The diversity of feeding habits recorded for water boatmen ( Heteroptera : Corixoidea ) world-wide with implications for evaluating information on the diet of aquatic insects

Food webs are of crucial importance for understanding any ecosystem. The accuracy of food web and ecosystem models rests on the reliability of the information on the feeding habits of the species involved. Water boatmen (Corixoidea) is the most diverse superfamily of water bugs (Heteroptera: Nepomorpha), frequently the most abundant group of insects in a variety of freshwater habitats worldwide. In spite of their high biomass, the importance of water boatmen in aquatic ecosystems is frequently underestimated. The diet and feeding habits of Corixoidea are unclear as published data are frequently contradictory. We summarise information on the feeding habits of this taxon, which exemplify the diffi culties in evaluating published data on feeding habits in an invertebrate taxon. It is concluded that Corixoidea are, unlike other true bugs, capable of digesting solid food, but their feeding habits are still insuffi ciently known. The dominant feeding strategy in this taxon is zoophagy, but several species consume other foods, particularly algae and detritus. Only members of the subfamily Cymatiainae seem to be exclusively predators. In other subfamilies, the diet of different species and different sexes or populations of a single species may vary depending on the food available or is still unknown. We conclude, that a multi-method approach is needed to elucidate the feeding habits of aquatic insects and invertebrates in general. * Present and corresponding address: Brandenburg State Offi ce of Environment, Tramper Damm 2, 16225 Eberswalde, Germany; e-mail: christian-haedicke@gmx.de INTRODUCTION Studies on feeding niches in different insect groups are an essential backbone for understanding terrestrial and aquatic ecosystems. Aquatic insects are remarkably diverse in terms of their bionomics and specifi c morphological adaptations (Usinger, 1956; Wallace & Merritt, 1980; Winterbourn & Gregson, 1981; Wesenberg-Lund, 1989; Wichard et al., 2013). Aquatic insects occupy a large number of feeding niches and the different taxa are placed into particular guilds [i.e. shredders, scrapers, fi lterers, etc.; Cummins (1973)]. This simplifi cation into functional feeding groups is a fundamental part of the river continuum concept (Vannote et al., 1980) and some systems to evaluating water quality (Palmer et al., 1996; Pavluk et al., 2000). Information on feeding habits in aquatic insects and other aquatic invertebrates is obtained using different methods, from simple observations (i.e., organisms hanEur. J. Entomol. 114: 147–159, 2017 doi: 10.14411/eje.2017.020

Although often neglected, water bugs in general and Corixoidea in particular are of major importance in aquatic ecosystems (Papáček, 2000(Papáček, , 2001)).Especially in aquatic habitats that lack fi sh (e.g.rock pools, mining lakes and heavily acidifi ed lakes), species of Corixoidea are important predators dominating the food webs (Henrikson & Oscarson 1981;Wollmann 1997Wollmann , 2000;;Wollmann & Deneke, 2002;Soldán et al., 2012).On the other hand, species of Corixoidea are a signifi cant food compound in fi sh (Frost & Macan, 1948).The potential of European Corixoidea as bioindicators of stagnant water has been outlined (Macan, 1938(Macan, , 1954;;Jansson, 1977;Savage, 1994;Hufnagel et al., 1999;Skern et al., 2010).Several studies underpin the potential value of the Nepomorpha as mosquito control agents (Darriet & Hougard, 1993;Ohba & Nakasuji, 2006;Saha et al., 2007;Sivagnaname, 2009).DNA-examinations of the gut contents of invertebrate and vertebrate taxa in Kenyan wetlands revealed Nepomorpha being more effi cient than amphibians in reducing mosquitoes of human importance (Ohba et al., 2010).The role of Corixoidea as primary consumers is, however, widely unknown.Therefore, reliable information on their role in matter exchange in aquatic habitats is unknown (Wollmann, 1997;Wollmann & Deneke, 2002).Corixoidea are positively buoyant and frequently visit the water surface to breathe.Due to their relative independence of dissolved oxygen and feeding on the bottom ooze, Corixoidea are probably crucial for the exchange of organic matter between hypo-and epilimnion, especially during summer stagnation.
