A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages

JGLR-01279; No. of pages: 9; 4C: Journal of Great Lakes Research xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Great Lakes Research journal homepage: www.elsevier.com/locate/jglr

A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages Andrya L. Whitten ⁎,1, Jose R. Marin Jarrin 2, A. Scott McNaught Department of Biology, Institute for Great Lakes Research, Central Michigan University, Mount Pleasant, MI 48859, United States

a r t i c l e

i n f o

Article history: Received 16 October 2017 Accepted 5 November 2017 Available online xxxx Communicated by William D. Taylor Keywords: Quagga mussels Zooplankton Mesocosm Direct consumption Lake Michigan

a b s t r a c t Dreissenid mussels are known to disrupt the base of the food web by filter feeding on phytoplankton; however, they may also directly ingest zooplankton thereby complicating their effects on plankton communities. The objective of this study was to quantify the effects of quagga mussel feeding on the composition and size structure of Lake Michigan zooplankton assemblages. Two mesocosm (six 946 L tanks) experiments were conducted in summer 2013, using quagga mussels and zooplankton collected near Beaver Island, MI, to examine the response of zooplankton communities to the presence and absence of mussels (experiment 1) and varying mussel density (experiment 2). Mesocosms were sampled daily and zooplankton taxa were enumerated and sized using microscopy and FlowCAM® imaging. In experiment 1, the presence of quagga mussels had a rapid negative effect on veliger and copepod nauplii abundance, and a delayed negative effect on rotifer abundance. In experiment 2, mussel density had a negative effect on veliger, nauplii, and copepodite abundance within 24 h. Multivariate analyses revealed a change in zooplankton community composition with increasing mussel density. Ten zooplankton taxa decreased in abundance and frequency as quagga mussel density increased: except for the rotifer Trichocerca sp., treatments with higher mussel densities (i.e., 1327, 3585, and 5389 mussels/m2) had the greatest negative effect on small-bodied zooplankton (≤128 μm). This study confirms results from small-scale (≤1 L) experiments and demonstrates that quagga mussels can alter zooplankton communities at mesoscales (~1000 L), possibly through a combination of direct consumption and resource depletion. © 2017 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Introduction Quagga mussels (Dreissena rostriformis bugensis) colonized Lake Michigan in 1997, and within ten years, their populations increased and surpassed zebra mussel (D. polymorpha) populations (Nalepa et al., 2010), an earlier invader. The expansion of quagga mussels in Lake Michigan coincided with the decrease in zebra mussels from a previous maximum of 2064 live individuals/m2 (16–30 m depth contour) in 1999 to zero live individuals/m2 in 2008 (Nalepa et al., 2010). Between 1997 and 2007, quagga mussels replaced zebra mussels in nearshore regions, established populations offshore (starting in 2004), and spread to waters N100 m in depth (Nalepa et al., 2010). By 2008, mean density of Lake Michigan quagga mussels at 16–30 m was 19,000 individuals/m2 (Nalepa et al., 2010). ⁎ Corresponding author. E-mail address: [email protected] (A.L. Whitten). Present address: Illinois Natural History Survey, Prairie Research Institute, University of Illinois, Illinois River Biological Station, 704 North Schrader Avenue Havana, Illinois 62644, USA. 2 Present address: Charles Darwin Research Station, Charles Darwin Ave. n/a, Puerto Ayora, Galápagos, Ecuador. 1

Dreissenid mussels (zebra and quagga) are able to disrupt the base of the food web in aquatic ecosystems by filtering phytoplankton from the water (Hecky et al., 2004; Holland, 1993; Nicholls and Hopkins, 1993; Vanderploeg et al., 2002). A few years after large populations of zebra mussels were established in the nearshore region, of Saginaw Bay, Lake Huron, phytoplankton abundance decreased by 60% (Holland, 1993). More recently, southeastern Lake Michigan experienced an 87% decrease in phytoplankton biomass during the spring isothermal mixing period between 1995 and 1998 and 2007–2008 (Fahnenstiel et al., 2010), coinciding with the establishment of large populations of quagga mussels. Dreissenid mussels can also divert energy and nutrients from the pelagic to benthic zone (Hecky et al., 2004), which enhances benthic organism density and taxonomic richness (Ward and Ricciardi, 2007) and negatively affects the health of zooplankton populations (Hecky et al., 2004; Vanderploeg et al., 2002). In general, the abundance of zooplankton declines in the presence of zebra and quagga mussels (Bridgeman et al., 1995). In addition to their indirect effect on planktonic animals, dreissenid mussels can also directly ingest zooplankton (Shevtsova et al., 1986; MacIsaac et al., 1991; MacIsaac et al., 1995; Wong and Levinton, 2005). Dreissenids consume zooplankton that fit into their siphon, and

https://doi.org/10.1016/j.jglr.2017.11.005 0380-1330/© 2017 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

