This section deals only with fish species in which aquatic macrophytes constitute part of their diet. It does not deal with species which, as adults, feed on phytoplankton or filamentous algae only.
Opuszynski (1992) and Opuszynski and Shireman (1995) classify herbivorous fish as those in which food constitutes more than 50% plant material by weight or volume, at least in some period of their life. Most species are opportunistic, changing their diet to whatever food items are abundant or available. Food habits also change with the age of fish and seasons, which may determine food availability. A generalised conceptual model gives major reasons why a herbivorous fish selects its food (Fig. 4).
Fig. 4. Conceptual model of how diet composition is determined in freshwater invertebrates. (From Lodge et al., 1998)
Opuszynski and Shireman (1995) mention that 24 families of fish contain freshwater herbivorous representatives. In total 37 species are listed as feeding on macrophytes, and two on “plants material”. Twenty of the macrophyte-feeding species belong to the family Cyprinidae, eight to the family Cichlidae. Among these, grass carp is the only fish used on a large scale for aquatic weed control. The authors do not indicate whether under macrophytes are meant aquatic macrophytes, or terrestrial macrophytes (such as flooded grass and other herbaceous plants of floodplains), or both.
Submersed, emergent and floating-leaved aquatic macrophytes are all subject to substantial grazing losses, which often alter relative abundance of macrophyte species and sometimes affect plant productivity (Lodge, 1991). Prejs (1984) gives examples of intensive grazing on Elodea canadensis and Ceratophyllum demersum by rudd and tench in Polish lakes. Most fish are opportunistic, feeding on a wide variety of food items according to their availability. This is also true of herbivorous fish which frequently ingest aquatic invertebrates dwelling on aquatic plants. The reason is that many aquatic grazers prefer bacteria and microscopic algae because these groups have higher protein and lower cellulose and lignin and ash contents (Lamberti and Moore, 1984). This is valid for grass carp which feeds on submersed, floating and emergent plants, as well as, e.g. mahseer (Tor tor), which scrapes moss and algae from rocks in rivers of India and Pakistan, or feeds on aquatic macrophytes. Opuszynski (1992) pointed out the scarcity of information on the proportion of animal food in herbivorous fish, mentioning microorganisms as the most protein-rich food in nature, and easier to digest than plant cells. Juveniles of herbivorous fish depend on high-protein animal food to survive, shifting gradually to mixed animal-plant food as they develop and grow. Opuszynski mentioned that when grass carp were fed only lettuce in experiments, they were hardly able to grow, but considerable improvement in growth was achieved when the fish were fed earthworm diet.
Plant food is less digestible than animal food because of the cellulose, which is difficult to break down. Grass carp has a muscular gizzard stomach where plant material is ground together with sand grains. Other herbivorous fish use pharyngeal teeth (pharyngeal mill) situated in the throut, to cut, tear and grind the plant material Both methods assist in breaking down the plant cell walls. Herbivorous fish compensate for the shortage of protein in plants (duckweed contains 31%, water hyacinth 24%) as compared to animal food (up to 80% protein) by ingesting higher quantities of food, the daily intake of which, in grass carp, may exceed their body weight. Many aquatic macrophytes have secondary compounds that may reduce grazing. Lodge (1991) in a brief summary of the literature, listed the major groups of compounds: alkaloids, flavanoids, steroids, saponins, phenolics (including tannins), glucosinolates. The digestibility of protein in plants decreases with increasing concentrations of tannins (Mitchell, 1974). Plants containing more than 6% to 7% tannin (dry weight) would be so low in digestibility as to be of little food value; most higher aquatic plants contain less than 2% to 3% tannin when harvested before maturity (Boyd, 1968), but tannin values tend to increase with age.
This chapter leans especially on the information contained in papers and publications by Schuytema (1977); Opuszynski and Shireman (1995); van der Zweerde (1990); Welcomme (1985); grass carp synopsis by Shireman and Smith (1983); special section on grass carp in the United States, as appeared in the Transactions of the American Society 1978, Vol. 107, No. 1; Proceedings of the Grass Carp Symposium (Anon., 1994). This is supplemented by material gleaned from a number of other papers. The section dealing with grass carp is not exhaustive, and the reader is advised to seek more information in the relevant publications listed above. Twelve other herbivorous fish species/genera are dealt with in more detail.
Adult grass carp eat exclusively plant material, but when young, grass carp are often considered facultative carnivores. When Tubifex worms were fed to fry (Haniffa and Venkatachalam, 1980) and fingerlings (Fischer, 1972) of grass carp a large proportion of energy intake was retained. When fingerlings were fed lettuce they grew very little, due to poor digestion and absorption; and when fed pellets made from pondweed they did not grow and had low absorption efficiencies. In comparison, larger grass carp fed Potamogeton crispus also had a large faecal loss but retained 12% of the consumed energy (Wiley and Wike, 1986).
While the use of grass carp for biological control of aquatic plants has been found to be advantageous in many situations, especially where chemical or physical control for whatever reason cannot be applied, Schuytema (1977) pointed out at a number of negative impacts, the possibility of which should be assessed prior to the introduction of grass carp. Among them are: uncertainty of the effect on native fish, possibility that removal of plants may eliminate endemic fish food and cover and waterfowl food; lack of knowledge of proper stocking rates; difficulty of live capture; possibility of natural spawning; possibility of introducing a disease vector, lack of knowledge of local plant preferences; and nutrients released into the water by excretion leading to increased primary productivity.
Whatever the negative aspects of using grass carp for aquatic macrophyte control, given existing problems associated with eutrophication and exotic plant species, large-scale aquatic vegetation control programmes using grass carp are likely to increase in the future (Bettoli et al., 1993). In the Netherlands the introduction of grass carp has been considered a success, regarding its control of aquatic weeds (Van der Zweerde, 1990). The area stocked with grass carp has steadily increased since 1977. In 1987 there were 1092 sites with grass carp and 820 users (a total of 270 tons in 2300 ha). Water Boards are the major users of grass carp, with other users having a larger number of relatively small sites.
Numerous studies have dealt with the aquatic plant preferences of grass carp. Because of food selectivity, the grass carp occasionally increases some stands of plant species by preferential feeding on one or more favoured plants. This situation may continue until the preferred food supply is exhausted, and then it usually starts to feed on less palatable species. Small-sized grass carp utilise soft vegetation while older fish can utilise harder, more fibrous plant material. Water hyacinth is usually avoided, as is the cattail (Typha). In some water bodies grass carp may prefer to feed on filamentous algae such as Cladophora and Spirogyra. Elsewhere, Nitella may be avoided. In total, however, grass carp have been reported to consume over 170 different species of aquatic macrophytes (Redding and Midlen, 1992). They also reported that grass carp consumed more softer and succulent submersed aquatic macrophytes when water temperatures were below 12–15°C. Prabhavathy and Sreenivasan (1976) reported that in India grass carp is known to ignore all aquatic vegetation in the presence of Hydrilla. Bhukasvan et al. (1981) reported that grass carp preferred submersed macrophytes such as Najas graminae and Hydrilla and the floating macrophytes (e.g. Azolla pinnata).
Vincent and Sibbing (1992), who fed grass carp various aquatic plants, followed how the fish chooses and chews its food. Given that leaf toughness, availability of particular leaves, and spacial availability of each particular species of macrophytes are characteristics that affect the energy costs of ingestion, the amount obtained by bite and the potential nutritive value are factors affecting the benefit derived from ingestion (Prejs, 1984). Vincent and Sibbing (1992) detected a clear selection for some plants because of the type of chewing mechanism of grass carp, i.e. pharyngeal teeth. Elodea canadensis, grass Glyceria fluitans, and filamentous alga Zygnema sp. were readily taken. Lemna minor, Potamogeton lucens and leaves of Typha latifolia were not acceptable. The authors suggested the possible limiting factors: Potamogeton and Elodea were probably avoided as the fish has difficulty in extracting a single, manageable stem. Grass cannot be swallowed quickly enough, and chewing is limiting with Typha as the leaves are too tough. Lemna is rejected perhaps because it has an unpleasant taste.
Grass carp was introduced into India in 1959, and since then it has been used in composite culture of carps, particularly in ponds infested with aquatic weeds (Nandeesha et al, 1989). It controls there Hydrilla, Najas, Ceratophyllum, Wolffia, Lemna, Spirodela, Ottelia, Vallisneria, Potamogeton, Utricularia, Trapa, Myriophyllum, Limnophila, Nechamandra, and algae Nitella, Spirogyra, Pithophora. It has been found ineffective in controlling the floating plants Eichhornia, Pistia, Salvinia and Nymphaea. However, Eichhornia, Hydrilla and Najas are used by grass carp for attachment of its eggs. Due to the conversion to excreta of aquatic macrophytes using grass carp in composite fish culture reduces the cost of fish production. In India, grass carp has also been used for control of aquatic macrophytes with limited success in the Chambal irrigation system.
In Egypt, experiments with grass carp were carried out in a small drainage canal in the Province of Giza near Cairo, with favourable results: submersed weeds were effectively controlled. Further experiments and management measures covered several other canals in the country. Grass carp also found a ready use as human food, with poaching of fish from the canals on a large scale at night (Van der Zweerde, 1990). The use of grass carp in one of the Egypt's main irrigation canals improved irrigation conveyance efficiency through decreasing the percentage of submersed aquatic weeds from 35 % to 6 % (Bakry, 1996), although not affecting the floating water hyacinth (Photo 1).
In pond experiments in Alabama, grass carp reduced water hyacinth plant size by eating the roots. When stocked at 16 per 0.04-hectare pond water hyacinth was eliminated within 30 days after the fish were stocked (Kilgen, 1978).
In the USA, Chara and Hydrilla seem to be the favourite submersed plants of grass carp, in Europe it is the exotic Elodea canadensis. For more information on plant preferences as food of grass carp see Schuytema (1977) and Shireman and Smith (1993). The selective feeding of grass carp may lead eventually to the establishment of an unwanted aquatic plant, such as Ranunculus in the Karakum canal in Turkmenistan (Charyev, 1984). Triploid grass carp is a functionally sterile strain of grass carp that have proved to be effective in controlling aquatic plants in southern California. Pine et al. (1990) followed the impact of grass carp on aquatic macrophytes in static and flowing water in canals in northern California in winter, spring and summer. Potamogeton pectinatus, Myriophyllum spicatum and Potamogeton nodosus produced longer shoots in canals with flowing water than with static water. During spring and summer, triploid grass carp showed a distinct variation in the preference of aquatic weed types in flowing and static water. Accessibility and ease of mastication were found to be more important in determining preference than nutritional quality of the plants.