Ecological preferences of the 600 described species of Corixoidea worldwide (Polhemus & Polhemus, 2008; Table 1) are only known for a few European (Macan, 1938(Macan, , 1954;;Wróblewski, 1980;Bosman, 1982;Tully et al., 1991;Hutchinson, 1993;Wachmann et al., 2006) and North American species (Hungerford, 1919;Dodson, 1975;Applegate & Kieckhefer, 1977;Kelts, 1979;Lauck, 1979;Hutchinson, 1993).Scudder (1976) and Hutchinson (1993) summarise the information on corixoid species of brackish waters and inland salt lakes and comment on their occasional occurrence in marine habitats.Several species of Corixoidea benefi t from their role as pioneer species in quickly colonising new habitats (Bröring & Niedringhaus, 1988;Savage, 1989;Wollmann, 2000).The information available on the biology of endemic Australian [Diapre-Table 1. Diversity of water boatmen recorded in different zoogeographical regions, with number of described species and quantity of information on the diet of individual taxa.Information on the diet of Corixoidea is mainly based on Old World species.However, with the exception of the Diaprepocoridae and Cymatiainae (Corixidae), information is available for less than half of the known species.In particular, information on the diet of Neotropical species is limited.

Region
No. of species of Corixoidea has been studied for more than a century but is still uncertain.Different methods seem to prove, at least partially, a carnivorous diet, but according to Baptist (1942) Corixa punctata only produces the gut enzymes invertase and amylase.Digestion of animal food is not possible with these enzymes and there are no indications that they have symbiotic bacteria.Although these results have not yet been confi rmed by more advanced techniques, they question carnivory in Corixoidea.Surprisingly, the predatory Belostomatidae also produce amylase in their salivary glands (Swart et al., 2006), which is said to digest plant matter in the gut of these predators prey.Although well known in terrestrial insects (reviewed by Terra & Ferreira, 1994), information on digestion enzymes in aquatic insects is sparse [except for aquatic Heteroptera, Trichoptera (Martin et al., 1981;Bärlocher & Porter, 1986) and Plecoptera (López-Rodríguez et al., 2012)].Insights from digestion physiology not only complement traditional gut analyses (López-Rodríguez et al., 2012), but are a useful tool for critically evaluating data on feeding habits in aquatic insects.
When studying data on feeding habits it should always be remembered that observations of an organism handling prey do not indicate consumption, and even the presence of food residues in the alimentary tract detected by microscopic examination or serological or DNA-based methods (i.e.PCR) also does not necessarily prove that either.Representatives of Potamanthidae (Ephemeroptera), for instance, were traditionally placed in the functional group collectors/gatherers, but McCafferty & Bae (1992) proved they are fi lter feeders; the presence of particles of chironomid larvae in the gut of Potamanthus luteus are considered to be a result of accidental ingestion (Fenoglio et al., 2008).In addition, not all the nutrients found by advanced techniques are digested; for example, in grasshoppers (Acrididae) only monosaccharides and proteins in the alimentary canal are assimilated, while polysaccharides are not (Beier, 1972).A single method is therefore frequently insuffi cient for assigning a certain group to a particular functional feeding group (Palmer et al., 1993a;Tomanova et al., 2006).
Feeding habits of Corixoidea attracted much attention in the past century.The present paper summarises and criti- buoyancy, thus they need to be attached to objects under water.A -most Corixidae and Micronectidae rest attached to the bottom and search for food in the bottom ooze (detritus, algae and small benthic animals) or graze on periphyton.B -species of Cymatia have distinct predatory tendencies, involving certain morphological adaptations (i.e.modifi ed fore tarsi ("palae"), rather globular compound eyes).Often they rest attached to water plants in the littoral from which they ambush passing prey (e.g.ephemerid larvae).C -some species of the subfamily Corixinae (e.g.Glaenocorisa sp. and Graptocorixa sp.) also have more globular compound eyes than the majority of Corixoidea, and pursue prey in the pelagial and below the water surface.
cally evaluates data in the literature on the feeding behaviour and feeding habits of this superfamily globally, with the aim of improving our understanding of their ecological signifi cance.In addition, we explore the diffi culties in evaluating data on feeding habits in this insect group and invertebrates in general.
