Ecology of Aquatic Invertebrates

J.-M. Elouard and F.-M. Gibon Translated from the original French by S. M. Goodman and L. Wilme

,The richness and geographic distribution of the freshwater at lower elevations, a humid region at higher elevations, a organisms of Madagascar cannot be understood without subhumid to semiarid region, and a dry region. taking into account the geomorphology and climate of the Interactions between geomorphology and climate gener­ island, as well as the vegetation patterns that result from the ate varied hydrologic regimes in Madagascar. Aldegheri interaction between these two factors. Madagascar is a (1972) has established a hydrologic classification system, mountainous island whose principal summits (Tsaratanana which has been updated by Chaperon et al. (1993). Nine ~.. and Andringitra) reach well over 2500 m. The island's relief principal regimes were underlined. These various regimes, If shows an east-west asymmetry, folJowing a continental di­ together with elevation and therefore water temperature, all r""yide that runs along the axis of the island (north-northwest have an influence on aquatic fauna and flora. f~o south-southeast). The divide is located one-third of the Last, geomorphology and precipitation influence the r ,JSland's width from the east coast and two-thirds from the vegetation itself. Mapping of major vegetational zones has ~,.,est coast, leading to an abrupt eastern slope and a much been established by Humbert (1955), Humbert and Cours­ ~Bentler western slope. The eastern-draining rivers slope Darne (1965), and Faramalala (1988). Vegetationcover also '.~ply down to the sea and have short estuaries, whereas influences aquatic fauna and flora because it affects insola­ ~}he western-draining rivers are of more moderate slope and tion, inputs of organic matter, turbidity, and, particularly, .'havelonger estuaries. hydrologic variability. Together these biotic and abiotic fac­ .• Because trade winds carrying precipitation blow from tors create, in Madagascar, a multiplicity of habitats that Southeast, the relief of the island generates a climatic are favorable to a rich and varied aquatic fauna and flora. f4lIl>YD:unetry. The eastern slopes, where annual precipitation iesfrom 2 to 4 m and sometimes reaches up to 10 m, re­ 've most of the atmospheric moisture. Consequently, Knowledge alI is much lower in the west (1200 to 2000 mm annu­ ~. y) and especially in the southwest «500 mm). These dif­ Until very recently, our knowledge of the systematics of rlcrences create four major climatic regions: a humid region Madagascar's freshwater organisms was variable from one \ .,

