Introduction

Elton (1927) formulated nearly all principles on which modern ecological theory is based, and described complex animal communities as interconnected food chains that form food webs. He further outlined that smaller animals are more abundant than larger ones to represent the “pyramids” ordered in increasing organismal size so that material flows through the community from small to larger organisms. Recently, those two facets of trophic ecology (i.e. predator–prey interactions and energy flows) have attracted more attention within the fast-growing field of ecological research: Trophic ecology emerged as a promising ground in an attempt to merge complex concepts such as natural selection, ecological networks, ecological stoichiometry and ecosystem metabolism (see Garvey & Whiles, 2017). This trend has been fostered by considerable improvements and developments in methods to observe feeding interactions and quantify the assimilation and transfer of organic matter (Majdi et al., 2018).

Nevertheless, meiofaunal communities are composed of inconspicuous assemblages of microscopic animals needing advanced taxonomical expertise and specific methodologies to examine their feeding interactions. This creates regional biases and knowledge shortfalls especially concerning species taxonomy (i.e. Linnean shortfall), abundance/biomass patterns (i.e. Prestonian shortfall), functional traits (i.e. Raunkiaeran shortfall) and feeding interactions (i.e. Eltonian shortfall) (see e.g. Fontaneto et al., 2012; Hortal et al., 2015). Although current research grants do not really promote thorough taxonomic works on species level, technological improvements of microscopy- and molecular-based techniques may reduce costs and/or offer alternatives to unravel large-scale diversity of species and their morpho-functional traits (Fonseca et al., 2018; Neury-Ormanni et al., 2019; Schenck & Fontaneto, 2019). Missing data about the functional traits and feeding interactions of the meiofauna will persist but there are some empirical research that highlight the role of meiofaunal-sized animals within the functioning of ecosystems (e.g. in Freshwater ecosystems: Schmid-Araya et al. 2002a, b, 2016; Majdi et al., 2012b; Majdi & Traunspurger, 2017. Marine ecosystems: Nascimento et al., 2012; Bonaglia et al., 2014; Braeckman et al., 2018. Soil ecosystems: Pausch et al., 2016; Maboreke et al., 2018).

Here, we review the experimental approaches and analytical tools useful to study feeding interactions and energy fluxes of the metazoan meiofauna. Further, we summarize the limitations of these methods, their advantages, and their potential while assessing meiofaunal diets. Although we mainly aim to highlight methodologies adapted to freshwater samples, we propose application examples from freshwater, marine and soil environments because (a) of the scarcity of examples from the freshwater only, and (b) most methodologies can be used across different habitat types.

Indirect evidences

Correlative evidences of trophic linkages

Many meiofaunal studies are primarily descriptive, reporting spatial and/or temporal variations of abundances. Reviewing 93 studies dealing with the trophic ecology of freshwater nematodes, Majdi & Traunspurger (2015) found that 42% of the studies suggested the existence of a trophic linkage through correlative evidences. For instance, several studies have found positive correlations between Chlorophyll-a concentration and the abundance of meiofauna in epilithic biofilms, suggesting a trophic linkage (Peters & Traunspurger, 2005; Esser, 2006; Gaudes et al., 2006; Vidakovic et al., 2011; Majdi et al., 2012a; Schroeder et al., 2012; Weitere et al., 2018). However, field patterns are not necessarily conclusive about trophic interactions and energy flows because correlations between species can be obscured by many factors (e.g. successional state of the assemblages, constraints of the abiotic environment). Furthermore, another problem with inferring trophic interactions from community structure correlatives is that the approach is sensitive to the Linnean shortfall, a shortfall that often characterizes meiofaunal and microbial datasets. For example, the meiofauna is often loosely determined to the phylum-level, while microbial communities are estimated through raw counts or biomarkers. Sometimes, one meiofaunal phylum is described up to genus- or species-level (often nematodes), but studies examining the spatiotemporal dynamics of all meiofaunal species in a freshwater habitat are very rare (but see Reiss & Schmid-Araya, 2008). Moreover, it is also rare that microbial biomass, or any other microbial biomarkers are measured together with the abundance and biomass of meiofauna. Hence, inferring trophic interactions from correlations in the community structure is also sensitive to the Prestonian shortfall (but see Traunspurger et al., 2012).

