Interactions between stoneworts (Charales) and waterbirds

(1) Sources of food for waterbirds provided by Charales

Charales plant bodies are differentiated into nodes and internodes (McCourt et al., 2017). Nodes bear regular whorls of branches and because of the very dense nature of the green biomass presented in this way within the water column, these vegetative structures form one major source of food biomass, especially for herbivorous grazing birds (e.g. Noordhuis, van der Molen & van den Berg, 2002). Such green material is attractive to waterbirds due to several intrinsic properties of the Charales. First, these species (especially those forming dense vegetation) tend to have a comparatively long growing season (Pereyra-Ramos, 1981; Nichols, Schloesser & Geis, 1985), although some species, such as Sphaerochara intricata and Nitella capillaris may be spring annuals with comparatively short growing seasons, while others may be restricted by summer desiccation. Many other Charales species continue to grow well into the autumn where this is possible, retaining green biomass into winter in some situations, whereas the above-ground parts of other sympatric flowering plant macrophytes may have disappeared by the end of August (van Wijk, 1989; van Dijk & van Vierssen, 1991). Secondly, although the surface calcium carbonate deposition results in relatively low overall caloric value [although this can be up to 12 kJ g⁻¹ dry mass (DW); Palma-Silva, Albertoni & Esteves, 2004], the low fibre content [less than half that of other vascular submerged macrophytes growing sympatrically in British gravel pits (Fox et al., 1994; A. D. Fox, unpublished data)] makes Chara potentially relatively easy to digest by herbivorous waterbirds compared to other plant material richer in structural carbohydrates (assuming the avian gut can dissolve the calcium carbonate without high metabolic costs). The higher content of cellulose compounds in the structural tissues of higher plants are not just hard to digest but inhibit digestion in the relatively simple avian gut, characterised by high throughput rates and rapid absorption of accessible soluble proteins and carbohydrates (e.g. Prop & Vulink, 1993). In Danish brackish lakes, Chara spp. had half the structural fibre (27–36% neutral detergent fibre of ash-free dry mass) and twice the non-structural soluble (i.e. potentially easily digested) carbohydrates of other submerged macrophytes in the same systems (Ruppia cirrhosa and R. maritima; Holm, 2002). Dry mass and calcium carbonate encrustation varies among Chara species and are related to water chemistry, water depth and light penetration (Pukacz et al., 2016; Sviben et al., 2018). Intriguingly, carbonate encrustations of Charales of the same species are significantly greater in freshwater than brackish systems even when calcium concentrations and pH are similar in their environment (Herbst & Schubert, 2018), a fact that potentially could affect their attractiveness as waterbird forage in different situations. Because the dissolution of calcium carbonate during digestion requires copious sources of acid, it has also been suggested that these encrustations offer protection from herbivores, yet the evidence for this is not compelling (McConnaughey & Whelan, 1997). Moreover, as we will show below, there are numerous avian herbivores [and indeed many fish species (e.g. Pípalová, 2006; Dibble & Kovalenko, 2009)], that habitually graze upon Charales above-ground biomass, suggesting that these species at least have adapted to these features of the plant ‘defence’ systems. There remains much debate about whether Charales gain structural benefit and/or protection from herbivores by their calcified deposits. Many Charophytes grow without calcifying in soft or Ca²⁺-depleted water. Although carbonate encrustations present unattractive properties to consumers, such as physical hardness, acid-neutralising capacity and their displacement of other nutritious components of the food quality, it seems that researchers overrate all of these qualities. Indeed, Charales seem to constitute highly profitable forage to some herbivores (McConnaughey & Whelan, 1997). The dense, compact growth form of most Charales results in very high local densities of plant biomass, which enable head-dipping, up-ending and diving waterbirds to harvest abundant local biomass rapidly and persistently (e.g. Schmieder et al., 2006), suggesting that these are highly attractive food resources.

The asexual reproduction of Charales can occur through unusual contracted starch-filled apices of branches, where the primary branchlet segments and internodes become packed with starch grains, similar in morphology to the turions of some aquatic angiosperms (Casanova et al., 2007). Asexual reproduction also occurs more commonly through starch-filled outgrowths of the rhizoids called bulbils which may become abundant on plants and eventually are shed and become buried within the substrate, persisting throughout the year (Wood & Imahori, 1965; Casanova, 1994) and which may ultimately germinate separate from the thallus. Still more common is asexual reproduction (and hibernation) by starch-filled node cells, as in the case of Nitellopsis obtusa (Bharathan, 1987). As is the case for many starch-packed Charophyte structures, if a consumer organism is able to dissolve and free their contents from calcified structures, they are potentially highly attractive to organisms browsing such structures or dabbling in substrates where they can accumulate (for example due to water currents) in some abundance as an easy source of concentrated energy. Water depth seems to play an important role in determining the degree to which Charophytes invest in bulbil production, with deeper perennial water bodies stimulating greater investment in such vegetative reproduction whereas the same species reproduce sexually in shallow water (e.g. García, 1994; Bociąg & Rekowska, 2012; Soulié-Märsche & García, 2015). This may explain the specialisms of diving waterbirds such as common pochard Aythya ferina, red-crested pochard Netta rufina and coot Fulica atra to feed on these diaspores of the Charales (see Section III.2). According to Bonis & Grillas (2002) there had only been one investigation of bulbil density in lake sediments at the time of their review, that of van den Berg (1999), who found 5 × 10³–19 × 10³ m⁻² of buried Chara bulbils in the top 20 cm under permanently flooded conditions in a lake in The Netherlands. Conversely, sexual reproduction seems more frequent in shallow waters (perhaps related to the threat of desiccation in ephemeral arid-zone water bodies) so studies have shown a direct relationship between oospore production and receding water depth (Casanova, 1994; Asaeda, Rajapakse & Sanderson, 2007).

