Introduction

Cladoceran subfossils in lake sediments have been used to examine the long-term biological implications of a number of environmental stressors, including major fish kills (Amsinck et al., 2005), lakewater calcium decline (Jeziorski et al., 2008), and lake acidification (Paterson, 1994). However, the disarticulation of Cladocera into their component parts upon death, and the taxonomic difficulties inherent in the differential preservation of species remains in lake sediments, present a challenge for the interpretation of cladoceran subfossil records. This is particularly true for species of the genus Daphnia, because typically identifications of daphniid subfossil remains are made using only postabdominal claws and ephippia. As a result, taxonomic resolution of sedimentary Daphnia remains is coarse at best, as it is not currently possible to attribute a particular postabdominal claw to an individual species.

The current practice in both European and North American sedimentary cladoceran analyses is to differentiate Daphnia claws into two species complexes. The first complex contains species with uniform spinules along the postabdominal claw (Fig. 1a), and is referred to by many different names in the paleolimnological literature including the Daphnia longispina complex in European studies (Korhola, 1999; Bigler et al., 2006; Kamenik et al., 2007), and in North American studies as either the D. rosea complex, (Leavitt et al., 1994; Bos et al., 1999), the D. dentifera complex (Bredesen et al., 2002), or the D. ambigua complex (DeSellas et al., 2008). Here, we refer to this group as the Daphnia longispina complex, in order to better harmonize the European and North American literature on subfossil Cladocera. The second daphniid complex is universally referred to as the Daphnia pulex complex, and consists of species with stout spines on the middle comb of the postabdominal claw (Fig. 1b; Brugam & Speziale, 1983; Bigler et al., 2006; DeSellas et al., 2008).

Fig. 1
figure 1

a A postabdominal claw with uniform spinules representative of the Daphnia longispina complex, and b A postabdominal claw with stout spines on the middle comb representative of the D. pulex complex

Full size image

Ecological requirements and relative sensitivity to invertebrate and fish predation vary among Daphnia species (e.g. Gliwicz, 1990). Therefore, there is considerable information potentially available to paleolimnologists if a reliable method was developed for identifying Daphnia species based on postabdominal claws. For example, D. mendotae (within the D. longispina complex) has been shown to survive the invasion of the predatory cladoceran Bythotrephes longimanus into North American lakes while other smaller Daphnia species are lost (Lehman & Cáceres, 1997; Yan et al., 2001). The isolation of D. mendotae based on postabdominal claws from lake sediments would therefore be useful for paleolimnological studies examining the spread of Bythotrephes. In addition, in eastern Canada, D. ambigua (D. longispina complex) and D. catawba (D. pulex complex) are relatively acid-tolerant (Keller & Pitblado, 1984), and are also tolerant of low calcium (Ca) concentrations (Cairns 2010), well below the 1.5-mg/l fitness threshold identified for D. pulex (Ashforth & Yan, 2008). Therefore, by grouping species together into complexes, we reduce the effectiveness of cladoceran subfossils as paleolimnological indicators of environmental stressors such as lakewater acidification and aqueous Ca decline (e.g., Jeziorski et al., 2011).

Hebert & Finston (1997) showed that the number of stout spines on the middle comb of the postabdominal claw was a useful feature for separating D. pulex from D. catawba; D. catawba typically has 3–4 stout spines, and D. pulex ordinarily has 5+. This distinction between the number of stout spines present on the middle comb has occasionally been applied to Daphnia subfossils (e.g. Bredesen et al., 2002; Bos & Cumming, 2003); however, there are disparities in the literature that question the validity of the number of stout spines as a diagnostic character in a paleolimnological context, when no additional diagnostic features can be used to validate identifications. For example, Paterson (1994) identified claws in the sediments of lakes in Adirondack Park (New York, USA) with 3–5 stout spines as belonging to D. catawba, Schwartz et al. (1985) designate D. pulex as having 4–9 stout spines, and examples of D. pulex postabdominal claws with four stout spines can be found in Hebert (1995). Furthermore, postabdominal claws have been recovered from lake sediments in south-central Ontario that clearly show four stout spines on one claw, and five stout spines on the other (Fig. 2).