FEEDING BEHAVIOUR OF CORIXOIDEA Walton (1938) described the feeding behaviour of Micronecta sp. as follows: "Leaning forward, the Micronecta diligently turns over the bottom refuse, seeking for minute plants and animals; with its palae it holds these against its mouth.Filaments of algae are moved along with a handoverhand motion while the contents of the cells are sucked out.Small pieces of alga may be swallowed whole and can be found in a lacerated condition in the stomach" (Fig. 1A).Based on laboratory observations, Micronecta species do not suck fi lamentous algae, but manipulate small grains of sand and plant particles using their palae (Wróblewski, 1958).This behaviour suggests that they scrape a biofi lm consisting of periphytic algae and other organisms from the surface of the particles, including sessile protozoans and Rotifera.The palae of Neotropical Synaptogobiinae, however, show distinct morphological differences to those of other Micronectidae (Nieser & Chen, 2006), which probably indicates different feeding habits.
Different feeding behaviours of European species are summarised by Sutton (1951) and Popham et al. (1984).Most species throw up the bottom ooze and feed on detritus, animals and algae by alternating movements of their palae (Hungerford, 1919).Detritus in the suspended bottom ooze is moved to the mouth and ingested by movements in the cibarium.Algal fi laments and active animals are impaled on the stylets and held by both palae above the functional mouth orifi ce; partly digested food is sucked into the food canal (Fig. 1A).The palae may also sweep over stones, dead animals etc., and detach material that is sucked into the food canal (Buchanan White, 1873).Corixini may occasionally also catch prey with their palae (Popham et al., 1984).

DIET OF THE CORIXOIDEA
No information is available for the feeding habits of some genera of Micronectidae, Stenocorixinae and Heterocorixinae.
Published information on the diet of Micronecta is controversial.In early reports, M. ovivora is described handling and presumably feeding on fi sh eggs in India (Westwood, 1871), however, no individual was actually seen to suck an egg (Hutchinson, 1993).Wróblewski (1960) suggests that they feed on detritus and microorganisms.Micronecta spp., unlike Corixidae, do not feed on freezer-preserved or freshly killed aquatic invertebrates in captivity.Providing fresh bottom material from their actual habitats 2-3 times a week successfully kept them alive (Jansson, 1986).Micronecta quadristrigata and some Malayan species of Micronecta are reported to feed on algae (Fernando & Leong, 1963).Li & Zou (2005) report M. sahlbergi consuming detritus and the fi lamentous alga Ulothrix sp.
In contrast, M. grisea is menioned capturing and feeding on insect larvae, including mosquito larvae (Diptera: Culicidae).The predation of 3 rd instar larvae of the mosquito Stegomyia aegypti by M. grisea is confi rmed by laboratory experiments (Amrapala et al., 2009).A DNA-analysis of the gut contents of Micronecta spp.from Vietnam and Kenya using PCR also provides evidence for the consumption of mosquito larvae (Aedes, Anopheles) (Ohba et al., 2010(Ohba et al., , 2011)).
In contrast, Wachmann et al. (2006) consider C. coleoptrata to be both carnivorous and herbivorous.A study of another species, the Asian C. apparens, based on a microscopic examinations of gut contents, revealed a mixed diet, containing algae, plant matter and dead fi sh (Li & Zou, 2005).
Corixidae: Corixinae: Agraptocorixini Hale (1924) states that Agraptocorixa eurynome feeds on mosquitoes in captivity.Information on A. hyalinipennis is contradictory.Although there are reports of adults feeding on a diet of boiled potatoes (Quadri, 1951), larvae of this species do not survive if fed algae (Fernando & Leong, 1963).

Corixidae: Corixinae: Corixini
There are claims that the diet of Corixini range from detritivory and herbivory to carnivory.Apart from these, Kirby (1983) describes a case of fungivory.Early observations indicate predation on freshwater crustaceans (e.g.Abbott, 1912).Hale (1922) kept several Australian species of Corixinae in aquaria for months, and during that time they were fed only with larvae of the mosquitoes Culex quinquefasciatus (= C. fatigans) and Scutomyia notoscripta.Even newly-hatched Corixinae can capture tiny mosquito larvae and increasingly larger individuals were taken during the successive stages of metamorphosis.