The sexual reproductive structures of the Charales are also highly attractive to waterbirds. Unfertilised female oogonia consist almost entirely of starch grains (Groves & Bullock-Webster, 1920–1924), so even prior to fertilisation are an attractive source of digestible stored energy to consumer species. The zygote is again filled with starch, surrounded by smooth walls overlaid with the oospore protective calcified walls (see Fig. 2 in Soulié-Märsche & García, 2015). Oospore production can occur throughout spring, summer and autumn (Casanova, 1994). Oospore production and ‘rain’ can be spectacular (e.g. 21,520 m⁻² day⁻¹ in Nitella mucronata in open field collections; van Onsem, Rops & Triest, 2018) leading to massive oospore densities in the sediment (e.g. 1.7 × 10⁶ oospores m⁻² after 6 years of continuous Chara aspera presence in Lake Veluwe; van der Berg, Coops & Simons, 2001). Some studies have shown the production of oospores to be salinity dependent in halophytic species, with production of ripe oospores only occurring in Lamprothamnium papulosum between 20‰ and 40‰ in populations growing in a salinity gradient from fresh water (0‰) to hypersaline (60‰) (Soulié-Märsche, 1998, 2008). At higher salinities, plants invested in starch-filled gametangia, and while the oospores were not fertilised they remained attractive to predators because of their high carbohydrate content.

(2) Which waterbirds species exploit which Charales parts, when, how, and why?

The literature review, based largely on analyses of gut contents supplemented by direct observations, revealed 17 European waterbird species that feed on Charales, primarily swans and dabbling ducks, plus a few specialist diving waterbirds like common and red-crested pochard, as well as the coot (see Table 1). Charales tend to achieve maximum green biomass in late summer (at least in boreal and temperate regions), although some species continue to attain greater biomass in autumn (e.g. Fernández-Aláez, Fernández-Aláez & Rodríguez, 2002) while many species remain green through the winter, albeit at lower biomass than in summer/autumn (e.g. Sanderson et al., 2008). Charales are distributed throughout mainland Europe to northern latitudes, where they may attain some of the highest biomass values known for Charales [for example in Poland (Pereyra-Ramos, 1981) and Sweden (Blindow, 1992b)]. However, in North America, their extent and biomass tends to be greater further south (Mann, Proctor & Taylor, 1999). Many northern-breeding (i.e. Arctic and boreal forest) waterbirds arrive at temperate regions in autumn, and the range of available studies suggest that 80% of these species forage on Charales in autumn and winter (Table 1). With due prudence, given the small sample size and biased seasonal distribution, it seems possible that this is the period when most waterbirds likely exploit Charales as a food resource, given the low biomass of these plants in spring, especially in relation to the availability of other new green macrophyte biomass at this time.

Herbivorous waterbirds gain access to submerged (but often near-surface) Charales by browsing green biomass on the surface, by head-dip feeding below the surface or by upending to gain access from the surface. Herbivorous dabbling ducks are reticent to dive, so access green biomass on the surface and up-end in shallow water to access standing crop below the surface, while having limited access beyond these levels. Both gadwall Mareca strepera and Eurasian wigeon Mareca penelope are quintessential surface-feeding ducks that never dive (Table 1). Despite this, both species have been recorded kleptoparasitising diving ducks (especially red-crested pochard) and coot and longer-necked species (such as swans) that bring submerged macrophytes to the surface (Hudson, Pierce & Taverner, 1960), including Chara species (e.g. Bodensee, Germany; Berthold, 1961). Berthold (1961) described the Charales vegetation as growing in deeper water, i.e. in situations inaccessible to surface-feeding ducks without the assistance of diving and long-necked waterbird species upon which they could kleptoparasitise. Finally, specialist diving herbivores such as coot, red-crested and common pochard access biomass by active diving below the water surface (see Fig. 1), while the former two species are also known to dive to access abundant diaspores where these occur.

Receding water levels may also change accessibility of Charales to waterbirds. Many Charales are adapted to thrive and reproduce in ephemeral shallow water lakes in arid zones, which may dry out later in the season. For example, in the shallow saltwater coastal lagoons of South Australia, Lamprothamnium papulosum invests heavily in bulbil and oospore production, but vegetative growth becomes increasingly stunted as salinity increases, failing to grow at all above 60‰ (Delroy, 1974). During late spring and summer, these same saltwater lagoons provide vital feeding areas for many species of waterbirds, but especially three dabbling duck species (grey teal Anas gibberifrons, chestnut teal A. castanea and mountain duck Tadorna tadornoides) displaced from ephemeral desiccated freshwater feeding areas elsewhere in Australia. These birds harvest the superabundant bulbils and oospores accumulated in shallow waters at this time because of receding water levels and it seems highly likely that this cycle of events supports similar patterns of seasonal production and avian exploitation in lagoons in Europe and elsewhere.