2

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

siphon diameter is directly related to shell size (Wong and Levinton, 2005). The typical size range of particles taken in by zebra mussels is 5–45 μm (Gergs et al., 2009; Sprung and Rose, 1988; Ten Winkel and Davids, 1982; Wong and Levinton, 2005). This particle size range corresponds to the size of algal cells and decaying matter. Adult dreissenid mussels, however, have been reported to filter particles as large as 0.4–1.2 mm (Shevtsova et al., 1986; Horgan and Mills, 1997). Larger particles could include rotifer and microcrustacean zooplankton. Small-bodied zooplankton have been shown to decrease in abundance in the presence of dreissenids in controlled laboratory experiments (i.e., rotifers and small cladocerans; MacIsaac et al., 1991; Shevtsova et al., 1986; Wong et al., 2003) and monitoring studies in the Hudson River, New York (i.e., rotifers and copepod nauplii; Pace et al., 1998; Strayer et al., 2014). Laboratory observations were made in short-term (2–24 h), small-scale (50–500 mL beaker), high density experiments with up to 10 zebra mussels per vessel (MacIsaac et al., 1991; Shevtsova et al., 1986; Wong et al., 2003). Furthermore, MacIsaac et al. (1991) determined that consumption of specific zooplankton species may depend on zooplankton body shape and mobility. Some zooplankton species have shell spines which deter ingestion or are able to rapidly swim away from predators (Wallace and Snell, 1991). Therefore, zooplankton taxa may respond differently to dreissenid predation. Over the past ten years, the biomass of zooplankton – especially cladocerans, cyclopoid copepods, and copepod nauplii – has declined in Lake Michigan, and this decline has been attributed to a decrease in phytoplankton availability due to dreissenid feeding (Vanderploeg et al., 2012). However, the results of laboratory experiments and monitoring studies on other systems suggest that the decline might be, in part, caused by direct consumption of zooplankton by dreissenid mussels. With the increase in density and expansion of quagga mussels into deeper water, it is important that we quantify their ability to affect zooplankton assemblages through direct consumption as well as indirect resource competition if we are to fully understand their effects on aquatic food webs. The goal of this study was to quantify the effects of quagga mussels over time and at a mesoscale (~ 1000 L) on the composition and size structure of zooplankton assemblages (rotifers and microcrustaceans) collected from Lake Michigan. We conducted a quagga mussel presence/absence experiment and a quagga mussel gradient experiment in mesocosms with natural organism densities. If quagga mussels are able to alter zooplankton assemblages through direct composition at mesoscales, we predict 1) there will be a decrease in the abundance of small, soft-bodied zooplankton with low mobility; 2) changes in zooplankton abundance and composition will occur rapidly due to high mussel filtration rates; and 3) changes will be more pronounced at high relative to low quagga mussel densities.

Table 1 Repeated measures ANOVA output for abundance (number per liter) of zooplankton groups (veligers, nauplii, rotifers, cladocerans, and copepods) with Dreissena rostriformis bugensis treatment and time as factors. Statistics include degrees of freedom (d.f.), sum of squares (SS), mean squares (MS), and F-statistic (F). Degrees of freedom for time and interaction were adjusted using Greenhouse-Geisser correction. Bonferroni corrected α of 0.01. Significant differences in bold. Taxa

Source

d.f.

SS

MS

F

p-Value

Veligers

Mussels Time Interaction Error Mussels Time Interaction Error Mussels Time Interaction Error Mussels Time Interaction Error Mussels Time Interaction Error

1 1.65 1.65 6.6 1 1.82 1.82 7.28 1 1.56 1.56 6.25 1 1.73 1.73 6.93 1 1.47 1.47 5.87

14,500 7620 6740 2420 594 313 80.4 60.8 540 109 221 267 194 29.3 62.1 20.8 19.5 3.19 1.81 3.89

14,600 4620 4090 367 594 172 44.2 8.36 540 69.5 141 42.7 194 16.9 35.8 3 19.5 2.17 1.23 0.662

121.6 12.6 11.14

0.000 0.007 0.009

129.3 20.59 5.28

0.000 0.001 0.040

30.47 1.63 3.32

0.005 0.263 0.109

56.13 5.63 11.93

0.002 0.038 0.007

12.12 3.28 1.86

0.025 0.116 0.233

Nauplii

Rotifers

Copepods

Cladocerans

the months of July and August (17–23 °C, NOAA CoastWatch, Great Lakes Node, http://coastwatch.glerl.noaa.gov, accessed on 01/12/2016). For each of the two experiments, quagga mussels and zooplankton were collected from St. James Harbor, Beaver Island, Lake Michigan. Quagga mussels on rocks were collected through snorkeling and transported to the laboratory in plastic crates containing ambient water. A foil weighing method was used to determine the number of mussels colonized per rock for distribution in the mesocosms (Mackie, 2004). We calculated a mean mussel density to foil weight ratio by first covering three 10 × 10 cm areas with foil, counting the number of mussels in each area, and then weighing the 10 × 10 cm piece of foil (Mackie, 2004). Then, mussel colonies on each rock were covered in foil and the weight of the foil was used to estimate mussel abundance per rock. Separate zooplankton assemblages were collected for each experiment using a 0.5 m diameter net with a small enough mesh (64 μm) to collect both rotifers and crustaceans (Barbiero et al., 2012). Fourteen vertical net tows were taken during daylight hours through the entire water column (15 m) and combined into one 18 L bucket and transported back to the laboratory for equal distribution between the six mesocosms after an initial zooplankton sample (1 L) was collected. The initial stocking density was approximately 7 times the estimated zooplankton density in St. James Harbor from a single 15 m net tow.

Methods Mesocosm experiments

Experiment 1

Two six-day mesocosm experiments were conducted in June and July 2013 at Central Michigan University Biological Station on Beaver Island, Michigan, USA. Six 946 L cylindrical mesocosms (surface area = 1.13 m2) were used to mimic Lake Michigan natural conditions in both experiments. Filtered, epilimnetic, Lake Michigan water was directly pumped into the mesocosms and mixed with 850 gph Hydor Koralia circulation and wave pumps. Overhead high-intensity, fullspectrum lights were available to provide 4 h of artificial light to supplement room lighting for a 15:9 light:dark cycle. Artificial lights heated the water and mean water temperature increased from 18.85 ± 0.13 °C on day three to 20.00 ± 0.10 °C on day six in experiment one. In experiment two, mean water temperature increased from 17.8 ± 0.15 °C on day one to 22.85 ± 0.08 °C on day six. These water temperatures are within the range of values observed for Lake Michigan between

Quagga mussel effects on zooplankton community composition were first tested using a presence/absence experiment. In a systematic design, three mesocosms were stocked with 5319 ± 364 (mean ± 1SD) mussels/m2 and three mesocosms were stocked with rocks having no mussels. We diluted the zooplankton assemblage collected from St. James Harbor to 21 L then equally stocked each mesocosm with 3.5 L of concentrated mixed zooplankton. The experiment lasted six days with one zooplankton sample collected daily (2000 h) from each mesocosm. Prior to sampling, water circulators were removed and each mesocosm was thoroughly mixed using a meter stick. Samples were collected using a 10 cm diameter, 64 μm mesh net. The net was submerged and set on the bottom of the mesocosm for 20 s before collecting each sample. All samples were preserved with 10% sugar formalin in 250 mL plastic bottles.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

3

Fig. 1. Density (number per liter) of veligers, nauplii, rotifers, copepods, and cladocerans in control and Dreissena rostriformis bugensis treatment mesocosms over time. Control mesocosms represented by closed circles, treatment mesocosms represented by open circles, and initial sample represented by open triangles. Error bars calculated with the standard error. Note different y-axis scales.

