Version classiqueVersion mobile
OpenEdition Books

The inland water fishes of Africa

Didier Paugy
Christian Levêque
Olga Otero


Didier Paugy, Christian Levêque et Fabrice Duponchelle


The reproductive cycle of a species implies a set of physiological and behavioural processes influenced by various factors in the abiotic and biotic environment. Traditionally a distinction is made between gametogenesis and spawning. Gametogenesis is the physiological process involved in gonad development that leads to the formation of gametes (oocytes and spermatozoa). In tropical regions, gametogenesis is controlled by tiny changes in environmental factors, and its duration varies according to species (Munro, 1990). Spawning is the critical moment of the reproductive cycle, which includes the sequence of events that leads to laying of the gametes (maturation of oocytes, ovulation, spermiogenesis), under the influence of specific stimuli (Billard & Breton, 1978).
Gametogenesis is the longest of these processes. It generally lasts several months, but reproductive success depends to a large extent on the success of a set of behavioural activities. These include the search for partners, parental care before and after spawning, whether or not migration is required to find suitable sites for eggs and larvae, synchronization of spawning with changes in environmental factors, synchronization of the genital products laying, etc. These activities may be accompanied by a set of phenological adaptations, such as the development of secondary sexual characters with regard to morphology or colouring.
The study of the fish reproductive cycle thus requires a multidisciplinary approach that involves ecologists, physiologists, ethologists, and geneticists. Coordinating all these disciplines is not always a straightforward affair, and explains why data is often fragmentary. Yet we need studies that include them if we want to understand the cause-and-effect relationships between environmental factors, physiological processes, and reproductive behaviour of species.

Texte intégral

Gonad maturation and fecundity

1Reproductive effort is the fraction of absorbed energy that an organism devotes to its reproductive activities. This investment can be partially measured through the amount of material or energy stored in the gonads, keeping in mind that it would also need to add all the energy spent on eggs or larvae care, or on mating displays.

Gonado-somatic index

2To describe the growth of gonads in relation to body mass, scientists frequently use the gonado-somatic index (GSI) which is the ratio of gonad weight to body weight excluding gonads:

3where GW is the weight of the gonads and TW the total weight.

4The growth of ovaries during the gonad maturation cycle reflects the growth of oocytes through yolk accumulation. The GSI of mature females varies widely from one species to another. It can reach 20 to 30% prior to spawning in some species, but only remain at a few% in others (tables 10.I and 10.II).

5In many tropical species, testes develop much less than ovaries, and for African species, the GSI of mature males rarely exceeds 2% (table 10.I; see also box “Gonado-Somatic Index (GSI)”). While the reasons for such a difference are unknown, males generally reach maturity before females and remain mature longer.

TABLE 10.I. Comparison of female and male gonado-somatic index for different species in the Baoulé River, Upper Senegal in Mali. Maximum observed values quoted for different species (from Paugy, 2002).




Polypterus bichir



Hyperopisus bebe



Mormyrus rume



Mormyrops anguilloides


Marcusenius senegalensis



Petrocephalus bovei



Hydrocynus forskalii



Alestes baremoze



Brycinus nurse



Brycinus leuciscus



Brycinus macrolepidotus



Labeo senegalensis



Labeo coubie



Chrysichthys auratus



Schilbe intermedius



Schilbe mystus


Clarias anguillaris


Synodontis schall



Synodontis ocellifer



Sarotherodon galilaeus


6The gonado-somatic index is a simple but rough estimate of sexual activity. A more refined assessment of gonad development requires a histological study, such as an investigation into the frequency distribution of oocyte size, provided that the relationship between oocyte size and the physiological state of gonads has been predetermined. The GSI is a good tool for estimating reproductive seasons of species that spawn only once per year or per reproductive season. However, it becomes inexact for species with multiple spawning events, because a low GSI (for instance between 0.5 and 1.5% for Oreochromis niloticus) can characterize very different stages of ovarian maturity, such as a gonad in early maturation (stage 2 according to the scale by Legendre & Écoutin, 1989) or a post-spawning gonad (stage 6). The resulting loss of information can be detrimental, particularly in attempts to correlate the reproductive season with environmental factors involved in the regulation of sexual cycles, and it is advisable to use a gonad maturity scale rather than the GSI for species that spawn several times during a reproductive season (Duponchelle et al., 1999).

Gonado-Somatic Index (GSI)
In the African inter-tropical zone, testes weight is always lesser than that of ovaries (table 10.I). This is also often the case in temperate zone. Thus, in Gasterosteus aculeatus (Gasterosteidae), while the GSI of the female reaches or even exceeds 20%, the fully mature male’s GSI is below 2% (Wootton, 1984). Nonetheless, in some brackish water species, higher values can be observed. In very favourable conditions, the GSI of male Fundulus heteroclitus (Cyprinodontidae) can reach 4.7% (Taylor, 1990). Even higher values can be found in some marine species. The male Arctic cod Boreogadus saida (Gadidae) can have testes weighing 10 to 27% of the body weight (Craig et al., 1982). Theory suggests that testes size (and thus weight) could be linked to the mode of fertilization used by the species. Thus, in the case of “coupling” with partners, the amount of sperm needed is less than in the case of dissemination in the physical system without an actual meeting of the progenitors, as with numerous marine species. In the latter, the chances of an encounter are greater if more sperm is released.
Other “solutions” can be found, such as variations in the sex ratio. When a species lays all its eggs in one instance and leaves them to sink to the bottom, where they become attached on submerged surfaces, a large number of spermatozoa is needed to increase chances of fecundation.
Consequently, testes are as developed as ovaries in these species, and males must be at least as numerous as females (high sex ratio). In other cases, meanwhile, spawning is spaced out and eggs float, so spermatozoa found in the same waters can be less abundant. As a result, testes are less developed then ovaries and there can be fewer males than females (low sex ratio). As the reproductive behaviour of tropical species is still poorly known, we cannot yet state if this strategy has been verified in this zone.
In any case, the energy allocated for the constitution of genital products appears to be lower in males than in females.

7Maturity scales allow rapid characterisation of the different developmental states of ovaries and testes. In particular, they take into account the macroscopic appearance of gonads (shape, colour, vascularization, transparency) and the GSI. Such scales have been established for Alestes baremoze (Durand & Loubens, 1970), Schilbe mystus, S. uranoscopus (Mok, 1975) and Sarotherodon melanotheron (Legendre & Écoutin, 1989). The maturity scale proposed for the last species has been applied successfully to Oreochromis niloticus (Duponchelle & Panfili, 1998, Duponchelle et al., 1998, 1999, 2000a; Duponchelle & Legendre, 2000, 2001) and to numerous Haplochromis species (Duponchelle et al., 2000b). It appears to be suitable for the majority of Cichlidae.


8The ovaries contain oocytes of different sizes corresponding to the different stages of development. A unimodal distribution suggests single spawning, while polymodal structures suggest that the species under study has multiple spawning events (Albaret, 1982) (figure 10.1). In all cases, absolute fecundity (F) corresponds to the number of mature oocytes found in the ovary immediately prior to spawning. For species with only one annual spawning, F equals total annual fecundity. In species that spawn several times, F is the number of oocytes that will be released in the next spawning event (these oocytes are the largest in size), and annual fecundity is the sum of all oocytes released over the year in successive spawning events.

FIGURE 10.1. Distribution in oocyte diameters of two species from Côte d’Ivoire (from Albaret, 1982).

FIGURE 10.1. Distribution in oocyte diameters of two species from Côte d’Ivoire (from Albaret, 1982).

9Brosset (1982) calculated the annual fecundity for three species of Cyprinodontiformes found in the Ivindo River (Gabon). The number of eggs released annually by a female is the same, that is, 200 to 300, regardless of the strategy adopted during the reproductive season: one egg a day in Diapteron sp., 8 to 15 eggs every 8-10 days in Aphyosemion hertzogi, 30-75 eggs a month in Aphyosemion cameronense and A. punctatum.