Among the diving duck species, the red-crested pochard is often considered an obligate Charales specialist (e.g. Allouche, Roux & Tamisier, 1988; Cramp & Simmons, 1977), almost entirely confined to clear-water lakes with abundant Charales, especially Nitellopsis obtusa (van der Winden et al., 1994; Ruiters, Noordhuis & van den Berg, 1994). In Camargue, southern France, red-crested pochard feed in winter on both the vegetative parts (stems and side shoots) and oospores (Allouche et al., 1988). In Austria and Switzerland, the occurrence and abundance of Charales (related to water quality) are essential factors influencing habitat use by the red-crested pochard (Aubrecht & Winkler, 1997; Schneider-Jacoby, 1998–1999; Keller, 2000a,b). There, females of the species dive to bring green biomass to the surface to consume and share with their ducklings during the breeding season (see Fig. 2). Diving for food is energetically costly relative to surface foraging for green plant biomass, so there may be energetic benefits from mate-feeding by red-crested pochards recorded in the Netherlands (Fig. 3), another subject worthy of further study. The common pochard is a fresh and brackish water diving duck that has long been known to specialise on feeding on Charales (e.g. Stevenson & Southwell, 1890; Ticehurst, 1932; Szijj, 1965a; Dementiev & Gladkov, 1952; Cramp & Simmons, 1977). This is especially the case in autumn and winter, when plant material occurred in all sampled birds, with Charales constituting up to 86% of total food volume in 22 common pochard examined in the UK (Olney, 1968). Among the dabbling ducks, which sieve water through the lamellae of their bills to extract animal and vegetable matter from the water column, species such as Eurasian teal Anas crecca, mallard Anas platyrhynchos, northern pintail A. acuta, northern shoveler Spatula clypeata, garganey Spatula querquedula and marbled duck Marmaronetta angustirostris have all been shown to concentrate diaspores of Charales (e.g. Cramp & Simmons, 1977; Pirot, 1981; Green, Figuerola & Sánchez, 2002a), often in very large amounts – Olsen (1944) reported a dissected gizzard from a Eurasian teal containing 10.5 cm³ ripe Charales oospores, also taken in great abundance in the Camargue (Tamisier, 1974). To what extent these oospores have been actively selected from the water column or have become concentrated in the gut by the resistance of such structures to decomposition compared to the green parts remains unknown and in need of confirmation. However, all of these duck species are omnivores that specialise in filtration of seeds and small invertebrates from the water column and show few traits of being herbivorous, suggesting active filtration of oospores as a source of food. Diving ducks submerging to forage on bottom sediments likely undertake similar sieving of the substrate to obtain diaspores concentrated within the benthos. While it seems highly likely that surface-grazing herbivorous species, such as swans, and diving species, such as common and red-crested pochard, regularly (and unselectively) consume large quantities of green above-ground production from Charales, it is the case that, at the same time, they coincidentally take in other Charales diaspores that are energetically more profitable, potentially elevating the attractiveness of these species over other submerged macrophytes.

Perhaps surprisingly, despite a wealth of studies, we still know relatively little about the true breadth of the diet of many Anseriformes. There are very few reports of geese feeding on Charales, with the notable exception of the bean goose Anser fabalis in Hungary (Table 1). However, this is likely partly the result of the relative lack of feeding studies of geese on natural and semi-natural wetlands. There has been a major increase in the numbers of geese feeding on agricultural habitats in Europe during the last 50 years, due to the greater profitability of feeding on food resources available there (Fox et al., 2017; Fox & Madsen, 2017; Fox & Abraham, 2017), effectively isolating many current generations of geese from formerly exploited natural habitats. The current concentration of geese foraging on terrestrial vegetation and the large number of wild goose dietary studies in recent years on such habitats makes it unlikely that we underestimate the degree to which geese currently feed on Charales, although we cannot rule out that this may have been different in the pre-agricultural past.

All three European swan species – mute swan Cygnus olor (sedentary in a large part of distribution area) and the more migratory whooper Cygnus cygnus and Bewick's C. columbianus, have been reported as grazers of submergent aquatic vegetation, frequently including green biomass of Charales (see Table 1). Mute swans feed on Charales throughout the annual cycle when biomass is sufficient and accessible (e.g. Holm, 2002) and have been reported doing so even as an introduced alien in North America (Bailey, Petrie & Badzinski, 2008). While both whooper and Bewick's swans are typically winter consumers, they are also apparently dependent upon overwintering Chara material at spring staging areas (such as Matsalu Bay in Estonia) to replenish diminished energy and nutrient stores en route to breeding areas (Rees & Bowler, 1991; Nolet et al., 2001). This may be especially critical for staging Bewick's swans in late springs in the Russian White Sea when sea ice covers more profitable Zostera marina beds and substrates containing Stuckenia pectinata tubers (such as in 1996). This is because the Charales grew in places where freshwater streams rapidly melted the ice in coastal ecosystems when no other green vegetation was available (Nolet et al., 2001). To compensate for the relatively low availability of such forage when restricted by ice during the vernal thaw, the swans consume large amounts of biomass and therefore have the potential to make a major impact on the local standing crop. The coot is unusual for being an omnivore that also shows an herbivorous feeding habit, feeding on many submerged macrophytes including Charales (Cramp & Simmons, 1977) and their oospores (Olsen, 1944) throughout their range (including Nitella in North Africa; Madon, 1935).