Fig. 2
figure 2

A Daphnia pulex complex claw recovered from the surface sediments of Dunbar Lake in south-central Ontario that has four stout spines on one claw, and five stout spines on the other

Full size image

Here, our objective is to perform a detailed examination of the features of postabdominal claws from several Daphnia species in south-central Ontario, Canada (including claw length, width, curvature, and spine/spinule length), to determine whether subtle differences exist that can be used to aid taxonomic resolution within species complexes. A set of four lakes was selected from a broad lake survey of pelagic zooplankton conducted by the Canadian Aquatic Invasive Species Network (CAISN) that contained only a single daphniid species from a given species complex. Subsequently, surface sediments were collected from the four lakes, with the assumption that the daphniid postabdominal claws recovered in these samples belonged to the taxon identified in the modern-day net tows. Using this method, we compare D. ambigua to D. mendotae within the D. longispina complex, and within the D. pulex complex, we compare D. catawba to D. pulicaria, a taxon that is morphologically identical to D. pulex and distinguished from it based only on genetics (Hebert et al., 1993) or habitat type (D. pulex is a strict pond dweller; Dudycha, 2004, Hebert, 1995). This exploratory study will not only evaluate whether it is possible to reliably distinguish between Daphnia species based on claws in lake sediments from south-central Ontario, but as differentiation of Daphnia claws is a concern for paleolimnologists globally, our regional findings will have much wider implications. We show that subtle differences in claw morphology do exist between D. ambigua and D. mendotae, as well as D. pulex and D. catawba, however the application of these differences to the identification of subfossil daphniid postabdominal claws will be labor-intensive, and some uncertainty in identifications will still be present.

Materials and methods

Lake selection and field methods

From mid-June to late August in 2005 and 2006, 311 lakes in south-central Ontario were sampled for zooplankton as part of a broad study to document the spread of the invasive species Bythotrephes longimanus by the Canadian Aquatic Invasive Species Network (Cairns et al., 2007). Duplicate vertical net hauls were taken from a deepwater station along the fetch of each lake using a 63 μm mesh conical net tow, and a composite sample that contained a minimum of 250 individuals was counted and used to determine daphniid species presence/absence (Cairns 2010). From this dataset, we selected three lakes based on their relatively simple daphniid communities: Oudaze Lake, which contained only D. mendotae from the D. longispina complex and no species from the D. pulex complex, Dunbar Lake, which contained D. ambigua from the D. longispina complex and D. catawba from the D. pulex complex, and Camp Lake, which contained several species from the D. longispina complex, but only D. pulicaria from the D. pulex complex (Table 1; Fig. 3). In addition, we also selected Young Lake, as this lake is sampled regularly by the Dorset Environmental Science Centre (Ontario Ministry of the Environment) and currently only contains D. mendotae.

Table 1 Daphniid species presence/absence data obtained from Cairns et al. (2007) for the four study lakes
Full size table
Fig. 3
figure 3

A map of south-central Ontario showing the locations of the four study lakes (black circles), Dunbar Lake (1), Oudaze Lake (2), Young Lake (3), and Camp Lake (4). Inset shows the location of Ontario within Canada

Full size image

In summer 2007, sediment cores were collected from the deepest basins of Young, Dunbar, and Oudaze lakes using a Glew (1989) gravity corer (Jeziorski et al., 2011). A Glew (1988) vertical extruder was used to extract the top 0.25 cm of sediment from each core, which in this region typically represents the last 0–3 years of sediment accumulation (e.g., Mills et al., 2009). A surface sediment sample was collected in an identical manner from Camp Lake in September 2010. To account for the 4-year gap between modern zooplankton sampling in 2006 and sediment collection in 2010, both the 0–0.25 and 0.25–0.5 cm sediment intervals were analyzed to ensure that the material included the 2006 sampling season.

