Effects of dreissenid mussels, chironomids, fishes, and zooplankton on growth of round goby in experimental aquaria

Abstract

There is a widely accepted paradigm supported by early field and laboratory observations that the adult round goby (Neogobius melanostomus) is highly adapted to, and primarily survives on, dreissenid mussels. However, more recent stable isotope and diet analyses indicate that the round goby may not rely on dreissenid prey to the extent that was previously believed. We conducted a feeding experiment where round gobies were provided with an excess of one of four naturally occurring diets for 25 days – dreissenid mussels, juvenile fish, chironomids, or zooplankton. Round gobies fed dreissenids had significantly lower growth (−0.04 g day⁻¹) than individuals fed fish and chironomids (0.13 g day⁻¹) and displayed the same weight loss as round gobies fed only zooplankton. Although dreissenids are often consumed by round gobies, this likely happens only when more profitable prey such as fish and non-dreissenid invertebrates are lacking or difficult to capture. Additionally, field observations of round goby diets that have supported the paradigm may overestimate the importance of dreissenid prey due to the longer retention time of shells in round goby guts compared to other soft-bodied prey. Our results provide direct evidence supporting recent findings that dreissenid mussels may not be as essential to round goby survival as previously considered.

Keywords:

• round goby
• dreissenid mussels
• growth rate

Introduction

The round goby (Neogobius melanostomus) is an invasive fish species that is causing considerable ecological damage within the Laurentian Great Lakes (Jude et al. 1995; Charlebois et al. 1997). The largest consequence of the round goby invasion has been the alteration of energetic pathways through the consumption of dreissenid mussels (Dreissena polymorpha and Dreissena bugensis; Johnson et al. 2005; Dietrich et al. 2006). Round gobies also compete with and consume native fish and macroinvertebrate species (Jude et al. 1995; Janssen and Jude 2001; Steinhart et al. 2004; Balshine et al. 2005).

Prior to the round goby invasion, exotic dreissenid mussels invaded the Great Lakes. These mussels, through their exceptionally high-water filtration rates, have had a significant impact on the food web energy flow relationships within the Great Lakes by taking pelagically derived energy from phytoplankton and small zooplankton and sequestering it in the benthos (Conroy et al. 2005; Johnson et al. 2005; Maguire and Grey 2006; Zhu et al. 2006). Here, as in their native range, round gobies feed heavily on dreissenid mussels (Jude et al. 1992; French 1993; Charlebois et al. 1997; Ray and Corkum 1997), and it has been suggested that the earlier establishment of dreissenids facilitated the establishment of round gobies (Ricciardi 2001).

Although it has been well-established that round gobies consume dreissenid mussels, the importance of other prey items has also been documented. Round gobies in the Great Lakes consume small fish and fish eggs (Chotkowski and Marsden 1999; Steinhart et al. 2004), zooplankton (Cooper et al. 2009), and non-mollusk invertebrates including chironomids, amphipods, and trichopterans (Jude et al. 1995; Taraborelli and Shaner 2002; Lederer et al. 2006). The degree to which round gobies consume dreissenid mussels is possibly based on the seasonal abundances of other non-dreissenid prey resources, since round gobies have been shown in laboratory and field experiments to prefer non-dreissenid prey (Kuhns and Berg 1999; Diggins et al. 2002). Furthermore, Cooper et al. (2009) found that for discrete time periods round goby diets were dominated by zooplankton, likely in response to periodic zooplankton hatches. Stable isotope and energy density analyses further demonstrate the importance of non-dreissenid prey to wild round gobies (Barton et al. 2005; Ruetz and Strouse 2009).

As round gobies have a diverse diet and may actually prefer non-dreissenid prey, the paradigm that adult round gobies heavily rely upon dreissenid prey is questionable. Therefore, we conducted a laboratory experiment to evaluate the growth rate of round gobies on four exclusive naturally occurring prey items – dreissenid mussels, small fish, chironomids, and zooplankton – to test the hypothesis that round goby growth would be greater on an exclusive dreissenid diet over other naturally occurring diet types.

Materials and methods

Round gobies were collected by hook-and-line from St. James Harbor, Beaver Island, MI and transported to the laboratory at the Central Michigan University Biological Station located on Beaver Island. Twenty-four 76-L aquaria were arranged in four rows of six and covered on all sides to minimize stress. All aquaria contained airstones and were lined with aquarium gravel and small rocks for shelter. Aquaria were filled with filtered (153 µm) water from Lake Michigan, and the light regime was set to 14 h light and 10 h dark, consistent with natural lighting during the trial time period (July and August, 2008). Round gobies were weighed and measured (Table 1), placed one per aquarium, and randomly assigned an exclusive prey treatment of either dreissenid mussels, small fish, chironomids, or zooplankton (n = 6 per prey group). Round gobies were acclimated for 24 h after being placed in aquaria before the first prey additions. All aquaria were supplied with excess prey to allow maximum growth. The trial lasted 25 days, and upon completion round gobies were weighed to determine daily growth rates. Growth rates were compared between treatments using an analysis of variance (ANOVA) with subsequent Tukey's multiple pair-wise comparisons.

