic declines in North Atlantic diadromous fishes. Bioscience 59:955–965CrossRefGoogle Scholar
  26. Lotze HK, Milewski I (2004) Two centuries of multiple human impacts and successive changes in a North Atlantic food web. Ecol Appl 14:1428–1447CrossRefGoogle Scholar
  27. Naiman RJ, Bilby RE, Schindler DE, Helfield JM (2002) Pacific salmon, nutrients, and the dynamics of freshwater and riparian ecosystems. Ecosystems 5:399–417CrossRefGoogle Scholar
  28. Paine RT (1969) A note on trophic complexity and community stability. Am Nat 103:91–93CrossRefGoogle Scholar
  29. Palkovacs EP, Dion KB, Post DM, Caccone A (2008) Independent evolutionary origins of landlocked alewife populations and rapid parallel evolution of phenotypic traits. Mol Ecol 17:582–597CrossRefGoogle Scholar
  30. Post DM, Frost TM, Kitchell JF (1995) Morphological responses by Bosmina-longirostris and Eubosmina-tubicen to changes in copepod predator populations during a whole-lake acidification experiment. J Plankton Res 17:1621–1632CrossRefGoogle Scholar
  31. Post DM, Palkovacs EP, Schielke EG, Dodson SI (2008) Intraspecific variation in a predator affects community structure and cascading trophic interactions. Ecology 89:2019–2032CrossRefGoogle Scholar
  32. Power ME, Tilman D, Estes JA, Menge BA, Bond WJ, Mills LS, Daily G, Castilla JC, Lubchenco J, Paine RT (1996) Challenges in the quest for keystones. Bioscience 46:609–620CrossRefGoogle Scholar
  33. Rudstam LG, Lathrop RC, Carpenter SR (1993) The rise and fall of a dominant planktivore—direct and indirect effects on zooplankton. Ecology 74:303–319CrossRefGoogle Scholar
  34. Salo J, Walls M, Rajasilta M, Sarvala J, Rasanen M, Salonen VP (1989) Fish predation and reduction in body size in a cladoceran population—paleoecological evidence. Freshw Biol 21:217–221CrossRefGoogle Scholar
  35. Schindler DE (1992) Nutrient regeneration by sockeye salmon (Onchorhynchus-nerka) fry and subsequent effects on zooplankton and phytoplankton. Can J Fish Aquat Sci 49:2498–2506CrossRefGoogle Scholar
  36. Schindler DE, Scheuerell MD, Moore JW, Gende SM, Francis TB, Palen WJ (2003) Pacific salmon and the ecology of coastal ecosystems. Front Ecol Environ 1:31–37CrossRefGoogle Scholar
  37. Vanni MJ, Layne CD, Arnott SE (1997) ‘‘Top-down’’ trophic interactions in lakes: effects of fish on nutrient dynamics. Ecology 78:1–20Google Scholar
  38. Vanni MJ (2002) Nutrient cycling by animals in freshwater ecosystems. Annu Rev Ecol Syst 33:341–370CrossRefGoogle Scholar
  39. Walters AW, Barnes RT, Post DM (2009) Anadromous alewives (Alosa pseudoharengus) contribute marine-derived nutrients to coastal stream food webs. Can J Fish Aquat Sci 66:439–448CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  1. 1.School of Forestry and Environmental StudiesYale UniversityNew HavenUSA
  2. 2.Department of Ecology and Evolutionary BiologyYale UniversityNew HavenUSA
  3. 3.Department of Ecology and Evolutionary BiologyCornell UniversityIthacaUSA

Personalised recommendations