Provided that coherent reference databases are available, molecular-based methodologies may dramatically improve the identification of meiofaunal assemblages from field samples and thus overcome the Linnean shortfall, and in some cases, the Prestonian shortfall when semi-quantitative or quantitative methodologies are used (see e.g. Schenck & Fontaneto, 2019). Molecular-based methodologies may provide a more detailed account of species’ occurrences in field samples, and thus may be helpful to suggest the most probable trophic interactions. For example, the detection of consistent associations between meiofaunal species or between meiofaunal species and microbial species may indicate potential trophic or symbiotic associations. Following this approach, Rzeznik-Orignac et al. (2018) recently observed associations between bacteria and nematodes in deep-sea canyons. They reported significant correspondences between microbial communities associated to different ecosystem functions and some species of bacterivorous nematodes (e.g. Nitrospirales bacteria correlated with Daptonema spp., Deltaproteobacteria with Dorylaimopsis spp., sulphate-reducing bacteria with Terschellingia spp.). Correlative evidences may ideally be merged with other lines of evidences such as morphology of the buccal cavity, stable isotopic signatures or composition of the microbiome to confirm trophic linkages (Derycke et al., 2016; Vecchi et al., 2018).

Trait-based inferences

Species’ biological traits help describing biodiversity by focusing on functional features instead of taxa (Gravel et al., 2016). In turn, the data entailed on matrices of biological traits are used to infer the ecological functions carried out by a community of species. Indeed, the usefulness of trait-based inferences depends upon the accuracy of biological-traits matrices available (Jardim et al., 2016). Some biological traits need empirical knowledge on species behavior and life history, and thus they often missed meiofaunal-sized organisms because their autecological data are difficult to measure in the field, but can be examined using laboratory populations (e.g. Fueser et al., 2018; Majdi et al., 2019). There are, however, three ways to overcome the Raunkiaeran shortfall:

(1) It is possible to fill sparse trait databases using models that consider the mechanisms causing missing data (e.g. Rubin, 1976). Imputation models considering species’ phylogenetic information seem to be the most relevant because related species have more probabilities to share common traits (Guénard et al., 2013; Jardim et al., 2016). Nevertheless, it is important to consider the pattern of missing data and then to use appropriate modelling methods in order to reduce potential misinterpretations when using imputed trait databases only (Jardim et al., 2016).

(2) Complex databases of traits can be developed from literature and from microscopic observations of meiofaunal-sized organisms (Ristau et al., 2015; Neury-Ormanni et al., 2019) by classifying species based on a large set of morphological, behavioral, life-history and feeding traits (e.g. body size and shape, characteristics of the feeding apparatus and of locomotory organs, reproduction mode). When combined to estimates of standing stocks, one can infer the ecological role of the community whenever one can find correlations between the prevalence of specific traits and the trophic status of an ecosystem (e.g. lake eutrophication, see Ristau et al., 2015) or the magnitude of an ecosystem process like primary production (Neury-Ormanni et al., 2019). Developing standardized measurement protocols leading to coherent databases of traits is a challenging task for disparate species assemblages such as the meiofauna, but there are now extensive repositories of traits including some meiofaunal species (e.g. Degen & Faulwetter, 2019). Furthermore, advances in cell-sorting methodologies and image treatment automatization have the potential to foster the inclusion of meiofauna in trait-based ecology by providing a high-throughput of morphological data from a known community (e.g. Kydd et al., 2018).

(3) Another way to reduce the Raunkiaeran shortfall is to focus on a subset of traits in a relatively ubiquitous and numerically dominant group of the meiofauna like nematodes. As a preeminent functional feature of nematodes, the feeding type (i.e. a quantitative trait considering the size, morphology and anatomical structures of the digestive tract and of the mouth cavity) may give insights about diet (Yeates et al., 1993). This simplification has been widely used to counterbalance the lack of other species-level information on feeding behavior. For instance, nematologists have developed coherent feeding-type classifications for terrestrial, marine and freshwater nematodes (see Wieser, 1953; Yeates et al., 1993; Traunspurger, 1997). In the case of meiofauna where hundreds of species with similar feeding types may coexist in small patches, it is however questionable whether feeding-type classifications alone can really help to better understand the diet spectrum and its resource dependence (Schmid & Schmid-Araya, 2002). Furthermore, experiments with bacterivorous nematodes suggest that species expected to occupy the same trophic niches do not really seem that redundant (De Mesel et al., 2004, 2006; dos Santos et al., 2009; Gingold et al., 2013; Gansfort et al., 2018). Also, some meiofauna possess suction-feeding stylets (e.g. tardigrades, dorylaimid nematodes, water mites), protruding pharynges (some catenulid microturbellarians) or mandibles and ligula (e.g. tanypod chironomids) enabling them to feed on a wide range of prey. In those cases, feeding type (and body size) may not always conform to patterns of trophic positioning. Nevertheless, it is conceivable that feeding-type structure may help to infer the most probable interactions occurring in a community: Recently, Sieriebriennikov et al. (2014) used the functional diversity of nematodes as a useful tool for the diagnoses of soil food webs. Also, Traunspurger et al. (2019) found a correspondence between the abundance of nematodes with large mouth cavities and the trophic state of lakes. However, we recommend that inferences using feeding types should be carried out cautiously or should also include direct measures of diet such as stable isotopes (Estifanos et al., 2013) and gut content analyses (Kazemi-Dinan et al., 2014).