10In the Nothobranchiidae species Nothobranchius guentheri, a mature female weighing 1 g produces 20 eggs a day, which represents a reproductive effort equivalent to 27% of its weight in a month. In this case, in a 4-5 month period, the total egg production can be equivalent to the female’s weight. This is much higher than the energy investment usually made by species that have single spawning (table 10. II).

11To compare the fecundity of fishes of different sizes or from different physical systems, we often calculate the relative fecundity which is the number of eggs per unit of body weight, expressed in g or in kg. However, in some species including a lot of Cichlidae, a correlation (generally negative) exists between relative fecundity and female weight (Sarotherodon melanotheron, Legendre 1992; Oreochromis niloticus, Duponchelle, 1997; Duponchelle et al., 2000a), and we must thus proceed by comparing lines of regression between fecundity and female body weight.

TABLE 10.II. Selected data on mean female size at first sexual maturity (mm), egg diameter (mm), relative fecundity expressed as number of eggs per kilogram of female body weight, maximum gonado-somatic index (GSI in %) and maximum size observed (MSO in mm) for different African fish species.

TABLE 10.II. Selected data on mean female size at first sexual maturity (mm), egg diameter (mm), relative fecundity expressed as number of eggs per kilogram of female body weight, maximum gonado-somatic index (GSI in %) and maximum size observed (MSO in mm) for different African fish species.

12There is an inverse relationship between the number of eggs released at each spawning event and the size of the oocytes. Very prolific species that produce small eggs mostly belong to the Cyprinidae, Alestidae, and Schilbeidae families. Most are pelagic species, and are often migratory. Species producing a few large eggs mainly belong to the Bagridae, Mormyridae (which only possess the left gonad), and Cichlidae families. For the latter, substrate layers (H. bimaculatus, H. fasciatus, C. zillii) have smaller and more numerous eggs than mouth brooding species (S. galilaeus, O. niloticus).

13Mouth brooders generally have a limited number of large eggs, as shown in the results obtained for various Cichlidae species of the African great lakes (table 10.II).

Relationships between fecundity and length or weight of females

14The fecundity of a species is most often a function of the size of the females. The larger they are, the more eggs they lay. It is customary to determine the relationship between fecundity and size of females, expressed either by length (often the standard length) or weight (table 10.III).

15The relationship is generally F = a SLb.

Reproductive strategies

16Each individual possesses a set of biological characteristics related to reproduction that are the heritage of evolution and phylogeny. The reproductive strategy of a fish species in a given environment is indeed a set of biological traits such as age and size at first reproduction, fecundity, gonad development and gamete size, reproductive behaviour including the existence of parental care, reproductive season, etc.

17But an individual can also develop tactics, which are in fact variations on the typical reproductive schema of the species, to adapt to changes in environmental factors. This adaptive behaviour to special ecological conditions is intended to ensure the survival of the species, as we have already seen. This is the case for instance of Oreochromis niloticus in artificial reservoirs in Côte d’Ivoire, whose reproductive characteristics (reproductive season, size at first sexual maturity, fecundity, egg size, etc.) vary from one year to another depending on environmental conditions (Duponchelle & Panfili, 1998; Duponchelle et al., 1998, 1999 et 2000a).

Main modes of reproduction

18Most fishes are oviparous. Oocytes and sperm are released into the water and fecundation takes place immediately. There are also cases of intra-buccal fecundation, particularly in Cichlidae.

19Depending on the demographic strategies implemented by fishes and the ecological adaptations to certain physical systems, we can distinguish different reproductive modes or styles. The general classification proposed by Balon (1975 and 1990), based on the degree of parental care (ethology) and reproductive sites (ecology) is generally accepted. The author identifies three broad sets within which there are etho-ecological subsets, regardless of the phyletic origin of the species (table 10.IV):

  • fishes that do not guard their eggs once these have been released, among which we can distinguish fishes that lay eggs on open substrates or those that provide a modicum of protection by concealing them. These are generally fishes that are highly fecund and have an indirect form of development;
  • fishes that guard their eggs in nests, either on pre-selected or pre-constructed substrates. These species generally have a few relatively large eggs, to which they provide different forms of parental care;
  • fishes that bear their eggs for at least part of the embryonic period. In external brooders, the eggs are carried on the body surface or in the mouth, or in structures that open out. In internal brooders, the eggs are fertilized and transported within the body.

TABLE 10.III. Relationships between fecundity (F) and standard length (SL in mm) or weight (W in g) for different African fish species.


F vs SL

F vs W



Marcusenius ussheri

F = 80 SL-9742

F = 45 W + 592

Albaret, 1982

Petrocephalus bovei

F = 55 SL-3335

F = 105 W-180

Albaret, 1982


Hepsetus odoe

F = 172 SL-41581

F = 59 W-14250

Albaret, 1982

Hepsetus odoe

F = 10.8 SL-803.0

Merron et al., 1990


Alestes baremoze

F = 657.4 SL-117813.5

F = 323 W-19122

Durand, 1978

Alestes baremoze

F = 288 W-7078

Paugy, 1978

Alestes baremoze

F = 391 SL-51456

F = 253 W-1884

Albaret, 1982

Brycinus imberi

F = 6994.6 SL-73028.3

Marshall & van. der Heiden, 1977

Brycinus imberi

F = 176.7 SL-11768.6

F = 186 W + 140

Paugy, 1980a

Brycinus imberi

F = 328 SL-24686

F = 356 W-2800

Albaret, 1982

Brycinus longipinnis

F = 119 SL-6633

F = 224 W-540

Albaret, 1982

Brycinus longipinnis (Bandama)

F = 124.3 SL-6834.6

F = 323 W-1295

Paugy, 1982b

Brycinus longipinnis (Cavally)

F = 70 SL-3818.9

F = 108 W-151

Paugy, 1982b

Brycinus macrolepidotus

F = 814 SL-137900

F = 195 W-1980

Albaret, 1982

Brycinus macrolepidotus

F = 812.8 SL-138120

F = 187 W-1460

Paugy, 1982a

Brycinus nurse

F = 448.6 SL-34638

F = 423 W-2406

Paugy, 1980b

Brycinus nurse

F = 491 SL-39871

F = 450 W-3967

Albaret, 1982

Hydrocynus forskalii

F = 1325 SL-319084

F = 187 W-25065

Albaret, 1982

Micralestes acutidens

F = 19.93 SL-539

F = 0.23 W-70.3

Lek, 1978


Barbus ablabes

F = 122 SL-5266

F = 258 W + 413

Albaret, 1982

Barbus kimberleyensis

F = 0.222 SL3-495.2

Gaigher, 1976

Barbus sublineatus

F = 534 SL-27819

F = 958 W-1849

Albaret, 1982

Barbus trispilos

F = 150 SL-6325

F = 439 W + 87

Albaret, 1982

Labeo capensis

F = 1220 SL-367774

F = 518 W-214717

Mulder, 1973

Labeo parvus

F = 721 SL-69461

F = 424 W-5933

Albaret, 1982

Labeo umbratus

F = 1324 SL-381914

F = 625 W-215164

Mulder, 1973


Bagrus docmak

F = 617 SL-278389

F = 43 W-16827

El Sedfy & El Bolock, 1987


Chrysichthys maurus

F = 45 SL-5265

F = 18 W + 138

Albaret, 1982


Schilbe mandibularis

F = 353 SL-44425

F = 157 W + 5970

Albaret, 1982

Schilbe mystus

F = 1.5 SL-185.1

F = 207 W-1292

Lévêque & Herbinet, 1980

Schilbe uranoscupus

F = 528 SL-98770

F = 250 W-7475

Mok, 1975


Amphilius atesuensis

F = 3.8 SL-104

F = 40 W + 2

Albaret, 1982


Synodontis schall

F = 201 W-7841

Albaret, 1982

Synodontis schall

F = 174.9 W + 5784

Ofori-Danson, 1992


Alticorpus ‘geoffreyi’