In contrast to these carnivorous tendencies, there are many observations on Corixini feeding on detritus and plant matter.First evidence was provided by Hungerford (1917), who observed different Corixini larvae and adults consuming detritus from bottom ooze or dead leaves, consisting of tiny bits of organic matter, diatoms, desmids, Oscillatoria, sometimes veins of leaves, more often dead fi laments of Zygnema, unicellular Euglena, Paramecium, Chlamydomonas, spores of various algae, and cysts of Euglena and other unicellular "plants" and "animals".The subsequent examination of their digestive tract revealed fi laments of the cyanobacteria (Oscillatoria) 9-10 mm long, and bits of Zygnema fi laments consisting of as many as seven cells.However, an attack on a blood worm by a starved Corixini is also recorded.In laboratory experiments, different species of Corixini were provided with different foods in aquaria and subsequently their fore guts were examined.This revealed remnants of various algae (e.g.Ankistrodesmus, Gonatozygon, Mougeotia, Tetraspora, Zygnema), desmids, diatoms, blue-green algae (Oscillatoria), oligochaetes, nematodes, rotifers and protozoa.He also reports corixines piercing the cells of fi lamentous Spirogyra and sucking out their contents and mentions predation on chironomid larvae.However, Hungerford (1919) did not provide any evidence that they could successfully complete their life cycle on any of these diets.
Examination of the midgut contents of different European Corixini supplied with various foods indicate a distinct preference for general detritus and algal fi laments (Sutton, 1951).The species were also provided with different animals, involving chironomid and tanypodid larvae, tanypodid pupae, may fl y larvae and Tubifex were offered.Ad-dtionally, corixines were observed feeding on daphniids, Asellus (Crustacea: Isopoda) and other corixines present in pond water placed in the aquaria.Third to 5 th instar larvae and adults fed upon all these animals, but fi rst and second instar larvae were only observed feeding on small daphniids.Corixinae frequently feed on their dead or moribund fellows.The results of examinations of the midgut contents of corixine feeding upon detritus are as follows: Corixa punctata: brown and green granular masses, remains of rotifers, Cyclops, algal fi laments and protozoa; Corixa panzeri: brown and green granular masses, amoebae, fl agellates, ciliates, nematode eggs and bacteria; Sigara striata: thin algal fi laments, intact fi laments and remains of Anabaena and allied algae, remains of Scendesmus, diatoms, desmids, Cyclops and one pollen grain of Pinus sp.; Sigara falleni: 90% brown and/or green granular masses, 10% consisting of algal fi laments (cells from full and normal to completely empty; predominantly Spirogyra spp.), setae of oligochaetes (e.g.Tubifex), remains of crustacean appendages, nematocysts of Hydra, diatoms (e.g.Navicula), desmids, cysts of plants, animals and protozoa (e.g.Euglena).Zwart (1965) determined the survival of Corixinae offered different food sources: Tubifex, daphnids, chironomid larvae, dead fresh water animals, water plants, living or decaying algae, natural or laboratory made detritus and in a control without any food, in the laboratory.Corixa punctata, Sigara distincta, and Arctocorisa germari fed on algae, water plants and detritus usually with similar high levels of mortality.Both adults and larvae exhibit higher survival rates when fed animal food.Reynolds (1975) examined the feeding habits of Cenocorixa bifi da and C. expleta consumed crustaceans (Diaptomus, Ceriodaphnia, Daphnia), chironomid larvae and zygopteran larvae (Enallagma), juvenile and adult Hesperocorixa laevigata accepted dead and living chironomids, amphipods and zygopterans.In addition, a serological analysis (C. bifi da, C. expleta, H. laevigata, Callicorixa audeni, Sigara spp.) records frequent reactions with chironomid and zygopteran, followed by ephemeropteran, chaoborid, diaptomid, daphniid and amphipod antiserums; the fewest reactions were detected with pulmonate (only in Cenocorixa spp.) and plant antisera (blue-green algae, Cladophora, macrophytes only in C. bifi da).Popham et al. (1984) visually examined the gut contents in 21 species of Corixinae collected in the fi eld.They found considerable variability among the species but with a preference for a mixed diet; only Corixa dentipes and C. panzeri seemed to be exclusively carnivorous, and Sigara lateralis, S. limitata, and S. stagnalis exclusively detritivorous.In S. falleni, the sexes had different feeding habits: high prevalence of an algal component in males and of a mixture of foods in females.