512 Invertebrates Aquatic

taxonomic group to another. Sorne groups were relatively cially in the eastern humid primary forests, where the high. weil known, even if sorne of their species still remained est richness is found but which is experiencing high rates of to be discovered or cryptic sister taxa could not be dis­ deforestation, presumably leading ta extinction of species. tinguished without molecular techniques. Among these It thus appears that beyond systematic inventory, consider­ groups, the aquatic plants (see Ranarijaona, this volume), able work remains to be done to understand the aquatic fishes (see Sparks and Stiassny, this volume), macrocrus­ fauna from an ecological point of view and to assess the taceans (see Short and Doumenq, this volume; Crandall, threats induced by various disturbances such as pollution this volume; and Cumberlidge and Sternberg, this volume), deforestation, and hydrologic modification. ' Odonata (see Donnelly and Parr, this volume), and Het­ eroptera are good examples. Certain groups were fàirly weil known from a systematic point of view, such as the Richness Coleoptera, among which certain families have been rather exhaustively inventoried, whereas others have been only Virtually aIl macrocrustaceans and aquatic insects found in little studied. In the same way, Diptera (see Irwin et aL, this Africa are also present in Madagascar, including Epheme­ volume), including the Culicidae (see Duchemin et aL, "Cu­ roptera, Odonata, Plecoptera, Megaloptera, Neuroptera: culidae, Mosquitoes," this volume) and Simuliidae (see Sisyridae, Heteroptera, Coleoptera, Lepidoptera, and Dip­ Elouard, "Simuliidae, Black Flies," this volume), have only tera. The Hymenoptera couId be considered the only ex­ been partially inventoried. Last, among the aquatic insects, ception, as it is not known if Agriotypidae, parasites of Tri­ the Ephemeroptera (Sartori et al. 1999) and Trichoptera choptera larvae and pupae, occur in the rivers of the Island, were almost ignored; only a few species were described, but this group constitutes only a marginal cornponent of the whereas these orders now include 200 and 500 species, re­ aquatic fauna. spective/y (see Elouard et aL, this volume, and Gibon, this volume). It should be noted that smaller groups such as Ple­ Macrocrustaceans coptera (see Elouard, "Plecoptera, Stoneflies," this vol­ The fauna of Madagascar is not particularly rich with ume), Megaloptera (see Penny, "Megaloptera, Fishflies and regard ta noncrayfish rnacrocrustaceans (crabs, atyid Stoneflies," this volume), Neuroptera (see Penny, "Neu-: shrimps, and Macrobrachium; Holthuis 1980). Moreover, roptera, Lacewings," this volume), Diptera: Blephariceri­ the majority of these groups have a circum-Pacific distribu­ dae (see Courtney, this volume), Culicidae (see Duchemin tion. On the other hand, Madagascar is one of the rare et aL, "Culicidae, Mosquitoes," this volume), and Cerato­ countries of the intertropical zone ta have native crayfishes pogonidae are in need of thorough revision. (Parastacidae, genus Astacoides), with six species known During the 1990s, the systematics of Malagasy freshwa­ (Hobbs 1987; see Crandall, this volume). ter organisms received renewed interest and was the subject of many publications by various groups of researchers. At the present time, their taxonomy is better known, although Insects another ten years will be required to complete descriptions Families of the material collected to date. There is no endernic insect farnily in Madagascar. On the In the same way, until the past decade, the majority of other hand, many of the families occurring in sub-Saharan described species were known only from the type locality African are not present on the Island. Among the orders and sometimes from one or two additional sites. It was thus listed in table 8.4, ten families present in Africa are not difficult to assess whether the species were rare or abun­ known frorn Madagascar. These are prirnarily Odonata (an dant, broadly distributed or showing marked microen­ archaic order) and Trichoptera (a recent order). Thus, at the demism. The field sampling carried out during the past de­ family level the Malagasy fa una is poorer than the African cade allows estimations of the geographic ranges of many of fauna. these organisms as weil as their major ecological prefer­ ences. Nevertheless, certain regions of Madagascar remain Genera poorly sampled, even unknown, including the Sarnbirano, The genera sarnpled to date on Madagascar amount ta the Masoala Peninsula, and the Tsaratanana region. slightly more than one-third of those known from Africa Finally, it should be noted that, with few exceptions, (110/280 = 39.3%; table 8.4). Even if we consider that very little is known about the biology and development ail the genera of Madagascar are not yet described (cf. cycles of the majority of freshwater taxa in Madagascar. Ephemeroptera: Leptophlebiidae), the difference is notable Considerable work thus remains to be carried out, espe- (sorne African genera are probably not described). At the J,-M. Elouard and F.-M. Gibon 513

Table 8.4. Number of families. genera. and species present on continental Africa and Madagascar

Africa Madagascar Common ta both regions

Order Families Genera Species Families Genera Species Families Percentage Genera Percentage

Ephemeroptera 11 71 295 11 42 111-172 10 90.9 19 26.8

Odonata 14 120 468 9 54 196 9 64.3 52 43.3 Megaloptera 2 4 8 2 2 4 2 100 0 0 Plecoptera 2 6 49 7-10 50 0 0 Trichoptera 20 76 840-985 15 44 663-700 15 75 38 50 Diplera: Simuliidae 3 120 40 100 33.3 Total 50 280 1780-1925 39 144 1021-1122 38 76 110 39.3