F = 2.4365 SL-171.76

F = 1.7836 W + 28.037

Duponchelle et al., 2000b

Alticorpus macrocleithrum

F = 3.6323 SL-259.3

F = 3.3229 W + 14.744

Duponchelle et al., 2000b

Alticorpus mentale

F = 2.4897 SL-211.7

F = 0.9359 W + 63.115

Duponchelle et al., 2000b

Alticorpus pectinatum

F = 2.2238 SL-114.87

F = 3.2395 W + 12.796

Duponchelle et al., 2000b

Aulonocara ‘blue orange’

F = 1.0608 SL-33.755

F = 3.3845 W + 9.6689

Duponchelle et al., 2000b

Aulonocara ‘cf. macrochir’

F = 1.5397 SL-87.623

F = 1.6714 W + 18.2

Duponchelle et al., 2000b

Aulonocara ‘rostratum deep’

F = 2.6948 SL-191.6

F = 2.3872 W + 1.3981

Duponchelle et al., 2000b

Buccochromis lepturus

F = 4.7556 SL-768.67

F = 0.9143 W + 22.481

Duponchelle et al., 2000b

Buccochromis nototaenia

F = 1.9659 SL-105.22

F = 0.8115 W + 96.468

Duponchelle et al., 2000b

Chromidotilapia guntheri

F = 2.6 SL-64

F = 3.4 W + 25

Albaret, 1982

Copadichromis quadrimaculatus

F = 0.684 W-0.8577

Duponchelle et al., 2000b

Cynotilapia afra

F = 0.6252 SL-17.179

F = 1.9679 W + 6.3959

Duponchelle et al. (unpublished)

Diplotaxodon apogon

F = 0.6738 W-1.0371

Duponchelle et al., 2000b

Diplotaxodon argenteus

F = 0.306 SL-11.866

F = 0.3006 W + 6.1836

Duponchelle et al., 2000b

Genyochromis mento

F = 1.6346 SL-94.359

F = 2.203 W + 0.8898

Duponchelle et al. (unpublished)

Labeotropheus fuelleborni

F = 0.6331 SL-21.182

F = 1.1687 W + 7.7613

Duponchelle et al. (unpublished)

Labeotropheus trewavasae

F = 2.857 W-15.549

Duponchelle et al. (unpublished)

Lethrinops argenteus

F = 2.5193 SL-185.09

F = 1.7255 W + 21.72

Duponchelle et al., 2000b

Lethrinops ‘deep water albus’

F = 2.8885 SL-178.91

F = 2.8411 W + 22.771

Duponchelle et al., 2000b

Lethrinops gossei

F = 2.2518 SL-147.31

F = 1.6832 W + 19.458

Duponchelle et al., 2000b

Lethrinops longimanus

F = 2.7948 SL-225.41

F = 2.9578 W-43.716

Duponchelle et al., 2000b

Lethrinops macrochir

F = 3.6833 SL-267.1

F = 2.7129 W + 12.68

Duponchelle et al., 2000b

Lethrinops ‘oliveri’

F = 1.7657 SL-72.585

F = 4.5704 W + 3.0726

Duponchelle et al., 2000b

Lethrinops polli

F = 1.5365 SL-83.52

F = 2.5654 W + 2.6796

Duponchelle et al., 2000b

Mylochromis anaphyrmus

F = 2.7755 SL-194.73

F = 1.9783 W + 25.864

Duponchelle et al., 2000b

Nyassachromis ‘argyrosoma’

F = 0.9259 SL-29.05

F = 3.2835 W + 10.481

Duponchelle et al., 2000b

Oreochromis niloticus

F = 111.92 SL-993.64

F = 2.91 W + 277.7

Duponchelle et al., 2000a

Oreochromis niloticus

F = 53.507 SL-308.55

F = 2.52 W + 164.4

Duponchelle et al., 2000a

Oreochromis niloticus

F = 129.9 SL-1254

F = 3.64 W + 197.1

Duponchelle et al., 2000a

Oreochromis niloticus

F = 85.852 SL-695.1

F = 3.61 W + 103.3

Duponchelle et al., 2000a

Oreochromis niloticus

F = 116.2 SL-1095

F = 2.59 W + 337.2

Duponchelle et al., 2000a

Oreochromis niloticus

F = 81.871 SL-633.54

F = 2.01 W + 368.6

Duponchelle et al., 2000a

Oreochromis niloticus

F = 122.43 SL-1156.2

F = 3.14 W + 260.5

Duponchelle et al., 2000a

Oreochromis niloticus

F = 77.809 SL-504.14

F = 3.78 W + 152.5

Duponchelle et al., 2000a

Oreochromis niloticus

F = 99.584 SL-775.71

F = 2.69 W + 317.8

Duponchelle et al., 2000a

Otopharynx ‘productus’

F = 2.8372 W-7.8469

Duponchelle et al., 2000b

Otopharynx speciosus

F = 3.547 SL-452.51

F = 1.0452 W-25.644

Duponchelle et al., 2000b

Pallidochromis tokolosh

F = 0.7106 SL-78.13

F = 0.5496 W-4.5233

Duponchelle et al., 2000b

Petrotilapia ‘fuscous’

F = 2.1274 W-6.1381

Duponchelle et al., (unpublished)

Placidochromis ‘long’

F = 1.333 SL-57.675

F = 6.8445 W-12.583

Duponchelle et al., 2000b

Placidochromis ‘platyrhynchos’

F = 1.8798 W + 14.85

Duponchelle et al., 2000b

Pseudotropheus ‘aggressive blue’

F = 1.0151 SL-46.016

F = 1.7042 W + 6.5604

Duponchelle et al. (unpublished)

Pseudotropheus callainos

F = 1.1263 W + 5.8628

Duponchelle et al. (unpublished)

Pseudotropheus heteropictus

F = 0.9346 SL-34.827

F = 4.0141 W-10.079

Duponchelle et al. (unpublished)

Pseudotropheus livingstonii

F = 3.1033 SL-128.38

F = 6.4275 W + 1.2343

Duponchelle et al., 2000b

Pseudotropheus tr. ‘lilac’

F = 0.7236 SL-23.059

F = 1.4092 W + 9.8755

Duponchelle et al. (unpublished)

Pseudotropheus tr. ‘red cheek’

F = 0.6259 SL-22.07

F = 2.0579 W-4.9336

Duponchelle et al. (unpublished)

Pseudotropheus tropheops ‘blue’

F = 1.6927 SL-89.196

F = 3.596 W-17.192

Duponchelle et al. (unpublished)

Pseudotropheus williamsi

F = 1.8937 SL-104.3

F = 3.0137 W-5.8688

Duponchelle et al. (unpublished)

Pseudotropheus zebra

F = 0.8112 SL-27.739

F = 1.9246 W + 6.6811

Duponchelle et al. (unpublished)

Pseudotropheus zebra ‘black dorsal’

F = 1.4866 SL-74.477

F = 1.7705 W + 9.2223

Duponchelle et al. (unpublished)

Pseudotropheus zebra ‘red dorsal’

F = 0.9901 SL-36.543

F = 2.1674 W + 6.981

Duponchelle et al. (unpublished)

Pseudotropheus zebra ‘yellow throat’

F = 1.1536 SL-53.814

F = 2.3995 W-1.8984

Duponchelle et al. (unpublished)

Sarotherodon melanotheron

F = 0.0008 LF 2.449

F = 1.72 W-15.0

Legendre & Écoutin, 1989 (unpublished)

Sarotherodon melanotheron

F = 2.7238 LF-160.01

F = 1.948 W + 119.08

Panfili et al., 2004a

Sarotherodon melanotheron

F = 2.8459 LF-156.76

F = 2.111 W + 128.11

Panfili et al., 2004a

Sciaenochromis alhi

F = 2.4762 SL-200.07

F = 4.4172 W-55.843

Duponchelle et al., 2000b

Sciaenochromis benthicola

F = 1.0481 SL-74.232

F = 1.264 W-0.3265

Duponchelle et al., 2000b

Stigmatochromis ‘guttatus’