Measuring trophic interactions and their consequences

Observation of feeding

The most straightforward and oldest approach to study feeding interactions relies undoubtedly in observations of the feeding of living animals (Giere, 2019). For meiofaunal-sized organisms, most observations need to be performed under a microscope in the laboratory, thereby introducing inevitable bias in comparison to field observations of the feeding behavior of large animals. Nonetheless, laboratory observations also allow to measure feeding responses under standardized conditions and thus, to test the effects of variables such as temperature, water velocity or food type, on feeding rates. A classical example of laboratory observation of feeding is the study of Duncan et al. (1974), producing one of the few experimental measures of bacterial-grazing rates by the free-living freshwater nematode Plectus palustris de Man, 1880. In their study, Duncan et al. (1974) used a mixture of observations under the microscope (counting pumping rates of the oesophageal bulb) with measures of 14C assimilation through the consumption of radio-labelled bacteria. They estimated a mean grazing rate of 5000 cells min−1 and a gut-filling time in the range between 3 and 10 min. They concluded that P. palustris females could daily consume on average 650% of their body weight, which was similar to the 1000% found for the pelagic rotifer Brachionus plicatilis Müller, 1786 by Doohan (1973). Later, Moens & Vincx (1997) successfully observed the feeding of many species of free-living marine nematodes using an inverted microscope and agar spot plates with tiny patches of plant-detritus or sediment. They were able to observe the consumption of food items (such as detritus, bacteria, diatoms, protozoa, other nematodes and meiofauna), confirming that only a few species were confined to a narrow diet (e.g. only bacteria). In freshwater, biofilms were grown directly in micro-flow chambers and observed live under a microscope, Esser (2006) estimated the individual grazing rates of one chromadorid nematode as 93 chlorophytes and 58 diatoms per day. Food-choice (aka cafeteria) experiments have been carried out successfully with meiofauna allowing to examine food-selectivity under various constraints as well as the role of volatile organic compounds operating as attractors towards a given food patch (Höckelmann et al., 2004; Weber & Traunspurger, 2013; Wilden et al., 2019). Furthermore, video-microscopy can be successfully applied in micro-flow chambers to continuously monitor behavior and grazing events (Weitere et al., 2018). Other recent developments in microscopic imaging have the potential to reveal directly the movements of animals within sediment columns (i.e. X-ray microtomography: Johnson et al., 2004) or the effects of micro-grazers on the 3-dimensional structure of microbial aggregates (i.e. confocal laser scanning microscopy: Neu & Lawrence, 2015).

Incubation and food clearance experiments

Incubation experiments are popular among meiobenthologists, because meiofaunal groups are generally well-suited for experimental work: From sandy/silty habitats, some groups can be easily retrieved (e.g. within a sediment core) and obtained in large numbers and directly exposed to different experimental treatments in the laboratory (e.g. Bell, 1988; Hägerbäumer et al., 2015). Another advantage of incubations is a relatively low degree of invasiveness, and the possibility to measure assemblage- to ecosystem-level effects, thus, determining meiofaunal feeding in rather realistic context of multispecies interactions. Various enclosure/exclosure experiments have been designed to examine the effects of the presence of meiofauna on microbial abundances and processes: (a) static or flow-through sediment cores of different sizes (e.g. Borchardt & Bott, 1995; Traunspurger et al., 1997; Nascimento et al., 2012; Bonaglia et al., 2014; Liu et al., 2017); (b) flow-through chambers of various sizes (e.g. Perlmutter & Meyer, 1991; Esser, 2006; Kathol et al., 2009), and (c) full-grown microbial biofilms exposed to different meiofaunal abundances (Mathieu et al., 2007; Peters et al., 2007; Liu et al., 2015; D’Hondt et al., 2018). Other experiments have measured prey disappearance as a function of prey number to examine the shape of predator–prey functional responses using turbellarians, oligochaetes, chironomids, tardigrades, nematodes or copepods as predators and algae, ciliates or nematodes as prey (e.g. Taylor, 1980; Goldfinch & Carman, 2000; Mohr & Adrian, 2000; Bergtold et al., 2005; De Troch et al., 2005; Hohberg & Traunspurger, 2005; Muschiol et al., 2008; Reiss & Schmid-Araya, 2011; Ptatscheck et al., 2015; 2017; Kreuzinger-Janik et al., 2018, 2019).

