Distribution and Biology of Chambo (Oreochromis spp.)
in Lakes Malawi and Malombe.

1. INTRODUCTION

The biology of chambo was studied from August 1990 to December 1991. Investigations were concentrated on those aspects most relevant to fisheries management, i.e. stock identification, distribution in relation to life history patterns, and reproductive biology.

Dietary studies were also carried out to investigate the relationship between limnological parameters and the distribution, growth and reproduction of chambo. Only preliminary results of this work are available: it will form the PhD thesis of N.C. Mwanyama.

In this work, chambo is taken as a general term for three closely related species of tilapiine cichlids, presently placed (Trewavas 1983) in the subgenus Nyasalapia of the genus Oreochromis: O.karongae (=O.saka), O.squamipinnis and O.lidole.

2. METHODS

2.1 Systematics and Identification

Systematics and identification were studied by reanalysis of morphometrics and meristics of 206 fish collected on the ODA-funded Oreochromis lidole project. To fully document the taxonomic status of chambo species, samples from outside the project area (SE Arm, Upper Shire and Lake Malombe) are required.

Measurements were made by dial callipers accurate to 0.1mm. Measurements of external body proportions taken were: standard length (SL), head length (HL), maximum body depth (BD), lower jaw length (LJ), horizontal eye diameter (ED), snout length (SN), and interorbital width (IO). The lower pharyngeal bone was dissected and measurements made of the total bone length (PL), total width (PW), blade length (PB), midline length of the toothed area (PT) and the length of the toothed area at a point halfway between the midline and the lateral extremity (Pharyngeal Arm Width- PA). The effect of size on body proportions was removed by regression of the log of each measurement on the log of standard length for each population or species. The regressions for each population or species were then averaged to produce a common-within groups regression (Thorpe 1976), and the residuals calculated. Statistical analyses were performed on residuals. As, in some cases a large number of statistical tests were performed, the possibility of type I error (i.e. finding a significant difference between populations by chance) is high. Therefore, instead of the standard 5% level, the level of significance accepted was determined by dividing 0.05 by the number of independent comparisons tested.

2.2 Distribution, Diet and Reproduction

Distribution, diet and reproductive biology were investigated in a year-long stratified programme of sampling both artisanal and commercial fisheries throughout the project area. 11,510 chambo were collected, identified, measured and weighed. Sex and gonad state were recorded for all fish above 20cm, and subsamples were taken for fecundity determination, dietary analysis and ageing through opercular bone reading. A three-day experimental trawl survey in July 1991 was used to investigate depth preferences of different sizes of fish, while a twelve-month experimental gillnetting (thrice weekly) and fyke-netting (daily) programme was used to study movements of fish through the Upper Shire River (Seisay et al. 1992a).

3. OTHER TILAPIA SPECIES

Apart from the chambo group, there are also two other species of tilapiine cichlids found in Lake Malawi: Oreochromis shiranus and Tilapia rendalli. Both are inshore species, favouring shallow swampy areas.

O.shiranus supports important fisheries in Lakes Chilwa and Chiuta, and is caught in substantial numbers in Lake Malombe. This species is extremely common in swamps, and temporary pools and streams, but it also found throughout the lake, including offshore islands such as Mumbo, and the Maleris. It is occasionally taken by pair trawlers, which fish in shallow water. O. shiranus is easily distinguished from chambo by its golden body colour, strong horizontal bars, deep caudal peduncle and four anal spines. Like two of the chambo species, ripe males are black, but the margin of the dorsal fin in O.shiranus is red, rather than white. Lowe (1952) records a size at first maturity of 17cm for females and 22cm for males. It feeds on epiphytic algae, plankton and detritus.

Tilapia rendalli is of less commercial importance: it is a deep-bodied fish of a generally reddish-brown colour, with a white belly. Mature fish are greenish with pink or reddish bellies. A prominent spot in the dorsal fin is present, even in large fishes. Unlike the Oreochromis species, there is no difference in colour between the sexes, even when breeding. It is not a mouthbrooder. Both male and female guard the eggs which are laid in a burrow, and continue parental care for some time after the fry become freeswimming. The diet consists of macrophytes and detritus.

4. CHAMBO SYSTEMATICS AND IDENTIFICATION

The three chambo species, although closely related and difficult to distinguish, differ in a number of features of their life history and must thus be considered separately, unless it can be shown that such differences are irrelevant for stock assessment and management purposes.