generic level and for every order, the Malagasy aquatic en­ sumed that either ail the families existed before the breakup tomofauna is poorer than that of Africa. Nevertheless, this of Gondwanaland or recent families have subsequently col­ result has to take into account the difference between the is­ onized the island. The Lirst alternative seems most likely land's area and that of sub-Saharan Africa. In addition, given the lack of higher-level endemism on the island and many Malagasy genera are endemic. The difference at the that the prevailing southeastern trade winds hinder disper­ genus level between African and Malagasy faunas is thus sai from Africa. The colonization of Africa by Malagasy grearer than shown. winged insects seems more probable for those groups that are able to fly long distances; this is almost certainly not the Species case for the orders Ephemeroptera, Plecoptera, Megalop­ For the major orders of insects, the Malagasy species rich­ tera, and even Neuroptera: Sisyridae. In fact, only certain ness is about 30-50% of that of the African fauna (table Odonata (Anisoptera), sorne Diptera, and certain Trichop­ 8.4), for a total area 50 times smaller. An exception, how­ tera seem capable of crossing the Mozambique Channel. ever, is observed with regard to the Trichoptera, for which At the generic level, both archaic and recent genera the number of Malagasy species exceeds haH of the richness can be found on Madagascar. Thus, among mayflies, the of the currently known African fauna. Ir should be noted derived Tricorythidae clade (genera Madecassorythus and that we do not have enough information on the aquatic Spinirythus) (Oliarinony 2000) or the primitive Caeni­ Coleoptera and Hemiptera to assess their richness in Mad­ dae genus Madecocercus (Malzacher 1995) is encountered agascar. One will notice that, at the species leve1, the dif­ along with recent groups such as the carnivorous genera be­ ference between African and Malagasy faunas is consider­ longing to the Baetidae: Herbrossus, Nesoptiloides, and able, with endemism in Madagascar varying from 90% to Guloptiloides (Gattolliat and Sartori 1999a, 2000a). The 100%, depending on the taxonomic group. high levels of endemism on the island are the result of a differential impoverishment between Madagascar and the rest of Gondwanaland, and more particularly sub-Saharan Endemism Africa, and of a later speciation of certain lineages in Mad­ agascar. With regard to the carnivorous Baetidae, the Owing to its isolated evolution over a long period, the genera Barnarmus and Centroptiloides (Lugo-Ortiz and fauna of Madagascar has a remarkable originality (Millot McCafferty 1998a) differentiated in sub-Saharan Africa. 1952; Paulian 1972). The freshwater fauna is not an ex­ Evolution after the separation of India and Madagascar is ception to this cule and is characterized by a high level of also observed in the loss of flight in the genus Cheirogene­ endemism at the species level, which decreases as one moves sia (PaIingeniidae), whereas ail other members of this fam­ up toward the supraspecific levels. Two opposite factors ily from Eurasia have a flying stage during reproduction. influence the endemism of the Malagasy fauna: the age of the group and its ability to fly. The Importance of Flight It is cIear that insects with long adult lives (>1 month) and The Importance of Evolutionary Age with good flying abilities show lower levels of endemism at As mentioned earlier, there is no endemic family among the both specific and supraspecific levels. Conversely, the Malagasy aquatic macroinvertebrates. It can thus be as- shorter the adult Iife span and t!le lower the flying ability,