In summary, most Corixinae consume a mixture of different foods with distinct preferences for either plant or animal food.Variation in food preferences is not only evident between species, but also between sexes (e.g. S. falleni) and developmental stages.A reliable evaluation of the feeding habits of individual species requires detailed studies on their autecology [e.g.Cenocorixa spp., Reynolds & Scudder (1987a, b)].Additional information on the diet of individual species of Corixinae is given below.

Corixidae: Corixinae: Corixini: Corixa
Capture of culicid and chironomid larvae is described for Corixa (Poisson, 1935) and of chironomid larvae by C. panzeri (Sutton, 1945).Corixa punctata, collected in a detritus-rich pond, was kept and survived in an aquarium containing only a layer of bottom ooze.The fi rst two instars were observed to feed on Spirogyra by Benwitz (1957).Jaczewski (1961) describe C. punctata attacking and sucking out the larvae of Chironomidae and mayfl ies (Ephemeroptera).
Corixidae: Corixinae: Corixini: Cenocorixa Scudder (1966) successfully reared Cenocorixa bifi da (fi rst instar to adult) and C. expleta (the fi rst three larval instars) on a diet of young brine shrimps (Artemia salina, Artemiidae).Scudder (1969a, b) found that C. bifi da hungerfordi and C. expleta fed almost exclusively on planktonic diaptomids (Diaptomus nevadensis, D. sicilis) in saline lakes in British Columbia.Other species of Cenocorixa and members of other genera of Corixini were successfully reared through several generations on frozen brine shrimp (Jansson & Scudder, 1972).Feeding niches of C. bifi da and C. expleta were studied by Reynolds & Scudder (1987a, b).Both species feed on living as well as dead prey, most often on larval chironomids, but also on larvae of beetles, chaoborids, ephemeropterans and zygopterans, larval corixines and eggs, amphipods, fairy shrimps, daphniids and diaptomids.Only adults of both species were observed feeding on aquatic angiosperms.All serological reactions with angiosperm antibody were negative, serological reactions to algal antibodies were positive but weak.The authors therefore conclude that the evidence for feeding on algal or angiosperm material is slight and past reports of corixines feeding on aquatic plants may have been just probing for potential prey.In both species, no signifi cant differences were found in relation to age, sex, location, or season, but there are some differences in feeding habits between the larvae and adults in both species, e.g. in their preference for living or dead prey (Reynolds & Scudder, 1987a, b).

Corixidae: Corixinae: Corixini: Sigara
Most of the information on the diet of Corixini is based on this genus alone.A considerable variation occurs not only between different Sigara species but also between the sexes (Popham et al., 1984).
Somewhat isolated is an observation of fungivory.Kirby (1983) observed captive individuals of Sigara dorsalis moving their palae rapidly over the surface of dead bodies of other water boatmen covered with a growth of fungal hyphae.Subsequent dissection revealed broken hyphal fragments in the crops of four out of twelve the specimens examined.The hyphae were probably not the primary target of the bugs as there was a large population of ciliate protozoans among the hyphae.
Sigara lateralis is reported feeding on Daphnia pulex in an aquarium (Walton, 1943).Jaczewski (1961) reports ob-serving S. falleni sucking out mayfl y larvae (Cloeon spp.).In some Sigara species, interspecifi c predation regulates population densities.Sigara scotti attacks and sucks 2% and S. dorsalis up to 40% of their own eggs in laboratory cultures (Young, 1965).Species of Sigara can cause significant harm to fi sh populations (Sokol'skaya & Zhiteneva, 1973).Despite the presence of Potamogeton and fi lamentous algae in the aquarium, adults and larvae of S. lateralis and S. striata prefer larvae (younger than 10 days) of Cyprinus carpio and Hypophthalmichthys molitrix.In S. falleni, these authors did not always detect predation on juvenile fi sh, but the consumption of the Daphnia added to the experimental cages.In some experiments 6-day old larvae of H. molitrix were killed by females of S. falleni in whose oviducts mature eggs were found.All three species examined consumed fi sh eggs in large numbers (Sokol'skaya & Zhiteneva, 1973).Gut examinations of three species of Sigara (S. falleni, S. lateralis, and S. striata) revealed animal body fl uid (S. striata 95.3%, S. falleni 89.5%, S. lateralis 61.0-99.3%);small amount of algae (S. striata 4.5%, S. falleni 10.25%, S. lateralis 0.3-38.5%)and negligible amounts of detritus (Bakonyi, 1978).However, seasonal variation in the ratio of the animal and algal component is recorded in S. lateralis.Feeding experiments by Henrikson & Oscarson (1981) and Nyman et al. (1985) reveal that S. distincta, and S. scotti are predators of Cladocera, Copepoda and Chaoborus.