Oreochromis lidole was originally recognised by Trewavas (1942) by its slender lower pharyngeal bone, which supports a small toothed area. In the field, it could be easily distinguished at sizes above 15cm. Characteristics are given in Table I. Morphologically, the species shows no statistically significant variation between locations (Tables II, III, IV).

Oreochromis squamipinnis is distinguished from O.lidole and O.karongae by its distinctive male breeding colours (Table I). Morphologically, females and immature males are very similar to O.karongae: in Lake Malawi, fish below 20cm were difficult to distinguish, while in Lake Malombe, only adult breeding fish could be reliably identified. Both external body proportions and pharyngeal bone shape show significant geographical variation between populations (Tables II, III), but the number of tooth rows in the lower jaw is not variable (Table IV).

O.karongae is distinguished from O.lidole by its smaller head and mouth, and heavier dentition, and from O.squamipinnis by its male breeding colours (Table I). O.karongae shows geographic variation in external and pharyngeal bone proportions (Table III) and in the number of tooth rows in the jaws (Table IV). However, although there are many statistically significant differences, most measurements show substantial degrees of overlap- thus, although there are statistical differences between populations, no single proportion is diagnostic of any population (Table V). Note that the most variable characters are feeding structures, and it is likely that differences in the types of food eaten and in the kind of substrate the fish feed from could be responsible for this variation. In other cichlid species, if a single brood of fry is split in two and each given a different diet, similar differences in feeding structures are produced (Meyer 1987, 1990). Thus, there does not need to be a genetic basis to such differences: alternatively similar morphological differences can be genetically-based, but the two morphs may interbreed (Kornfield et al. 1982).

Note also that while the proportions of the holotypes of the two ‘species’ are very different, those of the type series generally overlap a little, but that larger samples from the type localities (Karonga and the SE Arm) show much greater overlap, while their ranges contain entirely all measurements of the types (Table V). Note also that the sample from Nankumba which is geographically intermediate, contains almost all the variation of both type localities in a single sympatric population (Table V), without any apparent reproductive isolation (Turner et al 1991b).

For these reasons, the southern populations of O.karongae, which had been believed to be a different species (Oreochromis saka- Lowe 1952, 1953; Trewavas 1983) are better regarded as geographic variants of O.karongae (Turner & Robinson 1990).

O.lidole is morphologically distinct from the other two species (Turner & Robinson 1990). On average, in the SE Arm of Lake Malawi, O. squamipinnis has a deeper body, longer jaw and smaller pharyngeal toothed area and arm width than O.karongae (Table VI). No differences were found between the two species in Lake Malombe.

Species identification was difficult and involved consideration of number of characteristics, none of which was in itself diagnostic of all specimens of each species. In addition to the characteristics listed in Table I, the arrangement of lower jaw teeth was found to be of considerable value in identification. In O.lidole, the small number of tooth rows are clearly separated by wide gaps. The first row of teeth in O.squamipinnis is usually clearly demarcated from the inner teeth which are closely-packed and usually smaller. Oreochromis karongae generally has longer teeth, which are, like the inner teeth of O.squamipinnis, closely-packed and not arranged in clear-cut rows. However, O.karongae in Lake Malombe has a tooth arrangement more like O.squamipinnis than is typical of Lake Malawi forms.

Figures are numbers of fish used for external measurements/pharyngeal bone measurements/ tooth row counts. The sample from Nankumba was taken from artisanal catches around the Nankumba Peninsula. The sample from Karonga was taken from the extreme north end of the lake between the Tanzanian border and Mulale Lagoon.

Statistical analysis: analysis of covariance on slope and elevation of log-transformed measurements regressed on logarithm of standard length. Significance level was taken as 0.1%, as 22 comparisons were performed for each species. See Table II for details of samples analysed.

Statistical Analysis: Kolmogorov-Smirnov Two Sample Test, significance level of 1% was taken, as 6 comparisons were made. No significant differences between tooth row counts of different populations were found for O.squamipinnis or O.lidole.

Statistical Analysis: Analysis of Covariance for external and pharyngeal bone measurements, Kolmogorov-Smirnov Test for distribution of tooth row counts (TR). 0.1% level was accepted as statistically significant, as 23 comparisons were made between each species in each location.

4. REPRODUCTION

4.1 Breeding Behaviour

Chambo, like other Oreochromis species, are maternal mouthbrooders. That is, females carry the eggs and fry in their mouths until the young have developed to a stage where exogenous feeding is possible.