All large (N0.2 mm) zooplankton taxa were identified and counted using a Nikon dissecting microscope (30×) by concentrating an entire sample and placing it in a Bogorov counting tray for identification and counting. Small (≤0.2 mm) zooplankton taxa were identified and counted using a Nikon compound microscope (100 ×). Each sample bottle was filtered and diluted to 10 mL, and a 1 mL sub-sample was pipetted into a Sedgwick Rafter cell for identification and counting. This method was repeated 3 times to calculate small zooplankton abundance. All organisms were identified to genus using Balcer et al. (1984) and Ward and Whipple (1959). A parametric repeated measures ANOVA with the Greenhouse-Geisser adjustment for violation of the sphericity assumption was performed in SPSS software (statistics 21) on specific zooplankton taxa (nauplii, veligers, rotifers, cladocerans, and copepods) to look for differences in abundance between treatments and through time (Bonferroni corrected α = 0.01). If time was significant, a least significant difference (LSD) post hoc analysis was performed to identify which time intervals differed from the others.

Experiment 2 Quagga mussel density effects over time on zooplankton community composition were tested using a mussel gradient experiment. Quagga mussel densities 0, 194, 760, 1327, 3585, and 5389 mussels/m2 were randomly distributed among six mesocosms. The mesocosms were then equally stocked with 3 L of concentrated mixed zooplankton from an 18 L assemblage collected from St. James Harbor, Beaver Island, Lake Michigan. The experiment lasted six days with two zooplankton samples collected every twelve hours for the first two days and then once daily for the remaining four days. Zooplankton sample collection and abundance calculations were conducted using the same method as in experiment 1. We used linear regression (model 1) to analyze the effect of mussel density on major taxonomic groups at three times early in the experiment. Data were log transformed when assumptions of linearity were violated.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

4

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

Table 2 Relationship between Dreissena rostriformis bugensis density and abundance of dominant zooplankton taxa (number per liter) in 946 L mesocosms at three time points in a 6 day experiment. Significant differences in bold (α = 0.05). Taxon

Time

Slope

Intercept

R2

p-Value

Log veliger

Hour 12 Hour 24 Hour 36 Day 1 Day 3 Day 6 Day 1 Day 3 Day 6 Day 1 Day 3 Day 6 Day 1 Day 3 Day 6

−0.00024 −0.00025 −0.00013 −0.00119 −0.00075 −0.00078 −0.00054 −0.0006 −0.00037 −0.0002 −0.00028 −0.0006 −0.00098 −0.00112 −0.00076

1.38 1.13 0.565 9.95 3.97 4.18 5.50 4.26 2.68 1.04 1.65 2.90 6.18 6.94 12.0

0.773 0.582 0.439 0.758 0.504 0.395 0.724 0.854 0.292 0.434 0.443 0.515 0.440 0.325 0.039

0.021 0.078 0.152 0.024 0.114 0.181 0.032 0.008 0.269 0.155 0.149 0.108 0.151 0.238 0.709

Nauplii

Copepods

Cladocerans

Rotifers

To identify the effects of mussel density and time on zooplankton communities, we used three nonparametric statistical analyses: blocked multiple response permutation procedure (MRBP), non-metric multidimensional scaling (NMDS), and blocked indicator species analysis (blocked ISA). The MRBP and ISA were blocked to account for time (n = 8). All three multivariate analyses were performed using the statistical software package PC-ORD (version 6, McCune and Mefford, 2011). The abundance data used for all three multivariate analyses were first log10 (x + 1) transformed because of large density differences between the most abundant zooplankton taxon and the rest of the taxa. The MRBP analysis was used to determine if there was a difference between the zooplankton communities in the six quagga mussel density treatments (groups). Within the MRBP, the Euclidean distance measurement was used along with the within group agreement value (A) and the Monte Carlo test with 4999 randomizations to calculate a p-value. The NMDS ordination was performed to identify differences among zooplankton communities across mussel density and time. A threedimensional solution was found for the quantitative data using the Bray-Curtis distance measure on the ‘slow and thorough’ autopilot mode (stress = 14.11, instability = 0.00000, p-value = 0.004). The Bray-Curtis distance measure was used because of its low sensitivity to heterogeneous data sets and outliers in community data (McCune and Grace, 2002). Pearson correlations were performed to look for association between ordination axes, mussel density, and time. A blocked ISA was used to identify zooplankton taxa indicative of quagga mussel density. The data were grouped by quagga mussel density (n = 6) and blocked by time (n = 8). This analysis combines information on relative species abundance and relative frequency (faithfulness of occurrence in a group) to express indicator values for species as a percentage ranging from zero (no indication) to 100 (perfect indication; McCune and Grace, 2002). Relative abundance is the average abundance of a given species at a given mussel density divided by the average abundance of the species in all mussel densities expressed as a percent. Relative frequency is the percent of time a given species is present at a given mussel density. The observed indicator value is the product of relative abundance and relative frequency for a given species. Zooplankton community size structure was assessed using FlowCAM® digital imaging where zooplankton samples were enumerated and sorted into seven preselected size classes—64, 96, 128, 192, 256, 384, and 512 μm—identified by the lower boundary and based on estimated spherical diameter. These size classes were chosen to encompass the size range of taxa in our Lake Michigan samples including veligers (70–130 μm), rotifers (70–300 μm), nauplii (80–300 μm), copepodites (300–1000 μm), and cladocerans (200–1000 μm). Size class distributions were compared across mussel densities at each sampled time interval (12, 24, 36 h) using a Chi-squared test for association in Minitab (version 16).