More recently, Alahmed et al. (2009) studied predation of S. hoggarica on larvae and pupae of the mosquito Culex quinquefasciatus, both in an aquarium and in the fi eld.The results show that the predatory effi ciency of S. hoggarica was highest when attacking fi rst instar larvae and decreases as the size and age of the prey increases.In experiments on foraging effi ciency, Klečka (2014) shows that S. striata prefers feeding on zooplankton, in particular chironomid midges and signifi cantly less so on Culex sp., Cloeon sp. and Daphnia sp.Sigara alternata fed on frozen brine shrimps do not reproduce, but do so after a week on a diet of freshly killed mayfl y larvae (Jansson & Scudder, 1972).This example may indicate that the nutritional value of different animals differ in terms of being suitable for reproduction.Quadri (1951) kept specimens of S. promontoria alive for several weeks on boiled potatoes mixed with water, suggesting herbivory in this species.
Based on laboratory feeding experiments with S. striata involving detritus, plants (algae belonging to the genera Zygnema, Cladophora, Rhizoclonium and apical parts of Ceratophyllum and Myriophyllum leaves) and animal food (Tubifex, planktonic crustaceans, chironomid larvae), Puchkova (1969) proposes that this species has a mixed diet based on herbivory, with optional consumption of animal food; however, if plant food is not available, it can survive on animal food.A mixed diet of algae, plants, detritus and dead fi sh is also reported for S. bellula (Li & Zou, 2005).

Corixidae: Corixinae: Corixini: Trichocorixa
According to Sailor (1948), all species of Trichocorixa feed on fl occulent bottom ooze, diatoms and algae.This author frequently observed both T. calva and T. kanza feeding on Spirogyra, thrusting its stylets into each cell, and sucking out the contents.Carnivory is indicated for some species of Trichocorixa, which were observed to seize their own eggs, pierce them with their stylets and suck them dry.A serological analysis of the gut contents of T. verticalis interiores revealed Chironomidae are the main food source (Scudder, 1976).Trichocorixa verticalis sellaris preys on certain animals (Chironomus, Ceratopogonidae and Oligochaeta) and on detritus and algae (Kelts, 1979).Trichocorixa verticalis is described as a predator of the brine shrimp species Artemia franciscana in Great Salt Lake (Utah) by Wurtsbaugh (1992).These result were confi rmed in a microcosm experiment.
Corixidae: Corixinae: Corixini: Callicorixa and Arctocorisa Sailor & Lienk (1954) observed predatory behaviour in males and females of Callicorixa audeni and C. alaskensis kept in jars and conclude that they are important in limiting the abundance of the mosquito Aedes communis.
Corixidae: Corixinae: Corixini: Trichocorixella 1 st and 2 nd instar larvae of Trichocorixella mexicana develop well on autoclaved mud and algae, but third to fi fth instar larvae of the same species require an additional nutritional source in the form of Tubifex worms if they are to complete their development (Peters & Ulbrich, 1973).

Corixidae: Corixinae: Glaenocorisini
Glaenocorisa cavifron is reported feeding on Daphnia pulex in an aquarium (Walton, 1943).Jansson & Scudder (1972) reared Glaenocorisa sp. on a diet of mosquito larvae.Nyman et al. (1985) describe G. propinqua as a predator of Cladocera, Copepoda and Chaoborus.Hrdličková (2014) analysed the gut contents of G. propinqua in three different acidifi ed, fi sh-less lakes in the Bohemian Forest Mts and compared these results with the zooplanktonic community.Different planktonic crustaceans (Copepoda and Cladocera) were identifi ed as the main food source, but there were remarkable differences in the relative representation of planktonic crustaceans in food of different populations of G. propinqua in the lakes studied.In one lake, Hrdličková (2014) found a large amount of algae in the gut of G. propinqua and suggests that these algae were ingested while feeding on herbivorous crustaceans.