Males construct nests on sandy or muddy bottoms and display their conspicuous colours to passing females. Eggs are laid on a central raised platform of the nest and are immediately taken into the female's mouth. The female then mouths at the male's elongate genital papilla, which is believed to have developed as an egg-mimic: fertilisation takes place inside the female's mouth. Females rapidly leave the nesting areas after spawning. Males play no part in parental care, and will begin to court another female immediately after spawning. Thus, fishing on nesting sites mainly catches ripe males: since males are polygamous, high mortality of nesting males makes little difference to total population fecundity.

4.2 Size at Maturity

As with many other tropical species with long breeding seasons, the proportion of ripe chambo caught is generally small, even at the largest size classes during the height of the breeding season. Consequently, the size at 50% maturity is taken as the size at which the percentage of ripe fish is half of the maximum for the species (leaving out size classes with small sample sizes). It was found that there were no consistent difference in sizes of ripe males and females. O. squampinnis matures at a smaller size than the other species (Table VII). Note that females of both O.squamipinnis and O.karongae are slightly larger in Lake Malombe than in the SE Arm. Thus, there is no justification for the present regulations where minimum mesh sizes for gillnets and chambo seines are smaller in Lake Malombe.

Shown are means, ranges and sample sizes of ripe fish examined

4.3 Depths of Nest Sites

In a study of chambo nesting sites at Cape Maclear (Nankumba Peninsula), Turner et al. (1991b) found that Oreochromis karongae nests at depths of 0.5 to at least 30m, while all O. lidole nests were found at depths in excess of 17m. During the present project, commercial trawl catches from as deep as 48m were found to have caught large numbers of running ripe male O.lidole.

Very few ripe male O.lidole were collected in Lake Malombe, which is nowhere deeper than 17m: it is likely that this lake is shallower than the preferred breeding depth of the species.

Lowe (1952) observed catches of ripe O.squamipinnis from as deep as 22m, but Tweddle (pers. comm.) collected this species from 8m at Karonga, and Turner et al. (1991b) recorded a territorial male from 4m in Monkey Bay. Since large numbers of ripe O.squamipinnis are caught in Lake Malombe, it appears that this species, like O.karongae, nests at a wide range of depths.

In general O.lidole builds larger nests than the other species (Turner et al. 1991b)- this may be a factor in preventing hybridisation.

4.4 Seasonality and Breeding Areas'

Oreochromis karongae breeds mainly during the hot season before the rains, from September to December (Fig. 1). In Lake Malombe, reproduction begins earlier than in Lake Malawi. In 1992, some ripe individuals were seen in Lake Malombe as early as May. Few ripe O.karongae were taken in Area A, South of Boadzulu Island. Lowe (1952) had found this to be an important breeding area for the species, and reports observing males building nests at Palm Beach. It is likely that extensive seining has eliminated the breeding grounds of this species in Area A. At present, the main breeding areas are in Lake Malombe and North of Boadzulu Island. In the North of the SE Arm, O.karongae continues to breed until March, as Turner et al. (1991b) also reported for the Nankumba Peninsula. In Lake Malombe, O.karongae ceases reproduction in December, as Lowe recorded for the population in Area A (1952).

Ripe O.lidole were mainly found in the north of the SE Arm (Fig.1), but they have also been seen in large numbers in the SW Arm and at Salima. Occasional ripe fish were taken in Lake Malombe, mainly near the beginning of the breeding season, and few in Area A. The breeding season continues until March, which is consistent with Turner et al.'s observations at Cape Maclear (1991b), but longer than previously reported (Trewavas 1983).

Reproduction in O.squamipinnis begins somewhat later than the other species, in November to December, and continues until April. This species breeds in all areas surveyed, although Turner et al. (1991b) found that it did not nest on the steeply-shelving rocky coasts of the Nankumba Peninsula.

4.5 Fecundity

Comparisons of egg production against length were made using analysis of covariance tests on logarithmically transformed data.

Oreochromis karongae from Lake Malombe and Lake Malawi did not differ significantly in their fecundities (F=0.67, 1,74 df, P=0.42), but it was found that O.squamipinnis from Lake Malombe produce more eggs than similarly-sized individuals from Lake Malawi (F=43.9, 1,66 df, P<0.001).

There were significant differences in fecundity in all comparisons between species, including those made with each of the two O.squamipinnis populations. At all sizes O.lidole has the lowest and O.squamipinnis from Lake Malombe the highest mean fecundity (Fig.2).