F = 1.1159 SL-87.92

F = 1.4769 W-8.0045

Duponchelle et al., 2000b

Taeniolethrinops praeorbitalis

F = 2.0847 SL-135.28

F = 0.9672 W + 74.587

Duponchelle et al., 2000b

Tilapia guineensis

F = 0.1458 LF 1.9899

F = 18.61 W + 2018.0

Legendre & Écoutin, 1989 (unpublished)

Coptodon zillii

F = 49 SL-2987

F = 28 W + 610

Albaret, 1982

Trematocranus placodon

F = 1.5364 SL-62.449

F = 1.2303 W + 50.512

Duponchelle et al., 2000b


Mastacembelus nigromarginatus

F = 8.4 SL-1468

F = 187 W + 67

Albaret, 1982

20Each etho-ecologic section includes different reproductive guilds, of which we provide only a few examples (table 10.IV). These guilds are based on a combination of morphological, behavioural, and ecological characteristics.

TABLE 10.IV. Simplified classification of reproductive styles in fish (from Balon, 1975).

TABLE 10.IV. Simplified classification of reproductive styles in fish (from Balon, 1975).

21An estimate of the proportion of different eco-ethological guides in African continental waters reveals a much higher number of guarders and bearers in the East African great lakes, which are characterized by relatively predictable physico-chemical systems. Meanwhile, a large percentage of non-guarder species are found in the rivers and wet zones, which have less predictable physical systems (Bruton & Merron, 1990).

22To complete the reproductive guilds, we should also take into account the frequency of spawing and the period(s) during which spawning occurs. This usually takes place when the environmental conditions are, in principle, most conducive to the survival of eggs and larvae, and many species have a seasonal reproductive cycle. It is nonetheless possible to distinguish several options in reproductive behaviour (table 10.V).

TABLE 10.V. Examples of main types of reproductive behaviour in African fishes (adapted from Lowe-McConnell, 1987).

Type of fecundity



Migratory behaviour/parental care


once in the lifetime


long distance migrations, no parental care

total spawners

highly seasonal


anadromous migrations

(very high fecundity)

(generally throughout the flood)


open water spawners


no parental care

partial spawners

throughout the flood

many cyprinids



bottom nester


egg garders


floating nest (plant material)


floating nest (foam)


small brood spawners

sometimes throughout the year


nest constructors and parental behaviour

(low fecundity)

sometimes throughout the flood


mouth brooders

during the filling of ponds


annual species with resting eggs

23One very rare behaviour is the “big bang” (Lowe-McConnell, 1987) which, like in eels, consists of reproducing once in a lifetime and then dying.

24Several species, meanwhile, have a single short reproductive period every year. This is the case for many riverine species that practice reproductive migrations and spawn at the beginning of or during the flood, when conditions are optimal for juveniles that can find shelter and food in the flooded plains. This periodic strategy does include risks tied to climate variability and variations in flooding, making it difficult to predict the optimal period for reproduction. The species in question often release a very large number of eggs, and are sufficiently long-lived to reproduce for several consecutive years, thus minimizing the consequences of climate-related variations on the recruitment.

25For other species, the reproductive season is spread out over the year. Scenarios differ depending on the group. Either individuals of a given species do not reach maturity at the same time, or eggs are spawned in batches at more or less regular intervals. This strategy tends to minimize the risks of a single spawning that may occur at an unfavourable period, and tends to benefit species with short lifespans.

26Finally, other species spawn practically throughout the year. Most of these multiple spawners provide parental care, like the Cichlidae. Although a full reproductive cycle cannot be completed in 24 hours, certain phases follow a daily cycle. In particular, spawning can take place at certain hours, as has been observed in many fishes such as B. macrolepidotus which appears to reproduce before dark (Paugy, 1982a)

27Cichlidae belonging to the genera Tropheus and Simochromis that live in the same habitats in Lake Tanganyika (between rocks, a few metres depth) and feed on Aufwuchs do not spawn at the same time of day. The former spawn early in the morning, while the latter spawn during daylight (Nelissen, 1977). Differences in spawning activity rhythms can contribute to limiting interspecies competition.

28In the Nothobranchiidae Nothobranchius guentheri, reproductive behaviour follows a day-long cycle (Haas, 1976b). Reproductive activity is low very early in the morning, but increases 2.5 hours after dawn and reaches a peak 6 hours after dawn before waning.

29The influence of lunar cycles has also been identified in the reproductive behaviour of certain fish species. The proportion of females ready to spawn among Lake Victoria Oreochromis niloticus is higher during the full moon than during the new moon (Okorie, 1973). In the Ethiop River in Nigeria, which has particularly clear waters, there is a marked synchronization of Tilapia mariae spawns with the lunar cycle, as 85% of spawning occurs during the last quarter of the lunar cycle, that is, between 0 and 6 days before the full moon (figure 10.2). Spawning before the full moon could contribute to improving larval survival owing to more efficient parental guarding during moonlit nights (Schwanck, 1987).

FIGURE 10.2. Daily counts of egg batches in Tilapia mariae (from Schwanck, 1987).

FIGURE 10.2. Daily counts of egg batches in Tilapia mariae (from Schwanck, 1987).

Synchronization between reproduction and changes in the environment

30Reproduction in fishes is a cyclic phenomenon, synchronized in many species, with seasonal environmental fluctuations that have an impact on fish physiology, as they can stimulate or inhibit certain stages of gametogenesis or spawning. Both are controlled by endocrine processes that are influenced by environmental factors such as temperature, photoperiod, rainfall, etc. Each species or species group can react differently to the environmental factors that affect their physiological rhythms.

31This synchronization between the reproductive cycle and predictable changes in the environment is an adaptive process. As they evolved, fishes adapted their sexual cycle in order to reproduce at periods most favourable to the survival of their eggs. How did fishes program their reproductive cycle? What factors are responsible for gonad maturation in such a way that species are ready to spawn when conditions are favourable for the survival of eggs and larvae, thus granting them a selective advantage? We are still far from knowing the answers to all these questions, even though we suspect certain ecological factors of playing an important role, as reproduction is a complex phenomenon in which several abiotic factors are involved. In all likelihood, the chronology of the reproductive cycle is in fact a compromise between stimuli from different sources. Gametogenesis thus starts several months before spawning season, when conditions are not yet favourable to the survival of the young. Whatever the case, the synchronization of the reproductive cycle with environmental changes confers an adaptive advantage to species.

Single spawning species

32Several authors have attempted to identify which factors in the physical system were involved in triggering maturation as well as spawning.

33In the Niger basin, Bénech & Ouattara (1990) investigated the role of various environmental factors in the ovarian growth of Brycinus leuciscus by comparing the reproductive cycle in different natural conditions (different flood periods) between the sites of Mopti and Niamey. In Mopti, gametogenesis begins in week 20 at the same time as the climate changes induced by the movement of the intertropical front. The same thing occurs in Niamey in week 22. In both stations, the start of gonad development occurs once the water temperature has attained a maximum (29° C) in week 19, and water conductivity has stopped increasing with the appearance of the first rains. The gonado-somatic index reaches its peak in week 30 in Mopti, after which spawning begins and continues up to week 39. In Niamey, meanwhile, ovarian growth appears to be interrupted in week 22, then restarts again six to nine weeks later to reach a peak in week 35, with spawning occurring until week 45. In both sites, gonad development coincides with a decrease in conductivity after a period of warming and the start of the flood. Spawning occurs as water rises, a few weeks before the water level reaches its peak. It is nonetheless difficult to establish precisely which factor is responsible for triggering the different reproductive processes.