FEEDING HABITS IN CORIXOIDEA AND THEIR ASSIGNMENT TO FUNCTIONAL FEEDING GROUPS
As demonstrated by this review, in spite of more than a century of research, the diet of Corixoidea has remained uncertain.The only two groups for which relevant information is available are the Cymatiainae and Corixini.Cymatiainae seem to be almost exclusively predators, whose hunting strategy includes waiting while perched on vertical surfaces (usually macrophytes), followed by quick attack and prey capture (e.g.Walton, 1943).Species of Cymatiainae are thus considered to be predators.Indications of herbivorous tendencies (Li & Zou, 2005;Wachmann et al., 2006), however, need to be confi rmed.
On the other hand, the diet of Corixini and Glaenocorisini is less clear and the evidence is often contradictory.For example, the European Sigara lateralis is recorded as either exclusively detritivorous (Popham et al., 1984), omnivorous but overwhelmingly zoophagous (Bakonyi, 1978), or mainly preying on juvenile fi sh (Sokol'skaya & Zhiteneva, 1973).Considering the fl aws in the various methods used (see below), a carnivorous diet is well documented for Corixini, fi rst of all by successfully rearing on such a diet (e.g.Jansson, 1969;Jansson & Scudder, 1972) and numerous observations of corixines attacking and consuming various invertebrates and fi sh larvae, including corixine eggs and their own nymphs (for references see above).However, there is also some evidence of them consuming algae, such as observations of them sucking out Spirogyra cells (Hungerford, 1919) and a minor but positive reaction in serological studies (Reynolds & Scudder, 1987b).The consumption of detritus seems to be the least convincing feeding strategy as it usually concerns animals kept under artifi cial conditions and is probably a result of starvation, or the consumption of detritus might be merely accidental during feeding on detritus-living protozoans, rotifers and nematodes, exactly the opposite to what Hungerford (1948) suggests, who considers the presence of small animals in guts of Corixini to be a result of accidental consumption during feeding on detritus.The fact that neither herbivory nor detritivory could be the main feeding strategy of Corixini seems to be supported by the lack of any successful breeding experiments using such diets.
Another factor complicating the determination of the diet of Corixini might be the variability in the nature of the food consumed by the different stages, sexes and in different seasons, as suggested by some of the investigations (Bakonyi, 1978;Popham et al., 1984;Reynolds & Scudder, 1987a).Based on the available information, species of Corixini could be either carnivorous or omnivorous with an animal component in their food.With the exception of the Glaenocorisini, which are predators, it is currently not possible to place Corixinae into a specifi c functional feeding group.There are uncertainties about the digestive enzymes in this group (Baptist, 1942) and the differences in the feeding habits of the sexes [S.falleni; Popham et al., (1984)] and different developmental stages (Reynolds & Scudder, 1987a, b).It is, however, obvious that regarding all Corixoidea as gatherer-collectors does not refl ect the diversity of feeding habits in this group.
Agraptocorixini seem to be predators; the food of Synaptogobiinae, Stenocorixinae and Heterocorixinae is unknown; while that of Diaprepocoridae and Micronectinae is mostly a matter of speculation.The experience of Jansson (1986) that European Micronecta (body length < 2 mm) may be kept in aquarium and only provided with fresh bottom material from their habitats indicates where to search for the microscopic prey of these animals.Also the observation of Wróblewski (1958) of European species of Micronecta handling small grains of sand and plant particles indicates they may scrape the biofi lm of periphytic algae, bacteria and other organisms from the surface these particles.The fi rst good evidence for consumption of mosquito larvae by tropical Micronecta spp.(body length usually 2-4 mm), obtained using PCR analysis of their gut contents, has recently been published (Ohba et al., 2010(Ohba et al., , 2011)); it is also is in accordance with laboratory observations (Amrapala et al., 2009).However, observations of Micronecta consuming algae and detritus also exist (Fernando & Leong, 1963;Li & Zou, 2005).Thus, placing one of these species in a certain functional feeding group is premature.
It is evident that Corixoidea have a diverse range of feeding strategies, with different species being assigned to different functional feeding groups (Table 2).The limited information on the biology and ecology of Diaprepocoridae, Micronectidae and some Corixidae (e.g., Stenocorixinae and Heterocorixinae), however, greatly hampers the defi nitive placement in a particular functional feeding group.

