The invasive ctenophore Mnemiopsis leidyi has a strong reputation as a threat to fish stocks. Apart from competitive relationships between M. leidyi and fish larvae, direct predation by the ctenophore on both eggs and larvae is considered an important factor linking ctenophore and fish populations. We therefore estimated abundances of both the ctenophore and its potential prey in the spring of 2008. No significant correlations were detected between ctenophore numbers and the abundance of fish eggs. We further carried out stable isotope analyses to investigate the trophic position of M. leidyi. Carbon and nitrogen stable isotope signatures of three potential prey groups (fish eggs, small plankton and larger plankton) showed that M. leidyi primarily feeds on plankton, while fish eggs are of minor importance. Mnemiopsis leidyi was located roughly one trophic level above the indigenous ctenophore Bolinopsis infundibulum, whereas its trophic position was more similar to another native ctenophore, Pleurobrachia pileus. A feeding selection experiment, with fish eggs and copepods offered in the same proportion, corroborated these findings. Mnemiopsis leidyi ingested significantly more copepods; feeding on fish eggs was not significantly different from zero. Based on these experiments, we conclude that in the North Sea, M. leidyi has no serious potential as a direct predator of fish eggs, but individuals of this species might compete for food with larval fish as well as with the native ctenophore P. pileus.
Invasive species are a threat to the ecosystems worldwide. These species can disturb ecosystem structures and even cause economic damage (Chandra and Gerhardt, 2008). The ctenophore Mnemiopsis leidyi (AGASSIZ 1865) invaded the Black Sea (Vinogradov et al., 1989), the Mediterranean (Uysal, 1993) and the Caspian Sea during the last decades of the 20th century (Shiganova, 2000). A few years ago, this lobate ctenophore was discovered in the Baltic Sea (Javidpour et al., 2006) and the first studies are now underway to assess the effects of these recent invasions (e.g. Haslob et al., 2007; Lehtiniemi et al., 2007; Javidpour et al., 2009, 2010; Gorokhova and Lehtiniemi, 2010). In the North Sea, where it was first recorded almost at the same time (Faasse and Bayha, 2006; Boersma et al., 2007; Tendal et al., 2007), its ecological impact is still completely unexplored.
The invasion of M. leidyi has been connected causally to a sharp decline in fish populations and fisheries yields in several areas, but there is still considerable discussion on whether M. leidyi caused the declines in population densities, or whether the depleted fish stocks as a result of overfishing allowed a successful invasion and establishment of the ctenophore. One of the reasons for this ongoing discussion is the conflicting information whether M. leidyi actually consumes fish eggs and/or larvae or not. Some authors have reported that this species preys on fish eggs (Purcell, 1985; Monteleone and Duguay, 1988; Purcell et al., 1994; Shiganova, 1998; Haslob et al., 2007), whereas others state that fish makes only a minor contribution to the diet of M. leidyi (Burrell and Van Engel, 1976; Javidpour et al., 2009) and account for only up to 4% of the gut content in natural catches (Mutlu, 1999). Nevertheless, direct predation on ichthyoplankton is considered an essential part of M. leidyi's capability to harm fish populations (Shiganova and Bulgakova, 2000; Purcell et al., 2001; Oguz et al., 2008). As such, there is an urgent need to investigate the potential role of M. leidyi in recently invaded areas such as the North Sea.
Without detailed knowledge about M. leidyi's diet preferences, we have only correlational evidence of predator densities with egg densities (Shiganova, 1998; Shiganova, 2000). Stomodaeum content analysis of free-living animals (e.g. Williams and Johnson, 2005) or laboratory trials are common approaches to investigating food selectivity (e.g. Monteleone and Duguay, 1988; Cowan and Houde, 1993). Mnemiopsis leidyi preys on various types of zooplankton (Burrell and Van Engel, 1976; Mutlu, 1999) and feeds selectively on slower moving organisms (Williams and Johnson, 2005; Javidpour et al., 2009). So, non-moving eggs are potentially a very attractive prey, and positive selection for fish eggs has indeed been reported for many gelatinous predators (Purcell, 2001).