Results Experiment 1 Quagga mussels had a significant negative effect on the abundance of veligers, nauplii, rotifers, and copepods, but not cladocerans (Table 1). Cladoceran densities increased over time in the control mesocosms only (Fig. 1), but the p-value (p = 0.025) for the effect of mussels on cladocerans was not significant (Bonferroni corrected α of 0.01). Veligers and nauplii declined significantly in the treatment and control over time (Fig. 1). There was a significant interaction between time and quagga mussel treatment on the abundance of veligers and copepods. There was no interaction between time and mussel treatment on the abundance of nauplii, rotifers, and cladocerans (Table 1). Quagga mussels had a rapid negative effect on the abundance of veligers and nauplii (Fig. 1). The initial veliger abundance (69 individuals/ L) declined 57% in the quagga mussel mesocosms but changed very little in control mesocosms within the first 24 h. Initial nauplii abundance decreased 81% in the quagga mussel mesocosms but only 34% in the control mesocosms within the first 24 h (Fig. 1). The percent reduction of veligers and nauplii in mussel versus control mesocosms was 59% and 66% after 24 h. Veliger and nauplii abundance decreased in all mesocosms after day two. Quagga mussels also had a negative effect on the abundance of rotifers, and the greatest difference in rotifer abundance between control and treatment mesocosms was apparent after 4 days (Fig. 1). After an initial increase in treatment and control mesocosms (likely due to sampling error given the low densities), rotifer abundance continued to increase 111% in control mesocosms but declined 76% in the quagga mussel mesocosms by day 6. There was a 77–89% reduction in rotifer abundance in quagga mussel mesocosms relative to control mesocosms on days 4–6. The original rotifer assemblage consisted of Euchlanis sp. (4.9%), Synchaeta sp. (1.6%), Trichocerca sp. (47.5%), Trichotria sp. (1.6%), Notholca acuminata (13.1%), Keratella cochlearis (9.8%), Lecane sp. (1.6%), Kellicottia longispina (1.6%), Lepadella sp. (9.8%), Monostyla sp. (3.3%), and unknown rotifer (4.9%). The most abundant rotifers at the beginning of the experiment were Trichocerca sp., Notholca acuminata, Keratella cochlearis, and Lepadella sp. Polyarthra vulgaris and Keratella quadrata were not present in the initial sample but were found sporadically over time in zooplankton subsamples from both control and treatment mesocosms. The most abundant rotifers at the end of the experiment were Lepadella sp., Trichocerca sp., Trichotria sp., and Euchlanis sp. As with rotifers, quagga mussels had a negative effect on the abundance of copepods and the effect was greatest later in the experiment (Fig. 1). Initial copepod abundance decreased 51% in the treatment mesocosms and 17% in the control mesocosms within the first 24 h. Following day 1, copepod abundance increased in control mesocosms and decreased in treatment mesocosms. By day 6, there was a 95% reduction in copepods in quagga mussel mesocosms relative to control mesocosms (Fig. 1). The initial copepod assemblage included both cyclopoid (1.63 individuals/L) and calanoid (1.27 individuals/L) copepods. Throughout the experiment cyclopoids were more abundant than calanoids in all mesocosms. Harpacticoid copepods were present, but their abundance was very low and decreased over time in all mesocosms (data not presented). Quagga mussels had no detectable effect on cladoceran abundance over time (Fig. 1). Cladocerans, which consisted of Chydorus sphaericus and Bosmina longirostris, had an initial abundance of 0.52 individuals/ L. In control mesocosms, C. sphaericus, the dominant cladoceran, increased from 0.52 to 2.38 individuals/L and B. longirostris increased from 0.01 to 0.11 individuals/L (data not presented). In the treatment mesocosms, C. sphaericus and B. longirostris maintained very low densities with high variability. C. sphaericus decreased from 0.52 to 0.11 individuals/L while B. longirostris increased from 0.01 to 0.02 individuals/L.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

5

Fig. 2. Relationship between Dreissena rostriformis bugensis density and abundance of dominant zooplankton taxa in 946 L mesocosms. Abundance of crustaceans and rotifers is shown for days 1 (triangle, dotted line), 3 (open circle, dashed line) and 6 (closed circle, solid line). Log abundance of dreissenid veligers is shown for hours 12 (triangle, dotted line), 24 (open circle, dashed line) and 36 (closed circle, solid line).

Experiment 2 Mussel density had a negative effect on the abundance of veligers, copepod nauplii, and copepodites (Fig. 2, Table 2). Increasing mussel density caused a significant reduction in veliger abundance within 12 h and a significant reduction in nauplii and copepodite abundance within 24 h (Table 2). Copepodite abundance continued to be negatively affected by mussel density after 3 days (Table 2). Abundance of veligers, nauplii, and copepodites decreased in all mesocosms over time and the negative effect of mussels became less apparent (Fig. 2). After 36 h, veliger abundance was zero in the 1327 and 3585 mussels/m2 treatments. Rotifer and cladoceran abundance increased over time, particularly in low density mussel treatments. Twenty-one zooplankton taxa were identified in mesocosm samples, and their abundance and diversity changed over time and with increasing mussel density (Table 3). Four taxa (Trichocerca sp., Lepadella sp., Alona sp., and harpacticoid copepods) decreased initially then

increased to an abundance greater than their initial density over time. Two taxa (Filinia sp. and Ascomorpha sp.) appeared only once after the first sampling time. There was a significant difference in zooplankton communities exposed to increasing quagga density (MRBP analysis, n = 8, A = 0.17, p b 0.0001). Pairwise comparisons revealed that zooplankton communities were significantly different among all quagga mussel densities (all p b 0.03) except between density one (0 mussels/m2) and two (194 mussels/m2) (p = 0.164). Taxon diversity decreased substantially as mussel density increased. Five taxa (mostly aloricate rotifers) disappeared between the zero and 194 mussels/m2 treatments (Table 3). The NMDS ordination produced a significant 3-dimensional solution with axis 1 strongly correlated with quagga mussel density (r = 0.61), axis 2 correlated with time (r = 0.45), but axis 3 was not a strong representation of either variable and therefore only a 2-dimensional solution was presented (Fig. 3a). Samples taken from increasing quagga mussel densities were located along axis 1 from right to left (Fig. 3a).