FIGURE 10.3. Hydrological cycle and changes in the gonado-somatic index (GSI) of different fish species from the Baoulé River, a tributary of the Upper Senegal basin. Spawning occurs before the flood peak (from Paugy, 2002).

FIGURE 10.3. Hydrological cycle and changes in the gonado-somatic index (GSI) of different fish species from the Baoulé River, a tributary of the Upper Senegal basin. Spawning occurs before the flood peak (from Paugy, 2002).

34In aquarium conditions, Kirshbaum (1984) triggered Mormyridae gonad maturation by diminishing water conductivity, simulating rainfall, and increasing the water level, all factors identical to those that appear to be involved for B. leuciscus.

35What stimuli trigger spawning? In large tropical rivers, the hydrological regime (or, to be precise, the set of climatic conditions prevailing at the start of the flood) appears to be the main regulator of reproduction. For many species, spawning coincides with the flood and can occur from the start of the flood up to its peak, but rarely during the recession. As such, most of the species present in the Baoulé (upper course of the Senegal) reproduce before the flood (figure 10.3), and this phenomenon has been seen in numerous species in various physical systems (Bénech & Quensière, 1985; Munro, 1990).

36However, reproduction and flood are not always synchronized, which casts some doubt on the true role of this factor. In the Okavango delta, the flood occurs in the dry season, during winter months, as the flood wave takes a while to go from southern Angola via the Okavango River. In these conditions, Hepsetus odoe spawns in summer, with activity reaching a peak when waters are low (Merron et al., 1990), which appears to indicate that the flood is not the primary stimulus triggering spawning in this species. A comparable situation exists in Lake Chad, where the lake floods in winter (December-January), several months after the flood peaks in the Chari River (September). Yet reproduction in most species takes place during the rainy season (July-August) which immediately follows the hot season (Bénech & Quensière, 1985). There are exceptions to this trend, as in Brycinus macrolepidotus and Hydrocynus forskalii of the ‘Sahelo-Sudanian’ basins found in the same latitudes. It is moreover customary in numerous equatorial species which indivudually have a single annual spawn, but without a precise spawning period when the entire population is considered. Brycinus imberi (figure 10.4) and B. longipinnis of the Côte d’Ivoire basins demonstrate such behaviour (Paugy, 1980a and 1982b).

FIGURE 10.4. Long breeding period of the populations of Brycinus imberi from Côte d’Ivoire. Throughout the year, more than half of fish are mature (from Paugy, 1980a).

FIGURE 10.4. Long breeding period of the populations of Brycinus imberi from Côte d’Ivoire. Throughout the year, more than half of fish are mature (from Paugy, 1980a).

Multiple spawning species

37For species that have a reproductive period spread out over time (interval spawning or spawning throughout the year), it is rather difficult to identify the factors influencing the reproductive cycle (Jalabert & Zohar, 1982). This is especially the case for Cichlidae, which are theoretically capable of reproducing uninterruptedly in equatorial regions, though with a tendency to seasonal behaviour depending on latitude. In the Ébrié lagoon (Côte d’Ivoire), lagoon Tilapias (Sarotherodon melanotheron and Tilapia guineensis) reproduce throughout the year without displaying any clear seasonal variation in the intensity of sexual activity (Legendre & Écoutin, 1989). A high proportion of maturing fishes is observed throughout the year. However, in populations reared in parks in the lagoon, the sexual activity of S. melanotheron decreases significantly during the rainy season, from June to September (see box “Spawning frequency in cichlids”). Meanwhile, the sexual activity of Oreochromis niloticus shows marked seasonality in artificial reservoirs of Côte d’Ivoire, with a more spread-out season in the centre and south of the country than in the small agro-pastoral reservoirs of the north (Duponchelle et al., 1999). In both north and south, periodicity of the reproductive season is nonetheless regulated by the photoperiod (see box “Influence of photoperiod on the reproductive cycle of cichlids”). On the other hand, in O. niloticus, fecundity rates are correlated with trophic availability in Côte d’Ivoire’s artificial reservoirs. The highest fecundity rates are recorded when chlorophyll a concentrations are at their highest, just before the flood (Duponchelle et al., 2000a).

Spawning frequency in cichlids (from Legendre & Écoutin, 1989)
In the lagoons of Côte d’Ivoire, where the two species reproduce throughout the year, the interval between two spawns is about two weeks in Sarotherodon melanotheron, and around three weeks in Tilapia guineensis (figure 10.5). Nonetheless, the frequency of spawning appears to be less regular in T. guineensis, which can go for nearly four months without spawning.
For both species, the frequencies provided are averages that should be adapted, as relative frequency varies according to the seasons as shown in the figure below, where it is clear that a minimum is reached in both species during the rainy season, that is, June-July.

FIGURE 10.5. Seasonal changes in the spawning frequency for Sarotherodon melanotheron and Tilapia guineensis in concrete tanks (redrawn from Legendre & Écoutin, 1989).

FIGURE 10.5. Seasonal changes in the spawning frequency for Sarotherodon melanotheron and Tilapia guineensis in concrete tanks (redrawn from Legendre & Écoutin, 1989).

38Several observations appear to reveal the existence of reproductive seasonality that is linked to the abundance of available food. In most of the zooplanktophagous haplochromines of Lake Victoria, there is a small proportion of mature females throughout the year, but there is a peak in reproductive activity during the dry period (June-October) when plankton production is highest owing to the mixing of waters (Goldschmidt & Witte, 1990). This situation is comparable to that of Lake Malawi where the zooplanktophagous ‘utaka’ spawn between March and July, when the waters mix (Iles, 1960).

39A recent study in Lake Malawi confirmed that the reproductive season for these species occurs between March and December, with an additional peak of activity in December. These periods of sexual activity correspond to the ones where plankton production is highest (Duponchelle et al., 2000b). Meanwhile, for the 37 other benthic or pelagic species studied that do not belong to the ‘utaka’ group, no correspondence has been found between the reproductive season and the annual cycles of temperature, photoperiod, oxygen concentration (for deep water species), conductivity, or phytoplankton abundance. Some species are capable of reproducing throughout the year, while others do so during well-defined seasons, regardless of the depth they inhabit. It appears that the reproductive modalities observed in these species, most of which feed on benthic invertebrates, are probably influenced by variations in the abundance of their preferred prey (Duponchelle et al., 2000b). This also appears to be the case for species confined to rocky zones or ‘mbunas’ (Marsh et al., 1986). These authors have noted that ten Cichlidae species, sexually active throughout the year, living in rocky physical systems, had a peak of reproductive activity during the spring (August-October) and a second one in early autumn (February-March), but were less active during cold periods (May-June). The spring peak coincides with significant algal development, whereas the autumn peak corresponds to a large biomass of epilithic algae.

Influence of photoperiod on the reproductive cycle of cichlids
The influence of the photoperiod on the regulation of fish reproductive cycles is well-known in temperate systems, but it is often neglected in tropical and sub-equatorial environments where it remains relatively constant.
Yet in the artificial reservoirs of Côte d’Ivoire, the periodicity of the Oreochromis niloticus reproductive cycle is strongly correlated with the annual cycle of the length of day (figure 10.6). Analyses have shown that among the various environmental factors (precipitation, variations in water level, water temperature, chlorophyll a concentration) that are potentially involved in the regulation of the reproductive cycle of this species, photoperiod is the only factor that contributes significantly to the model.
Indeed, it accounts for 66 to 85% of the variation in reproductive seasonality, depending on the reservoir (Duponchelle et al., 1999).
A good correlation was also observed with the annual cycle of chlorophyll a concentrations (used as an indicator of trophic availability), or with water temperature.
However, in experimental conditions, with optimized regular feeding, reproductive activity ceased when the photoperiod decreased (Baroiller et al., 1997; Desprez & Mélard, 1998), even in optimal temperatures. These various results, in both natural and experimental conditions, led to the following hypothesis.
The reproductive season in Tilapias could be controlled
at two levels: an upper level, controlled by a constant astronomical factor (photoperiod) which determines the periodicity of the reproductive season, and a lower level, regulated by the combination of different environmental variables (temperature, precipitation, availability of resources, etc.) which affect the inter-annual variations of reproductive duration and intensity (Duponchelle et al., 1999).