Besides numerical techniques such as abundance estimations and feeding experiments, stable isotope analysis is a useful approach in the analysis of dietary composition (Pitt, 2008). Stable isotope signals integrate over time and allow the tracking of an organism's predominant diet independently of the instantaneous gut content or current feeding preference (Peterson and Fry, 1987). By measuring the relative accumulation of 15N to 14N isotopes in the tissue, the trophic level of an organism can be estimated (De Niro and Epstein, 1978). In contrast, photoautotrophs accumulate 13C in a species-specific proportion and their carbon isotope ratio is passed almost unmodified to the consumers feeding on them. Carbon isotope composition can therefore be used to identify the original source of energy in a food chain (De Niro and Epstein, 1978). Interestingly, only a few publications have utilized stable isotope analysis to reveal trophic interactions of ctenophores (e. g. Toyokawa, 1990; Montoya, 1991).
Hence, the aim of our study was to assess the potential predatory impact of M. leidyi on the pelagic ecosystem of the North Sea. We pursued this aim by a combination of field surveys and experiments: we carried out a study in which we estimated the densities of ctenophores and fish eggs, carried out direct feeding experiments with M. leidyi, and investigated the trophic positioning of field-caught gelatinous zooplankton, using stable isotope analysis.
Ctenophores and fish eggs were caught from the Helgoland Roads station (54°11.18′N, 07°54′E) from a small research vessel at a speed of 1.5 km, from January to July 2008. A CalCOFI ring trawl (500 µm mesh size, aperture 100 cm, length 400 cm, equipped with a flowmeter and solid net bucket) was used, sampling a volume between 110 and 617 m3. Sampling was discontinued in July as around this time, densities of fish eggs and larvae drop to zero around Helgoland (Malzahn and Boersma, 2007). The samples were transferred to the laboratory within 2 h. To obtain distinct plankton fractions as potential prey for the ctenophores, plankton nets with 55 µm and 150 µm mesh sizes were towed through the water column and the samples were sieved by 150 µm and 1000 µm, respectively, resulting in a small (55–150 µm) and a large (150–1000 µm) zooplankton fraction. These fractions were used for the isotopic analysis of the base of the heterotrophic food chain. Fish eggs, fish larvae and ctenophores were removed from each fraction, to avoid contamination of isotopic signatures. The samples (around 3.5 mg dry weight) were rinsed in purified water and directly pipetted into 9 × 5 mm tin capsules. These were dried for 3 days in a drying chamber at 60°C (in the case of the smaller fraction) and stored folded in a desiccator for later analysis. The larger plankton fraction was pipetted into a 2 mL reaction tube, frozen at −80°C for up to 5 months and then freeze-dried for 24 h to constant weight. Ctenophores (minimum length without lobes 13 mm, maximum length 48 mm, mean 23 mm) were frozen at −80°C for up to 6 months, freeze-dried for at least 12 h to constant weight and subsequently ground to fine powder using a plastic pestle (Eppendorf). Before weighing on a microbalance (Sartorius), the samples were kept for up to 3 h at 60°C in a drying chamber. For longer storage, they were kept in a desiccator. In the case of humidification during weighing, samples were put into the drying chamber for another hour. Dried and encapsulated samples were analysed by mass spectrometry at the UC Davis Stable Isotope Facility, California. δ values were calculated by ; R is the ratio of the amount of the heavy to the lighter isotope, Rx indicates the sample values and Rstd the standard ratio. Zero is defined by the isotope composition of air standard (δ15N) and PDB (δ13C). The long-term standard deviation of the analysis was 0.3‰ for carbon and 0.2‰ for nitrogen. A three-source mixing-model (small zooplankton, large zooplankton, fish eggs) with two stable isotope markers (15N and 13C) was then used to estimate the diet composition of gelatinous zooplankters in the samples (e.g. Phillips et al., 2005), under the assumption that 15N is enriched by 2.2‰ per trophic level and 13C by 1‰ (Minagawa, 1983; McCutchan, 2003). In cases where the mixing model yielded fractions of the diet lower than 0 or higher than 1, these were set to 0 and 1, respectively, and the total of the three resources scaled to 1.