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

6

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

Table 3 Relative abundance, relative frequency (parentheses), indicator value (Observed IV), and identified Dreissena rostriformis bugensis density group (Mussel Group) for 21 zooplankton taxa in 6 mesocosms defined by quagga mussel density. Zooplankton communities were sampled eight times over 6 days. IV is the product of relative frequency and relative abundance. Monte Carlo test of observed IV based on 4999 randomizations produced mean IV and p-values for the hypothesis of no difference between mussel densities. The p-value is based on the proportion of randomized trials with IV equal to or exceeding observed IV. Significant differences in bold (∝ = 0.05). Zooplankton taxa

1 Nauplii 2 Veliger 3 Asplanchna sp. 4 Ascomorpha sp. 5 Euchlanis sp. 6 Filinia sp. 7 Keratella sp. 8 Lecane sp. 9 Lepadella sp. 10 Monostyla sp. 11 Polyarthra sp. 12 Synchaeta sp. 13 Trichocerca sp. 14 Trichotria sp. 15 Alona sp. 16 Bosmina sp. 17 Chydorus sp. 18 Ostracod 19 Calanoid 20 Cyclopoid 21 Harpacticoid

Quagga mussel density 0

194

760

1327

3585

5389

23 (100) 30 (63) 100 (13) 100 (13) 19 (63) 0 14 (50) 29 (75) 15 (63) 24 (88) 100 (38) 100 (50) 10 (38) 32 (63) 32 (50) 33 (13) 23 (63) 100 (13) 38 (100) 22 (100) 7 (13)

22 (100) 51 (88) 0 0 33 (100) 0 12 (75) 35 (75) 23 (50) 30 (63) 0 0 4 (13) 55 (88) 4 (25) 38 (25) 36 (100) 0 19 (100) 18 (100) 0

14 (100) 8 (50) 0 0 15 (63) 0 39 (88) 6 (25) 6 (38) 19 (50) 0 0 0 2 (13) 3 (25) 11 (13) 10 (100) 0 15 (100) 18 (100) 0

18 (100) 2 (25) 0 0 9 (38) 0 15 (75) 15 (50) 17 (50) 5 (25) 0 0 0 4 (13) 19 (63) 0 15 (88) 0 12 (88) 16 (100) 40 (25)

12 (100) 1 (13) 0 0 15 (88) 0 9 (25) 12 (63) 11 (50) 6 (25) 0 0 63 (100) 3 (25) 16 (50) 11 (13) 14 (100) 0 7 (63) 15 (100) 30 (25)

12 (100) 7 (50) 0 0 9 (63) 100 (13) 10 (38) 2 (13) 28 (75) 16 (50) 0 0 23 (63) 5 (50) 26 (88) 7 (13) 3 (38) 0 10 (63) 12 (88) 23 (25)

Samples taken through time were located diagonally from the bottom right corner to the top left corner of the graph (Fig. 3a). Among the 21 zooplankton taxa encountered, the blocked ISA identified eleven taxa as indicators of mussel density (Table 3, ESM Tables S1 through S4). Soft-bodied zooplankton, such as nauplii, calanoid and cyclopoid copepods, and aloricate rotifers Synchaeta sp. and Polyarthra sp. were indicators of no mussels. These results are related by the decrease in abundance of nauplii and copepods as mussel density increased and the absence of the rotifers Synchaeta sp. and Polyarthra sp. from all mesocosms with mussels. The cladoceran C. sphaericus, rotifers Trichotria sp. and Euchlanis sp., and veligers were indicators of the lowest mussel density (194 mussels/m2). They were present in all quagga mussel treatments but decreased in abundance as mussel density increased above 194 mussels/m2 (Table 3). The small, mobile rotifer Keratella sp. was present in all mesocosms and indicated a moderate mussel density (760 mussels/m2). The rotifer Trichocerca sp. was an indicator of high mussel density (3585 mussels/m2; Table 3). Mussel density had a direct effect on the abundance and distribution of zooplankton size classes. Abundance of all zooplankton size classes except for the two largest size classes, 384 and 512 μm, decreased as quagga mussel density increased (Fig. 4). The two largest size classes remained relatively stable throughout the experiment, fluctuating around 0.80 individuals/L. There was a significant difference in the distribution of FlowCAM® size classes across mussel density (Chi-squared, d.f. = 30, p12 = 0.001; p24 = 0.000; p36 = 0.000) at each time interval sampled (12, 24, and 36 h). Mussel density had a greater effect on smallbodied than large-bodied zooplankton. Size class 96 μm was most abundant in the control mesocosm, but in the low-density mussel treatment (194 mussels/m2), size class 128 μm was most common after 36 h. As mussel density increased, the first three size classes (64, 96, and 128 μm) decreased in abundance and larger size classes became more prominent. Size classes 192 μm and 256 μm exhibited the greatest relative abundance at high quagga mussel densities (3585 and 5389 mussels/ m2) after 36 h. The effect of quagga mussels on the abundance of individual size classes varied with time (12, 24, and 36 h). Small size classes (64, 96, and 128 μm) were depressed after only 12 h in high density (1327, 3585, and 5389 mussels/m2) mussel treatments (Fig. 4). Moderate size classes (192 and 256 μm) required more time to be decreased in abundance in high mussel density mesocosms. At low to moderate

Mussel Group

Observed IV

p-Value

0 194 0 0 194 5389 760 194 5389 0 0 0 3585 194 5389 194 194 0 0 0 1327

22.6 45.0 12.5 12.5 32.5 12.5 34.1 26.6 20.6 20.7 37.5 50.0 62.8 48.1 22.6 9.4 35.7 12.5 37.8 21.6 10.0