Table 1. Mean (±SE) TL, and mean (±SE) weight of round goby fed one of four prey treatments at the beginning and end of the 25-day trial period.

	Beginning		End
Prey treatment	TL (cm)	Weight (g)	TL (cm)	Weight (g)
Fish	9.0 (±0.4)	8.42 (±1.24)	9.7 (±0.4)	11.62 (±1.86)
Chironomids	9.9 (±0.7)	12.14 (±2.75)	10.3 (±0.6)	14.70 (±2.80)
Mussels	9.3 (±0.5)	9.43 (±1.67)	9.4 (±0.5)	8.46 (±1.55)
Zooplankton	9.0 (±0.3)	8.40 (±1.08)	9.1 (±0.3)	7.35 (±0.99)

Dreissenid mussels were collected by hand from St. James Harbor, Beaver Island, in Lake Michigan. Dreissenid mussels with shell lengths ranging between 4.5 and 20.8 mm (median = 11.9 mm) were added to provide a naturally occurring range of mussel sizes. Prior to the experiment, we developed a regression equation to predict soft tissue wet weight from dreissenid mussel shell length (soft mass = 0.000008 × shell length^3.3297, n = 50, R ² = 0.914). Dreissenid shells were measured before being added to aquaria, and the regression equation was used to calculate the amount of digestible wet weight that was provided. Dreissenid mussels were supplied as needed to maintain at least 30 mussels of various sizes per tank. We supplied on average 3.6 ± 0.4 (standard error; SE) dreissenid mussels tank⁻¹ day⁻¹ (Table 2). The most commonly provided mussels ranged from 5.1 to 15.0 mm. The initial mean total length (TL) of round gobies fed dreissenid mussels was 9.3 cm (Table 1).

Table 2. Mean number of dreissenids (across the six replicate tanks) by size class (SC) added throughout the experiment.

	SC (mm)
Experimental day	<5	5.1–10	10.1–15	15.1–20	>20.1	Mean total
1	0	0	6.3	13.2	0.5	20
4	0	2.3	0.7	0	0	3
5	0	0.5	0.5	0	0	1
11	0.2	9.7	5.0	0	0	14.9
14	0	1.8	3.2	0	0	5
16	0	1.2	5.0	3.8	0	10
17	0	0.7	3.7	0.5	0	4.9
19	0	3.2	11.2	1.0	0	15.4
22	0	1.8	2.3	0	0	4.1
24	0.2	2.5	6.5	0	0	9.2
Mean total added	0.3	23.7	44.3	18.5	0.5	87.4
Proportion each SC of total added	<0.01	0.27	0.51	0.21	0.01	–
Number remaining at the end of trial	0	4.2	35.5	16.5	0.5	56.7
Proportion consumed	1.00	0.82	0.20	0.11	0.00	0.65

Small fish (i.e., fry and juveniles) were collected from inland lakes on Beaver Island by seining. Prey fish consisted of banded killifish (Fundulus diaphanous), largemouth bass (Micropterus salmoides), and Johnny darter (Etheostoma nigrum) that were between approximately 15.0 and 40.0 mm TL. Fish were weighed in bulk and added to aquaria as they were depleted. On average, 2.59 g ± 0.05 (SE) of fish were added daily. The initial mean TL of round gobies fed fish was 9.0 cm (Table 1).

All chironomids used in the experiment were purchased frozen from a local pet store because a natural source of live chironomids was unavailable. Chironomids were weighed in bulk and provided to round gobies as they were depleted. On average, 1.39 g ± 0.10 (SE) of chironomids were added daily. The initial mean TL of round gobies fed chironomids was 9.9 cm (Table 1).

Zooplankton was collected from inland lakes on Beaver Island using a 153 µm Wisconsin plankton net and primarily consisted of Bosminidae (approximately 88%) but also included Daphniidae, Sididae, copedods, and copepod nauplii. Zooplankton was added to round goby tanks at least once per day (twice per day during the second two weeks of the trial). Prior to each feeding event, a bottle of concentrated zooplankton was gently stirred and equal volumes were added to each aquarium receiving a zooplankton treatment. One portion of zooplankton was retained for analysis, and all zooplankters in this sample were identified and enumerated. From these counts, we estimated the number, density, and wet weight of zooplankters that were added each day. A mean of 2633 ± 344 (SE) zooplankters were added to aquaria daily which equates to a mean of 34.6 ± 6.5 (SE) zooplankters L⁻¹ added each day. Assuming that each zooplankter provided to round gobies weighed approximately 0.25 µg (Haberman et al. 2007), the mean daily wet weight provided to round gobies was approximately 0.66 mg. The initial mean TL of round gobies fed zooplankton was 9.0 cm (Table 1).