Incubation experiments have unravelled interesting facets of meiofauna–microbe interactions such as the apparent stimulation of bacterial (or algal) activity and nutrient/organic matter cycling with increasing meiofaunal densities (Traunspurger et al., 1997; Mathieu et al., 2007; Nascimento et al., 2012; Bonaglia et al., 2014; Liu et al., 2015, 2017; D’Hondt et al., 2018). Based on the above results, it is necessary to examine whether enhanced microbial activity may be due to either: (a) micro-bioturbation of meiofauna increasing the porosity of microbial mats to nutrients and light, (b) grazing pressure that optimizes growth rates of microbial populations or (c) a combination of both. Some studies also point out the indirect role of meiofaunal secretions products (mucus trails, faecal pellets or gut flora), stimulating microbial growth locally (Riemann & Helmke, 2002; Moens et al., 2005; De Troch et al., 2010; Hubas et al., 2010; Cnudde et al., 2011; Gaudes et al., 2013).

Gut content analysis

Gut content analysis (GCA) is a practice to characterize individual diet and to draw species-interaction networks. While GCA is commonly used for macroinvertebrates and fishes, very few studies have examined gut contents of meiofauna to infer its diet and position in food webs (but see e.g. Schmid-Araya & Schmid, 1995; Schmid & Schmid-Araya, 1997). A possible rationale might be the severe difficulties for non taxonomic experts to detect specific remains of meiofaunal organisms such as trophies of rotifers (e.g. Figure 1a, b), pharinges and claws of tardigrades, scales and spines of gastrotrichs, chaetae of small oligochaetes, pharinges and stylets of microturbellarians, mouth-parts of chironomids.

Fig. 1
figure 1

Illustrative examples of various items found (rotifers, ostracods and chironomids) in gut contents of meio- and macrofaunal consumers (Photos Schmid & Schmid-Araya). aDicranophorus luetkeni (Bergendal, 1892) (Rotifera) trophy found in guts of Tanypodinae (Chironomidae- Diptera) larvae. b Bdelloidea trophy found in the gut of the caddisfly Plectrocnemia conspersa (Curtis, 1834) (Trichoptera). c A series of ostracods eaten by the larvae of Macropelopia sp. (Tanypodinae, Chironomidae- Diptera). d The gut content of a tanypod chironomid larva that had eaten an orthocladinae larva and a tanypod chironomid

Full size image

Nevertheless, GCA proves especially useful to document the (a) dietary composition and predator–prey relationships (Schmid & Schmid-Araya, 1997), (b) food web topology (Schmid-Araya et al., 2002a, 2016) and (c) patterns of food web connectance (Schmid-Araya et al., 2002b). GCA also offers the possibility to assess prey size or developmental stage giving more detailed information on diet (Fig. 1). For instance, using GCA, Schmid & Schmid-Araya (1997) found that three species of stream predatory tanypods fed on numerous different meiofaunal prey (41 benthic rotifer species and 23 chironomid species, see e.g. Figure 1). The predatory tanypods switched diet from rotifers in early instars to chironomids and diverse other meio- and macrofaunal taxa in later instars, so growing tanypods expanded their upper size thresholds but continued to include smaller prey species in their diet. These larval tanypods consumed on average 1.32 prey individuals per predator type and prey consumption varied with sediment depth layer with higher prey consumption in the upper 20 cm of the streambed. GCA has also revealed that lotic food webs contain a high proportion of species (60–80% of total community) in the meiofaunal size range and many of these species belonged to the intermediate category (species with both prey and predator) improving the food web completeness (Schmid-Araya et al. 2002a, b, 2016). However, discrepancies in the taxonomic distinctness of the approach are inevitable since the assessment of diet is only based on prey items hard enough to persist in guts: Soft-bodied prey or tissues being poorly recognizable or quickly digested can be underestimated. Also, GCA represents a snapshot of diet, and thus, it should be highly replicated to provide more robust conclusions about diet spectrum.