0.0170 0.0024 1.0000 1.0000 0.0056 1.0000 0.0232 0.1500 0.4413 0.3603 0.0284 0.0064 0.0002 0.0030 0.2444 0.8658 0.0022 1.0000 0.0002 0.0078 0.6131

quagga mussel densities (194 and 760 mussels/m2), 24 or 36 h were necessary before the mussels had an effect on zooplankton abundance. Discussion Results from our two, six-day mesocosm experiments indicate that quagga mussels can directly or indirectly alter the abundance and composition of zooplankton assemblages. Three lines of evidence suggest that mussels are directly consuming zooplankton. (1) There was a rapid (b 24 h) decrease in the abundance of small taxa (veligers and copepod nauplii) in quagga mussel mesocosms compared to control mesocosms with no mussels (experiment 1). (2) The short term (b24 h) effect of quagga mussels on the abundance of dominant taxa (veligers, copepod nauplii, and copepodites) was more pronounced at high rather than low mussel densities (experiment 2). (3) Quagga mussels were able to alter the size distribution of all zooplankton within 12 h by removing small individuals (≤128 μm). Because no other variable was different between treatment and control mesocosms, some level of quagga mussel consumption is the most parsimonious explanation for these results. The only potential predators in the system were cyclopoid copepods, and their densities did not differ between treatment and control mesocosms during the first 24 h of experiment 1. Indirect effects, such as resource competition and zooplankton metamorphosis, likely played a role in altering the abundance of zooplankton in our experiments and made it difficult to discriminate between direct and indirect effects. Phytoplankton resources may have become severely limited after 1 or 2 days as a result of mussel filtration in treatment mesocosms. Although many zooplankton taxa have the ability to survive 4–10 days without food (Kirk, 1997; Threlkeld, 1976), poor survival of rotifers, cladocerans, and calanoid copepods after 4 days in treatment mesocosms can be explained by lack of phytoplankton resources. Phytoplankton resources in control mesocosms were sufficient to sustain the growth of some zooplankton taxa. Calanoid and cyclopoid copepods, Euchlanis sp., Lepadella sp., and Trichotria sp. increased in abundance in the control mesocosms during experiment 1. The observed decrease in nauplii abundance and increase in copepodite abundance in experiment 1 control mesocosms are most likely driven by metamorphosis of stage six nauplii into juvenile copepods. The duration of experiment 1 (six days) would have been enough

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

Fig. 3. NMDS ordination of zooplankton communities (a) in six Dreissena rostriformis bugensis density treatments over eight time intervals with quagga mussel density separated along axis one, and (b) with 21 zooplankton taxa identified in samples separated along axis 1 (Na = Nauplii, Ve = Veligers, Ap = Asplanchna sp., As = Ascomorpha sp., Eu = Euchlanis sp., Fi = Filinia sp., Ke = Keratella sp., Le = Lecane sp., Lp = Lepadella sp., Mo = Monostyla sp., Po = Polyarthra sp., Sy = Synchaeta sp., Tc = Trichocerca sp., Tr = Trichotria sp., Al = Alona sp., Bl = B. longirostris, Cs = C. sphaericus, Os = Ostracod, Ca = Calanoid, Cy = Cyclopoid, Ha = Harpacticoid).

time for some late stage nauplii to develop into adult copepods (Geiling and Campbell, 1972). The development time of copepods from egg to adult varies between 1 and 3 weeks (Wallace and Snell, 1991), depending on light and temperature conditions (Dussart and Defaye, 1995). An extended photoperiod and an increase in temperature can increase development in most species of copepods (Dussart and Defaye, 1995). Therefore, late stage nauplii could have developed into adult copepods, in both control and treatment mesocosms in experiment 1, indirectly influencing copepod densities and making it difficult to determine the effect of mussel filtration on adult copepods. Our study on quagga mussels yielded similar results to small scale laboratory experiments with zebra mussels (MacIsaac et al., 1991, 1995; Shevtsova et al., 1986), and gives support to the suggestion that small-bodied zooplankton (i.e., rotifers and veliger larvae) are more vulnerable to dreissenid consumption than large-bodied zooplankton (MacIsaac et al., 1991; Shevtsova et al., 1986). Unlike previous studies on zebra mussels, our study demonstrates that quagga mussels also have a negative effect on copepod nauplii (experiments 1 and 2) and copepodites (experiment 2). Furthermore, similar to prior studies, our results suggest that the vulnerability of some zooplankton to dreissenids is influenced not only by zooplankton body-size but by morphology and mobility (Jack and Thorp, 2000; MacIsaac et al., 1991; Shevtsova et al., 1986).

7

Body shape and the presence of shells or spines can influence vulnerability of zooplankton to consumption by dreissenid mussels. Shevtsova et al. (1986) found that zebra mussels generally consume zooplankton with elongated body shapes, such as Euchlanis dilatata. Organisms with hard shells or those capable of producing spines in the presence of predators, such as the rotifer Keratella cochlearis, could be less vulnerable to direct predation by dreissenids (Wallace and Snell, 1991). In experiment 2, Euchlanis sp. and Keratella sp. had similar relative abundance and frequency throughout the experiment, but Keratella sp. was an indicator of a higher mussel density (760 mussels/m2) than Euchlanis sp. (194 mussels/m2) suggesting that they are more resistant to mussel filtration. Soft-bodied rotifers with no protective shell or spines, such as Synchaeta sp. and Polyarthra sp., may be consumed easily by dreissenids. In experiment 2, Synchaeta sp. and Polyarthra sp. were present only in the no mussel treatment. Similarly, Shevtsova et al. (1986) noted the disappearance of Polyarthra vulgaris in the presence of zebra mussels, and MacIsaac et al. (1991) reported a decrease in Polyarthra sp. abundance over time in the presence of zebra mussels. Previous studies have shown that less mobile zooplankton taxa, including rotifers, small cladocerans, and veliger larvae, are unable to avoid the inhalant current of mussels; whereas mobile taxa, such as adult copepods, are capable of avoiding mussel predation (MacIsaac et al., 1991; Shevtsova et al., 1986). In this study, copepod nauplii and copepodites, in addition to veliger larvae and rotifers, decreased in abundance in the presence of quagga mussels and across a mussel density gradient. Vertical downward movement in response to light may have increased encounter rates between copepodites and inhalant currents at the bottom of the mesocosms, particularly in high density mussel treatments. Not all rotifer genera were equally vulnerable to quagga mussels. Polyarthra sp., an aloricate, weak-swimmer, were present only in mesocosms without mussels (Wallace and Snell, 2010). Mobile rotifers, such as Euchlanis sp., and Trichocerca sp. were indicators of low and high mussel densities (194 and 3585 mussels/m2). Euchlanis sp. is a strong swimmer (Rico-Martínez and Snell, 1997), which could aid in avoiding a mussel feeding current. Trichocerca sp. is the only taxon that increased in abundance at high mussel densities. The rotifer Trichocerca sp. is considered a moderate swimmer and inhabits the littoral zone along with dreissenids (Rico-Martínez and Snell, 1997). Trichocerca sp. also has morphological features (lorica, long toes) that may allow it to avoid ingestion. Veligers are weak swimmers. Equipped with small tufts of cilia (Mackie, 1991), they have little means of escaping consumption by mussels in well mixed nearshore waters. Adult dreissenids are cannibals (MacIsaac et al., 1991), that can easily consume their own young through filtration of the water. In experiment 2, veligers were indicators of the lowest mussel density (194 mussels/m2) where they were most abundant, and their abundance and frequency decreased as mussel density increased. The lowest mussel density (194 mussels/m2) had more veligers than the control mesocosm (0 mussels/m2) most likely because mussels might have supplemented the initial veliger abundance with new offspring. Mesocosm experiments were conducted in late spring during a period that quagga mussels reproduce (Claxton and Mackie, 1998). Direct consumption of veligers by mussels must have exceeded reproductive efforts because veliger abundance decreased over time in the lowest mussel density (194 mussels/m2). Spatial and temporal limitations of mesocosms prevent direct application of our experimental results to lake ecosystems. The overlying water in the mesocosms was 1 m deep, similar to that of a shallow littoral zone. However, water movement and resource replenishment characteristic of the littoral zone of lakes was not replicated in our mesocosm experiment. We monitored zooplankton populations for only 6 days in the presence and absence of quagga mussels. This short time period is adequate to distinguish between direct and indirect effects of mussel filtration on zooplankton populations, but it is not adequate to show long term consequences of direct and indirect effects.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