Although some authors imply that Charales are not important to American waterbirds (e.g. Anderson, 1958), they have long been reported to contribute to the diet of at least 14 different species there (e.g. McAtee, 1915; see also Knapton & Pietrie, 1999). Trumpeter swans Cygnus buccinator have been reported feeding on Chara spp. in the Yellowstone Ecosystem (Squires & Anderson, 1995). Chara spp. have also been reported in the oesophageal contents of northern pintail, northern shoveler and blue-winged teal Anas discolor in Mexico (Thompson, Sheffer & Balassarre, 1992). McAtee (1915) reported finding 1100–15,000 oospores in the guts of many dissected ducks, so this is a food source that appears important to several species, as well as to the invasive alien mute swan (Bailey et al., 2008) and young white-winged coot Fulica leucoptera in Argentina (Nores & Cerana, 2010). Non-breeding American flamingoes Phoenicopterus ruber ruber also feed on the tubercules and oogonia of Chara fibrosa on the Yucatan Peninsula of Mexico (Schmitz et al., 1990; Arengo & Baldassarre, 2002).

(3) Waterbird aggregative responses

Several studies have shown relationships between either biomass or extent of submerged Charales and the numbers of feeding-specialist waterbirds that have been reported in the literature to exploit these plants as source of food. For instance, van der Winden et al. (1994) and Ruiters et al. (1994) linked the red-crested pochard expansion in range and increase in abundance to improvements in freshwater lake quality and the concomitant expansion in the distribution and abundance of the Charales. Similarly, increasing numbers of moulting red-crested pochard on the Ismaninger reservoir and fishponds, Germany, fed on Chara vulgaris and several Potamogeton/Stuckenia species, which have spread following reductions in water column dissolved nutrients at the site, associated with the restoration of a clear water column (Köhler et al., 2009). In the Cotswold Water Park, England, densities of wintering common pochard on multiple flooded gravel pits of uniform depth, all characterised by dense Charales cover (mostly Chara globularis) were related to the extent of open water area, but these relationships were reduced on lakes with bank-side access or water-based recreation compared to nature reserves with restricted access, suggesting a numerical response modified by human disturbance (Fox et al., 1994). Hargeby et al. (1994) found correlations between the biomass and extent of Chara tomentosa stands and numbers of mute swans, coot and dabbling ducks in a Swedish lake following the natural recovery of submerged macrophytes. At the Dutch Lake Veluwe (3400 ha), after a period of severe eutrophication, the aquatic vegetation, mainly Chara aspera beds, recovered and began to increase gradually. The development of the vegetation coincided with a strong decrease in the bream Abramis brama population, probably the result of introducing a commercial bream fishery there during 1993–1997, which reduced fish biomass by ca. 65%, reducing sediment disturbance and enhancing transparency in the water column (Noordhuis et al., 2002). Despite abundant Potamogeton beds at the site in the 1980s, waterbirds only really returned in large numbers following the re-establishment of Chara. Of all correlations between filamentous algae, narrow-leaved Potamogeton/Stuckenia, zebra mussel Dreissena polymorpha and Chara biomass, eight of 12 herbivorous and omnivorous waterbirds (predicted to benefit from the presence of Chara) showed significant positive correlations with Chara biomass during 1987–1998, with mute swan (virtually absent before the return of Chara) and coot showing the strongest relationships (Noordhuis et al., 2002). Northern pintail and red-crested pochard numbers were correlated with Chara biomass and all of the recoveries in waterbird numbers were linked to improvements in water transparency, responding positively to increases in densities and extent of Dreissena and Chara and decreases in filamentous algae and narrow-leaved pondweeds, mainly Stuckenia pectinata (Noordhuis et al., 2002). Holm & Clausen (2006) found herbivorous waterbirds choose to forage at greater densities at one site where the more energetically profitable Chara species were superabundant compared to another adjacent site with the less-profitable Ruppia cirrhosa, in waterbodies that otherwise differed little (e.g. in water depth or nutrient status) from each other. Following drastic changes in water quality, Chara species disappeared from many Lithuanian lakes in the 1980s. In this region, Švažas, Stanevičus & Čepulis (1997) linked marked declines in breeding mute swan populations in SW Lithuania to the extinction of the Charales, their most important food plants, as was the case for coot reductions during post-breeding and autumn staging for the same reasons (Stanevičius, 1999). Likewise, when an invasive fish species, the common carp Cyprinus carpio was introduced to the Tablas de Daimiel National Park (Spain), exclosure and other experiments showed that enhanced levels of phosphorus and the presence of these invasive herbivorous fish (and not the alternative hypothesis of the presence of herbivorous waterbirds) were responsible for the loss of Chara meadows (Laguna et al., 2016). These submerged macrophyte beds formerly supported a year-round abundance of red-crested pochard, a species that completely disappeared following the loss of Chara. Armitage, Holloway & Rehfisch (2000) found that high densities of coot, mute swan and common pochard fed extensively on Chara papillosa, to the extent that there were significant positive correlations between Chara height and the densities of all three species at Hickling Broad, East Anglia, UK. Numbers of herbivorous mute swan, coot and dabbling ducks breeding at Lake Krankesjön, Sweden, during a period when conditions flipped between a clear and a turbid water column, were very highly correlated with biomass and area of Chara tomentosa, which thrived in years characterised by improved water transparency (Hargeby et al., 1994).