Laboratory methods

Approximately, 1.0 g of wet sediment was deflocculated in 10% KOH for ~30 min at 70–80°C, following the general techniques outlined in Korhola & Rautio (2001). The solution was then passed through a 37 μm sieve, and the residue retained on the sieve was transferred to a vial with deionized water. Safranin–glycerol solution was added to color the remains, and 10% ethanol was added to prevent fungal growth. Slides were then mounted using glycerin jelly.

For D. catawba (Dunbar Lake), D. mendotae (Young and Oudaze lakes combined), and D. pulicaria (Camp Lake), a minimum of 50 claws were recovered and analyzed at 400× magnification on a Leica DMRB microscope using bright field optics. We selected 50 claws per species consistent with previous paleolimnological investigations into daphniid postabdominal claw length (PCL, Manca et al., 2000; Korosi et al., 2008, 2010). Only 46 claws were recovered for D. ambigua from Dunbar Lake. On all claws, the PCL, postabdominal claw width (PCW), spinule length (SL), and two different angles (Ø1 and Ø2) were measured (Fig. 4a) using Northern Eclipse Image Analysis Software version 6 (Empix Imaging Inc.) For D. catawba and D. pulicaria, comb lengths (CL) and spine/spinule length (SL) for the proximal, middle, and distal portions of the claw were also measured (Fig. 4b), and the number of stout spines present on the middle comb was recorded.

Fig. 4
figure 4

a Measurements used to characterize Daphnia postabdominal claws. PCL postabdominal claw length, PCW postabdominal claw width, SL spinule length. A claw belonging to the D. longispina complex is shown here, but these measurements were also taken for claws in the D. pulex complex. b Additional measurements taken for claws belonging to the D. pulex complex, including proximal (prox.) comb length (CL) and spinule length (SL), distal (dist.) CL and SL, and middle (mid.) CL and spine length (SLmin and SLmax)

Full size image

Statistical analysis

To determine whether D. ambigua could be separated from D. mendotae, and D. pulicaria from D. catawba, a nonparametric classification and regression-tree (CART) analysis was conducted using the rpart package (Therneau et al., 2009) for the R software environment (R Development Core Team, 2011) to assess the minimum number of claw traits that would allow for robust differentiation between species. This recursive approach has previously been successfully applied in a paleolimnological context to differentiate between morphologically similar pollen grains (Lindbladh et al., 2002; Barton et al., 2010).

Results

Daphnia longispina complex

Postabdominal claw lengths (PCL) significantly differed between D. ambigua and D. mendotae samples (t test, P < 0.0001). The size distribution of D. mendotae was skewed toward higher values compared to D. ambigua, which never exceeded a length of 150.0 μm (Table 2; Fig. 5). To estimate within-species variation, as well as inter-specific variation, we also compared D. mendotae claws recovered from Oudaze Lake to D. mendotae claws from Young Lake. Although the average length of D. mendotae claws from Oudaze Lake was significantly longer than D. mendotae claws from Young Lake (t test, P = 0.007; Fig. 6), the smaller D. mendotae claws from Young Lake were still significantly longer than D. ambigua claws (t test, P = 0.0001; Fig. 6).

Table 2 Descriptive statistics for the specimens analyzed from the Daphnia longispina complex
Full size table
Fig. 5
figure 5

Distribution of postabdominal claw lengths (PCL) for Daphnia ambigua and D. mendotae

Full size image
Fig. 6
figure 6

Boxplots comparing postabdominal claw length between Daphnia ambigua in Dunbar Lake, and D. mendotae recovered from Young, Oudaze, and Young + Oudaze lakes

Full size image

The classification tree produced by CART determined that, of the measured variables, D. ambigua and D. mendotae claws were best separated by postabdominal claw length (PCL) and spinule length (SL; Fig. 7a). Of the claws that were greater in length than 133.3 μm, only three were from D. ambigua, while 34 were from D. mendotae (Fig. 7a). Claws < 133.3 μm long were further divided based on SL, where amongst the claws with SL ≥ 4.3 μm, 32 were from D. ambigua and four were from D. mendotae. Of claws with SL < 4.3 μm and PCL < 93.1 μm, ten were from D. ambigua and six were from D. mendotae, and for claws with SL < 4.3 and PCL > 93.1 μm, six were from D. mendotae and one was from D. ambigua. In general, D. ambigua claws were shorter with longer spinules than D. mendotae claws, although a degree of uncertainty is present.