All unconsumed fish, chironomids, and dreissenid mussels were removed at the end of the experiment. These were weighed to estimate the wet weight of prey that was actually consumed by round gobies. The density of the remaining zooplankton was estimated by taking three 1L samples from each aquarium after the completion of the experiment, and all zooplankters were identified and enumerated.

Results

Fish- and chironomid-fed round gobies gained weight, whereas round gobies fed dreissenid mussels and zooplankton lost weight. Mean growth rates for the small fish and chironomid treatments were significantly higher than those on dreissenid and zooplankton diets (F _3,21 = 41.35, p < 0.0001). One replicate each of the fish and dreissenid treatments was removed from the analysis because these round gobies escaped their aquaria. One round goby in the chironomid treatment died after the first 5 days of the experiment and was weighed upon discovery and replaced by a newly weighed individual. Data for both of these fish were included in analyses, and growth rates were based on the number of days that they were fed.

Round gobies lost a mean of 0.97 g ± 0.21 (SE) on the dreissenid diet over the 25-day study period for a daily mean weight loss of 0.04 g day⁻¹ ± 0.01 (SE). Round gobies consumed a mean of 34.0 ± 5.9 (SE) mussels throughout the trial and a mean of 1.4 ± 0.2 (SE) mussels day⁻¹ (Table 2). This corresponds to a mean soft tissue wet weight of 0.01 g ± 0.01 (SE) goby⁻¹ day⁻¹consumed. Dreissenids 10.0 mm and smaller were the most consumed size classes (SCs). Except for the 0.0 mm to 5.0 mm SC, round gobies were provided with more dreissenid mussels than were consumed.

Round gobies consuming chironomids grew at a mean rate of 0.13 g day⁻¹ ± 0.02 (SE). Mean weight increase of round gobies during the entire trial was 2.55 g ± 0.45 (SE) (adjusted to account for the fact that one round goby died after 5 days and was replaced by another that fed for 20 days). A mean of 1.30 g ± 0.10 (SE) of chironomids was consumed each day.

The mean growth rate of round gobies provided with small fish was also 0.13 g day⁻¹ ± 0.03 and the overall mean weight increase was 3.20 g ± 0.66 (SE) over the 25-day trial. Round gobies consumed on average 0.92 g ± 0.18 (SE) of fish each day.

Round gobies that consumed zooplankton had a mean weight loss of 1.05 g ± 0.12 (SE) and had a mean daily weight loss of 0.04 g day⁻¹ ± 0.01 (SE) over the trial period. Round gobies consumed a mean zooplankton wet weight of 0.01 g ± 0.001(SE) throughout the trial period. Round gobies were, however, provided with considerable zooplankton densities compared to natural conditions in the Great Lakes (Coulter et al. unpublished).

Discussion

We observed round goby weight loss on an exclusive dreissenid mussel diet, while round gobies fed fish and chironomid diets gained weight. Collectively, these results are not consistent with the prevailing paradigm that round gobies specialize on dreissenids. Dreissenid mussels alone appear to be an unprofitable prey item for round gobies, despite their prevalence in gut contents of wild round gobies. Dreissenid shell material, which has a longer retention time through the digestive track, may result in artificial satiation which would reduce subsequent feeding and ultimately lead to reduced energy intake and growth. There are probably also higher energetic costs associated with shell breakage and passage relative to energy gain from the soft inner tissue. In support of this hypothesis, Ruetz and Strouse (2009) analyzed the energy density of different SCs of round gobies collected from various habitats and found that larger round gobies in areas containing rock substrate (i.e., higher dreissenid populations) had lower energy density than individuals from other locations. They hypothesized that the reduced energy density of individuals sampled from the rock habitat was due to the inclusion of dreissenid mussels in their diet. Furthermore, stable isotope analyses of round gobies collected from Lake Erie revealed that even large individuals likely do not solely consume dreissenid mussels (Barton et al. 2005).