For the food selection experiments, M. leidyi were kept in a planktonkreisel (Greve, 1968, 1970) with a 60 cm diameter and a vertical current with a flow rate 2 L min−1. These animals were fed daily with 1- to 2-day-old Artemia salina nauplii for at least 5 days to assure a uniform feeding status. Before being used in the experiment, ctenophores were starved for 18–20 h. The animals were transferred individually into wide-necked 2 L polyethylene terephthalate glycol bottles, which were then filled bubble-free with GF/F filtrated sea water (salinity 32) and were slowly turned end over end at one rotation per 50 s on a plankton wheel (Sullivan and Gifford, 2007) to ensure a homogeneous distribution of the prey items. Feeding duration for 13 specimens of a size between 22 and 30 mm (exclusive of lobes) was 3 h, 2 specimens of 35 and 37 mm with particularly long lobes were allowed to feed for 2.5 h. Preliminary experiments revealed that these durations were sufficient to avoid limitation effects and enough to achieve an optimal reduction of prey items by 40% (Gifford, 1993). After acclimatization for 1 h, the experiment was started by adding 100 copepods (max. 1 mm in length) and 100 planktonic fish eggs. Control treatments did not contain ctenophores (n = 5). Fish eggs were obtained from CalCOFI catches at Helgoland Roads (Limanda limanda LINNÉ 1758 eggs of ca. 0.8 mm and some Pleuronectes platessa LINNÉ 1758 eggs of ca. 1.5 mm in diameter) and from a specimen of Ciliata mustela (LINNÉ 1758, a bottom-dwelling Gadiform with ca. 0.8 mm in egg diameter), which spawned in captivity. The eggs from C. mustela were not older than 48 h and stored at 8°C so that most of them stayed alive and neutrally buoyant, as the other eggs did as well. The experimental temperature was 8°C, with dim light. Feeding was terminated by removing the ctenophore. The contents of the feeding chamber were then poured through a 150 µm sieve, the remaining prey items were stained with Lugol's solution and counted in a Bogorov counting chamber. Samples were analysed using the equations developed by Frost (Frost, 1972). Grazing rates were calculated per hour and analysed using a paired t-test. Statistics were done using the software package Statistica 7.0 (StatSoft).
Mnemiopsis leidyi densities did not exceed 29 ind. 100 m−3 from January to July 2008 (Fig. 1), and maximum M. leidyi abundance was reached in summer. Mnemiopsis leidyi co-occurred with other ctenophore species, such as Pleurobrachia pileus or Bolinopsis infundibulum. Planktonic fish eggs and M. leidyi specimens >5 mm coincided regularly. Fish egg abundances fluctuated widely and showed several peaks of up to 884 eggs 100 m−3. Densities of M. leidyi and of fish eggs did not show any significant correlations (linear regression analysis, r2 = 0.06; P > >0.05), nor did the fish egg densities correlate significantly with total ctenophore numbers.
Mnemiopsis leidyi grazing rates in the laboratory experiment on the two equally abundant prey types (copepods and fish eggs) were significantly different from each other, and much higher for the copepods. The average grazing rate on copepods was 0.19 h−1 (SD 0.09, n = 15), while grazing rates on fish eggs were significantly lower at 0.003 h−1 (SD 0.05, n = 15), paired t-test, P << 0.001. In fact, grazing rates of M. leidyi on fish eggs were not significantly different from zero (t-test).