PHOTO 1: Irrigation canal infested with water hyacinth. Nile delta, Egypt.
Van der Zweerde (1990) enumerated and briefly commented on the factors which influence successful aquatic weed management with grass carp: water depth, regularity of previous control measures, water temperature, grass carp density, size of the grass carp, predation, selectivity of aquatic plants by the grass carp, surface area of grass carp stocking in a channel network, and proper integration of the use of grass carp in the overall management programme of a water body. A successful control of submersed aquatic macrophytes also requires a certain level of dissolved oxygen concentration, with the lethal point for grass carp being 0.5 mg L-1, a level which may be reached for example in dense growth of Hydrilla (Sutton, 1985).
When grass carp (stocking rate 74 fish ha-1) eliminated aquatic vegetation over 3 600 ha in the 8 100 ha Conroe reservoir in Texas (Figs. 5a and 5b), this resulted in an increase in nutrient (phosphorus and nitrogen) concentrations, and a decline in water transparency by 40% as a result of a build up of phytoplankton (Maceina et al., 1992). Once the vegetation was removed, autotrophic production switched from macrophytes to phytoplankton, resulting in a nearly twofold increase in algal biomass as measured by chlorophyll a. Zooplankton was relatively unaffected by plant control for a year and a half, which was followed by a decline in cladocerans as a result of an increase in planktivorous fish.
Fig. 5a. Aquatic vegetation map of Lake Conroe (20 October 1981 and 13 October 1983), based on colour infrared remote sensing survey. No submersed weeds were detected by the method employed. (From Martyn et al., 1986).
Fig. 5b. Aquatic vegetation map of Lake Conroe (20 October 1981 and 13 October 1983), based on colour infrared remote sensing survey. No submersed weeds were detected by the method employed. (From Martyn et al., 1986).
Vegetation removal by grass carp in Lake Conroe system represented the single largest reduction in aquatic vegetation by any means in North America (Bettoli et al., 1993). Prior to the biological control using grass carp, the dominant submersed plants were Hydrilla verticillata, Myriophyllum spicatum and Ceratophyllum demersum. The most notable declines were observed for several small, phytophillic Lepomis spp, for bluegill Lepomis macrochirus, and for crappie Pomoxis spp. Biomass of largemouth bass Micropterus salmoides did not decline but the density of age-1 and older fish did decline. Density of threadfin shad Dorosoma petenense increased nearly fivefold. There was a collapse in the stocks of three cyprinodont species and two other species, i.e. bantam sunfish Lepomis symmetricus, and brook silversides Labidesthes sicculus, while catches of inland silversides Menidia beryllina and threadfin shad increased significantly. The original largemouth bass - crappie - hybrid striped bass (Morone chrysops × M. saxatilis) fishery was replaced by a channel catfish -white bass - hybrid striped bass - largemouth bass - black crappie fishery after vegetation removal. Hence, the structure of the sport fish community changed. Several changes were likely linked to changes in structural complexity and primary production resulting from vegetation removal. The littoral fish community shifted from a sunfish and shad community to one that also included sizeable numbers of cyprinids, inland silversides, and channel catfish. The total number of species declined over the 7-year study period. Available prey:predator ratio remained similar after vegetation removal for large predators but increased an order of magnitude for small predators. The responses of some species to complete vegetation removal in Lake Conroe are consistent with known life history requirements.
Numerous studies exist on the feeding impacts of grass carp on other fish species. Grass carp consumes phytoplankton and infusoria during the first 2–4 days of life, changing to zooplankton after 5 days. The fish do not become phytophagous until they are about one month old and 2.5–5 cm in length. Aquatic macrophytes become a regular constituent of the diet when the fingerlings reach 55 mm TL (Watkins et al., 1981). In experiments it was found that 7 to 11 months old grass carp would eat aquatic invertebrates and trout fry in the presence of palatable plants when there was no cover for the prey. Elimination of aquatic vegetation may have disastrous consequences on native aquatic invertebrates, fish and aquatic birds, as the balance between plants and the regular local fauna dependent on them is disturbed. Kilgen (1978), referring to the work of other authors, mentions that grass carp increased bluegill (Lepomis macrochirus) reproduction, reduced largemouth bass (Micropterus salmoides) populations, and reduced bluegill production in ponds. A study on more of 100 lakes in Arkansas, where grass carp was introduced as a biological weed control agent, showed that plant removal by grass carp did appear to improve the condition factor of largemouth bass, bluegill and redear sunfish (Lepomis microlophus), and Bailey (1978) suggested that the introduction of grass carp will not always improve or harm fish populations. This has also been confirmed in the study of 8 lakes in Florida, which have experienced the long-term removal of all aquatic vegetation with grass carp (10 to 15 years): there was no consistent trend with aquatic macrophyte removal. Some grass carp lakes supported excellent fish populations, and some did not. From this study Canfield and Hoyer (1992) concluded that the long-term loss of aquatic macrophytes from a lake ecosystem will not necessarily lead to a decrease in a lake's total or sport fish population, but there exists a potential for such a decrease.
Reviewing the management of aquatic vegetation with grass carp in Iowa over the period 1973–1993, Mitzner (1994) pointed out that in the 2 000 private waters that have been stocked with grass carp, partial control of vegetation is rarely attained: normally there is an all-or-none response. In lakes with 100% control of vegetation there are only minor changes in sport fish populations. According to Mitzner, sport fishing is not adversely impacted by grass carp introductions and in most cases catch has increased because of more available access to shore by anglers fishing for bluegill and crappie.
The beneficial influence of grass carp on tench in the Lancaster Canal, England, is described by Petridis (1990). Grass carp, through moderate consumption of aquatic plants (dominant species Elodea nuttallii) produced better conditions for the exploitation of benthic organisms by the benthophagous tench (Tinca tinca), because tench (>35 cm fork length) were able to search for food in the open areas created. This also enabled the fish to make use of zooplankton, mainly the large species Eurycercus lamellatus whose restricted phytophilic habitat made it vulnerable to predation. Epibenthic animals, in particular gastropods and Asellus aquaticus, suffered from the predation, which caused severe reduction of them in the site totally depleted of vegetation. In this site, tench utilized a diet in which red chironomids were the dominant prey group together with a large amount of fine detritus. Petridis (1995) suggested that a moderate control of vegetation in temperate climates, using the grass carp, could be a satisfactory management technique for the improvement of aquatic habitats. Reviewing literature, the same author mentions an increase in the growth rate of bream (Abramis brama) when stocked with grass carp. Benthivorous species are generally considered to increase their production and growth, benefiting chiefly from the direct utilization and nutrient release of the undigested grass carp faeces. In England, tench can be stocked with grass carp at high densities, because both species occupy different ecological niches, and both species benefit each other through feeding activities.
The changes in the natural habitat of some Arkansas lakes (USA) stocked with grass carp resulted in replacement of Lepomis microlophus and Esox niger by largemouth bass (Micropterus salmoides), bluegill (Lepomis macrochirus) and the gizzard shad (Dorosoma cepedianum) (Baily, 1975). Grass carp stocking also led to a decrease in crayfish (Procambarus clarkii) production in small ponds (Forester and Avault, 1978). The interference seemed to arise from the competition for plant food. The grass carp affects other fish through interference with their reproduction, broadening or narrowing their food base, and decreasing refugia (Shireman and Smith, 1983). In the former Soviet Union grass carp introductions have adversely affected pike, perch, crucian carp (Carassius carassius), and roach (Rutilus rutilus). In a USA reservoir, the stocked grass carp through the subsequent elimination of vegetation disrupted the macroinvertebrate food base and this resulted in a 50 % reduction of centrarchid biomass (Shireman and Smith, 1983). In the Austrian lake Neusiedler grass carp was introduced alongside six other alien species to enhance the fisheries. Its presence has resulted in almost the complete disappearance of submersed macrophytes, an important spawning habitat for a number of fish species (Mikschi et al., 1996). While it is not possible to separate the impact of this fish from other impacts, such as water level regulatory measures, which led to a loss of habitats within the reed belt due to drying out, and eutrophication, it is believed that its presence contributed to the disappearance of some of the indigenous fish species from Neusiedler Lake.
In Lake Marion, South Carolina (USA), removal of Hydrilla verticillata by triploid grass carp led to an increase in littoral fish, especially Centrarchidae, as hydrilla decreased from a maximum of 4 700 ha (approximately 50% of the surface area) to less than 100 ha. The mean lengths of most littoral species were similar during the study period, during which, in spite of a substantial decline in hydrilla, other forms persisted. This provided an intermediate level of structural complexity, preventing any detectable negative effects on the littoral fish (Killgore et al., 1998).
In reservoirs, aquatic vegetation may conflict with other reservoir uses, such as boating, swimming, and electric power generation. Some reservoirs are important habitats for migrating and wintering waterfowl which is foraging on aquatic vegetation. The use of grass carp to control aquatic vegetation in some reservoirs may reduce habitat quality for waterfowl because plants preferred by grass carp are also important as waterfowl food and as habitat for invertebrate food items (McKnight and Hepp, 1995). In Guntersville Reservoir in Alabama, introduced grass carp reduced the presence of native aquatic macrophytes by 44%, but at the same time the invasive exotic milfoil (Myriophyllum spicatum) increased by 15%, as grass carp finds this plant less palatable. Although milfoil is eaten by American coots (Fulica americana) and some ducks during autumn and winter, a mixture of native plants and milfoil is more desirable. Native aquatic plants provide better habitat than does milfoil for aquatic invertebrates, which are important dietary components for a variety of waterfowl species (Krapu and Reinecke, 1992). Milfoil also is higher in fibre than are native aquatic species typically eaten by waterfowl in southeastern USA and most ducks and geese are limited in their ability to digest and assimilate fibre. Use of grass carp to control aquatic vegetation in Guntersville Reservoir reduced the quality of waterfowl habitat by reducing the biomass of the native aquatic plants.