While we predict numerical responses of species known to feed on the green biomass and diaspores of Charales to increase with the extent and biomass of these plants, it is important to remember that Charales beds themselves also offer refuge to abundant invertebrate biomass seeking refuge from predatory fish (e.g. Diehl, 1992; Marklund, Blindow & Hargeby, 2001). Invertebrate biomass in Chara beds was more diverse and seven times greater than that simultaneously sampled in Stuckenia pectinata stands and areas of bare substrate from the same lake (Hargeby et al., 1994). The presence of such biomass associated with Chara beds presents a food resource for omnivorous waterbirds that might also be predicted to show relationships with changes in Charales biomass, especially since invertebrate species show diurnal movement cycles within such beds (Kuczyńska-Kippen, 2001) that increase availability. At the same time, very dense Chara growing to the surface reduces invertebrate availability to dabbling and diving ducks and thus contributes to their reduced use of such areas, as was found by Marklund (2000) for omnivorous mallard and tufted duck. These experiences may explain why omnivorous duck species benefit from the restoration of beds of Charales vegetation without showing specific numerical responses to the extent of the submerged macrophytes. This was the case in the study of Noordhuis et al. (2002), where mallard showed no relationship to Chara biomass and tufted duck increases were more related to increasing extent of Dreissena (a favoured prey species that also benefited from clear water conditions at Lake Veluwe) than Chara per se.

Freshwater fish use macrophyte beds (including Charales) for shelter and refuge, spawning, nesting and nursery sites, and in some cases as direct and indirect sources of food (Petr, 2000). However, a number of factors complicate the relationships between piscivorous waterbirds and Charales. For instance, the abundance and size class distribution of different fish species vary enormously with multiple factors, many of which are unrelated to submerged macrophytes, but piscivorous birds (fast-pursuit visual foragers) generally need a clear water column and an abundance of fish of adequate biomass, regardless of species. So, interpreting changes in piscivorous bird abundance can be challenging without a detailed understanding of the fish species composition, abundance and size-class distribution. For instance, studies at Lake Tåkern have shown that piscivorous species such as great cormorant Phalacrocorax carbo and goosander Mergus merganser were more abundant in years with limited Chara growth (Milberg et al., 2002). However, while the same authors speculated that dense submerged plants presented both a refuge to fish and a mechanical obstacle to avian fish pursuit, they also explained that the great cormorant had shown major expansions in range and numbers that also affected local abundance. A third piscivorous species in their study, the great crested grebe Podiceps cristatus, showed no strong correlation with submerged macrophyte abundance. Ironically, highly turbid anoxic waters can kill fish stocks and attract numerous heron and egret species, as well as cormorants for short periods of feeding when dead and dying fish are abundant (as in the case of over-manured fishponds in the Czech Republic, A. D. Fox, personal observations).