Since dreissenid mussels appear to be of low profitability to round gobies they likely consume dreissenids when other, more profitable prey are scarce or difficult to capture. Diggins et al. (2002) found that round gobies switched from the preferred amphipod to dreissenid mussels when the ability to capture amphipods was reduced, for example due to habitat complexity which provided refuge for amphipods. Round gobies also selected dreissenids when higher turbidity reduced visibility and decreased their ability to efficiently detect and capture amphipods. Also, Walsh et al. (2007) found that round gobies in the profundal zone of Lake Ontario switched from consuming primarily dreissenid mussels to Mysis relicta with increasing density of Mysis (and water depth) even though dreissenids of appropriate edible size were present. Furthermore, the importance (e.g., percentage biomass) of dreissenid mussels in round goby stomachs may be artificially inflated (Barton et al. 2005) due to the longer retention time of dreissenid shells relative to other soft-bodied prey items.

It is possible that the weight loss exhibited by round gobies fed dreissenids in this experiment may, to some degree, have been due to too few mussels provided to each individual. However, dreissenids were never completely depleted in any of the treatments at any time during the experiment. Although there were insufficient numbers of very small mussels (<5 mm) provided to the round goby, 78% of the mussels added were between 5.1 and 15 mm (the two SCs that were most consumed). There were ample mussels within these SCs remaining at the end of the study, and on a daily basis these were the SCs that were replaced at the greatest rate. Although it is difficult to compare the size distribution of dreissenids we provided to natural size distributions, we believe that we probably provided more small dreissenids than round gobies would typically have access to. Data presented by May and Marsden (1992) indicate that the majority of dreissenids in the vicinity of Rochester, New York, Lake Ontario were less than 20 mm, with a peak around 12 mm (consistent with the distribution we provided) but at Cape Vincent, New York they ranged from less than 8 to 40 mm with a small peak at approximately 16 mm and a larger peak from 28 to 36 mm, much larger than the distribution we provided. Similarly, Mills et al. (1993) sampled several sites across Lake Erie and Lake Ontario and the upper St. Lawrence River and found mean shell lengths ranged (among sites) between 4.0 and 29.2 mm, minimum shell lengths ranged from 2 to 12 mm, and maximum shell lengths ranged from 17 to 36 mm. These previous distributions were estimated prior to the round goby invasion; thus, we would likely expect current size distributions to be skewed even further toward larger individuals, assuming that round gobies select smaller dreissenids. Therefore, we believe that we very likely provided more small dreissenids than they would generally have available in the wild.

Search time and capture efficiency were also likely not responsible for the observed weight loss. Because of the small bottom area of each aquarium, search time for dreissenids, including the most desirable sizes, should have been minimal, especially given that dreissenids are stationary prey. Also, since the dreissenids were added on a regular basis, few had sufficient time to become attached to the substrate. Therefore, round gobies did not need to expend energy detaching clumps of mussels. This is in direct contrast to the fish diet treatment, where capture of this prey certainly had the lowest success rate and the highest energy expenditure. Regardless of the effort exerted to capture fish prey, round gobies in this treatment had the highest observed growth. Thus, it is not reasonable that searching for a stationary prey item (i.e., dreissenids) within a restricted area would severely reduce fish growth.

A strictly zooplankton diet also resulted in round goby weight loss. Cladoceran zooplankton, similar to the species that we offered to round gobies, weigh no more than 0.25 µg each (Haberman et al. 2007), so in order to obtain 1 g of cladoceran biomass, round gobies would need to consume at least 4,000,000 individuals. Therefore, it is highly unrealistic to believe that under most circumstances (i.e., excluding local seasonal hatches) round gobies could survive and grow only on zooplankton. As with dreissenid mussels, round gobies probably consume zooplankton only when other more profitable prey are unavailable or when zooplankton hatches produce extremely high densities.

The results indicate that both small fish and chironomids were the most beneficial to round goby growth. The negative growth observed for round gobies consuming stationary dreissenid mussels is further emphasized considering that round gobies fed fish gained weight after pursuing and capturing a highly mobile prey item. The mean growth exhibited by round gobies consuming chironomids in this experiment may be artificially high compared to natural conditions, since we were unable to obtain a large supply of live macroinvertebrates. Therefore, round gobies did not spend energy sorting through the substrate searching for live chironomids as would occur in natural conditions. Regardless, these findings support our hypothesis that round gobies tend to consume dreissenids in order to fill the void when fish and non-mollusk macroinvertebrates are in low densities or are unavailable.

Acknowledgements

This project was funded by a Faculty Research and Creative Endeavors (2008) grant to DGU and BAM. We thank the Central Michigan University Biological Station on Beaver Island, with special thanks to Jim Gillingham and Doug Tilly for providing logistical support and laboratory space. We also thank Jacqueline Upshur and Molly Ramsey for laboratory and field assistance and Alison Coulter for reviewing early drafts. This is a contribution from the Central Michigan University Biological Station.