In Florida Lake in South Africa, which for a number of years experienced a major invasion of Potamogeton pectinatus, Lagarosiphon muscoides and L. major, triploid carp was introduced to control these aquatic plants (Venter and Schoonbee, 1991). Within a period of one year the mean wet mass standing crop of the weeds in the lake was reduced to one sixth of the initial value. Reduction in the weed cover coincided with changes in the populations of aquatic plant-eating birds, such as coot (Fulica cristata), Egyptian goose (Alopochen aegyptiacus), and yellowbilled duck (Anas undulata). Owing to the reduction in plant presence by grass carp, small fish became more exposed to fish-eating birds such as reed cormorant (Phalacrocorax africana) and whitebreasted cormorant (P. carbo), which have both increased in numbers at the lake following the clearing of aquatic macrophytes from the surface.
For the impact of grass carp on sport fish see Section 11.
Stocking rates of grass carp and the age/size of the stocked fish differ from water body to water body and would be determined by many factors, including the amount and species of aquatic plants present, water temperature, and the objective of the use of grass carp as biological control agent. In Russia the stocking rates could be 79 kg ha-1 of one-year fish to clear ponds of vegetation, and 254 kg ha-1 to keep canals clear of aquatic plants. In India, effective stocking rates ranged from 654 to 5200 ha-1 for fish of various weights. In fish culture ponds in the USA a rate of 247–1235 fingerlings ha-1 efficiently cleared aquatic weeds from ponds, and in lakes twenty-five of 20–25 cm fingerlings ha-1 were sufficient (Schuytema, 1977).
Stocking rates for grass carp for the control of aquatic macrophytes are usually estimated from the surface area of macrophyte coverage. But this reduces the predictability of the outcome because this approach does not account for plant biomass, which can vary with plant depth and the plant species in question (Killgore and Payne, 1984). Osborne et al. (1982) determined the stocking rate based on whole-lake plant biomass: 14–20 grass carp per metric ton (wet weight) of plants would eliminate Hydrilla verticillata, and the 20 grass carp per metric ton would do this in less than 4 months. For the lakes in the Pacific Northwest of the USA, where water temperatures are lower than in Florida, Bonar (1990) estimated grass carp stocking rates of 4.7–15.1 fish per metric ton. Cassani et al. (1995) tested in eight small warm-water impoundments (0.8–45.3 ha) of Florida various stocking rates for triploid grass carp to identify a threshold rate at which macrophytes were suppressed but not eliminated. Stocking rates of 4–8.4 grass carp per metric ton of vegetation (wet weight) resulted in gradual reduction to zero macrophyte biomass in 8–17 months. The plants in question were Najas guadalupensis, Hydrilla verticillata, Ruppia sp., Chara sp. (Fig. 6). All these plants are readily consumed or preferred by grass carp, but slight differences in chemical composition of different populations of the same plant species can potentially affect preference and consequently consumption rates. According to Bonar et al. (1993) grass carp consumption was positively correlated with the concentration of calcium and lignin and negatively correlated with the concentrations of iron and cellulose. In the Cassani et al. (1995) experiments shifts in macrophyte species composition took place at sites where macrophytes were suppressed or began to regrow after a period of zero macrophyte biomass. However, the number of species present and composing at least 4% by weight of the total macrophyte biomass did not decline. Osborne and Sassic (1981) who studied the impact of grass carp on hydrilla infestation in a central Florida lake concluded that only small grass carp should be stocked, and that one can expect the fish become ineffective as a means of weed control as they approach a body weight of 14 kg, due to their reduction in the absolute, as well as the relative weight of vegetation consumed with increasing size.
In New Zealand incremental stocking of 60 grass carp (total 218 kg live weight ha-1) in a small 1.9 ha eutrophic sand dune Lake Parkinson led to the apparent eradication of Egeria densa, and to a more than 99% reduction in the cover and biomass of other macrophytes. Grass carp numbers were then gradually reduced by trammel netting and seining, with the last fish being removed using rotenone. Five years after the removal of grass carp predominantly native macrophytes re-established themselves in the lake. The emergent Eleocharis sphaceolata returned to the margins, and the submersed communities dominated by Potamogeton ochreatus and Nitella hookeri grew to a maximum depth of 5 m (Tanner et al., 1990).
In most water bodies outside its area of natural distribution spawning of grass carp must be induced artificially. However, a number of countries reported natural spawning where the fish has been transferred. In the Tone River in Japan grass carp has spawned since 1947, in a reservoir in Taiwan the escape of fingerlings from a pond resulted in its establisment and natural reproduction. Natural reproduction has taken place in a river and lake system in Mexico after the fish were released following water hyacinth control experiments. In Turkmenistan, grass carp reproduces in irrigation canals. Some countries (e.g. New Zealand) have regulated the introduction of grass carp, and some states of the USA have banned its use.
In some countries the ban on the use of grass carp has been eased under condition that a sterile diploid or triploid form of the grass carp is used. In New Zealand, only sterile fish are released for weed control (De Zylva, 1996). In the laboratory, triploid grass carp is known to exhibit a slight reduction (10%) in ingestion rate (Wiley et al., 1986).
A number of fish diseases has been introduced with the introduction of grass carp into Europe (for more details see Shireman and Smith, 1983).
Fig. 6. Macrophyte species composition at three sites before and after triploid grass were stocked. (Cassani et al., 1995).
Environmental impact of grass carp on the aquatic ecosystem cannot be disputed: by suppressing some species of macrophytes other species may increase; by grazing off some plants the spawning substrata of important fish may disappear; without a counterbalance of phytoplankton-feeding fish, water quality deteriorates. Aquatic systems function best with a moderate abundance of aquatic macrophytes and introduction of grass carp could assist in reaching such equilibrium.
When using grass carp in a multi-use lake (in North America) Leslie et al. (1987) made the following recommendations:
A grass carp symposium in Florida, USA (Anon., 1994) met to discuss the effects of stocking grass carp for aquatic plant control in the USA. Even after 25 years since stocking grass carp for the purpose of plant control, some management concerns remain open and require better answers. Grass carp has been used especially to suppress the invasive hydrilla, which has rapidly established itself in many water bodies of the USA. Only in Florida, well over 2000 water bodies have been stocked with grass carp. Haller (1994) gave two possible scenarios on the use of grass carp to suppress hydrilla in a landlocked prime sportfishing lake of 2000 ha area. In the first scenario, by the end of the second year, grass carp, after consuming hydrilla, including the regrowth of the second year, starts feeding on native plants, and in the fifth year there are no aquatic macrophytes left in the lake. The lake remains plant free for 20 years, and revegetation is attempted to improve sport fishing, but is unsuccessful because there is a residual grass carp population still in the lake. In the second scenario, Haller shows a combination of application of herbicide and grass carp stocking, but also shows the presence of adverse conditions such as heavy predation on the stocked fish, low water level promoting regrowth of hydrilla, poor impact of herbicide in repeated treatments in the following years. With further stocking inputs of grass carp to make control more effective, accompanied by rising water level, the result is as in the scenario one. In both situations one ends with a lake with very little emergent vegetation, and no submersed vegetation for 15–20–25 years. According to Haller, at present, there are no clear-cut guidelines for stocking rates of grass carp which could be applied to all the water bodies, and little is known on the long-term effects of grass carp on the ecosystem of such water bodies. At the same symposium Brakhage (1994) pointed out that at present every system will have to be treated separately with a commitment to monitor the effects of the control programme and modify the approach. The debate will continue over the appropriate amount of aquatic vegetation needed in a lake system, if any, to maintain healthy fisheries. This should also involve considerations of wildlife as numerous birds, reptiles, amphibians, and mammals rely on wetlands for their survival. The majority of these require aquatic macrophytes, and the boundaries of their desirable habitat are generally delimited by the occurrence of aquatic macrophytes within a system.
Despite the extensive use of grass carp for control of aquatic plants, the impact of the control is still not easy to fully predict. Variations in nutrient loading, macrophyte seasonality, climate among the variety of water bodies, as well as variations in consumption rates and preference for certain plant species complicate vegetation management with grass carp, especially where mixed species communities occur (Cassani, 1995). Use of stocking rates that do not account for these differences are largely responsible for the unpredictable impacts of stocking grass carp. Other factors that contribute to the unpredictability of using grass carp for vegetation control are inconsistent survival rates associated with size at stocking, poor water quality, and variable predation pressure. Resource managers will need to place more value on macrophyte suppression rather than elimination so that the low cost of using grass carp is balanced with potentially long-term impacts (Cassani, 1995).
When 5 to 8 cm long Tilapia zillii shifts from predominantly animal to predominantly herbivorous diet. In Africa, T. zillii is known to feed readily on Hydrilla (Pieterse, 1981). In Lake Naivasha, Kenya, Siddiqui (1977) reported that macrophytes represent 67.7% of its diet, but a more recent information by Muchiri et al. (1995) for the same lake has shown that this species is omnivorous (Fig. 7). They also compared the food web of T. zillii with that of another common Lake Naivasha tilapia Oreochromis leucostictus (Fig. 8).
Fig. 7. Diet composition in Oreochromis leucostictus and Tilapia zillii in Lake Naivasha. (From Muchiri et al., 1995).
Fig. 8. Food webs of Oreochromis leucostictus and Tilapia zillii in Lake Naivasha. (From Muchiri et al., 1995).
In Volta Lake, Ghana, in T. zillii and T. rendalli higher plants (predominantly of terrestrial origin from flooded land, such as grass) formed 61.4% of the total food eaten (Petr, 1967). In Egypt, T. zillii fed upon both plant and animal material. Barnley (undated), who for the Uganda Medical Department prepared a publication on anti-malarial precautions in the management of fish ponds, cautioned that T. zillii can eat only leaves of comparatively soft plants, and as soon as they have eaten all, they will starve and die out unless an alternative source of vegetable food is supplied. In fish ponds, if present in sufficient numbers, it will keep the marginal grass eaten down to an extent at which it and other fish (e.g. Lebistes) can catch and eat the larvae of mosquitoes.