It is possible to improve GCA using biomarkers in order to highlight hardly recognizable or minor food items present in the guts of meiobenthic animals. For example, the auto-fluorescence of ingested chlorophyll or carotenoid pigments can be detected in the guts of rotifers under a confocal laser scanning microscope (Mialet et al., 2013). Additionally, ingested biomarker pigments can be extracted from guts and quantified using HPLC to assess diet at population scales or in relation to the temporal availability of algae in the habitat (Buffan-Dubau et al., 1996; Buffan-Dubau & Carman, 2000; Majdi et al., 2012c).

The genetic techniques of polymerase chain reaction (PCR)-amplification of DNA may also contribute substantially to unravel certain trophic relations particularly on soft-bodied meiofauna. Prey DNA can be amplified from digestive systems, faecal pellets or whole organisms mostly of large-sized invertebrates where barcodes are available (King et al., 2008). In some cases under controlled laboratory conditions, it was possible to detect from copepods and from their faecal pellets: (a) a model alga offered to filter-feeding copepods (Nejstgaard et al., 2003) or (b) copepod prey given to carnivorous copepods (Vestheim et al., 2005). As another example, laboratory and field experiments by Heidemann et al. (2011) found that the so-called ‘detritivorous’ gamasid and oribatid mites carried out predation and scavenging on nematodes in soils. Also, PCR-based approaches were successfully applied to highlight the diet of soft-bodied meiofaunal predators such as marine microturbellarians (Maghsoud et al., 2014) and the extension of the method can open up a vast venue for trophic analyses. The next challenge has been the development of real-time quantitative PCR (qPCR) already demonstrated by Nejstgaard et al. (2008) in marine zooplankton. They found that a target gene of phytoplankton varied with growth phase while developing a qPCR assay to target gene fragments to estimate copepod feeding. Their field studies using gut contents derived from qPCRs, gut pigment and direct microscopy (GCA) demonstrated a semi-quantitative relationship. However, absolute estimates of gut content based on qPCRs were lower than expected, probably due to the digestion of prey-species’ nucleic acids. Moreover, by sequencing the microbial 16S rRNA gene, Derycke et al. (2016) confirmed the existence of species-specific microbiomes of three cryptic species of the nematode Litoditis marina (Bastian, 1865) Sudhaus, 2011. More strikingly, Derycke et al. (2016) found that the food offered to these cryptic species affected their microbiomes, illustrating different feeding behavior between the cryptic species. These molecular approaches to highlight diet are far from being common practice for meiofauna and many disadvantages persist. Among the downsides of these methods are (a) risks of contamination, (b) the potential bias by different DNA degradation dynamics during digestion for quantitative assessment, (c) DNA extraction protocols can strongly affect comparison of results and (d) potential uncertainties while disentangling ingested microbes from the resident gut microbiome, parasites or symbionts.

Detecting energy fluxes and assimilation

Measuring nutritional status and metabolism

Body mass indices, life-history traits related to fitness (e.g. development rate, survival, reproduction success) as well as elemental composition and energy storage have long been used by ecologists to infer the nutritional status of individuals, or to evaluate the nutritional quality of a food source (e.g. Jakob et al., 1996; Raubenheimer et al., 2009). Although these approaches may be sensible to other triggers than food (e.g. temperature, light, ontogeny), they have been used extensively in meiofaunal research under standardized conditions to document community- or population-level responses to (a) nutrient enrichment (Ristau et al., 2012; Gaudes et al., 2013), (b) determinations of optimal food concentrations for population growth (e.g. Schiemer et al., 1980; Robertson & Salt, 1981; Muschiol & Traunspurger, 2007; Schroeder et al., 2010; Weber & Traunspurger, 2013). It is also possible to evaluate the nutritional status and the metabolism through measuring the protein content/composition of animals. For this purpose, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) can generate protein mass spectra from small individual specimen (e.g. harpacticoid copepods as little as 350 μm length) or even from copepod body parts (Rossel & Martínez Arbizu, 2018). Besides, protein composition can be used to perform chemo-taxonomy, i.e. link specific protein fingerprints to species identity (Rossel & Martínez Arbizu, 2018, 2019). Another way to collect information on the nutritional status of meiofauna is to combine measures of life-history traits at the individual-level (e.g. using hanging-drop cultures) with the determination of lipid contents in single nematode individuals using coherent anti-stokes Raman spectroscopy (CARS) (Fueser et al., 2018). This approach is well-suited for translucent, meiofaunal-sized organisms unravelling the three-dimensional distribution of lipid droplets across tissues.