Fig. 7
figure 7

Classification tree produced by (CART) analysis of a Daphnia ambigua and D. mendotae claws, and b D. catawba and D. pulicaria claws. Measurements are in μm. The numbers below species name refer to the number of claws correctly classified (bold), and the number of claws incorrectly classified (not bold)

Full size image

Daphnia pulex complex

No significant difference in claw morphology was observed between D. pulicaria claws in Camp Lake recovered from the 0–0.25 and 0.25–0.5 cm sediment intervals (ANOSIM, P > 0.05), therefore all claws from Camp Lake were grouped together for comparison with D. catawba claws from Dunbar Lake. D. catawba claws had 3–4 stout spines on the middle comb, and D. pulicaria had 3–7, but the majority of claws recovered for both species had four stout spines (Fig. 8). D. pulicaria claws were significantly longer than D. catawba (t test, P = 0.001); however, there was more overlap in size than compared with D. ambigua and D. mendotae (Table 3; Fig. 9).

Fig. 8
figure 8

A histogram showing the number of stout spines on the middle comb of Daphnia pulicaria and D. catawba postabdominal claws

Full size image
Fig. 9
figure 9

Distribution of postabdominal claw lengths (PCL) for Daphnia pulicaria and D. catawba

Full size image
Table 3 Descriptive statistics for the specimens analyzed from the Daphnia pulex complex
Full size table

Neither postabdominal claw length (PCL) nor the number of stout spines was identified as a predictor of species identity by the CART analysis. Instead, the best split was based on the length of the middle comb of the claw (CLmid) where the stout spines are located (Fig. 7b). Of claws with CLmid > 29.3 μm, 26 were from D. pulicaria and three were from D. catawba (Fig. 7b). Claws with CLmid < 29.3 μm were further split based on the length of the proximal comb of the claw (CLprox), and 31 claws with CLprox ≥ 13.2 μm belonged to D. catawba while nine were D. pulicaria. This portion of the regression tree was split again based on CLmid, where ten claws with CLprox < 13.2 μm and CLmid < 17.4 μm were from D. catawba and three were from D. pulicaria. The final split was based on the spinule length on the distal comb of the claw (SLdist), where 12 D. pulicaria claws and two D. catawba claws had PLdist < 6.2 μm, and five D. catawba and two D. pulicaria claws had SLdist ≥ 6.2 μm.

Discussion

Subtle differences in claw morphology were observed between Daphnia ambigua and D. mendotae in the D. longispina complex, as well as D. pulicaria and D. catawba in the D. pulex complex (summarized in Table 4). The application of these features as a taxonomic tool for subfossil Cladocera analyses, however, will be considerably more cumbersome than the traditional method of assigning complexes, and a small degree of uncertainty appears to be unavoidable.

Table 4 The main morphological characters of Daphnia postabdominal claws recovered from lake sediments in south-central Ontario that may be useful for differentiating between D. catawba and D. pulicaria within the D. pulex complex, and D. ambigua and D. mendotae within the D. longispina complex
Full size table

Daphnia longispina complex

As expected, a significant difference in postabdominal claw length (PCL) was observed between the two species, since D. mendotae is the largest daphniid species found in central Ontario (1.2–2.8 mm), while D. ambigua is the smallest (max. size 1.3 mm; Hebert, 1995; Witty, 2004). More than half of D. mendotae claws recovered had PCL > 133 μm, with several claws > 200 μm in length. Conversely, only three D. ambigua claws were > 133 μm, and none reached a maximum size > 150 μm. Shifts in claw length, therefore, have the potential to be used as a crude method for identifying shifts in the dominant daphniid taxa in paleolimnological analyses for some species. For example, it has been suggested that differences in D. longispina complex claw length across environmental gradients likely reflect Daphnia species turnover (Korosi et al., 2008), and a decrease in size within the D. longispina complex since pre-industrial times has been interpreted as an increase in the relative abundance of D. ambigua related to acidification in the region (Korosi et al., 2010).