Preferential feeding of T. zillii was observed by Buddington (1979). He found it to prefer Najas guadalupensis to Lemna, Myriophyllum and Potamogeton pectinatus. Saeed and Ziebell (1986) found it to prefer Chara, followed by Najas marina, Elodea densa and Myriophyllum exalbescens. T. zillii avoided bushy twigs or bulky stems of such plants as N. marina and E. densa and fed on leaves and soft slender stems which are easy to grasp and separate.
In the 1950s the cichlids Oreochromis leucostictus, Tilapia zillii and Oreochromis niloticus were introduced in Lake Kyoga, Uganda. Lake Kyoga, situated a short distance downstream of Lake Victoria, receives the Victoria Nile. Both species are captured in traps and gill nets set along the marginal papyrus mats or among aquatic macrophytes, while O. niloticus supports a very active fishery among floating islands of papyrus (Twongo, 1995). T. zillii occurs mainly under the cover of submersed and floating macrophytes such as Ceratophyllum, Myriophyllum, Potamogeton, Nymphaea and Pistia, and in sheltered bays, often close to the papyrus fringe. The wide use of seine nets has led to the reduction in macrophyte cover and this has been suspected as being one factor which has contributed to the decline in stocks of T. zillii. Another factor could be the competition for nursery grounds with other tilapias, i.e. O. variabilis and O. niloticus.
T. zillii was introduced in the USA for macrophyte control (Shireman, 1984). This species has established self-sustaining populations where water temperature (>10°C) is suitable for overwintering, such as in Arizona and California, or in thermal springs and heated power plant discharges. The results of the impact of this fish on aquatic macrophytes depended on the initial stocking density, the target macrophyte species, and the presence of native fish. The success of the use of T. zillii in controlling aquatic plants in ponds with sport fish was found to depend on high survival of tilapia young. Bickerstaff et al. (1984) found in experiments with T. zillii and the predator Lepomis macrochirus that tilapia survival was highest at 75% artificial plant cover. L. macrochirus and T. zillii, when they appear in the same water body, will occupy the shallow areas near submersed vegetation and are likely to interact with each other. Extrapolation of the survival data from the experiments indicated that too few T. zillii, when stocked with predatory sport fish, would survive to the size where they could adequately control aquatic plants. However, at the same time Bickerstaff et al. (1984) noted that tilapia fry never retreated into plant cover for protection; they appeared to depend on tight schooling and parental protection for survival.
In a North Carolina (USA) power plant cooling reservoir T. zillii rapidly established a reproducing population after inadvertent introduction in 1984. It eliminated the submersed macrophyte community, including a 57-ha infestation of Egeria densa, within one year. Just prior to macrophyte disappearance from the reservoir the tilapia density and standing crop were 1 080 fish ha-1 and 16.6 kg ha-1 respectively (Crutchfield et al., 1992). By 1985, nearly all macrophytes had disappeared, and after 1985 no submersed macrophytes were found in the reservoir. When macrophytes were scarce or absent, tilapia switched to a diet dominated by detritus. The ability of tilapia to change to alternate food permitted the population to continue expanding in the absence of macrophytes. Changes in water quality were minimal after macrophyte removal with no increased nutrient enrichment. Factors leading to the establishment of the tilapia were an overwintering refuge provided by continuous thermal discharge of water temperature above 10°C from the thermal plant, a paucity of predators largemouth bass (Micropterus salmoides) and bluegill (Lepomis macrochirus), and the species ability to utilise alternate food sources following macrophyte removal.
Macrophytes density decreased with no apparent effect on native fish in some southern California irrigation canals where T. zillii was introduced in 1971 (Hauser, 1975). In the Sonoran desert in California (Legner et al., 1975), T. zillii and O. mosambicus stocked at a high rate were found not to have an adverse effect on the sport fish largemouth bass. As T. zillii also consumes sprouting rice, one should very carefully consider its use in irrigation canals bringing water to such fields.
T. rendalli (formerly T. melanopleura) has demonstrated its potential in controlling aquatic macrophytes by removing them completely from some impoundments (Junor, 1969). However, in some situations the fish has had less impact, especially where the macrophytes are too dense. Higher aquatic macrophytes such as Hydrilla, Chara, Sparganium, Potamogeton, Leersia, Lagarosiphon, Carex, Typha, Cyperus papyrus and Paspalum have been found in the stomach of this species (De Bont et al., 1949). According to Caulton (1977) the diet of T. rendalli is strongly related to the abundance of aquatic plants. In Lake Kariba (Zambia/Zimbabwe), during the period of invasion of this reservoir by Salvinia molesta, rootlets of this plant appeared in 19.6% of the investigated stomachs, representing 6.7% of the total food taken, while vegetable detritus represented 60% (Mitchell, 1976). With the decline in salvinia from the 1970s this plant has been gradually replaced by increasing amounts of rooted macrophytes, especially Lagarosiphon ilicifolius, which grows down to a depth of around 10m. This has been accompanied by an increase in abundance and biomass of benthic invertebrates, with a spectacular increase in bivalves (Karenge and Kolding, 1995).
In experiments to find out whether T. rendalli is a selective feeder giving preference to some plants as compared to others, Chifamba (1990) fed the fish in tanks the following plants: Ceratophyllum demersum, Lagarosiphon ilicifolius, Vallisneria aethiopica and Najas pectinata. The fish preferred V. aethiopica, N. pectinata, L. ilicifolius and C. demersum in that order. In the wild, the fish feeding preference was as follows: V. aethiopica, L. ilicifolius, N. pectinata, C. demersum. In both cases, C. demersum was the least preferred food. The preference of the fish for V. aethiopica is explained as given by its high protein and ash content. However, other factors such as palatability, a lot of structural material and polyphenolic compounds, abundance of the particular plant, are also known to affect selection.
Small fish feed on cladocerans, and there is a shift to filamentous algae in fish larger than 50 mm (Munro, 1967). LeRoux (1956) observed a shift away from chironomid larvae at about 130 mm TL. In Malawi, the indigenous fish Tilapia rendalli has been used to control aquatic plants in rainfed ponds. The analysis of stomach contents has shown that juvenile fish of less than 150 mm TL prefer to feed on filamentous algae, followed by submersed macrophytes such as Myriophyllum and Vallisneria, and on softer emergent vegetation (Brummett, 1995). In Malawi, the use of this fish to control aquatic plants in fish ponds was not found to be a practical proposition, as only fish of an already harvestable size could do the job.
In the USA this species is considered potentially harmful because of its competitiveness and reproductive potential. It tolerates salt water and could spread along the coast of Florida from watershed to watershed. It was also feared that this fish could feed on rice seedlings if stocked in irrigation canals. (Schuytema, 1977, and references therein).
In Papua New Guinea, T. rendalli was introduced during 1990–92 into the Sepik-Ramu catchment on the New Guinea island with the purpose of enhancing the fish production on floodplains and in floodplain lakes (FAO, 1993). Many floodplain lakes and associated wetlands are rich in aquatic macrophytes. By 1992 the first fish were already entering the local fish markets.
Lahser (1967) reviewed the previous literature on its feeding preferences: the fish showed its preference for filamentous algae and ability to digest blue-green algae. In his own experiments Lahser (op.cit.) reported that leaves of aquatic plants are removed by O. mossambicus to obtain attached periphyton, and that the increase in turbidity caused reduced macrophyte densities. Close observations of feeding fish indicated that the consumption of many aquatic macrophytes was incidental to the removal of periphyton using the plants as a substrate. Leaves, stems, and roots were scraped or rasped to shreds; the plants were killed and consumed secondarily. But Lemna and Azolla were consumed in amounts equal to those of filamentous algae. Cabomba, Myriophyllum, Potamogeton, Vallisneria, and Najas were eaten extensively, Eichhornia, Brasenia and Ludwigia were killed through destruction of roots and stems but were not consumed in any appreciable amounts. Casual observations in the field showed that O. mossambicus would eliminate aquatic macrophytes (such as Najas, Heteranthera, Chara) and marginal vegetation (Zizania, Setaria, Paspalum, Echinochloa, Cyperus, Polygonum) from the bottom and margins of rearing ponds. This elimination is effected through grazing and through increasing the turbidity of the water during nest building.
Blue tilapia (O. aureus), stocked at 500 or 2500 adults ha-1 in small ponds in Oklahoma (USA) successfully controlled submersed aquatic vegetation dominated by Najas and Chara (Schwartz et al., 1986). The speed and degree of control were proportional to initial stocking density, with effective control observed in low density ponds and high density ponds within 120 and 90 days, respectively. Blue tilapia uprooted and deleafed plants, and there was an increase in turbidity, in water temperature and dissolved oxygen levels. The authors also reported a significantly lower stratification in the experimental ponds, perhaps a result of an increase in the wind force on the water surface without macrophytes, and this resulted in a better mixing. Schuytema (1977) cautioned against the use of O. aureus, as this species has spread widely through western and central Florida, where it is competing with the native fish species, and dominates the fauna in many eutrophic Florida lakes. In the state Oklahoma, where Schwartz et al. (1986) carried out their experiments, this danger does not exist, as the fish generally could not survive the winter temperatures.
In Israel, O. aureus has been found to contribute to cleaning water in reservoirs by feeding on organic sediments (Leventer, 1981).
Feeding experiments have shown that O. aureus will lose weight when fed pure plant diet (Okeyo, 1989). This has confirmed similar experiments on O. mossambicus and Tilapia zillii, and it shows that supplemental protein is required to support the growth of herbivorous fish.
The gourami (Osphronemus gourami) has been considered useful in controlling some submersed macrophytes in Asian ponds and reservoirs. Edwards (1980) reported this species to feed mainly on plant leaves. It was introduced into irrigation wells in India from Java to control submersed macrophytes. In India giant gourami has also a large appetite for Pistia stratiotes, on which mosquitoes transmitting filariasis, breed. Full grown gourami consumes 300 g of Pistia day-1, and can clear a 1-ha pond in a month (Anon., 1989).