(4) Contributions of Charales to maintaining water clarity

Charales form a common component of the littoral zone in clean, unpolluted oligo- to moderately eutrophic water bodies (McCourt et al., 2017). However, with increasing eutrophication, Charales often become out-competed by submerged angiosperm macrophytes [e.g. Potamogeton/Stuckenia species (Ozimek and Kowalczewski, 1984; Pieczyńska, Ozimek & Ryback, 1988; Blindow, 1992a)], all of which ultimately completely disappear from extremely turbid lakes. It is however, possible to reverse this process, restoring Charales communities following lake restoration measures (Blindow, 1992b; Simons et al., 1994; van den Berg et al., 1999). This can include applying biomanipulation techniques, such as removal of planktivorous and benthivorous fish (Bernes et al., 2015) or reduction of nutrients (especially phosphorus; Richter & Gross, 2013). Once stabilised, Charales can form very dense underwater canopies, often greater than in beds of vascular aquatic macrophytes, which are highly effective at trapping sediment and restricting resuspension, especially of soluble nutrients, a major source of internal nutrient loading in shallow lakes (Søndergaard, Kristensen & Jeppesen, 1992; van der Berg et al., 1998a; Casanova, de Winton & Clayton, 2002; Schneider et al., 2015). In doing so, they restrict aqueous sources of nutrients to planktonic algae and the periphyton, the growth of which they also inhibit through allelopathic secretions (e.g. Forsberg, Kleiven & Willén, 1990; van Donk & van de Bund, 2002; Berger & Schagerl, 2003; Mähnert, Schagerl & Krenn, 2017; Złoch et al., 2018). Several Chara species overwinter as green material, suggesting accumulation of plant biomass beyond one growing season, although it is known that some Charales can efficiently take up nutrients through their rhizoids (e.g. Wüstenberg, Pörs & Ehwald, 2011; Raven, 2013). Charales have been reported to decompose more slowly than their vascular plant counterparts prolonging nutrient (including carbon and phosphate) storage in plant biomass compared to other plant communities (Kufel & Kufel, 2002). In addition, Charales precipitate abundant calcite during periods of intensive photosynthesis, leading to major carbonate accumulation (Pełechaty et al., 2013) which can immobilise phosphorus within the crystalline structure and lock this and carbon into the sediment (e.g. Murphy et al., 1983). Generally, the relative abundance and spatial distribution of Charales is correlated to water clarity and concentrations of total phosphorus and chlorophyll a (e.g. Hutorowicza & Dziedzicb, 2008), often associated with distinct clear/turbid water phases in the lifetimes of lakes, to which waterbirds respond in relation to the food abundance available from Charales (e.g. Hargeby et al., 1994; Moreno-Ostos et al., 2007, 2008). The effects described above often mean that once restored (usually through measures to severely restrict phytoplankton abundance), a clear water column and dense established Charales beds represent a naturally low-trophic-status stable state, which is discrete from the previous turbid system (Scheffer et al., 1993; Jeppesen et al., 1998) that was far less attractive to feeding waterbirds (e.g. Noordhuis et al., 2002). It is not just the dense standing crop of green Charales biomass that attracts feeding herbivorous waterbirds. Charales oospore densities were 6–15 times greater in clear-water ponds compared with those in turbid waters from a study of Belgian ponds (van Onsem & Triest, 2018) and so more likely to attract specialist filter-feeding waterbirds actively foraging on diaspores. The presence of Chara during such changes has a major effect on the biomass, diversity and community composition of the associated invertebrate fauna. For example, Chara beds were dominated by chironomids and gastropods whereas amphipods dominated the community in the rooted-angiosperm beds (Hanson, 1990; van der Berg et al., 1997), so affecting the aquatic and waterbird community composition responding to such changes in food supply.

As described above, after a period of severe eutrophication, the aquatic vegetation of Lake Veluwe, in the Netherlands, began to recover, showing dramatic increases in Chara cover and enabling colonisation by invasive zebra mussels. The presence of both forms (i.e. a dense charophyte vegetation canopy and an abundant filter-feeding bivalve) contributed to creating a clear water column and supporting the further rapid expansion of macrophyte beds dominated by Chara aspera (Lammens et al., 2004). The presence of such previously absent food sources consequently attracted large numbers of specialist waterbird species exploiting Chara as a food resource, or sources of food associated with dense Chara beds (Noordhuis et al., 2002; see Section III.3).

(5) Avian effects on Charales

Bird herbivory may have contrasting effects on the standing crop of growing aquatic plants. Some authors reported negligible effects on living vegetation (e.g. Kiørboe, 1980; Perrow et al., 1997; Marklund et al., 2002), while others reported a strong negative impact of herbivory by waterfowl on plant biomass (e.g. Lauridsen, Jeppesen & Andersen, 1993; Hilt, 2006; Hidding et al., 2009), but clearly the season and levels of grazing affect plant abilities to use compensatory growth to make up for biomass lost to grazing. Bird herbivory may also induce changes in relative abundance of different species making up the aquatic vegetation (e.g. van Donk & Otte, 1996; Weisner, Strand & Sandsten, 1997; Santamaría, 2002; Rodríguez-Villafañe, Bécares & Fernández-Aláez, 2007), so herbivorous waterfowl may potentially shape the species composition of aquatic plant communities in shallow wetlands by selecting more palatable species over others. Hidding et al. (2010) showed no effect of below-ground herbivory in spring by whooper and Bewick's swans, combined with summer grazing by waterfowl and fish on the biomass and macrophyte community composition (dominated by Chara aspera, with subdominant Stuckenia pectinata and Najas marina) present in a shallow Baltic estuary, Matsalu Bay, in Estonia during 1 year. However, S. pectinata was more abundant in plots closed to spring and summer herbivores, N. marina was more abundant in grazed plots, whereas Chara spp. biomass remained unaffected. Below-ground propagules of both C. aspera and S. pectinata were likely consumed by swans but the authors considered that since C. aspera bulbils were so superabundant this may have compensated for the losses.