8

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

Fig. 4. Zooplankton size distribution (number per liter) across seven size classes (identified by the lower boundary) in six mesocosms with increasing Dreissena rostriformis bugensis densities (a–f: 0, 194, 760, 1327, 3585, and 5389 mussels/m2).

Given these limitations, our results demonstrate a potential impact of dreissenid mussels on zooplankton communities that may or may not be realized in actual lake ecosystems. Habitats where mussels are likely to have a significant negative effect on zooplankton communities include shallow lakes and protected embayments, where mussels are in close contact with overlying water and mussel densities exceed 760 m−2 (experiment 2). Although mussel density in many regions of Lake Michigan typically exceeds 760 m−2 (Nalepa et al., 2010), pelagic

zooplankton are separated from benthic mussels by many meters of water and have a refuge from direct mussel filtration. Additional research is needed to disentangle the direct and indirect effects of dreissenid mussels on Great Lakes zooplankton assemblages. To further understand how direct consumption by mussels is affecting the zooplankton community, and ultimately the pelagic food web, it is essential to include small zooplankton (rotifers, veligers, copepod nauplii) in regular monitoring programs as quagga mussels continue

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005

A.L. Whitten et al. / Journal of Great Lakes Research xxx (2017) xxx–xxx

to expand into deeper water. Few studies on Lake Michigan zooplankton from the Great Lakes report abundances of small zooplankton (b 150 μm). Mussel selectivity for various zooplankton taxa should be examined across gradients of mussel shell length and experimental vessel size and then compared to observations of zooplankton community shifts in the Great Lakes. By better understanding the relative importance of direct and indirect effects on zooplankton communities, we will be able to better predict future changes to the Lake Michigan ecosystem. Acknowledgments We would like to thank K. Buzalski, J. Gordon, M. Hass, J. Loughner, T. Malinich, A. McGrew, H. Preston, D. Schuberg, B. Schuler, and J. Works for their support with field and laboratory work during this study. We also thank B. Murry and D. Woolnough for providing assistance with study design and comments on previous drafts of this manuscript. Additionally, funding for this study was provided by Central Michigan University Department of Biology, College of Science and Technology, and the Institute for Great Lakes Research. This is contribution number 91 of the Central Michigan University Institute for Great Lakes Research. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.jglr.2017.11.005. References Balcer, M.D., Korda, N.L., Dodson, S.I., 1984. Zooplankton of the Great Lakes: A Guide to the Identification and Ecology of the Common Crustacean Species. University of Wisconsin, Madison, Wisconsin. Barbiero, R.P., Lesht, B.M., Warren, G.J., 2012. Convergence of trophic state and the lower food web in Lakes Huron, Michigan and Superior. J. Great Lakes Res. 38, 368–380. Bridgeman, T.B., Fahnenstiel, G.L., Lang, G.A., Nalepa, T.F., 1995. Zooplankton grazing during the zebra mussel (Dreissena polymorpha) colonization of Saginaw Bay, Lake Huron. J. Great Lakes Res. 21, 567–573. Claxton, W.T., Mackie, G.L., 1998. Seasonal and depth variations in gametogenesis and spawning of Dreissena polymorpha and Dreissena bugensis in eastern Lake Erie. Can. J. Zool. 76, 2010–2019. Dussart, B.H., Defaye, D., 1995. Copepoda: Introduction to the Copepoda. Academic, Amsterdam, The Netherlands. Fahnenstiel, G.L., Pothoven, S.A., Vanderploeg, H.A., Klarer, D., Nalepa, T.F., Scavia, D., 2010. Recent changes in primary production and phytoplankton in the offshore region of southeastern Lake Michigan. J. Great Lakes Res. 36, 20–29. Geiling, W.T., Campbell, R.S., 1972. The effects of temperature on the development rate of the major life stages of Diaptomus pallidus Herrick. Limnol. Oceanogr. 17, 304–307. Gergs, R., Rinke, K., Rothhaupt, K., 2009. Zebra mussels mediate benthic-pelagic coupling by biodeposition and changing detrital stoichiometry. Freshw. Biol. 54, 1379–1391. Hecky, R.E., Smith, R.E.H., Barton, D.R., Guildford, S.J., Taylor, W.D., Charlton, M.N., Howell, E.T., 2004. The nearshore phosphorus shunt: a consequence of ecosystem engineering by dreissenids in the Laurentian Great Lakes. Can. J. Fish. Aquat. Sci. 61, 1285–1293. Holland, R.E., 1993. Changes in planktonic diatoms and water transparency in Hatchery Bay, Bass-Island area, western Lake Erie since the establishment of the zebra mussel. J. Great Lakes Res. 19, 617–624.