This species does not feed directly on aquatic macrophytes, but the macrophytes represent an important link in its life cycle and indirect source of food. Mature fish build nests to spawn in the macrophytes. The fry, which hatches within 24 hours, after the absorption of yolk sac will feed on phytoplankton and zooplankton. The adult fish feed on periphyton and small invertebrates. In Thailand fish ponds aquatic weeds such as Eleocharis equisetoides, Paspalum conjugatum and Hymenachne myurus are cut and serve as fertilizer the presence of which results in zooplankton bloom in a short time (Boonsom, 1984). The fish is an important food species in southeastern Asia, where in the nature it is common in swamps, canals, lakes or in any shallow water and sluggish current. It is also commonly grown in rice fields.
Pathani (1980) and Pathani and Joshi (1980) found that in India mahseer (Tor tor and Tor putitora) feed on and control the growth of submersed plants such as Ceratophyllum demersum, Myriophyllum sp., Hydrilla verticillata and Vallisneria spiralis. In the Narmada River, India Desai (1970) found the food of Tor tor dominated by macrophytes, feeding mainly on aquatic macrophytes and filamentous algae, but it also consumed molluscs and insects. Desai (1970) believes that mahseer could be a useful fish for controlling both aquatic macrophytes and with them associated molluscs, intermediate hosts of trematodes causing parasitic infection of fish. In Lake Govindgarh in Madhya Pradesh fingerlings up to 160 mm were found to subsist mainly on macrophytes, while adult fish over 200 m preferred animal food, such as insects, molluscs and fish (Pisolkar and Karamchandani, 1984). A study on the food and feeding habits of Tor tor in Meghalaya, northeastern India, where Dasgupta (1990) collected the fish from the Simsang River, has shown that while the larger fish feed predominantly on algae and macrovegetation, in its juvenile stage they consume more insects. Dasgupta does not provide information whether they feed on submersed terrestrial vegetation during floods, or on aquatic macrophytes. In the more recent publication, Desai (1992) compared the food of Tor tor from the Narmada River with that of Govindgarh and Bhimtal lakes and the Bhagirathi River. The first two localities are in the plains of Madhya Pradesh, the last two are situated in the Himalayan uplands. He concluded that in lakes and reservoirs with clear water and the presence of aquatic macrophytes, the fish would do well and should be stocked. Mahseer will also feed seasonally on aquatic plants in lowland rivers, such as Narmada, where the clarity of water during summer and winter favours abundance of macrophytes and algae. In more turbid waters and during floods mahseer will become an opportunistic feeder taking macroinvertebrates, crustaceans, molluscs, attached algae, plant detritus etc.
The cyprinid Puntius (=Barbodes) species are generally omnivorous, with a tendency towards feeding on plants. Nandeesha et al. (1989) summarized the information available on the feeding habits of four species of Puntius in India. P. pulchellus, which reaches 8 kg weight in the Anjanapur reservoir in Karnataka and supports there a good fishery, was found to feed on Cyperus, Typha, Scirpus, Leersia, Pseudorphis, Hydrilla, Vallisneria, Lemna, and also on the roots of water hyacinth. According to Devaraj and Manissery (1979) this species shows a great promise in controlling aquatic weeds in ponds. Fingerlings stocked in cisterns with Lemna and Hydrilla fed on them at a rate of more than 50% of their body weight per day. Nandeesha et al. (1989) found P. dobsoni and P. sarana to feed on Chara, Vallisneria, Hydrilla, diatoms and green algae. P. kolus prefers planktonic algae and plant matter, but also takes molluscs.
In Sri Lanka, the diets of Puntius amphibius and P. dorsalis consist of 23 and 27 identifiable plant species, respectively, of higher plant leaves and animal matter (De Silva et al., 1980). In both species the major contribution comes from the plant material. In reservoir Parkrama Samudra in the littoral area with dense cover of Ceratophyllum, P. filamentosus was the dominant species in fish catches. This fish cuts the plant into pieces between 2 and 5 mm length, which, however, are poorly assimilated as the digestive system of the fish is unable to attack crude fiber. Hofer and Schiemer (1983) have suggested that the fish probably obtains much of its nutrition from the animal, bacterial and algal periphyton. Puntius sarana, another important species of the reservoir fisheries in Sri Lanka, appears there to be rather omnivorous than herbivorous.
In Bangladesh, Puntius javanicus was effective in controlling aquatic vegetation under experimental conditions. In Indonesia and a few other Asian countries it serves the dual purpose of fish production and weed control.
P. gonionotus, native to Thailand, Malaysia, Laos, Vietnam and Java (Indonesia) is now widely distributed throughout the Asian region, due to its use in aquaculture and introductions to establish commercial fisheries. It feeds on algae and aquatic macrophytes. It is used extensively for weed control in fish ponds (Jhingran and Pullin, 1985). Scattered references to its habits indicate that the species does not deviate from this essentially vegetable matter diet. Puntius gonionotus controlled a dense cover of Ceratophyllum in a 284 ha reservoir in East Java in Indonesia within 8 months of stocking (Schuster, 1952). In experiments carried out in Bangladesh on the diet and feeding ecology of the introduced Puntius gonionotus by Haroon (1998), macrophytes represented 39.2% of the gut content in small fish, and 15.7% in large fish. Haroon classified this species as macrophytophagous column feeder, depending on aquatic macrophytes with increase in size and development of pharyngeal mill, and with benthic foraging on tiny molluscs as the fish grow larger. This supports the findings by Ukkataweat (1979) who found this species feeding on macrophytes in Thailand at the size >12.5 cm.
To achieve fast results in submersed aquatic macrophyte control, Puntius gonionotus needs to be overstocked. Overstocking may also be required to prevent that not all this tasty fish end in the nets of commercial and subsistence fishermen prior to achieving the required results. The same concerns other aquatic weed-feeding fish, such as Tilapia rendalli and T. zillii.
Colossoma, a South American characid (Serrasalmidae), is an important fish in capture fisheries in most tropical South America. Petrere (1983) lists C. macropomum as important fish in catches in Amazonas State in Brazil, and Payne (1987) noted that C. macropomum and C. brachypomum are of major importance in catches from the Rio Mamore (Trinidad). C. macropomum is present in the Orinoco River system in Venezuela, and the Amazon system of Brazil, Colombia and Peru. The fish may reach over 30 kg weight and 1 m length. In the wild, Colossoma macropomum and C. brachypomum feed on plant seeds and fruits in inundated forests, and with the retreat of water they will feed on zooplankton, fish, insect larvae (Goulding and Carvalho, 1982). Colossoma can tolerate low concentrations of dissolved oxygen for short periods of time (see also Section 4.1), and a flap on the lower lip, when extended, allows the fish to skim the surface layer of water for more oxygen when necessary (Ginnelly, 1990). This adaptation enables it to survive through the dry season in pools which become isolated from the river as flood waters recede. Goulding (1980), who studied the food of this species captured from the Rio Machado (Venezuela) flooded forests, found that rubber tree seeds (Hevea spruceana) and palm nuts (Astrocaryum jauary) were the dominant food consumed. Goulding notes that these may be selected because they are hard and most other fish species cannot exploit them. Other fruits/seeds are probably competed for by hundreds of other fish species. Large fat reserves are built up during the flood season, as the dry season is a time of poor feeding. In varzea, juveniles of the characid Colossoma macropomum have been found in the floating meadows to feed mainly on filamentous algae and wild rice seeds (Goulding and Carvalho, 1982).
Araujo-Lima et al. (1998) estimated the contribution of the flooded forest to Colossoma macropomum production and the economic value of this contribution to the economy of Manaus, the largest city in Central Amazon. Flooded forest seeds were responsible for more than 41% of the carbon assimilated in over 80% of the examined fish, while carbon from aquatic grasses contributed less than 5%. The total gross revenue from the sale of this fish in Manaus in 1993–1994 was estimated at US$ 13 million, US$ 8.2 million (65%) of which came from C. macropomum produced with flooded forest carbon.
C. macropomum has been introduced into aquaculture. It will readily eat farm produce, and hatchery operations for breeding and growing this fish are not complicated (Ginnelly, 1990). This species has been cultured in Venezuela, Brazil, Colombia, Panama, on some Caribbean islands, as well as in some countries of Asia. More information on culture of C. macropomum is available for example in Martines Espinoza (1984) and Ginnelly (1990). Some other species of Colossoma, such as C. mitrei, an important fish of the Parana-Uruguay river system, have also been tested for the suitability for aquaculture (Merola, 1988). After the construction of hydroelectric dams on the Parana River this species has almost disappeared from the river. A restocking of the reservoirs will require that stocking material is produced in hatcheries.
C. bidens, highly adapted to eating fruits and seeds in flooded forest, feeds during the water level decline on leaves and grass, and hence it has a much higher mean stomach fullness in the dry season than C. macropomum. This species has been released in the Sepik-Ramu river system in Papua New Guinea with the purpose to enhance the fish production from the extensive floodplains and to make use of the fruits and seeds of the periodically inundated forests (Coates, 1997). (See also Section 9.1).
The best accounts on food and feeding habits of rudd and roach are presented in papers by Prejs (1984), and Prejs and Jackowska (1978). The diet of rudd, one of the most common littoral fish in eutrophic European lakes, includes 65–90% submersed macrophyte tissue (Prejs, 1984). The contribution of submersed macrophytes to the food of rudd in three Polish lakes investigated by Prejs and Jackowska (1978) increased with the size of fish, attaining over 90% of the total food weight of fish longer than 16 cm. According to Van Donk (1998) only larger rudd (age 1+) are herbivorous.
In Polish lakes, Elodea canadensis was found most frequently in the food of rudd and roach, although its biomass in lakes was lower than that of one or two or three of the dominant plant species. In Lake Mikolajskie and in Lake Warniak roach and rudd consumed during the period of intensive feeding 775 kg ha-1 and 651 kg ha-1 of submersed macrophytes, in the respective lakes (Prejs and Jackowska, 1978). This represented some 50% of the total weight of the macrophytes consumed. Second among the macrophytes consumed was Ceratophyllum demersum, and third Characeae and Potamogeton pectinatus, respectively. Hansson et al. (1987) consider rudd a typical periphyton feeder, which also feeds on invertebrates associated with aquatic macrophytes, and on zooplankton. Rudd only graze during the growing season of the macrophytes at water temperature >16°C (Van Donk, 1998), and, if having the opportunity of feeding on a variety of plants, it may avoid Ceratophyllum, which is calcareous in structure (Prejs and Jackowska, 1978).