A study on Lake Constance showed that feeding waterbirds (mainly coot, common and red-crested pochard) extensively depleted Charales biomass in shallower areas (<1 m water depth) in early winter, while deeper regions were only used later (Schmieder et al., 2006). At the end of winter, herbivorous waterbirds had almost completely removed available Charales biomass down to 2 m water depth. Enclosure experiments showed plant senescence had a negligible influence on Charales biomass loss until early February. Despite this major impact on winter biomass, avian herbivores likely had limited influence on subsequent Charales regeneration because of the production of innumerable oospores that contributed to the subsequent formation of dense Charales meadows at depths down to 4 m every year (Schmieder et al., 2006). In a study of the diet of summering and moulting waterbirds (dominated by coot) at the same site, comparison of vegetation in exclosure cages (protecting macrophytes from avian grazing) with unprotected grazed sites showed that herbivorous waterbirds depleted over 40% of Charales biomass at 1.5 m depth, although no effect was evident at greater depths (>2 m). Available food consisted mostly of Chara spp. (350 g m⁻²), since available animal food items were negligible, yet dense Chara beds persisted at the site despite such annual heavy grazing pressure (Matuszak et al., 2012). Long-term studies of the submerged macrophytes at Lake Krankesjön, Sweden also strongly suggest that waterbird herbivory on Chara spp. has no long-term effects on species composition, persistence or biomass of these macrophytes at this site (Hansson et al., 2010). Seventeen large-scale waterbird exclosure plots monitored over 2 years at Lake Botshol, Netherlands showed no significant increase in Charales biomass in the absence of grazing compared with the controls during transition from turbid to clear water and high Chara biomass, demonstrating that light limitation was the main factor controlling the collapse and return of Charales there (Rip, Rawee & de Jong, 2006). Charales biomass at Lake Veluwe, Netherlands was greater inside bird exclosures than outside as a result of overwintering waterbird exploitation, mostly after the end of the growing season, but this too did not affect the annual persistence of Charales at the site as a whole (van den Berg et al., 1998b). To what extent herbivorous birds may selectively exploit overwintering diaspores (including bulbils produced by Chara aspera) also remains unclear (van den Berg et al., 1998b). Migrating swans grazing on Chara spp. in spring did not seem to affect biomass later in the season in Matsalu Bay, Estonia, thought due to compensatory growth in summer when avian herbivory is limited and light attenuation high (Hidding, 2009).

At Lake Veluwe, a community of dabbling, up-ending and diving herbivores completely removed above-ground Chara biomass after arrival, large numbers of autumn-arriving birds concentrating conspicuously in parts of the lake where Chara was the dominant plant cover and hence provided an abundant food supply (Noordhuis et al., 2002). Ironically, reduction in biomass of some other submerged macrophytes through waterbird herbivory may actually be beneficial to Charales, as in the case of shallow lakes in the Netherlands and southern Sweden, where avian herbivory (especially by Bewick's swans depleting Stuckenia pectinata tubers which affects subsequent regrowth; Van Vierssen, Hootsmans & Vermaat, 1994) was shown to induce a shift in macrophyte dominance from S. pectinata to Chara spp. and Myriophyllum spicatum (van Wijk, 1989).

Outside of Europe, the experience is similar, autumn exclosure experiments carried out on Lake Erie (Canada/USA) showed that, compared to ducks and abiotic factors, two large herbivorous waterfowl (tundra swan Cygnus columbianus and Canada goose Branta canadensis) had no additional impact on annual above- or below-ground biomass of Chara vulgaris at the site (Badzinski, Ankney & Petrie, 2006). The population density of black swans Cygnus atratus feeding on Nitella hookeri at Hawksbury Lagoon, New Zealand correlated with plant biomass, but plant biomass removal by grazing was less than replacement by regrowth. However, nutrient recycling and bioturbation affected suspended solids mobilisation, phytoplankton abundance and light attenuation, which adversely affected plant abundance (Mitchell & Waas, 1996).

Existing studies suggest that Charales are not normally affected by avian herbivory (in the sense of suffering reduced annual persistence), even when autumn and winter grazing in shallow waters completely removes all biomass before the start of the following growing season. This effect may be because most intensive grazing in the observed systems tends to occur outside the main growth season (spring/summer/autumn) when densities of grazing avian herbivores are generally lower or absent. The effect is likely also mitigated by the release of very large numbers of sexual and asexual diaspores into the environment each season (e.g. Schmieder et al., 2006), which despite harvest by avian and other predators still maintains stock levels in substrates sufficient to maintain biomass and the dominance of Charales in the subsequent submergent plant community.

Indirect impacts of avian foraging on Charales should not be neglected. Bonis & Grillas (2002) claim that burial of oospores deeper than 2 cm inhibits their germination. Swans not only graze on macrophyte green parts, but also dig for tubers in shallow substrates, simultaneously removing thalli and potentially burying oospores to inhibit regrowth, but also bringing buried oospores towards the sediment surface to germinate. Such impacts are especially likely in sites where Charales grow interspersed by pondweeds such as Stuckenia pectinata or close to common reed Phragmites australis stands (A. Stīpniece, personal observation).

When water nutrient levels are not sufficiently low to prevent algal blooms, consumption of Charales by birds may trigger a change from the clear-water, macrophyte-dominated state to the turbid state discussed above (Van Donk & Gulati, 1995; Weisner et al., 1997). Occasionally waterbirds themselves can contribute to eutrophication of waterbodies, enhancing water turbidity and leading to the extinction of submerged macrophytes, potentially including Charales. This is especially the case where large-bodied colonial nesting birds (such as cormorants, gulls, herons and egrets) aggregate along lake shores to contribute high N and P loadings to water bodies sensitive to such external and internal inputs, although we could find no cases involving Charales specifically (see review in Klimaszyk & Rzymski, 2016).