Phenotypic plasticity places limitations upon the interpretation of changes in mean postabdominal claw length as shifts among daphniid taxa, as we observed D. mendotae claws recovered from Young Lake to be significantly smaller than those recovered from Oudaze Lake. Various environmental factors can influence daphniid body size, and lead to the differences in size structure observed between D. mendotae in Young and Oudaze lakes. Higher lakewater temperatures, for example, may increase the growth rate and moulting frequency of a daphniid, but lead to decreased feeding efficiency and smaller body size at maturity (Moore et al., 1996). Size-selective predation can also affect intra-species size variability, and invertebrate predators like Bythotrephes often select for larger Daphnia individuals (Manca et al., 2000; Yan et al., 2001) while fish predation selects for smaller daphniids (Beckerman et al., 2010). Daphnia that are under predation pressure by both invertebrates and fish may have a faster growth rate (to minimize vulnerability to invertebrates), but a smaller body size at maturity (Lynch, 1977). In Oudaze and Young lakes, D. mendotae claws belonging to the smaller size class tended to have shorter spinule lengths (SL) than D. ambigua; therefore, an approach that combines PCL and SL measurements may be more effective at detecting species shifts within the D. longispina complex in lake sedimentary records. No useful taxonomic information was provided by PCW or curvature, and performing these measurements is likely not a worthwhile time investment.

Within central Ontario, three additional members of the D. longispina complex exist that are each intermediate in size between D. ambigua and D. mendotae (D. dubia, 1.1–1.9 mm; D. longiremus, 0.6–2.4 mm, and D. dentifera, 0.9–2.2 mm; Witty 2004; Hebert, 1995). Future analyses could examine the qualitative features and size distributions of these taxa relative to D. ambigua and D. mendotae. Still, the results of this exploratory study are promising, given the importance of these two taxa for addressing current research questions in central Ontario relating to calcium decline (D. ambigua is tolerant of low [Ca]; Cairns 2010) and spread of Bythotrephes longimanus (D. mendotae is able to co-occur with Bythotrephes; Yan et al., 2001). Measuring claw and spinule lengths from lake sediments has the potential to reveal biologically relevant species trends within the Daphnia community unavailable from the species complex dichotomy alone, presenting an opportunity for resolving Daphnia species identity using subfossils.

Daphnia pulex complex

The number of stout spines on the middle comb of the postabdominal claw has been previously identified as a useful feature for differentiating between species of the D. pulex complex using whole specimens (Hebert & Finston, 1997), as well as in subfossil analyses using only the postabdominal claw (Bos & Cumming, 2003). In general, D. catawba and D. minnehaha are identified as having 3–4 stout spines, and D. pulex and D. pulicaria as having five or more (Hebert & Finston, 1997). While our data do support the assumption that D. catawba generally have 3–4 stout spines, D. pulicaria were frequently observed in this study to have <5 stout spines. Therefore, rather than a definitive diagnostic character, the number of stout spines at best provides an additional clue for species identification. This is useful for identification based upon entire animals, where a number of other morphological characteristics can be considered including the length of the tail spine, abdominal processes, and degree of pubescence (Hebert, 1995). However, application of this distinction to paleolimnological studies, without the benefit of additional taxonomic features, may underestimate the prevalence of D. pulex/pulicaria.

The postabdominal claws of D. pulicaria were significantly longer than those of D. catawba; however, the overlap in size between the two species prevent the use of postabdominal claw length (PCL) as a reliable taxonomic tool. The CART analysis identified the lengths of the middle and proximate combs of the claw as useful for distinguishing between the two species: in D. pulicaria the length of the middle comb of the claw is longer, whereas in D. catawba the proximal comb is longer. An indication of useful taxonomic differences exists, then, with respect to the relative lengths of the proximal, middle, and distal combs on the postabdominal claw, rather than postabdominal claw length.