The macrophyte Anacharis canadensis was eliminated in experiments in aquaria by the grazing activity of rudd (Hansson et al., 1987). Artificial, macrophyte-like structures were found to impair the foraging of juvenile roach, but were no obstacle to feeding on cladocerans by young rudd and perch (Perca fluviatilis) (Winfield, 1986). Winfield found that a dense stand of artificial Nuphar sp. usually increased the performance of rudd and perch relative to those in open water. Prejs (1978) interpolated his experiments on the daily intake of macrophyte food in relation to body size of rudd and roach to natural conditions, and concluded that the daily rations decrease linearly with the increasing size of fish. Fish of a body length 11 cm consumed 15.5 % of the body weight, while fish of 24 cm body length consumed only 7.7%. A similar decrease in food intake with age has been observed for example with grass carp (see Section 3.1.3). In Polish lakes in the food of larger rudd and roach Elodea dominated.
In Lake Zwemlust in the Netherlands the biomass of submersed macrophytes removed by rudd was less than that removed by coots (Van Donk, 1998). Both fish and bird grazing on macrophytes may affect the internal balance among autotrophic components of a water body, by reducing the biomass of macrophytes, thereby reducing their competition with algae for nutrients. Grazing pressure by rudd is, contrary to grazing by coots, unevenly distributed among macrophyte species. Due to the selective grazing by rudd on its preferred food Elodea it could have affected the shift in dominance from Elodea nuttallii to Ceratophyllum demersum (Van Donk et al., 1994). Coots, however, pull out whole plants and may influence the macrophyte composition and succession by especially removing plants still present during autumn and winter. After such grazing losses to Elodea, next spring other macrophyte species such as Ceratophyllum, which forms dormant buds during winter, can occupy the whole available area.
For more information on roach see also Sections 2.1.1, 2.1.3, 3.1.2, 3.3 (Birds), 4.3, 5.3, 6, and 11.
The habit of feeding on bottom sediments, which uprootes aquatic plants and stirs the sediment, which in turn leads to an increase in water turbidity, makes common carp an unwanted species in some water bodies, especially those which serve as a source of drinking water. At a density of 400 carp ha-1 in ponds the common carp activity controlled the submersed aquatic plants in Alabama. At a density of 448 kg ha-1 the carp destroyed submersed vegetation in enclosures placed in a Lake Erie marsh. Common carp is also very numerous in shallow bays of Lake Ontario, where it causes resuspension of sediments and uprooting of aquatic macrophytes (Crowder and Painter, 1991). In a 1215 ha reservoir in Arkansas about 10 tons of carp stocked over 3 years controlled Ceratophyllum demersum and Elodea (Schuytema, 1977).
In India, common carp is considered to be indirectly useful in controlling submersed rooted vegetation because of its roiling habit (Nandeesha et al., 1989). This leads to an increase in water turbidity which in turn suppresses the growth of submersed aquatic plants. The combination of common carp and grass carp was used to controll Vallisneria in irrigation canals of Tungabhadra project in Karnataka, and common carp was also used successfully for controlling Vallisneria in perennial ponds.
The presence of common carp and large bream (Abramis brama) at densities of 500–700g ha-1 in some shallow Dutch lakes is to have a major role in increased turbidity resulting from resuspension of sediments (Meijer et al., 1990). The authors constructed a model for assessing the direct impact of fish on the turbidity. According to the model benthivorous fish having a biomass of 600 kg ha-1 are able to reduce the Secchi disc transparency to 0.4 m (in absence of algae and at a depth of 1.0 m). Interestingly, in the Netherlands in experimental ponds with benthivorous bream (Abramis brama) and carp (Cyprinus carpio) it was the bream which resuspended more sediments than the carp (Breukelaar et al., 1994). Although carp can dig deeper and more efficiently than bream the concentration of suspended solids was 50% lower than in ponds with bream. A possible explanation for this difference is that carp also feeds on molluscs in contrast to bream and does not have to resuspend the sediments for feeding on these organisms. Unlike bream, the carp had no negative impact on the density of chironomids suggesting that it used different food than bream.
In South Africa and in Australia common carp is known to be one of the few fish species tolerant of high suspended sediment concentrations, such as phosphorus-rich colloidal sediments of some South African rivers (Bruton, 1985). In Australia, where many rivers have a persistent high turbidity due to suspensions of dispersed clay particles, Fletcher et al. (1985) found no impact of high carp densities on water transparency. The turbidity values at the Australian sites were generally above 50 NTU and could be considered high. In one instance, with a density of carp of 690 kg ha-1 the turbidity was found lower than at a number of other sites. In the USA, Robel (1961) also found no significant differences in turbidity in ponds at carp densities of up to 670 kg ha-1. On sandy substrates, much higher density of fish than observed in the Dutch lakes would be required to increase the turbidity (Robel, 1961). However, Meijer et al. (1990) suggested that if macrophytes are to become re-established, the reduction in benthivorous fish, such as carp, is essential.
In Australia, where the exotic common carp has spread rapidly through the Murray-Darling Basin since the early 1960s, the residents along these rivers attributed the decline in aquatic macrophytes since the early 1970s to the expansion of carp. Fletcher et al. (1985) identified a critical carp biomass of 450 kg ha-1, above which plant damage was possible, with those most likely to be affected being Potamogeton and Chara. These impacts are generally similar to those reported in Canada (McCrimmon, 1968) and USA (Crivelli, 1983). Roberts et al. (1995) re-examined the potential for carp to affect aquatic ecosystem structure and processes under Australian conditions, using small ponds. Out of five plants tested, the carp reduced only the abundance of the submersed Vallisneria and Chara, to zero in eight days. Crivelli (1983) also mentioned that Chara is particularly susceptible to carp foraging. The experiments by Roberts et al. (1995) also showed that the sediment surface around Juncus, Schoenoplectus and Myriophyllum were not worked over by carp. The presence of these plants seemed to effectively inhibit carp foraging in the substrate, hence reducing the turbidity or suspended sediment concentration, and therefore having a stabilizing role in the aquatic environment.
The biomass of fish communities of reservoirs in the central and southern USA are dominated by omnivorous fish such as gizzard shad (Dorosoma cepedianum) and common carp. According to Gallo and Drenner (1995) these omnivorous fish act as “nutrient pumps” when they consume sediment-bound nutrients and subsequently excrete them into the water column. Also clay particles with adsorbed nutrients may be resuspended by the fish feeding activities. These activities almost always lead to an increase in phytoplankton biomass and/or primary productivity. In pond experiments with eleven aquatic macrophytes, the presence of carp reduced the biomass of Najas guadalupensis, filamentous algae, and the number of macrophyte species. According to the authors, the presence of the macrophytes may have “dampened” the effects of carp on the water column. Carp-mediated increases in turbidity, total phosphorus, and chlorophyll a appeared to diminish with increased biomass of Myriophyllum suggesting that this plant may regulate the effects of carp on the water column.
In Argentina, a high common carp biomass is associated with high turbidity and high conductivity. Apart from the common carp, other fish found in such environment were pejerrey (Odontesthes bonariensis), madrecita (Astianax eigenmanniorum) and lisa (Mugil liza). In southern Argentina the common carp has spread rapidly through the irrigation systems of the lower valley of the Rio Colorado, and it is now found in 5400 km of irrigation canals and 3700 km of drainage canals. In spite of the canals being dried out for 2–3 winter months for maintenance, the fish re-enter the system every year from the river (Fernandez et al., 1998). Anecdotal evidence suggests that the originally clear water of canals has become gradually turbid under the impact of the invasion and increasing numbers of the common carp. The turbidity has led to the disappearance of submersed aquatic macrophytes, formerly infesting the canals. In those sections of canals with lowest conductivity and clearest water a high biomass of the plant Chara contraria tended to occur, and these sections also had the lowest carp biomass. Field experiments on the interaction of the common carp with Potamogeton pectinatus and other submersed plants in the irrigation canal system found a significant relationship between the reduced plant growth and an increase in water turbidity resulting from the bottom-feeding habit of the fish. In experiments in aquaria the presence of carp increased turbidity from 2.2 NTU to up to 73.4 NTU, which was accompanied by an up to 58% decrease in submersed plants. In the Argentinian irrigation canals the common carp was also found to feed on submersed plants. Fernandez et al. (1998) suggested a possibility of manipulating the size and age structure of the carp population of the irrigation canals for aquatic plant management purposes.
In 19 wetlands in the Great Lakes basin of Canada, waters with turbidity above 20 NTU contained less than five species of submersed plants, while a more diverse community existed in less turbid systems (Lougheed et al., 1998). By extrapolation and with support from the results with enclosures with different densities of common carp and monitoring the impact on aquatic environment, the authors suggested that in one particular water body water turbidity cannot be reduced below 20 NTU unless the common carp is excluded. In experimental ponds in New South Wales, Australia, high density of carp increased turbidity from approximately 7 NTU to 26 and 73 NTU by day 4, resulting in a complete loss of two (Chara fibrosa and Vallisneria sp.) out of three submersed aquatic plants (Roberts et al., 1995). Under low impact conditions the uprooting rate of Vallisneria was reduced to a third.
That common carp is not selective in its choice of substrate for attachment of eggs, has been well documented from the irrigation canals and storage reservoirs of the Murrumbidgee Irrigation Area in New South Wales, Australia. Carp deposited their eggs on dead and live perennial Persicaria sp., grass Pennisetum clandestinum, milfoil Myriophyllum papilosum, cattail Typha orientalis, on filamentous algae, but also on stones and broken off fresh eucalypt branches, and on raffia, which was used as an artificial substrate (Adamek and Adamkova, 1999).
Activities of some other aquatic or semi-aquatic vertebrates, such as turtles, birds, manatee, nutria, muscrat and cattle, through their grazing activity on aquatic macrophytes, may influence fish species composition, abundance and production.