IMPLICATIONS FOR PLACING AQUATIC INSECTS IN FUNCTIONAL FEEDING GROUPS
Controversial information is partly due to the different methods used to analyse the diet of certain species, some of them being less informative than others (Table 3).Macroscopic observations can give a fi rst hint, but alone can only result in mere speculation [e.g.feeding on fungal hyphae reported by Kirby (1983)].Hypotheses based on such observations can be tested against specifi c morphological and behavioural adaptations [e.g.fore feet and grinder in Corixoidea (Parsons, 1965;Martin, 1969;Popham et al., 1984;Hädicke, 2012), Ephemeroptera andTrichoptera (Wallace &Merritt, 1980;Merritt & Wallace;1981;Palmer et al., 1993b;Elpers & Tomka, 1994;Polegatto & Froehlich, 2001) and other aquatic insects (Wichard et al., 2013)].
Simple microscopic observations of the gut contents (e.g.colour of the gut contents) may also be misleading and certainly less reliable than specifi c serological methods (Reynolds & Scudder, 1987b).Gut fl uorescence (Cowan & Peckarsky, 1990) is a powerful tool for quantitative studies on functional feeding groups only if a diet of algae is already indicated (Glozier et al., 2000).Most recent advances in studies on animal diets have used DNA-analysis (e.g.Valentini et al., 2009).
While these methods are suitable for demonstrating the range of the diet, they do not indicate that such a diet is suffi cient for completion of the whole developmental cycle.For Periplaneta americana (Blattodea: Blattidae), Bignell (1982) points out that: "Essential nutrients are those required for indefi nite growth and reproduction ...".Successful laboratory breeding experiments in which animals are provided with a single type of food ad libitum do not exclude the possibility of a more variable diet under natural conditions.An insuffi cient supply of nutrients under laboratory and natural condition is maybe enough to keep an organism alive but does not guarantee successful reproduction and thereby the survival of the population.
Table 2. Assignment of different Corixoidea to functional feeding groups.Information on the gut contents and behaviour of Diaprepocoridae, Micronectidae and some Corixidae indicates assignment to the guild Scrapers/grazers, whether it is a general feeding habit needs further study.Representatives of Corixini are the only unambiguous example of gatherer-collectors, the functional morphology of other Corixoidea taxa also indicates an assignment to this guild.Available information on the diet, behaviour, functional morphology and breeding of Cymatiainae justify an assignment to the predator guild.
To improve our knowledge on the alimentation of Corixoidea in particular and aquatic insects in general, analysis of functional morphology, gut contents (serological and/or DNA-based methods) should be used to identify the food consumed under natural conditions (Ohba et al., 2010(Ohba et al., , 2011;;Klimaszewski et al., 2013), with subsequent evaluation of the results and determination of the nutritional value of the food consumed in laboratory breeding experiments.
Table 3. Correlation between different sources of information and their reliability.Reliability increases from mere observations to successful breeding.Every attempt to clarify feeding habits has its limitations.Thus, using a combination of different methods is more likely to result in reliable conclusions.

Source of information Limitations
Reliability observations on behaviour in nature and captivity a reliable differentiation between probing and feeding is diffi cult functional morphology/morphological adaptations only useful if considered in the context of behaviour microscopic examinations of gut contents fi ne organic matter impossible to separate, artifi cial uptake cannot be excluded serological and DNA-based examinations of gut contents artifi cial uptake cannot be excluded successful breeding experiments using specifi c diets infl uence of other ecological factors (biotic and abiotic) cannot be excluded

Fig. 1 .
Fig. 1.Schematic drawing of different feeding behaviours of European Corixoidea.Due to their plastron Corixoidea have a positivebuoyancy, thus they need to be attached to objects under water.A -most Corixidae and Micronectidae rest attached to the bottom and search for food in the bottom ooze (detritus, algae and small benthic animals) or graze on periphyton.B -species of Cymatia have distinct predatory tendencies, involving certain morphological adaptations (i.e.modifi ed fore tarsi ("palae"), rather globular compound eyes).Often they rest attached to water plants in the littoral from which they ambush passing prey (e.g.ephemerid larvae).C -some species of the subfamily Corixinae (e.g.Glaenocorisa sp. and Graptocorixa sp.) also have more globular compound eyes than the majority of Corixoidea, and pursue prey in the pelagial and below the water surface.