(6) Waterbirds as a means of dispersal of Charales

Many researchers have been fascinated by the ability of Charales to colonise new and novel aquatic habitats often very rapidly as such opportunities become available (e.g. Wade, 1990). Many have been convinced that waterbirds are a principal agent of dispersal (e.g. Ridley, 1930; Olsen, 1944; Proctor, 1962), as for instance in the colonisation of the Faroes, Iceland and Greenland from freshwater habitats further south (Langangen, 1996; Langangen, Hansen & Mann, 1996; Hrafnsdottir et al., 2019). The presence of Chara species on Svalbard (Langangen et al., 2020) has been considered totally dependent upon their transport there by geese, in the absence of any migratory duck species regularly using these polar areas with links to suitable Charales habitats further south. Malone (1966) reported that two out of five 6-month-old female mallards fed to repletion on Chara sp. on one occasion regurgitated 2.5 cm balls of loosely compacted food 45 min later, which had not been altered by compaction in the gizzard nor the digestive process. Proctor (1962) found no viability for Charales thallus fragments after passing through a duck. In the same study, Proctor (1962) removed intact oospores of six Chara species from the lower digestive tracts posterior to the opening of the caeca, in 16 of 47 shot waterbirds, of which 40–60% germinated from eight of these birds. Although Charalambidou & Santamaría (2005) found oogonia of Chara species intact in faeces of mallard, teal and coot (demonstrating an ability to pass through the guts of these waterbird species and potentially to being transported elsewhere by them), they did not measure propagule germination rate, so the degree to which they escaped digestion and remained viable remains unknown. In contrast to previous studies of Chara spp. (e.g. Proctor, 1962), Brochet et al. (2010b) found low germination rates and viability among Chara oospores. Nevertheless, they reported that teal oesophagi can be full of thousands of Chara oospores (among the smallest of identifiable prey regularly found in this species; Tamisier, 1971), which constituted the most abundant taxon amongst viable diaspores in the lower gut of teal shot in the Camargue (Brochet et al., 2010a). Oogonia of Chara (C. vulgaris, C. globularis and C. aspera) were most abundant in the rectum of sampled teal (60% of all intact propagules), suggesting that small size favours internal transport through the gut of this species (Brochet et al., 2010a). Two Charales oogonia have even been found in the faeces of a wading bird, the common redshank Tringa totanus (Lovas-Kiss et al., 2019), so perhaps we should look further afield to judge the full extent of potential dispersal by birds. Oospores can survive desiccation and rewetting over periods from 4 years (Proctor, 1967) to 50–150 years in ghost ponds (Alderton et al., 2017), so are likely extremely long-lived and resistant to desiccation, further enabling their survival during transit by birds to colonise new habitats (Wade, 1990). Waterbirds can migrate relatively fast, so there is at least the theoretical potential for long-distance dispersal. Brent geese Branta bernicla have been recorded flying 2600–3500 km in less than 94 h (Clausen & Bustnes, 1998; Green et al., 2002b), while a Eurasian teal was tracked migrating 1285 km in less than 24 h (Clausen et al., 2002). However, it is generally considered that retention times are relatively short within the avian alimentary canal, ranging from an estimated 6 h in the mute swan to as little as 1.7–1.9 h in Eurasian wigeon and greylag geese Anser anser (Clausen et al., 2002). The same authors considered it likely that many species would deliberately eject most if not all contents of the alimentary canal prior to departure on migration (to reduce payload flight costs). It may also be the case that migrating waterbirds potentially alter their gut architecture to reduce body mass and metabolic maintenance costs (as is the case for shorebirds; see Battley & Piersma, 2005) and/or at least eject faeces in flight well before arrival at final destinations to avoid carrying unnecessary mass in flight. This does not exclude the possibilities of small diaspores persisting in the caeca (although this is not normally a site of entrapment of small seeds passing through the avian gut; Kleyheeg, Claessens & Soons, 2018) and folds of the alimentary canal. Figuerola et al. (2010) found that smaller ingested angiosperm seeds showed longer avian gut retention times, higher survival and germination rates than larger seeds, so the very small size of Charales diaspores may also be advantageous.

It is well known that seeds with hard integuments or of small size are more capable of passing through the avian intestinal tract and are more likely to germinate when ejected (Krefting & Roe, 1949; deVlaming & Proctor, 1968; Lovas-Kiss et al., 2020). In this respect, the small size and the potential protective function provided by calcification of oospores suggest that these diaspores theoretically have some advantages over angiosperms for both passing through the guts of waterbirds and being viable on ejection. Although this review has found a rich literature on this subject, there remains enormous potential for field and (more particularly) manipulative captive studies on waterbirds to assess the levels of ingestion/ejection of viable Charales diaspores of different species and levels of calcification after passage through the gut of a variety of waterbird species.