The herbivorous turtles Kachuga tectum and Hardella thurgi feed on aquatic plants in India and Bangladesh (Chokder, 1967), with the first species found to feed in India on the following: Lemna, Ceratophyllum, Eichhornia, Hydrilla, and Ipomea. The species can be easily bred under controlled conditions and there is a good scope for their utilisation in the biological control of aquatic weeds (Nandeesha et al., 1989). In Florida, the turtle Pseudoemys floridiana is also herbivorous (Yount and Crossman, 1970). In Delaware waters (USA) Townsend (1979) estimated that fresh aquatic macrophytes formed about 13 % of the diet of the snapping turtle (Chelydra serpentina).
Many ducks and geese, domesticated or not, as well as swans and coots, are consumers of aquatic plants. The smaller species of ducks are rather selective, giving preference to duckweeds and other small plants with soft tissues. Among aquatic birds, only geese and swans are strict herbivores, and apart from aquatic macrophytes both consume a certain amount of terrestrial plants. Other aquatic birds consume seeds and/or fruits of aquatic plants, thus reducing their reproduction (Van Zon, 1976). McAtee (1939) found that a number of waterfowl species feed on Potamogetonaceae, Cyperaceae and Polygonaceae, Potamogetonaceae being the preferred food source. Grazing affects both the biomass and the growth of submersed macrophytes significantly. In the colonization phase waterfowl may prevent growth of submersed macrophytes (Moss, 1990).
Waterfowl often consume significant portions of annual aquatic macrophyte production: 40% peak standing crop of Potamogeton pectinatus in Delta Marsh, Manitoba, Canada (Anderson and Low, 1976); 30% peak standing crop of P. filiformis in Loch Leven, Scotland (Jupp and Spence, 1977); and about 50% peak standing crop of an assemblage of submersed macrophytes in Tipper Grund, Denmark (Kiorboe, 1980).
Shifts from high density in aquatic macrophytes to phytoplankton and back to macrophytes may occur naturally or under human influence. Such shifts may also lead to changes in species composition, biomass and migrations of aquatic birds. When a storm destroyed macrophyte beds in Lake Ellesmere in New Zealand, the population of black swans declined from 40 000–80 000 to 4 000, over a period of about 20 years (McKinnon and Mitchell, 1994). In southern Sweden in Lake Krankesjon, Potamogeton pectinatus and Chara tomentosa expanded spatially after more than a decade of phytoplankton blooms and sparse submersed vegetation (Hargeby et al., 1994). During the expansion of submersed plants the number of resting and breeding waterfowl increased. The increase was significant for the herbivorous coot (Fulica atra) and mute swan (Cygnus olor), and for omnivorous dabbling ducks (Fig. 9). Submersed plants, and especially their green parts, form the main food item of swans and coots, and the increase in abundance of these birds indicates that shortage of suitable food was limiting during the previous state of phytoplankton dominance. The increased abundance of dabbling duck was related to the increase in both plants and invertebrates. Further changes occurred, such as decrease in the density of planktonic Cladocera, increase in the species diversity of benthic invertebrates and the change in their dominant species. The increase in the mean size of perch (Perca fluviatilis) could be the result of a better food supply in form of invertebrates associated with the aquatic plants.
Fig. 9. Average daily numbers of mute swan (Cygnus olor), coot (Fulica atra) and diving and dabbling ducks in Lake Krakesjon in June-July 1985–1990. (From Hargeby et al., 1994).
Invasion of Currituck Sound, North Carolina, by Myriophyllum spicatum increased the total biomass of submersed macrophytes, and this in turn led to an increase in numbers of dabbling ducks (Anas rubripes, A. platyrhynchos, A. acuta, A. americana, A. crecca), ring-necked ducks (Athya collaris), and American coot (Fulica americana) (Wicker and Endres, 1995) (Fig. 10). The above examples suggest that population densities of water birds are often closely related to the abundance of aquatic macrophytes. But it is also true that an increase in macrophytes that results in more food for some water bird species may represent a decrease in available habitat or feeding opportunity for other species. A negative relationship between macrophyte cover and bird abundance has been recorded for several species of birds associated with Florida lakes (Hoyer and Canfield, 1994).
Fig. 10. Percent contribution of waterfowl in Currituck Sound to the Atlantic Flyway midwinter waterfowl survey counts during periods before (1961–1965), during (1968–1977) and after the infestation with Myriophyllum spicatum. (From Wicker and Endres, 1995).
Nelson and Kadlec (1984), in their review paper describe the interactions occurring among macrophytes, macroinvertebrates, and waterfowl in freshwater wetlands as a complex interdependency in which dynamic changes in the abundance and distributional pattern of macrophytes resulting from climatological fluctuations influence functional processes in litter decomposition and macroinvertebrate communities that, in turn, affect reproductive effort and success of waterfowl. Lilie and Evrard (1994) found that large, dense stands of submersed vegetation of Wisconsin wetlands offered greater shelter for macroinvertebrates from predation by birds, hence also had a higher biomass of invertebrates compared to less dense stands. Less dense macrophyte stands provide only limited refuge for macroinvertebrates, hence better opportunities for feeding by birds and fish on them. They found that macroinvertebrate densities below 5000 individuals m-2 resulted in low densities of breeding pairs of blue-winged teal (Anas discors), but this had no effect on mallard (Anas platyrhynchos) pair densities.
A number of studies has shown a direct effect of fish on benthic invertebrates, and in this way competition between fish and waterfowl. Winfield and Winfield (1994) provided further evidence in their study of interactions between fish populations and overwintering tufted duck (Aythya fuligula) of Lough Neagh, Northern Ireland. They leaned on previous studies, which found that such competitive interactions occur especially during the waterfowl breeding season when their protein requirements are high. Breeding hens utilise invertebrates as a source of proteins for egg-laying, and ducklings depend heavily on the consumption of invertebrates as a source of protein during their first few weeks of life. During such periods in life of the water birds, competition for food by fish may determine whether some water birds will select a particular water body for nesting and breeding. Erikson (1979) found that the distribution of fledged goldeneye (Bucepala clangula) was negatively associated with the presence of fish, mainly roach and perch. Giles et al. (1990) suggested that significant diet overlap exists between perch and ducklings of mallard (Anas platyrhynchos) and tufted duck. In the above interactions, the observed or postulated competition resulted in waterfowl being out-competed by fish (Winfield and Winfield, 1994). In Lough Neagh, the largest lake in the UK, molluscs (especially Lymnaea peregra and Valvata piscinalis) were the dominant prey in the diet of tufted duck, followed by chironomid larvae. Molluscs also formed a major component of the diet of roach (Fig. 11), while the diet of eel was overwhelmingly dominated by chironomid larvae. Tufted ducks are primarily carnivorous with molluscs dominating the diet and chironomid larvae typically assuming primary importance only when the former are unavailable. Interestingly, in other localities, roach often feed predominantly on macrophytes or filamentous algae (see review by Lammens and Hoogenboezen, 1991) and such a plant-dominated diet was found by Giles et al. (1990) to minimise the overlap between roach and waterfowl in a macrophyte-rich lake. The same authors also found that consumption of cladoceran zooplankton by adult roach minimised diet overflap between roach and waterfowl in a macrophyte-poor lake. In Lough Neagh it is likely that the benthivorous roach population increase during the winter may result in a decline in the overwintering numbers of tufted duck (Winfield and Winfield, 1994). This further reconfirms the observations by Andersson (1981) on increasingly eutrophic lakes in southern Sweden, where declines in waterfowl were probably caused by competition for benthic food resources with increasing cyprinid populations, principally roach and bream (Abramis brama).
Fig. 11. The relationship between the size of roach (Rutilus rutilus) and the consumption of molluscs. For each class, the frequency of occurrence of molluscs is shown by the closed portion of the bar. The numbers of fish in each size class are given at the top of each bar. (From Winfield and Winfield, 1994).
Food competition between a freshwater fish community and tufted duck was further confirmed through an experimental removal fish from a quarry lake (17 ha) in England (Giles, 1994). The fish community was dominated by adult bream and roach, with numerous small perch and abundant pike (Esox lucius). The fish employed several mechanisms in the suppression of macrophyte growth including uprooting by benthivorous species (bream, carp), browsing of germinating seedlings (roach), and probably increasing nutrient cycling favouring an algal dominated open water plant community. Prior to the fish removal this lake had very little submersed aquatic vegetation. After the fish removal, dense beds of several species of submersed macrophytes grew over the whole lake bed. The post fish-removal increases in snail biomass have probably been central to the increased brood size at fledging recorded in this lake (Giles, 1994). The numbers of herbivorous wintering wildfowl (mute swan, Cygnus olor, coot, Fulica atra, and gadwall, Anas strepera) using the lake have increased dramatically following fish removal and the consequent surges in submersed aquatic plants. The probable effects of fish on aquatic birds in the quarry lake in England are shown in Table 1.
Table 1. Probable effects of fish on gravel pit aquatic bird species and mechanisms of action. (From Giles, 1994)
|Fish species||Diet||Birds affected (and mechanisms)|
|Bream, carp, tench||Chironomid larvae, molluscs||All ducklings via food competition. Herbivorous waterfowl via depressed submersed plant growth|
|Roach||Zooplankton, algae, plant seedlings||Tufted ducklings via suspension of plant growth and reduced snail production|
|Perch||Chironomid pupae, crustaceans||All ducklings via food competition|
|Pike reduction||Fish, ducklings||Predation of ducklings by small pike but also in food competition from insectivorous fish via predation by pike of all sizes|
In Florida lakes, Hoyer and Canfield (1994) found that birds do not significantly impact the trophic status of the lakes under natural conditions, but lake trophic status is a major factor influencing bird abundance and species richness on lakes. They found no significant correlation between annual average abundance and species richness and macrophyte abundance. Bird abundance and species richness remain relatively stable as macrophyte abundance increases, but birds that use open-water habitats (e.g. cormorant Phalacrocorax auritus) are replaced by species that use macrophyte communities (e.g. duck Aythya collaris). On the other hand, Lodge (1991), referring to literature data, pointed out that waterfowl influence submersed and emergent plants locally and seasonally: snow geese (Chen caerulescens) substantially reduce macrophyte (predominantly emergent and wetland plants) biomass on their summer breeding grounds, at migration stop-over sites, and on their wintering grounds. Less gregarious waterfowl, e.g. breeding ducks and swans during the macrophyte growing season, also substantially reduce plants, perhaps over larger areas than colonially nesting geese.