- J .' r 514 Invertebrates Aquatic

the higher the rate of endemism. This is seen in the Ephe­ ity of the Macrobrachium (Palaemonidae) have a circum­ meroptera, which are poor fliers and therefore have a weak Pacifie distribution. The same applies ra a certain number dispersal capacity and which have life spa~s in their aerial of shrimps (Atyidae). However, sorne cave-dwelling species stages (subimago and imago) that vary from around ten are endemic co the island (see Short and Doumenq, this vol­ minutes to three days. These aspects of their natural history ume). èast, with regard to the crabs (six genera; see Cum­ suggest that no immigration or natural emigration is pos­ berlidge and Sternberg, this volume), the endemism seems sible when considering the distances that currently separate co be 100%, as in the crayfishes (genus Astacoides), which Madagascar from mainland Africa (600 km) or from the are absent from the African continent (see Crandall, this c10sest islands (300 to 1000 km). Consequently, for this or­ volume). der the rate of endemism is close, if not equal, co 100 % (see Elouard, "Ephemeroptera, Mayflies," this volume). Microendemism On the other hand, the Simuliidae (Diptera) are regarded as being strong fliers (see Elouard, "Simuliidae, Black In ail taxonomie groups there is a marked level of micro­ Flies," this volume). Migratory flights of 300 km have been endemism, particularly in the humid forest areas and on the observed during monsoonal winds (Philippon 1978). Not principal summits of the island: Tsaratanana, Andringi­ surprisingly, there are no endemic genera of Simuliidae in tra, Montagne d'Ambre, Ankaratra, and the Anosyennes Madagascar. In addition, several species, ail having broad MOlJntains. There is little microendemism in the south or geographic distributions in continental Africa, are present on the western slopes of the island. Exceptions exist, how­ in Madagascar (Pilaka and Elouard 1999a). At least spo­ ever, in the Mandrare River basin in the south with the radie exchanges within chis family certainly take place be­ mayfly genus Afrobaetodes and blackfly Simulium buck­ tween Madagascar and the African continent. leyi. In the west, the fauna is generally homogeneous from Within the Trichoptera (see Gibon, this volume), the Montagne d'Ambre in the north to the Efaho River basin in species rate of endemism is very high. However, certain spe­ the south. cies are shared between Madagascar and the African conti­ Geographie isolation, temperature, and vegetation barri­ nent, including Hydroptila cruciata and Catoxyethira mali ers preventing the movement of poorly flying individuals (Hydroptilidae), Chimarra dybowskina (Philopotamidae), from one summit to another can explain the microen­ and Amphipsyche senegalensis (Hydropsychidae). Thus, in demism found on the summits of Madagascar's mountains. the relatively recent past there were species able co cross the The opposite phenomenon is observed in certain Trichop­ Mozambique Channel and colonize the opposite land mass tera, within which the genus Chimarra is a good example. (Gibon 2001). Whereas the species of the eastern forests are generally very The Odonata present an intermediate situation. The localized, the forms that colonize the higher vegetational su border Zygoptera is generally represented by poor fliers, formations, above forest line and in the ericoid thickets, whereas the Anisoptera include families with very strong have very broad latitudinal distributions. This pattern can flying capacities and sometimes migratory tendencies. be explained by the decrease in annual average tempera­ Those strong flying species weil known from Madagascar tures during the last glaciation (Burney 1996). The upper include Tramea basilaris, T. limbata, Tholymis tillarga, limit of the forest would have been substantially lower than Pantala flavescens, Macrodiplax cora, Trithemis annulata, in modern times, and may have been around 1000 m co­ T. arteriosa, Diplacodes lefebvrei, Orthetrum brachiale, ward the end of the Pleistocene. This would have resulted in O. stemmale, Palpopleura lucia, Crocothemis erythraea, fauna associated with ericoid thickets occurring across a Hemianax ephippiger, and Anax imperator. The Zygoptera broad region. These taxa then progressively took refuge on are smaller, often confined to the vicinity of their larval the summits during subsequent c1imatic warming. habitat, and undoubtedly only passively dispersed. Sorne Microendemism in the eastern part of the island is more exceptions exist, however, within species present in Mada­ difficult co explain. Factors influencing this pattern couId be gascar, such as Ischnura senegalensis, distributed from those that act on the aquatic larval stages (e.g., habitat spe­ Senegal to the Philippines, and Ceriagrion glabrum, found cificity) or on aerial adult stages (e.g., food resources, flying from Palestine, across the complete African continent, to ability). Around small forest tributaries, the high density of Madagascar and Mauritius. Phaon iridipennis, Lestes och­ the vegetation forms a strong obstacle to flight in adult cad­ raceus unicolor, and Agriocnemis exilis occur on both disflies. Thus, for a given species there appears co be a rela­ Madagascar and the African continent. tion between increasing geographic distribution and in­ Dispersal in macrocrustaceans is related to their ability creasing distance from the headwaters. to tolerate marine waters. It would seem that the major- Limited aerial dispersal is very pronounced, in general, J.-M. Elouard and F.-M. Gibon 515