9

Horgan, M.J., Mills, E.L., 1997. Clearance rates and filtering activity of zebra mussel (Dreissena polymorpha): implications for freshwater lakes. Can. J. Fish. Aquat. Sci. 54, 249–255. Jack, J.D., Thorp, J.H., 2000. Effects of the benthic suspension feeder Dreissena polymorpha on zooplankton in a large river. Freshw. Biol. 44, 569–579. Kirk, K.L., 1997. Life-history responses to variable environments: starvation and reproduction in planktonic rotifers. Ecology 78, 434–441. MacIsaac, H.J., Sprules, W.G., Leach, J.H., 1991. Ingestion of small-bodied zooplankton by zebra mussels (Dreissena-polymorpha)-can cannibalism on larvae influence population-dynamics. Can. J. Fish. Aquat. Sci. 48, 2051–2060. MacIsaac, H.J., Lonnee, C.J., Leach, J.H., 1995. Suppression of macrozooplankton by zebra mussels: importance of mussel size. J. Great Lakes Res. 34, 379–387. Mackie, G.L., 1991. Biology of the exotic zebra mussel, Dreissena-polymorpha, in relation to native bivalves and its potential impact in Lake St-Clair. Hydrobiologia 219, 251–268. Mackie, G.L., 2004. Applied Aquatic Ecosystem Concepts. Kendall/Hunt, Iowa. McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. Mjm Software Design, Oregon. McCune, B., Mefford, M.J., 2011. PC-ORD. Multibariate Analysis of Ecological Data. Version 6. MjM Software, Oregon. Nalepa, T.F., Fanslow, D.L., Pothoven, S.A., 2010. Recent changes in density, biomass, recruitment, size structure, and nutritional state of Dreissena populations in southern Lake Michigan. J. Great Lakes Res. 36, 5–19. Nicholls, K.H., Hopkins, G.J., 1993. Recent changes in Lake Erie (north shore) phytoplankton: cumulative impacts of phosphorus loading reductions and the zebra mussel introduction. J. Great Lakes Res. 19, 637–647. Pace, M.L., Findlay, S.E.G., Fischer, D., 1998. Effects of an invasive bivalve on the zooplankton community of the Hudson River. Freshw. Biol. 39, 103–116. Rico-Martínez, R., Snell, T.W., 1997. Mating behavior in eight rotifer species: using crossmating tests to study species boundaries. Hydrobiological 356, 165–173. Shevtsova, L.V., Zhdanova, G.A., Movchan, V.A., Primak, A.B., 1986. Experimental interrelationship between Dreissena and planktonic invertebrates. Hydrobiological 22, 36–39. Sprung, M., Rose, U., 1988. Influence of food size and food quality on the feeding of the mussel Dreissena polymorpha. Oecologia 77, 526–532. Strayer, D.L., Cole, J.J., Findlay, S.E.G., Fischer, D.T., Gephart, J.A., Malcom, H.L., Pace, M.L., Rosi-Marshall, E.J., 2014. Decadal-scale change in a large-river ecosystem. Bioscience 64, 496–510. Ten Winkel, E.H., Davids, C., 1982. Food selection by Dreissena polymorpha Pallas (Mollusca: Bivalvia). Freshw. Biol. 12, 553–558. Threlkeld, S.T., 1976. Starvation and the size structure of zooplankton communities. Freshw. Biol. 6, 489–496. Vanderploeg, H.A., Nalepa, T.F., Jude, D.J., Mills, E.L., Holeck, K.T., Liebig, J.R., Grigorovich, I.A., Ojaveer, H., 2002. Dispersal and emerging ecological impacts of Ponto-Caspian species in the Laurentian Great Lakes. Can. J. Fish. Aquat. Sci. 59, 1209–1228. Vanderploeg, H.A., Pothoven, S.A., Fahnenstiel, G.L., Cavaletto, J.F., Liebig, J.R., Stow, C.A., Nalepa, T.F., Madenjian, C.P., Bunnell, D.B., 2012. Seasonal zooplankton dynamics in Lake Michigan: disentangling impacts of resource limitation, ecosystem engineering, and predation during a critical ecosystem transition. J. Great Lakes Res. 38, 336–352. Wallace, R.L., Snell, T.W., 1991. Rotifera. In: Thorp, J.H., Covich, A.P. (Eds.), Ecology and Classification of North American Freshwater Invertebrates. Academic, California, pp. 187–248. Wallace, R.L., Snell, T.W., 2010. Rotifera. In: Thorp, J.H., Covich, A.P. (Eds.), Ecology and Classification of North American Freshwater Invertebrates, 3rd edition Academic, California, pp. 173–235. Ward, J.M., Ricciardi, A., 2007. Impacts of Dreissena invasions on benthic macroinvertebrate communities: a meta-analysis. Divers. Distrib. 13, 155–165. Ward, H.B., Whipple, G.C., 1959. Fresh-water Biology. John Wiley & Sons Inc., New York. Wong, W.H., Levinton, J.S., 2005. Consumption rates of two rotifer species by zebra mussels, Dreissena polymorpha. Mar. Freshw. Behav. Physiol. 38, 149–157. Wong, W.H., Levinton, J.S., Twining, B.S., Fisher, N., 2003. Assimilation of micro- and mesozooplankton by zebra mussels; a demonstration of the food web link between zooplankton and benthic suspension feeders. Limnol. Oceanogr. 48, 308–312.

Please cite this article as: Whitten, A.L., et al., A mesocosm investigation of the effects of quagga mussels (Dreissena rostriformis bugensis) on Lake Michigan zooplankton assemblages..., J. Great Lakes Res. (2017), https://doi.org/10.1016/j.jglr.2017.11.005