The presence of black swans (Cygnus atratus) is closely bound with the presence of aquatic macrophytes. Black swans, introduced to New Zealand from Australia, are exclusively herbivorous. A shallow lake (Hawksbury Lagoon) benthic vegetation was for some time fully dominated by thick mats of filamentous green algae Spirogyra, Rhizoclonium and Enteromorpha intestinalis, with charophyte Nitella hookeri also dominating the biomass at times. The only angiosperm recorded was Ruppia (Mitchell and Wass, 1996). After a 2–3 year period of macrophyte and benthic filamentous algal dominance, the lake switched to phytoplankton dominance. The reason for this was not, however, the impact of swans through direct feeding on benthic algae, but rather the change in the depth penetration of light, which occurred as a combination of phytoplankton and other suspended matter increase. Under such conditions, the additional effect of black swans grazing might well be critical and lead to the decline in aquatic macrophytes (Mitchell and Wass, 1996). The same authors also assessed the impact of the faecal loading rates to nutrient dynamics. The data suggest that waterfowl faeces have low ratios of nitrogen to phosphorus, so that in lakes where their contribution is large, the phytoplankton will tend to be dominated by cyanobacteria. However, in Hawksbury Lagoon black swans faecal contribution to nutrient dynamics was slight (Mitchell and Wass, 1995), less than half of that estimated for Canada geese on a small Michigan lake. Geese feed on pasture, thus bringing into water nutrients from external source, rather than recycling material like black swans, feeding on benthic algae. Mitchell and Wass (1996) constructed a conceptual model for the roles of black swans and other non-seasonal factors in decline in macrophyte biomass in shallow lakes (Fig. 12).
In lakes where for whatever reason submersed aquatic macrophytes have disappeared, e.g. in eutrophicated water bodies, the regrowth of macrophytes may be delayed by waterfowl grazing (Lauridsen et al., 1993). Kiorboe (1980) stated that grazing by coots has only a minimal effect on macrophyte growth because grazing often takes place outside the growing season of the plants. Coots, however, pull out whole plants and may influence the macrophyte composition and succession by removing especially plants still present during autumn and winter. In experiments, under the impact of waterfowl grazing, especially coot (Fulica atra), plants growing on mud were found to have been pulled out of the sediment and most of them had disappeared. A high number of of coots appeared on the Dutch Lake Zwemlust when perennial macrophytes became dominant after biomanipulation. Van Donk (1998) suggested that high grazing pressure by coots on Elodea and Ceratophyllum during the first years after biomanipulation probably promoted the rise of Potamogeton berchtoldii. This plant was not negatively affected by coots because this species forms nongrazable overwintering structures. The impact of sequential grazing on aquatic macrophytes by rudd and coots in Lake Zwemlust, as reflected in changes in the total biomass of macrophytes, is shown in Fig. 13.
Fig. 12. Conceptual model for the roles of black swans and other non-seasonal factors in decline in macrophyte biomass in shallow lakes. Stimulatory → effects; inhibitory effects ⊥; effects examined in the study (1); x,y groups of confounded factors, combined effects of each group examined in the study. Detrital pathways, external nutrient and sediment loads, some components in internal nutrient loads, and food chain interactions affecting phytoplankton are omitted. (From Mitchell and Wass, 1996).
Fig. 13. a) Total biomass of submersed macrophytes (kg dry weight) in Lake Zwemlust estimated in March, June, August, October 1991 and February 1992; and b) estimates of herbivory (kg DW macrophyte per day) by rudd and coots in Lake Zwemlust (March 1991–Feb. 1992). (From Van Donk et al., 1994).
Information on the consumption of aquatic macrophytes by herbivorous birds is sparse: in Lake Zwemlust, Netherlands, the estimated annual consumption by coots was calculated as being between 30 and 1 200 kg. This compares with the consumption of 170 to 360 kg by the fish rudd (Scardinius erythrophthalmus) (Van Donk and Otte, 1996). Mitchell and Perrow (1998) suggested that birds may have a major impact on the dynamics of plant biomass in temperate waters only during a low macrophyte production - either early in the growth phase of seasonal species, when biomass is low, or in autumn, when seasonal plant growth has slowed or ceased. Arrival of migratory flocks may dramatically increase the grazing impact and lead to a substantial decline in plant biomass after seasonal growth has ended.
Olofson (1991) found significantly higher growth of fenced-in macrophytes (Potamogeton crispus) than of unfenced macrophytes in Lake Stigholm, Denmark. The fish grazing was of minor importance as most herbivorous fish were prior to that removed. He concluded this was the result of grazing on water plants by waterfowl. Jupp and Spence (1977) found that plants of Potamogeton filiformis, protected from waterfowl grazing, became up to 5.3 times larger than unprotected plants. Lauridsen et al. (1993) concluded that waterfowl will be able to delay the re-establishment of submersed macrophytes, especially in shallow parts, where waterfowl are able to reach the bottom when feeding. This might delay the recovery of eutrophic lakes.
Concluding, one can say that birds typically consume only a small fraction of the annual aquatic macrophyte production. Their major influence appears to derive not from what they consume but from the loss of the future growth potential of plants, compounded through the growth period. The macrophyte biomass parameters that are relevant to zooplankton or fish may not always be so for birds, as birds are highly mobile and sampling in - say - two or four weeks interval is quite inappropriate. Birds mobility means that the option of emigration to seek a better habitat is always available (Mitchell and Perrow, 1998). Where large generalist herbivores such as aquatic birds continue to increase, they may become an impediment to managing lakes for the clearwater state: a small increase in herbivory by such birds could tip the balance toward the turbid phytoplankton state (Lodge et al., 1998), and in turn this would impact on fish.
For the impact of grass carp on aquatic birds see Section 3.1.2.
In Florida and Guyana, Trichechus manatus is known to consume 36 genera of macrophytes, but not water hyacinth when other plants are available. It was reported that they efficiently clear canals when present in sufficient density. Various ponds and canals in Guyana have been kept clear of aquatic weeds by Trichechus inunguis for many years (Anon., 1974). This species, which inhabits the tropical rivers of south America, feeds at any level from bottom to surface and is able to feed also on the floating vegetation. Their preference is for succulent aquatic macrophytes, but they will consume almost any aquatic plant (Ronald et al., 1978). They prefer submersed to floating, and floating to rooted-emergent plants. In the Chagres River in Panama weed control using the manatee was unsuccessful (Schuytema, 1977, and references therein). Some authors pointed out some of the limitations and difficulties in attempting to utilize this species. While it is a voracious feeder it tends to graze selectively and, except in situations where it can be confined to a small area and thus forced to feed on all plants, it will bypass thick stands of one species to feed on another which it prefers. This
factor, coupled with its scarcity, its low reproductive rate and the high cost of managing it, suggests that it is unlikely to be widely used in weed-control programmes.
Estimates for wet weight food consumption by captive manatees (T. manatus) range from 5 to 11 % of body weight per day (Best, 1981). Etheridge et al. (1985) made a detailed field study of manatees of the Crystal River, Florida, where they determined that the longest animal feeding in a 24 hour period was 6 hours and 10 minutes, with an average of 5 hours per day, an equivalent of approximately 7.1% of their own body weight. The introduced hydrilla was the principal food of manatees there. The authors, using Haller and Shireman (1982) data on aquatic plant biomass for Kings Bay, where the plant covered 165 ha, concluded that 18 manatees per ha would be required just to maintain the hydrilla at a constant biomass by consuming daily productivity. This is nearly ten times as many manatees than are present in the area. They argue that previous assessments of efficiency of manatees in controlling aquatic plants might have been exagerated, unless the local plant productivity was much lower than that in Kings Bay, and conclude that manatees cannot be considered efficient control agents for aquatic macrophytes.
Trichechus senegalensis inhabits the west coast from Senegal to Angola in Africa, but nothing is known about its plant-feeding habits.
Schuytema (1977) reviewed the impact of nutria on aquatic plants. This animal, often introduced for its pelt and meat, is able to control the emergent Typha angustata and Phragmites australis in ponds in Europe and Israel, and water grass (Echinochloa) in Africa (Cameroon). This it does quite efficiently, and where nutria is present close to common carp ponds it can significantly increase fish production by destroying the emergent vegetation. The burrowing habit of nutria is a negative aspect of its presence as it may lead to damage to pond embankments, canal banks and dams. In Kenya nutria (coypu), which was imported for fur-farming in 1950, escaped and arrived at Lake Naivasha from 1965 onwards, where, by the early 1970s, there was a large population (Harper et al., 1990). Coypu are generally blamed for the disappearance of water-lilies (Nymphaea caerulea) from Lake Naivasha, on which individuals were widely observed to be feeding. Water-lily shoots make up a major part of the coypu's diet (Gibson, 1973). Water-lilies began to disappear in the eastern part of the lake at around the same time that the crayfish Procambarus clarkii populations were high (4 ind. m-2). The crayfish had a substantial effect upon the submersed vegetation, which was reported to be abundant in 1979 but had completely disappeared from the lake by 1983, the last year in which crayfish harvest exceeded 100 tons. It is likely that water-lilies disappeared under the combined grazing pressure of coypu and crayfish (Harper et al., 1990).
In the Norfolk Broads, UK, the highest rate of decrease in young reed shoots was found when the population of coypu was very high (Boorman and Fuller, 1981).
In the Czech Republic muskrat consumed or used in lodge construction 9–14% of the annual net biomass production of Typha latifolia (Pelikan et al., 1971). In a lake in northern Germany a population of muskrats consumed or damaged 0.27 ha of Typha, 0.15 ha Phragmites australis, 0.86 ha Glyceria, and 1.58 Scirpus. The damaged area of Scirpus and Phragmites did not recover even when the muskrats significantly decreased in number (Akkermann, 1975). In the USA muskrat in the Atchafalaya Bay, Louisiana, graze on submersed swards, for example Eleocharis (Fuller et al., 1984).