Samples for stable isotope analysis were collected over the whole sampling period when enough individuals were available (a total period of 4 months for M. leidyi and P. pileus, of 3 months for fish eggs and of 1 month for B. infundibulum). All stable isotope signals fluctuated over time (Figs 2 and 3), but these fluctuations were more or less parallel so that the data range of each group did not overlap with another one. The planktonic fraction values varied most in their stable isotope signals (Fig. 2A). The ctenophores showed consistently different δ15N signal values in the first part of our sampling season, until a strong decrease in δ15N values of P. pileus in April. After that, it held the same trophic level as M. leidyi. Based on the stable isotope signals and the mixing models, the main diet of all three ctenophores was the large zooplankton fraction. The time series of the inferred diet shows that there might be substantial predation on fish eggs by the indigenous P. pileus, especially in March/April. In contrast, fish eggs play only a minor role in the diet of M. leidyi, with only two sampling occasions in April–May where, based on the stable isotope signals, fish eggs may account for around 10% of the diet of M. leidyi (Fig. 4). The average values of the stable isotope signals corroborate this (Fig. 5). In fact, based on these averages, fish eggs did not occur in the diet of M. leidyi at all. Bolinopsis infundibulum has a very unusual isotope signal compared with its “potential” prey. It has the same δ15N signal as the plankton fraction, and the δ13C-signal is depleted compared to the plankton signals. Hence, there has to be one or more food source, which we did not sample, with lower δ15N and δ13C values to produce the isotopic composition of B. infundibulum.
A negative correlation between potential prey and predator densities could imply direct predation. This negative correlation was indeed found between anchovy eggs and M. leidyi in the Black Sea (Shiganova and Bulgakova, 2000) and between copepods and M. leidyi in Narragansett Bay, USA (Kremer, 1994). For the North Sea, our abundance estimations did not show this clear relationship. Even though our net sampling method is an appropriate method to sample sensitive ctenophore species, we surely did not overcome another typical problem of quantitative gelatinous plankton sampling: as Purcell (Purcell, 2009) points out, the uneven distribution and temporarily rare occurrence of ctenophores would have required a greatly extended sampling strategy, and, as it is, our results rather represent a rough estimation of ctenophore abundances. Due to ambient hydrography as well as the tidal currents, the distributions of fish eggs around Helgoland are also expected to be non-uniform and patchy. In fact, in our observations, fish egg densities showed strong fluctuations, which are probably also a result of patchy spawning events, or spawning by different species (Malzahn and Boersma, 2007). Since we did not identify the fish eggs to species, we cannot evaluate whether a certain fish species might be particularly endangered by M. leidyi, but the strong fluctuations in egg densities combined with the much lower ones in the ctenophore densities suggest that the driving factor of the changes in fish egg densities is not predation. In addition, M. leidyi densities were very low compared to those reported from the Black Sea (up to 180 specimens per m3, Mutlu, 1999) or the eastern US coast (peaks of 50 specimens per m3, Kremer and Nixon, 1976), so that a serious predatory impact is even more unlikely here.
Many jellyfish, especially scyphozoans, have been observed to prefer fish eggs (Fancett, 1988; Sullivan et al., 1994). Mnemiopsis leidyi feeds selectively too (Javidpour et al., 2009) and applies two different prey catching mechanisms which work simultaneously, adapted to fast (copepods) or low swimming abilities (eggs) of the prey, respectively, thereby preferring slow-moving prey items (Larson, 1987; Waggett and Costello, 1999). These mechanisms enable M. leidyi to feed on mobile and immobile prey items at the same time. Capture success for mobile copepods such as Acartia sp. is quite high with 74% of all encounters (Costello et al., 1999), and the capture success for immobile fish eggs should even be higher, as these lack escape abilities. However, in our study this was not the case. Fish eggs were grazed significantly less than fast moving copepods; in fact, grazing rates on fish eggs were not significantly different from zero in our experiment. Container effects caused by the relatively small 2 L jars used here might have limited ingestion (as suggested by Purcell, 2009), but it seems unlikely that this would have affected the different prey items differently. Hence, absolute predation rates on the different prey items might be incorrect, but the comparison between prey types is still valid. Smaller individuals of M. leidyi might feed differently from the older ones used in our experiments (Rapoza et al., 2005), but it seems unlikely that smaller ones eat the larger (eggs) prey. Thus, in our experiment, M. leidyi did not feed on fish eggs.