FIGURE 10.6. Correlation between photoperiod and the reproductive seasonality of Oreochromis niloticus in small agropastoral reservoirs of Côte d’Ivoire (adapted from Duponchelle et al., 1999).

FIGURE 10.6. Correlation between photoperiod and the reproductive seasonality of Oreochromis niloticus in small agropastoral reservoirs of Côte d’Ivoire (adapted from Duponchelle et al., 1999).

40It was also highlighted in the pelagic Clupeidae of Lake Tanganyika, Limnothrissa miodon and Stolothrissa tanganicae, which reproduce throughout the year, that the main period of juvenile production coincided with the period of maximum plankton development (Coulter, 1970). It appears that the maximum spawning period varies according to the lake’s regions, in relation to zooplankton production (Chapman & Well, 1978).

41In other cases, it is difficult to provide an interpretation based on food availability. For the three Tilapias endemic to Lake Malawi, which have a well-defined reproductive period (Lowe McConnell, 1987), Oreochromis saka and O. lidole spawn in shallow waters before the beginning of rains (October-November), in the hottest period of the year. Meanwhile, O. squamipinnis spawns during the rainy season (December-February) in deeper waters.

Flexibility of the spawning season

42A species’s spawning season can change if the conditions in the physical system change as well. This is the case for instance if a dam is built on a river. Thus, Pellonula leonensis (Clupeidae) which reproduces during the dry season in the Volta River, spawns all year round in Lake Volta. Likewise, Physailia pellucida (Schilbeidae), which spawns after the rainy season in rivers, also spawns throughout the year in the lake (Reynolds, 1974).

43Other species use different strategies depending on geographic region, and thus according to climatic zones. Thus, all populations of Brycinus imberi (Alestidae) reproduce during the flood in the Zambezi (tropical regime) (Marshall & van der Heiden, 1977), while communities in Côte d’Ivoire (equatorial or subtropical regime) reproduce throughout the year with no specific synchrony with the flood regime (Paugy, 1982b). Moreover, potential fecundities differ significantly. We thus note that Ivorian individuals that may find more favourable conditions at some point in the year are individually less fecund than specimens from the Zambezi. There is thus a reproductive strategy adapted to the local conditions of the physical system.

Adaptive strategies

44In some species, it has been noted that egg size is relative to parental size. In Sarotherodon melanotheron for example, the average weight of eggs is less than 5 mg for females weighing 50 g, but can reach 20 mg for 200 g females (Peters, 1963). A significant relationship between female size and egg size has also been observed in Lake Victoria Haplochromines (Glodschmidt & Witte, 1990). On the other hand, no relationship between egg size and female size was observed in Oreochromis niloticus in Côte d’Ivoire (Duponchelle et al., 2000a), nor in 39 species of Haplochromis in Lake Malawi (Duponchelle et al., 2000b).

45The number and size of oocytes can also vary within the same species, and for individuals of the same size. Peters (1963) showed, for instance, that in laboratory conditions, a female S. melanotheron can produce either a large number of small eggs or a small number of large eggs. In the lagoons of Côte d’Ivoire, we have also noted that for females measuring 180-220 mm, the average weight of oocytes was 28 mg in the lagoon (F = 329 eggs per spawn and per female), from 15 to 18 mg in ‘acadja’ parks (F = around 500 eggs), and only 12 g in intensive-breeding fish parks (F = 726 eggs) (Legendre & Écoutin, 1989). Nonetheless, regardless of the spawning technique, the relative weight of gonads in relation to body weight remained the same, which suggests that the ovarian quantity generated during the reproductive cycle is a specific constant. That said, while this character is the least variable in Oreochromis niloticus, some inter-population and intra-population differences in successive years have been reported in Côte d’Ivoire (Duponchelle et al., 2000a).

46In Senegal, S. melanotheron populations show different reproductive and growth characteristics depending on the salinity of the estuaries they are in (Panfili et al., 2004a). In the saltiest physical systems, size at first sexual maturity is smaller, relative fecundity is higher, oocytes are smaller, and growth slower. The most marked changes are observed in hyperhaline conditions (> 60 psu). Similar observations have been made in Ethmalosa fimbriata (Panfili et al., 2004b). In this Clupeid species, the lack of genetic differentiation in the studied populations suggests that variations in reproductive and growth traits are not due to genetic isolation. Rather, they are probably phenotypic responses to marked fluctuations in salinity (panfili et al., 2004b).

Changes in cichlid reproduction induced by environment
To test if the reproductive differences observed between Oreochromis niloticus populations from various artificial reservoirs in Côte d’Ivoire were induced by the environmental conditions encountered or if they had a genetic basis, two approaches were adopted:
• Fishes from two of the monitored reservoirs in Côte d’Ivoire, whose reproductive characteristics were among the most pronounced, were captured, marked, and placed in the same environment for five months. At the end of this period, they were sacrificed and their reproductive characteristics were compared.
• At the same time, the genetic characterisation of 300 specimens from each of the two populations was carried out using microsatellite markers.
No genetic difference was found between the two populations. After sharing the same environment for five months, no difference was found in the reproductive characteristics of the two populations.
Fecundity and oocyte size, which were very significantly different in their respective physical systems, were identical in the communal pond (figure 10.7). This led to the conclusion that the differences in reproduction observed between the two populations in natural conditions were a phenotypic response to their different environments. (Duponchelle et al., 1998).

FIGURE 10.7. Reproductive characteristics (fecundity and oocyte weight) of two populations of Oreochromis niloticus females in their respective reservoirs of origin in Côte d’Ivoire (a) and after 5 months of rearing in communal pond (b) (adapted from Duponchelle et al., 1998).

FIGURE 10.7. Reproductive characteristics (fecundity and oocyte weight) of two populations of Oreochromis niloticus females in their respective reservoirs of origin in Côte d’Ivoire (a) and after 5 months of rearing in communal pond (b) (adapted from Duponchelle et al., 1998).

47In Oreochromis niloticus, marked variations in size at first sexual maturation (Duponchelle & Panfili, 1998), reproductive season (Duponchelle et al., 1999), fecundity and egg size (Duponchelle et al., 2000a) have been observed between populations of different artificial reservoirs in Côte d’Ivoire. Significant variations in reproductive characteristics have also been observed within each population over successive years. Genetic analyses and experiments in controlled environments (see box “Changes in cichlid reproduction induced by environment”) have shown that the reproductive differences observed in the reservoirs were phenotypic responses to variations in the environmental conditions encountered (Duponchelle et al., 1998).

48The remarkable plasticity of the reproduction of S. melanotheron and O. niloticus depending on environmental conditions are illustrations of adaptive strategies that can be developed by Cichlidae. While the mechanisms involved in the regulation of their reproductive cycles have not yet been identified, a few elements can serve as a starting point. Vitellogenesis takes place very rapidly in Cichlidae. In O. niloticus, for example, a full gonadal cycle lasts 27 days on average and can be as short as 15 days in non-incubating females (Tacon, 1995). The rapidity of their gonadal cycle thus allows Cichlidae to adapt their reproductive characteristics very quickly in reaction to changes in their environment (see box “Rapid phenotypic changes in Nile tilapia, Oreochromis niloticus”).