Lucas & Watson (2002) defined animal metabolism as the collective processes of anabolism and catabolism. As heterotrophs, animals achieve biosynthesis (anabolism) at the expense of energy from organic matter that is consumed and degraded (catabolism). In aerobic organisms, measurement of the rate of oxygen consumption gives a measure of the energy expenditure in the normal processes of the body and the metabolic rate per unit body weight of the intensity of its metabolism (Duncan & Klekowski, 1975). In direct calorimetry, the amount of heat produced by the animal itself is measured. Recently, Ruiz et al. (2018) used multichannel isothermal micro-calorimeters to estimate in real-time the metabolic rates of cladocerans as heat flow while offering stoichiometrically balanced or unbalanced algal food. Ruiz et al. (2018) observed that, to maintain their stoichiometric homeostasis, the animals fed stoichimetrically unbalanced food showed higher metabolic rates at the expense of growth. This experiment also demonstrated that real-time micro-calorimetry was a powerful technique to obtain precise measures of metabolic rates at the scale of meiofaunal individuals.

Indirect calorimetry involves the measurement of oxygen uptake (i.e. respirometry), which has long been the conventional approach to measure the metabolism of meiofaunal-sized invertebrates (e.g. Schiemer & Duncan, 1974; Schiemer, 1982; Herman & Vranken, 1988). However, in comparison to direct calorimetry, indirect measures of metabolism like respirometry can lead to under-estimations of metabolism (e.g. Walsberg & Hoffman, 2005; Burnett & Grobe, 2013), as respirometry only measures aerobic heat production while calorimetry measures the sum of aerobic and anaerobic catabolism. Moreover, respirometry does not consider the storage of CO2 as bicarbonates and biochemical synthesis in the cells of tissues, and closed respirometers can produce further biases as concentrations of gases change during closure time (Malte et al., 2016).

Bulk stable isotopic analysis

The stable isotope composition of carbon and nitrogen in bulk tissues is one of the most popular method using trophic tracers (Majdi et al., 2018). In the case of the meiofauna, marine studies considering stable isotopic analysis (SIA) have flourished over the last two decades (Giere, 2019). Meiofaunal SIA have only been reported recently in freshwaters (e.g. Majdi et al., 2012b; Estifanos et al. 2013; Schmid-Araya et al., 2016; Majdi & Traunspurger, 2017). Since stable isotopic composition is based on the result of the assimilation of a diet over relatively long periods, stable isotopes (and other assimilation tracers) have the immense advantage of quantifying fluxes of biomass over meaningful periods of time. However, a general disadvantage of assimilation tracers is that the presence of a tracer does not only reflect resource consumption but it also depicts the pathway from the resource to consumer’s tissues (e.g. selective digestion). Another major limitation of SIA while assessing the diet of meiofaunal assemblages is the sensitivity of the conventional elemental analyser-stable isotope mass ratio spectrometers (EA-IRMS), which may force to sort and clean a large number of individuals (Fig. 2). Indeed, this step is time-consuming and needs taxonomic expertise, but great numbers of living specimen can be retrieved from the field which may drastically reduce sorting-time (see e.g. migration/extraction procedures described in Wu et al., 2019). Moreover, the sample-size limitation can be overcome by reducing the volume of EA-IRMS columns after Carman & Fry (2002) as used for freshwater meiofauna by Schmid-Araya et al. (2016). Less helium is required to transport gaseous samples from the EA to the IRMS, thus samples are less diluted during transport, and detection limit can be reduced to 2 μgC (Carman & Fry, 2002); but it is also important to reduce potential sources of contaminations by using smaller tin cups (Carman & Fry, 2002). Langel & Dickmans (2014) also describe improvements of the conventional EA-IRMS (termed μEA-IRMS) leading to detection limits as low as 1 μgC and 0.6 μgN. Further use of the μEA-IRMS revealed stable isotopic signatures of soil nematodes at unprecedented genus and/or family levels, as few as fifteen nematode individuals being sufficient to get a reliable signal (Melody et al., 2016).