within the Ephemeroptera, Plecoptera, and Megaloptera, dae). However, in general, such is not the case. If sorne which often show poor flying capacity, and even for certain affinities can be noted in each group mentioned above, a Trichoptera. This phenomenon is most pronounced within doser relation can be found with fauna from continental those Ephemeroptera that, in the adult stages, live only a Africa and more particularly that of South Africa. How­ few tens ôf minutes ta a few hours and that have limited en­ ever, the "Lemuria" fauna (Madagascar + the Dekkan Pla­ ergy reserves and low flying capacity. Generally, within teau) separated from Africa 120 miIlion years ago (Ma), these insects only the females are able to accomplish a slight whereas India separated from Madagascar between 90 and dispersal flight; these flights are generally in an upstream 80 Ma. The Indian fauna should thus show doser faunistic direction. Males achieve only copulatory flights, mainly in affinities ta the Malagasy fauna than tO that of sub-Saharan swarms. Within the Trichoptera, flying capacity depends on Africa. If we assume that the facts and dates concerning the family and genus. In sorne cases these insects live rela­ continental drift are precise, the majority of the genera tively long periods, up ta one or twO weeks. However, given must have existed weil before 120 Ma, and the fauna of their limited flying ability, the isolation of watersheds or the Dekkan Plateau has been largely modified. This hypo­ groups of watersheds or the effects of accentuated relief pre­ thesis appears more probable when one considers that sumably impose important limitations on dispersal. about 50 Ma the Dekkan Plateau passed from the south­ Explanations for the high levels of microendemism ernmost Tropics to the northern Tropics, thus undergo­ within insects, such as simuliids, that are known to migrate ing rapid and significant dimatic modifications (see Wells, long distances (30 ta 300 km) cannot be supported by re­ this volume). During this movement, the fauna of the Dek­ duced flying capacity. The range limits of certain species kan Plateau probably became impoverished. After the Dek­ could be due either ta physicochemical qualities of the kan Plateau's collision with Asia, an Asian fauna invaded water or ta the presence of particular hosts on which the it, and only rare representatives of the Gondwanan fauna hematophagous females feed. remained. Between 120 Ma (separation of Madagascar and India) and 50 Ma (collision of India with Asia), sorne groups Distribution could have passed from Asia tO Madagascar by way of ma­ Diversity of aquatic insects is much higher in the eastern hu­ rine regressions or, more likely, via the intermediate steps mid forests than in the western regions. This great diversity that were created by India and the Seychelles Plateau. This in the east is due, in great part, to e1evational and latitudi­ colonization route has been called the "lemurian stepping nal vicariance. In the west, a continuous distribution is of­ stones" (Schatz 1996, p. 77). Examples within the Trichop­ ten observed among certain taxa from the northern ta the tera indude the Asian and North American genus Po­ southern portions of the island, without any marked mi­ tamyia, absent from the African continent, and Leptocerus croendemism or regionalism. In the east, two different geo­ and Setodes, for which the Malagasy lineages are doser graphie patterns are found. Numerous taxa show marked to old Asian lineages than ta those of Africa (Gibon and leve1s of microendemism, whereas others belonging to the Randriamasimanana 2000; Randriamasimanana and Gi­ same genera are relative1y ubiquitous with a broad distri­ bon 2001). bution, sometimes extending across the whole east coast. For the moment, there is no satisfactory expIanation for such differences in distribution in the east. It is probable Uniqueness of the Malagasy Taxa that the answer is complex and variable within each taxo­ nomie group. Within the Trichoptera, for example, the In addition tO aspects concerning species richness, endem­ fauna of the primary forests is localized, whereas that of the ism, and microendemism, the Malagasy fauna presents natural open or human-induced habitats shows broader other original characters. Among these are gigantism, ar­ distributions. chaism, and sorne specifie adaptive radiations that have presumably been generated by the poverty of certain taxo­ Affinities nomie groups.

If one assumes that the families and majority of the genera Gigantism date from the period before the breakup of Gondwana, the Malagasy fauna should show considerable affinities with The mayflies belonging tO the family Polymitarcyidae in the the faunas of South America, Australia (crayfish), New genus Proboscidop/ocia are the largest in the world, the fe­ Zealand (certain Leptophlebiidae), and India (Palingenii- males reaching up ta 7 cm in length (exduding tail cerci). 516 Invertebrates Aquatic