Rapid phenotypic changes in Nile tilapia, Oreochromis niloticus Duponchelle & Legendre, 2001
In Côte d’Ivoire, one of the major characteristics of certain artificial lakes is the presence of a large number of dead tree stumps.
To find out if the resulting spatial structure can influence the reproductive characteristics of Oreochromis niloticus, fishes from the same parents were placed for two months in concrete tanks with three different types of spatial organization:
• bare concrete tanks serving as controls;
• tanks with a sand-covered bottom;
• tanks with an artificial reef created from concrete blocks.
After two months, analysis of the fishes showed that Tilapias were capable of adapting to the organization of their living space by modifying their living traits. Females had lower fecundity in the tanks with reefs than in the control tanks (figure 10.8). An opposing trend was noted for egg size. These results also suggest that spatial organization, through the increase in the number of shelters and nesting sites, exercises a greater influence on Tilapia reproduction than the intrinsic quality of the spawning substrate does.
Even more significantly, it is the first time than an experiment demonstrated such a swift adaptive response in fishes. It occurred within one to two reproductive cycles, that is, around four to six weeks for the species (Duponchelle & Legendre, 2001).

FIGURE 10.8. Fecundity of Oreochromis niloticus females and mean oocyte weight depending on rearing environment. For clarity, values were equated for 100 g standard females (redrawn from Duponchelle & Legendre, 2001).

FIGURE 10.8. Fecundity of Oreochromis niloticus females and mean oocyte weight depending on rearing environment. For clarity, values were equated for 100 g standard females (redrawn from Duponchelle & Legendre, 2001).

Parental care

49Parental care refers to any assistance provided by parents whith the purpose of ensuring better survival of eggs after fecundation. Such assistance may range from the construction of nests to the brooding of eggs and fry, including egg ventilation to ensure oxygenation or to remove silt.

50Only 22% of teleost families (Blumer, 1982), especially ones with freshwater presence, provide a form of parental care. Care is more often provided by males (11% of teleost families) than by females (7% of families), while biparental care is seen in 4% of families (Sargent & Gross, 1992) (see box “Types of parental care”).

Types of parental care
Parental care in fishes takes many forms:
• guarding eggs and larvae by chasing off other fishes and predators,
• construction and maintenance of a nest,
• incubation of eggs and larvae in the mouth or gill cavity,
• transport of eggs from one site to another, generally in the mouth,
• cleaning of spawning sites,
• ventilation of eggs using fins to ensure better oxygenation and remove deposited sediments.

51It has been suggested that parental care developed in fishes occupying physical systems characterized by their spatial and temporal unpredictability (Wootton, 1990) or densely-populated systems characterized by significant predation on eggs and juveniles. Such parental care practices helped reduce hazards, but this theory ought to be backed by observations. The main function of parental care is to protect the young from predators and a wide variety of behaviours exist in substrate brooders and mouth brooders.

52Substrate brooders keep their spawn on or near a substrate and watch over their youngs until they are independent, while mouth brooders transport their youngs in the mouth from spawning until they are also independent. There are paternal, maternal, and biparental incubators (see box “Incubation period for cichlids”).

53The number of eggs that can be incubated in the mouth varies depending on species and individual size. In O. macrochir, females measuring 300-350 mm LS incubate up to 800 eggs (Marshall, 1979). Similar values have been observed in the paternal mouth brooder S. melanotheron (Legendre & Écoutin, 1989).

54The practice of parental care is fairly widespread in African fishes (Blumer, 1982) and particularly in Cichlidae (Keenleyside, 1991). Many Cichlidae substrate brooder species are widely distributed in north and west Africa, and several species are also present in Lake Tanganyika, while the majority of mouth brooder Cichlidae species (over 70% of Cichlidae) are limited to the East African great lakes (see box “Biparental mouth brooder”).

Incubation period for cichlids
Substrate brooders
The sticky eggs are deposited on a hard surface, fertilized, and hatch a few days after the two parents guard the eggs with great vigilance. Eggs are smaller but more numerous than in mouth brooders. After hatching, the young larvae are hidden in a hole or in vegetation. When they can swim freely, they remain grouped near the substrate under the watchful eye of the parents.
Mouth brooders
There are three broad categories:
maternal mouth brooders constitute the most common system. Spawning takes place on a substrate often prepared by the male, and the non-sticky eggs, released individually or in small batches, are rapidly taken into the mouth by the female. The male releases its sperm as the female collects the eggs, or fertilizes them in the mouth. The female then incubates the eggs until they are fully independent. In some cases, the female abandons them from time to time in order to feed, then takes them back again into the mouth. This behaviour has been observed in numerous Cichlidae in the East African lakes;
• paternal mouth brooders are found in only a few species. It is the case for Sarotherodon melanotheron ;
biparental mouth brooders, that is, mouth brooding by both parents from spawning until the young are independent, is also rare in Cichlidae. It has been observed in Sarotherodon galilaeus.
There are of course many variants for each type of mouth brooding.
The number of eggs that can be incubated depends on the size of the oral cavity.
For S. melanotheron, a paternal mouth brooder, the number of eggs that can be incubated in relation to the number of eggs released depends on the size of the male compared with the female. It has indeed been shown that the volume of the male’s oral cavity increases faster with the male’s size than the size of the spawn with the female’s size (Legendre & Trébaol, 1993). As a result, smaller males cannot incubate all the eggs released by a female of the same size, as they would occupy 90% of the oral cavity. Incubation is thus more successful when females pair off with larger males.

Biparental mouth brooder
Xenotilapia spiloptera is an endemic Cichlidae species in Lake Tanganyika that is a biparental mouth brooder (Konings, 1992). The species forms shoals but, as the reproductive season approaches, each shoal separates into several couples that seek a rocky habitat. The couple is established by repeated courtship between the male and the female, and then settles in a small territory about 30 cm in diameter that is defended against other fishes. A few days prior to spawning, there is a noticeable increase in mutual courtship, and a slight change in colour pattern.
At the time of spawning, the female cleans the site and deposits a few eggs. The male, which usually waits behind her about 3 cm above the substrate, then positions himself above the eggs and fertilizes them. The female waits for the male to go away then takes the fertilized eggs in her mouth.
After a brief pause during which the male chases off intruders, the female swims to the substrate again, waits for the male to position himself behind her, and releases a new series of eggs. The spawn may have 40 eggs, but it appears that quantity depends on the buccal capacity of the female.
The first nine or twelve days after spawning, the female incubates the embryos and does not eat. After this period, the larvae are transferred to the mouth of the male, which incubates them for another ten days. At the end of incubation, the male releases the fry in the territory defended by both parents. In the first few days, the youngs can seek shelter in the male’s mouth, but they generally remain on the substrate and navigate within the territory. Youngs measure around 1.5 cm at the time of release, and it takes two years for them to reach the adult size of 10 cm.

55There are a few exceptions such as Oreochromis and Sarotherodon. The paternal mouth brooder S. melanotheron is a species common to all brackish waters along the West African coast, while the biparental mouth brooder S. galilaeus is widely distributed in the north of tropical Africa.

Fishes build nests

56A form of protection for eggs and larvae is to place them in a nest that may or may not be guarded by the parents. There is a wide variety of nests in fishes. Some are constructed depressions as with Heterotis niloticus) (see box “Nest of Heterotis niloticus”), or consist of burrows, as with Protopterus annectens (see box “Nest of Protopterus annectens”). Floating nests are a fairly common form in physical systems presenting a risk of anoxia. This is the case for nests of Gymnarchus niloticus built using plants, and Mormyridae nests that are also built using plant matter (see box “Nest of Gymnarchus niloticus”). In Pollimyrus petricolus, a species endemic to Niger, males actively guard the nest built from decomposing roots and branches of Echinocloa stagnina, which allows development of micro-invertebrates that serve as food for their young. Floating foam nests are also built by Hepsetus odoe.

Nest of Heterotis niloticus
The nests of Heterotis niloticus are built in herbaceous vegetation. They are comparable to small basins measuring about 1.2 m in diameter, with a slightly excavated centre located at a depth of about 30 cm.
The bottom is bare and flat.
The compact rims are about 20 cm thick at the top and projects slightly above the water surface.
They are constructed using plant stems that have been removed from the centre of the nest. Parents stay near the nest once the eggs have been released.
Eggs are fairly small (2.5 mm in diameter) and orange.
They hatch about two days after release.
Youngs have long, dark red gill filaments that extend outside the gill cover.
They rapidly form a shoal around 30 cm in diameter occupying the centre of the nest.
On the fifth or sixth day, they leave the nest, still in a dense shoal, and under the protection of the parents (Svensson, 1933).