Fig. 2
figure 2

Range of individual body masses in a meiofaunal community dwelling sandy streams (after Majdi et al., 2016), and compatibility with the detection limits of conventional EA-IRMS platforms

Full size image

SIA and its improvements have the potential to provide quantitative estimates of elemental fluxes between different trophic levels of a food web at a relatively low-cost and effort. Nevertheless, SIA cannot assess the number of links in a food web for which direct observations, such as GCA or in conjunction with other approaches are better suited. It is possible (and probably advisable) to combine complementary approaches like SIA and food-choice experiments (Moens et al., 2013), or SIA and GCA to document the two facets of meiofauna’s trophic ecology (interactions and energy flow). For example, combining SIA with GCA, Schmid-Araya et al. (2016) showed that between 28 and 44.2% of the top consumers (having prey but no predators) were meiofaunal species. Consequently, it was assumed that (a) small-bodied taxa do not only occur low in the food webs, and (b) trophic positions do not necessarily increase with body size (Schmid-Araya et al., 2016; Majdi & Traunspurger, 2017).

Recently, meiobenthologists have used SIA to explore particular trophic connection between chemoautotrophic bacteria and some marine nematodes or harpacticoid copepods, showing strongly depleted δ13C signatures typical for methane- or sulphur-oxidizing bacteria (Van Gaever et al., 2009; Vafeiadou et al., 2014; Cnudde et al., 2015). This trophic connection has not been investigated for freshwater meiofauna, although freshwater macroinvertebrates (those dwelling in back-water pools, hyporheic zones or soft sediments in lakes), may assimilate substantial amounts of methane-derived C (e.g. Kohzu et al., 2004; Deines et al., 2007b). It is also not clear whether meiofauna uses chemotrophic-derived C by consuming bacteria or by hosting symbiotic chemoautotrophs, an issue that may be examined using approaches described below.

Stable isotopic probing and compound-specific stable isotopic analysis

Injections of small quantities of isotopically enriched sources can be applied to trace C or N pathways throughout the consumer network (referred to as stable isotope probing; SIP). These studies follow a time-course of the added tracer and have proven powerful to document microbial involvement in biogeochemical processes and to quantify trophic transfers from bacteria or micro-algae to consumers (e.g. Middelburg et al. 2000; Radajewski et al., 2000; Witte et al., 2003; Crotty et al., 2012; Majdi et al., 2012b). 13C-enriched sodium bicarbonate has been used to label photosynthetic products and the excess 13C can be tracked through time in meiofaunal consumers to quantify “green” trophodynamics (Middelburg et al., 2000; Majdi et al., 2012b; Estifanos et al., 2013). 13C-enriched glucose can be added to trace the “brown” pathways and the consumption of heterotrophic bacteria by meiofauna (e.g. Van Oevelen et al., 2006).

High-resolution imaging secondary ion mass spectroscopy (SIMS) can be combined with SIP to localize 13C enrichments at the cellular level (ca. 50 nm), thereby ruling out sample-size limitations in conventional EA-IRMS. Thus, there are new opportunities to study feeding selectivity, resource routing and processing in tissues, and intraspecific variability in feeding at the scale of the meiofauna (Musat et al., 2016). For instance, using this approach, Volland et al. (2018) demonstrated that the epidermis of a colonial ciliate coated with thiotrophic symbiotic bacteria processed rapidly the 13C-bicarbonate in the presence of H2S. The ciliate host then assimilated labelled organic carbon compounds within 25 min or through phagocytosis of ectosymbiotic bacteria over longer periods of time.

Compound-specific isotopic analysis (CSIA) determines the isotopic ratios of specific proteins, metabolites, fatty acids, or amino acids and is widely applied by microbial ecologists in combination with SIP to overcome sample-size and taxonomic limitations they also face (Jehmlich et al., 2016; Lueders et al., 2016; Wegener et al., 2016). The amino-acid CSIA received great attention; Coupled with SIP, ribosomal RNA or DNA have been used as integrative tracers. The amplification and barcoding of sequences that have assimilated the 13C or 15N-tracer allows the identification of species that have assimilated the labelled resource (more details reading e.g. Neufeld et al., 2007). The approach is promising to trace specific functional guilds of microbes (e.g. methanotrophic bacteria) and their fate as prey for consumers (Lueders et al. 2004). Moreover, CSIA of poly-unsaturated fatty acids (PUFAs) has also been used together with SIP to reduce errors prone to unpredictable trophic enrichment factors and intermolecular variability of isotopic signatures within the same food source (Bec et al., 2011). Deines et al. (2007a) incubated lake sediment cores with 13C-labelled CH4 and added chironomid larvae. After exposure, chironomids were starved and PUFAs profiles demonstrated the pathway of CH4 to chironomids via consumption of Methylobacter.