Although B. infundibulum and M. leidyi are similar in size, in prey-catching strategy and occurred more or less simultaneously in the investigated area, based on the stable isotope data they did not compete for food. Their stable isotope signatures differed completely, and hence we conclude that they occupy completely different trophic niches. However, we have to consider our outcome with reservation. In scyphozoans, isotopic fractionation is different in ectodermis, mesoglea and entodermis (Pitt, 2008), and analysing whole animals as pooled samples will have lead to an averaging of the isotopic signatures, which if animals have strongly different turn-over rates might lead to misinterpretation of the exact trophic positioning. Nevertheless, B. infundibulum seems to feed on a different ultimate carbon source in the food web. The carbon signal is clearly more depleted than the carbon signals of both the two planktonic fractions as well as the two other ctenophores under investigation. Moreover, of the three ctenophores in this study, its location is lowest in the food web, and we obviously did not sample all of the potential food sources for this species. We observed B. infundibulum on only a few occasions. Hence, it could well be that these individuals originated from a different area where they had been feeding on other more depleted carbon sources. Alternatively, they may actually be feeding lower in the food chain (algae, protozooplankton; see also Sullivan and Gifford, 2007), which we did not sample for this study. In contrast, M. leidyi probably prefers metazoan prey, represented by the sampled plankton fractions (55–150 and 150–1000 µm). The same is true for P. pileus, whose prey spectrum overlaps with that of M. leidyi. Both P. pileus and M. leidyi have a strong preference for the large zooplankton fraction throughout the year. In contrast to M. leidyi, the diet of P. pileus may consist of fish eggs to some extent at specific times of the year, and based on this we have to conclude that if any ctenophore species is a threat to the fish populations around Helgoland, it is the endemic species Pleurobrachia pileus and not the invasive Mnemiopsis leidyi. It remains to be seen whether these species truly compete for resources and, if so, if this competitive interaction will have a winner, or whether the two species are going to co-occur by preying on different plankton groups characterized by size or mobility as proposed by Costello and Coverdale (Costello and Coverdale, 1998).
Fish eggs, with their clear maternal signal of feeding on higher trophic levels, are a minor food source for Mnemiopsis leidyi. Whether this is the result of active discrimination against fish eggs in the feeding process, which means the rejection of fish eggs in the case of a contact, as suggested by our feeding experiments, a limited seasonal overlap (many fish spawn very early in the year, Malzahn and Boersma, 2007) or simply by the dilution of the fish eggs by many more zooplankton prey remains unclear. Most likely, as even in very high density experiments, fish eggs were not consumed by M. leidyi, the ctenophores were rejecting fish eggs in favour of copepods.
In conclusion, our findings give important indications that fish eggs might not be an important food source for M. leidyi in the North Sea, nor are they for the indigenous ctenophore B. infundibulum, but more so for P. pileus. Food competition of the invasive M. leidyi with planktivorous fish and fish larvae is a more likely, but yet unexplored threat to North Sea fish stocks. However, densities of M. leidyi were low in comparison to areas such as the Black Sea, where values of up to 30 400 ind. 100 m−3 were observed (Vinogradov et al., 1989). Hence, we propose that M. leidyi is at present unable to affect the ecosystem in the North Sea in a similar way as it did in the Black Sea, and as a result might fit into the current food web without the catastrophic impacts observed in other invaded seas.
This study was partly financed by German Science Foundation (DFG AB 289/2-1). This study is part of the AWI Food Web project.
We thank the crew of the research vessel Aade for supplying us with the samples, and Katherina Schoo, Florian Hantzsche, Daniel Schütz, Simon Hillbrenner and Carsten Dittmayer for their help in the laboratory and fruitful discussion. We also thank three anonymous reviewers for their many helpful comments. This study complies with German law.