Also within the maytlies, those belonging to the genus Eato­ Cheirogenesia dig out vertical U-shaped burrows in the nica, though similar to their African relatives, are larger mud at the edge of rivers is an unusua.1 behavioral tcait. ft is and include sorne of the largest maytlies in the world. In the likely an adaptation to very low minimum flows that leave same way, two Malagasy species of maytlies in the genus the steep riverbanks exposed foc many months of the year. Prosopistoma (Prosopistomatidae) are much larger than other members of this genus elsewhere in the world. How can this gigantism be explained? We think that weak pre­ Future of the Malagasy Aquatic Fauna dation pcessure or absence of competitors might have driven the evolution and maintenance of these large forms. Various chapters in this volume present the specifie prob­ lems involved in the conservation and future of various tax­ onomie groups. However, with regard to aqu'atic inverte­ Archaism brates one must underline the following principal threats. Mil/ot (1952) considered the abundance ofarchaic forms as one of the most important characteristics of the Malagasy Deforestation fauna. Again, this appears linked to geographic isolation, as Clearance of vegetation is, at the present time, the major found on many other isJands, including Austcalia. ft also problem concerning the aquatic fauna and flora of the applies to the Parastacidae crayfishes as wel/ as to several eastern humid forests. Deforestation results in a number freshwater insect clades. For example, the geneca Madecas­ of physicochemical changes in aquatic ecosystems (e.g., in sorythus and Spinirythus are the most plesiomocphic of the water temperature, insolation, pH, turbidity, hydrology, tricorythid lineage, which is pcesent in continental Afcica, dissolved oxygen levels, and quantity and quality of organic on Madagascac, and on the Deccan Plateau. matter inputs), which will influence lotie communities. Many specialized taxa cannot tolerate these major changes and disappear. Because of the important levels of microen­ Adaptations demism in the aquatic ecosystems of the eastern portion of The most rheophilic (preferring fast-flowing watecs) of the the island, the effects of local forest clearing wil/ have more Ephemecoidea in the world ace found in Madagascac serious consequences foc aquatic invertebrates there than in within the genus Proboscidop/ocia. These taxa ace rela­ other aceas. Deforestation thus initial/y generates a reduc­ tively highly evolved insects from a larval point of view, in tion in distributional ranges and then involves the disap­ a group within which the Gondwanan lentic origins gener­ pearance of species or even genera. Because of endemism ated mostly fossorial (and few rheophilic) forms. They are and microendemism, many forms ace threatened. Various thus more evolved than the African genera Ephoron, Eato­ species of crayfishes have already suffered a marked con­ nico, Ephemera, and Afromera. The hyperdevelopment of traction of their distributions; in the same way, the st~eam the mandibles is an adaptation to life under the rocks; these insect fauna of the eastern rain forest is progressively being are used neither to grip nor to catch prey but act like an ex­ lost as deforestation continues (see Crandall, this volume). cavator to dig and move sand. Dorso-ventral flattening is It appears extremely probable that, within 40 years, at least another adaptation to this way of life. On the other hand, half, if not two-thirds, of the Malagasy Ephemeroidea will the forefeet do not show flattening or widening to aid in have become extinct. digging, once again differentiating this genus from other In the west, the stream invertebrate fauna is poorer and Ephemeridae. The eggs of Proboscidop/ocia ace also of a more widely distributed. These patterns attenuate the risks very original form compared with those of other represen­ of extinction, although deforestation of the central high­ tatives of the family (Sartori et al. 1999). lands has had notable effects on hydrology and sediment The wings of Cheirogenesia, with their undulated poste­ transport. However, study of certain groups reveals rare rior edge, bat-wing-shaped, are unique among the mayflies. and localized western species. These taxa are probably Males never fly but have an extraordinary "hydroplane" relier forms from a period during which the central high­ behavior (Sartori and Elouard 1999). The particular shape lands and their sedimentary basins showed denser vegeta­ of the anterior wing is certainly related to this ability, and tion coyer and more stable hydrologie regimes. The future this energy-saving adaptation may be related to weak pre­ of these forms relies less on specifie conservation activi­ dation pressure from fishes (Ruffieux et al. 1998); very few ties than on a general improvement in vegetation coyer insectivorous fishes populate the slow-flowing zones of (e.g., through agriculturaJ systems that control erosion and Malagasy rivers (Kiener 1963). Final/y, the fact that certain reforestation). R. Dolch 517