Fatty acids

Since four decades, poly-unsaturated fatty acids (PUFAs) have proved useful to trace fluxes and nutritional quality of food. The relevance of PUFAs as trophic biomarkers relies on specific values of their ratios found in resources and/or in producers (e.g. 16:1ω7/16:0 for diatoms). These are then transferred to the consumers’ tissues and can be detected assuming that consumers do not synthesize PUFAs de novo. The fact that some PUFAs can only be acquired by feeding on specific sources underlies the critical importance of (sometimes minor) food items for the development of animals, and thereby introduces the concept of nutritional quality. Although the use of PUFAs is popular in aquatic ecology (Arts et al., 2009), few studies have measured fatty acid profiles in meiofaunal-sized organisms (but see e.g. Caramujo et al., 2008; Leduc & Probert, 2009; Guilini et al., 2013; Braeckman et al. 2015; Wu et al., 2019). Indeed, the conventional gas chromatographs coupled to mass spectrometers (GC–MS) displays sample-size limitations as for EA-IRMS. Nevertheless it is possible to overcome sample-size limitations using comprehensive two-dimensional GC (aka GC x GC). The main advantage is that closely related molecules migrate over 2-dimensions and are thus better separated than with unidimensional GC, which also enable to detect smaller quantities. Akoto et al. (2008) describe a set-up (Direct thermal desorption-GCxGC-time-of-flight mass spectrometer) and a protocol compatible with meiofauna samples. Another way to reduce sample-size requirement is to couple GC to flame ionization detectors (FID) instead of MS (Hordijk et al., 1990; Boschker et al., 2001; Caramujo et al., 2008). Using GC-FID, Caramujo et al. (2008) were able to track PUFAs from a cyanobacteria or diatom diet to the tissues of an harpacticoid copepod using samples of 50–150 mature females. Copepods fed cyanobacteria showed reduced fatty acid content when compared to copepods fed with diatoms. Interestingly, copepod-fed cyanobacteria showed longer-chain PUFAs suggesting the existence of a mechanism by which fatty acids from a poor diet become elongated and desaturated by the freshwater harpacticoid. Rotifers and nematodes are also known to synthesize certain essential PUFAs de novo (Rothstein & Götz, 1968; Lubzens et al., 1985; Watts & Browse, 1999). More research is needed to determine which species can synthesize PUFAs de novo, and under what conditions (Bell & Tocher, 2009). If de novo synthesis is confirmed and widespread in meiofaunal organisms, it certainly has considerable implications for the way we should conceptualize sources of essential PUFAs in aquatic ecosystems.

Conclusion

Although mounting evidences support that meiofauna produce substantial amount of biomass and are thus important intermediaries in energetic transfers between freshwater biota, meiofaunal trophic relationships have long been, and still remain, a black box to freshwater ecologists. The most probable reasons for this gap of knowledge come from the cryptic nature of the meiofauna, the necessity to have a taxonomic expertise not usually found in all laboratories, and the fact that the minute size of most meiofauna makes difficult to meet the detection limits of some analytical platforms such as elemental analyzers or mass spectrometers. The latter drawback might be alleviated through methodological improvements and here we review the different methods compatible with the study of meiofauna’s trophic ecology, their requirements, routines, limitations and advantages (summarized in Table 1). It appears that many conventional methods can bring valuable information on diet specificity and ingestion rates (e.g. gut content analysis, observation of feeding, incubation experiments) with slight modifications of standard protocols. Elemental composition, stable isotopes and fatty acids can bring valuable information on assimilation and energy fluxes at the scale of the meiofauna provided minor modifications e.g. reducing the volume of combustion columns in EA-IRMS devices, or coupling gas-chromatographs to flame ionization detectors instead of mass spectrometers. The rapid development of microscopy, micro-spectroscopy and molecular-based techniques in the field of microbial ecology also opens interesting opportunities to study the interactions between microbes and meiofauna, and, as an example, those techniques could be used to better understand methane-based feeding channels in lotic and lentic systems.

Table 1 Summary of methodological approaches to examine meiofaunal trophic interactions
Full size table