Nest of Protopterus annectens
The lungfish burrows into the bottom to reproduce. Nests are generally built in shallow areas covered by herbaceous plants. The most common type is a U-shaped tube with two openings, one measuring 20 to 30 cm in diameter and the other 10 to 15 cm only. The nest is about 40 cm deep and the internal diameter is identical to that of the smaller opening. It contains a larger chamber (20-30 cm in diameter) containing the eggs and larvae.
There is some variability in nest shape. Some may have three openings, while others may be constructed on the bank and have only one.
The male guards the nest until the youngs have left it. He also ventilates the water inside the nest.
It has been observed that eggs inside the nest may be of different sizes. It is thus possible that the nest may be used by more than one female, or if not, that the female has intervals in spawning.

Nest of Gymnarchus niloticus
Gymnarchus niloticus is only found in the Nilo-Sudanian zone. It can reach a length of 1.60 m, and its pale yellow spherical eggs are very large (9 to 10 mm in diameter).
In the Niger floodplain, during the flood, Gymnarchus builds a floating nest in prairies of burgu millet (Echinocloa stagnina, alsonamed hippo grass) submerged in depths of 1.5 to 2.2 m. The nest is formed from burgu millet stalks that are 2 to 3 m long, cut at the base and folded in on each other to define a cavity of about 50 cm length by 20 cm width. It forms a sort of floating pouch that measures around 150 cm by 80 cm and floats on the surface of the water. The parents enter the nest through a tunnel and deposit about a thousand eggs. Hatching is thought to take place some five days later, and the young leave the nest after 18 days, when the yolk sac has been completely resorbed. When they hatch from the egg, youngs have external gills that extend out of the brachial cavity, but they are quickly resorbed.
Parents are never far from the nest in order to protect it from other fishes, particularly lungfish which appear to have a liking for Gymnarchus eggs (Daget, 1952).

Origin and evolution of parental care

57Fishes that do not provide parental care lay a large quantity of small eggs that are dispersed in the pelagic environment. This is the case for many species belonging to the Clupeidae, Alestidae, Cypriniformes, etc. families. In running water, these eggs drift downstream and the larvae colonize fluvial annexes where they find both food and shelter. There is little chance that pelagic eggs need to withstand anoxic conditions, but they are exposed to much predation. This reproductive mode may represent ancestral conditions. Encouraging dispersion could increase the chances of survival in an unstable environment.

58Reproduction by dispersion does not encourage the development of parental care, and it is possible that during evolution, some species found it more advantageous to recognize and select sites that were conducive to spawning. This is especially true for freshwater habitats that are very temporally and spatially heterogeneous, with some of the biotopes less favourable to egg development than others. In reality, many species that do not practice parental care do not lay their eggs at random, but deposit them among plants, in gravel, in holes, etc.

59Once favourable habitat has been identified and regularly used, there may be an adaptive advantage to staying there, which is the key step in the appearance of territorial behaviour as well as site preparation and defensive behaviour, in order to monopolise the sites that are most favourable to reproduction.

60Substrate brooding, which is seen in both riverine and lacustrine species, is considered an ancestral behaviour in Cichlidae. Their very adhesive eggs, compared with those of mouth brooders, tend to support the hypothesis of a more ancient origin on the phylogenetic level.

61Mouth brooding, meanwhile, is more specialized and more recent. It probably derives from an ancestral substrate brooding behaviour in which the act of taking the eggs in the mouth to transport them from one site to another could have become a fixed component of reproductive behaviour. The adaptive advantage is that it keeps embryos away from predators and limits competition for reproductive sites. We may also see a means of escaping the need for a substrate to reproduce, when benthic space is highly sought after (Balon, 1978).

62In the beginning, among substrate brooders, parental care was probably provided by both parents. Mouth brooding evolved independently in phylogenically distant groups such as Ariidae, various Anabantidae, Arapaimidae, and Cichlidae. For the great majority of Teleosts, the male is most often the mouth brooder (Blumer, 1982). In Cichlidae, meanwhile, the female provides the care, and this investment may be an innovation in this family (Stiassny & Gerstner, 1992). In some haplochromines of Lake Malawi, it has been observed that mouth brooding can last until the young reach a length of 3 cm (Turner, 1994; Duponchelle et al., 2000b). This long period of parental care given to the young after spawning, as observed in Cichlidae, is also unusual in teleost fishes.

63A special case is that of species whose parents place the eggs in floating nests made from plants or foam (for instance Hepsetus odoe). This behaviour may be interpreted as an adaptation to development in anoxic waters (as eggs are placed on the better-oxygenated water/air interface) but may also be considered an adaptation to an unstable environment.

Table des illustrations

Fichier image/jpeg, 16k
Titre FIGURE 10.1. Distribution in oocyte diameters of two species from Côte d’Ivoire (from Albaret, 1982).
Fichier image/jpeg, 84k
Titre TABLE 10.II. Selected data on mean female size at first sexual maturity (mm), egg diameter (mm), relative fecundity expressed as number of eggs per kilogram of female body weight, maximum gonado-somatic index (GSI in %) and maximum size observed (MSO in mm) for different African fish species.
Fichier image/jpeg, 461k
Fichier image/jpeg, 505k
Fichier image/jpeg, 609k
Fichier image/jpeg, 574k
Titre TABLE 10.IV. Simplified classification of reproductive styles in fish (from Balon, 1975).
Fichier image/jpeg, 150k
Titre FIGURE 10.2. Daily counts of egg batches in Tilapia mariae (from Schwanck, 1987).
Fichier image/jpeg, 45k
Titre FIGURE 10.3. Hydrological cycle and changes in the gonado-somatic index (GSI) of different fish species from the Baoulé River, a tributary of the Upper Senegal basin. Spawning occurs before the flood peak (from Paugy, 2002).
Fichier image/jpeg, 270k
Titre FIGURE 10.4. Long breeding period of the populations of Brycinus imberi from Côte d’Ivoire. Throughout the year, more than half of fish are mature (from Paugy, 1980a).
Fichier image/jpeg, 73k
Titre FIGURE 10.5. Seasonal changes in the spawning frequency for Sarotherodon melanotheron and Tilapia guineensis in concrete tanks (redrawn from Legendre & Écoutin, 1989).
Fichier image/jpeg, 90k
Titre FIGURE 10.6. Correlation between photoperiod and the reproductive seasonality of Oreochromis niloticus in small agropastoral reservoirs of Côte d’Ivoire (adapted from Duponchelle et al., 1999).
Fichier image/jpeg, 103k
Titre FIGURE 10.7. Reproductive characteristics (fecundity and oocyte weight) of two populations of Oreochromis niloticus females in their respective reservoirs of origin in Côte d’Ivoire (a) and after 5 months of rearing in communal pond (b) (adapted from Duponchelle et al., 1998).
Fichier image/jpeg, 80k
Titre FIGURE 10.8. Fecundity of Oreochromis niloticus females and mean oocyte weight depending on rearing environment. For clarity, values were equated for 100 g standard females (redrawn from Duponchelle & Legendre, 2001).
Fichier image/jpeg, 63k
Fichier image/jpeg, 230k
Fichier image/jpeg, 209k
Fichier image/jpeg, 153k


Research director at IRD. He has published numerous works on the systematics, distribution, and ecology of West African fresh water fishes, and expanded the collections of the MNHN in Paris.

Research director at IRD. He has published numerous works on the systematics, distribution, and ecology of West African fresh water fishes, and expanded the collections of the MNHN in Paris.

© IRD Éditions, 2017

Conditions d’utilisation :


Accès exclusif

open access

Offert par L’éditeur de ce site