Hydrologic Modification spected. What results is a reduction in stocks and yields. It is common to see Ti/apia on sale that are only 3-4 cm in Two major types of hydrologic modification act on lotic length, as weil as crayfishes less than 7 cm in length and fe­ communities in Madagascar. These are adjustments to wa­ males carrying eggs or juveniles. tershed hydrology and changes to river and stream chan­ neiS. The establishment of rice fields primarily causes changes ta watershed hydrology. Although rice agriculture probably has notable effects on river flow and on the tem­ Pollution perature and pH of the water, no study has yet documented Aquatic pollution is, for the moment, of minor importance them. Adjustments to river channels consist of the con­ and only affects limited areas. First, pollution occurs as a struction of dams and dikes. Dams, in addition to their result of urbanization, such as that found on the Ikopa known effects downstream, also influence populations of River in Antananarivo. This organic and chemical pollution macrocrustaceans and fishes, which migrate along the is, at the present time, diluted relatively quickly down­ channel as part of their reproductive cycle. Eels and Macro­ stream. Second, mining generates large amounts of sedi­ brachium shrimps often disappear from the zones upstream ment. The effects of mining can be seen at the graphite min­ of dams because the'y are unable to move up and down the ing operation near Andasibe and in streams used in the river. extraction of colluvial sapphires in Ilakaka. More serious, at the present time, is pollution by pesticides that are often used in massive quantities (e.g., in cotton plantations and Overexploitation of Stocks during locust outbreaks). This pollution is more difficult to The weakness of regulations associated with the collection evaluate because it is diffuse in space and time (nonpoint of freshwater fishes and macrocrustaceans is notable. Nei­ source). However, it is largely restricted to the western slope ther minimal sizes nor the dates of fishing seasons are re- of the island, which is less rich in aquatic insect species.

Insect-Plant Interactions: Their Importance for Biodiversity and Ecological Functioning

R.Dolch

Madagascar has deservedly gained fame as a global biodi­ insects and plants are of critical importance for ecosystem versity hot spot (Myers et al. 2000). Most recent biological functioning, since they encornpass ecological processes such research on the island concerns the vertebrate fauna (with as polIination and herbivory. Generally, more than two disproportionate primatological bias). Invertebrates, and trophic levels are involved, and interactions of herbivores insects in particular, have a much smaller lobby. Yet, if we with their predators and parasitoids directly affect biologi­ want to improve our knowledge of Madagascar's biodiver­ cal control. sity, we must study its insects. Arthropods form the bulk of Both Malagasy plants and insects are immensely species­ the world's biodiversity and account for more than half of rich. It is currently estimated that about 12,000 species ail known species (Stork 1991). Extensive insect inventories of plants occur on Madagascar, depending on what is con­ in Madagascar have largely focused on species of medical sidered a taxon, and about 92% are endemic (Koechlin importance (Doucet 1951) and, more recently, on aquatic et al. 1974). Sorne forests show intermediate levels of spe­ (Sartori et al. 2000) or saprophytic (Betsch 2000a) insect cies diversity in a pantropical comparison (Abraham et al. cornrnunities. Surprisingly, phytophagous insects have re­ 1996); others are florally among the most diverse in the ceived the least attention. Yet rnost insects depend on plants world (Dumetz 1999). Moreover, there is an enormous as resources, and together insects and plants account for al­ nurnber of insect species with high levels of endemisrn (e.g., rnOSt three-quarters of ail living organisms (Strong et al. Cassola and Andriamampianina 1998), of which a signifi­ 1984). Moreover, the majority of the interactions between cant proportion may still be unknown. Previous generations Elouard Jean-Marc, Gibon François-Marie, Goodman S.M. (trad.), Wilmé L. (trad.) (2002) Ecology of aquatic invertebrates In : Goodman S.M. (ed.), Benstead J.P. (ed.), Schütz H. (photogr.) The natural history of Madagascar. Chicago (USA) ; Chicago : University of Chicago Press ; Field Museum, 511- 517 ISBN 0-226-30306-3