There are a number of inherent assumptions made in our methodological approach to identifying useful features to distinguish between daphniid postabdominal claws. We assume, first, that Daphnia claws recovered from one lake are representative of the range of phenotypic plasticity displayed by that daphniid species over the entire region. Still, the comparison of D. mendotae claws from Young and Oudaze lakes demonstrated that although differences in size structure can exist within a single species between lakes, these differences were more subtle than the differences between species. Secondly, we assume that samples collected from modern-day point sample net tows are comparable to samples recovered from surface sediments that integrate material deposited over several years (typically 0–3 years). Realistically, however, differences in species seasonal and inter-annual abundances that would be captured by the sediment record are potentially missed in the modern samples. Still, we report differences between samples/daphniid species and identify morphological characters that can be used to better differentiate Daphnia based on subfossil remains recovered from lake sediments. To build on the ideas presented in this study, we recommend analysis of postabdominal claws from live animals recovered in net tows, where fewer opportunities for misidentifications exist. A large sample size taken from multiple lakes will ensure that results are representative of phenotypic diversity across a given region.

Although the approach we outline here is tedious and labor-intensive, differentiating between D. pulicaria and D. catawba postabdominal claws may be a useful technique for certain research questions, and warrants further investigation. D. catawba is an acid-tolerant taxon that can thrive in low-Ca lakes (Cairns 2010), and is one of the most common daphniid taxa of the naturally acidic, low-Ca lakes of the Canadian Shield (Hebert & Finston, 1997). D. catawba has been observed to replace D. pulicaria in Plastic Lake (Muskoka, Ontario) when Ca concentrations declined (N. Yan, York University, personal communication), and Jeziorski et al. (2011) identified coarse species resolution and differential species tolerances to low [Ca] to be detrimental to paleolimnological investigations on the response of daphniids to aqueous [Ca] decline. Furthermore, in areas such as Nova Scotia (Canada), where no long-term zooplankton monitoring datasets exist, our current understanding of how keystone daphniid communities have changed over time in response to acidification-related stressors relies solely on paleolimnological data (Korosi & Smol, 2011). The community dynamics within the D. pulex complex, then, are important to our understanding of how lakes have changed since pre-industrial times in North America.

Conclusions

The purpose of this exploratory study was to identify taxonomic differences in Daphnia postabdominal claws recovered from the sediments of softwater lakes in south-central Ontario (Canada) that could be useful for differentiating daphniid subfossil remains. Within the D. longispina complex, potential taxonomically useful differences exist in postabdominal claw length (PCL) and spinule length (SL) between D. ambigua and D. mendotae. D. mendotae are larger than D. ambigua and also have shorter spinule lengths. In our study region of south-central Ontario, three additional members of the D. longispina complex are present (D. dubia, D. longiremus, and D. dentifera), but were excluded from our analysis. It is unknown how the claw morphologies of these three species compare with those of D. ambigua and D. mendotae; however, as they are intermediate in size, it is likely they would complicate the use of PCL and SL measurements as a taxonomic tool. Adding further complexity is the potential for D. mendotae to hybridize with D. dentifera (Taylor et al., 1996), as even when identifying whole animals, species boundaries are not entirely distinct.

The number of stout spines present on the middle comb of a postabdominal claw is an unreliable characteristic to distinguish between species in the D. pulex complex using subfossil remains. Instead, this feature is more useful for identifying entire animals, where additional taxonomic features are also present to improve confidence in species identification. If species resolution is critical to a research question, there may be potential value in measuring the relative lengths of the proximal, middle, and distal combs of postabdominal claws (although opportunities for misidentifications would still be present).

Cladoceran subfossils are being utilized in a growing number of paleolimnological applications. Therefore, any possible advances in the taxonomic resolution of daphniid subfossils will allow researchers to address increasingly nuanced questions, and ultimately help to improve our understanding of how lakes change over time in response to complex multiple environmental stressors.