Length-fecundity equations are given on Table VIII

5. LIFE HISTORY AND MOVEMENTS

5.1 Size-Specific Depth Preferences

After leaving the nesting sites, females continue to brood eggs and larvae for about two weeks (Lowe 1952). When the yolk sac of the fry has been fully absorbed, brooding females move into shallow water, where they release and protect their fry. During this period, the female continues to protect them and to allow them to return to her mouth if they are threatened. During this fry-guarding stage, females and young remain in shallow water, where they are highly vulnerable to beach seining.

Large numbers of spent female O.lidole, many with fry, appear in Lake Malombe from December to April. Since ripe adults are rarely found South of Boadzulu Island, these females must have migrated from the northern part of the SE Arm, through the Shire River into Lake Malombe: a distance of at least 50km.

Fry continue to return to the females' mouths until they have reached large sizes: 24mm for O.karongae, 15mm for O.squamipinnis and 58mm for O.lidole (Lowe 1952). Thus, the fry guarding period must last for several weeks.

Observations made while snorkelling or diving indicate that chambo fry up to 7cm are generally found in large shoals at depths of less than 1m. Between 7 and 15cm, they remain in water of a few metres depth, being most abundant in the vicinity of weed beds or rocks.

Experimental trawling showed that this movement to progressively greater depths continues throughout the immature phase, with 20–22cm fish dominating catches at 40–50m. At around the minimum size at maturity (24cm), most chambo begin to return to shallower waters (Fig. 3). This is the pattern for O.squamipinnis and O.karongae. Oreochromis lidole was not caught by the experimental demersal trawl, and is rare in commercial demersal trawl catches. Immature O.lidole at sizes of 10–25cm were rarely seen, even in midwater trawl catches. The habitat preference of immature O.lidole could thus not be determined with certainty, but it seem likely that they may be in midwater.

In addition to the effect of age on depth preference, there is also a seasonal effect. In all shallow water fisheries, the catch per unit effort for chambo declines during the cold season. Underwater observations indicated that few chambo over 10cm were to be seen at depths of less than 20m from May to August.

5.2 Movements Through the Upper Shire River

Experimental gillnetting in the Upper Shire River (Seisay et al 1992) indicated that, like other shallow water areas, few chambo were to found during the cold season. There was a significant change in the direction in which fish entered the net: more fish were facing towards Lake Malombe during the earlier part of the breeding season, later most faced in the opposite direction. Thus there is evidence for a seasonal migration through the river. However, multivariate morphometric analysis indicates that more than 80% of adults of both O.squamipinnis and O.karongae from Lake Malombe could be distinguished from conspecifics in the SE Arm, arguing that the Malombe stocks of these species are largely sedentary.

Experimental fyke netting (Seisay et al. 1992a) showed that large numbers of 6–10cm juveniles are present in the river from May to July: the majority of them moving towards Lake Malombe. The timing, and their relatively large size suggests that they have moved from the drying swamps around the river, into the main river channel.

If adults from Lake Malawi have moved into Lake Malombe to spawn or, in the case of O.lidole, release fry, there should be a return migration of immature fish. This was not observed. Comparison of experimental seining at Mangochi in October 1991 and October 1973 indicated that the density of chambo in the river had dramatically declined (Seisay et al 1992), and migration patterns which formerly existed may have been disrupted by the collapse of the Lake Malombe and Upper Shire chambo fisheries.

6. DIET

Like other Oreochromis species, chambo feed mainly on algae, detritus and zooplankton. Turner et al. (1991a) found that the diet of adult (>17cmSL) and juvenile chambo at Cape Maclear were dominated by filamentous green algae (Mougeotia), but juveniles ate more attached benthic filamentous species, such as Calothrix, Cladophora and Spirogyra. Underwater observations indicated that the smallest juveniles (5–7cmSL) obtained most of their food by scraping the surfaces of rocks, switching more to surfaces of submerged plants at 8–12cm SL, but feeding on sediment and plankton at larger sizes (Turner et al. 1991a).

From samples taken from commercial trawlers in the SE Arm, adult chambo of all species were found to feed mainly on the filamentous diatom Aulacoseira (=Melosira), while diets in Lake Malombe were more varied, but included a higher proportion of Mougeotia and zooplankton (Turner et al. 1991a). There was little difference between the diets of different chambo species caught in the same place at the same time of year, suggesting that if food competition is avoided it is by means of habitat preference rather than food preference.

Turner et al. (1991a) did not quantify seasonal fluctuations in diet, nor the proportions of algae and zooplankton taken. During the present study, monthly samples of the stomach contents of O.squamipinnis were obtained from Lake Malawi and Lake Malombe. In the SE Arm of Lake Malawi, Aulacoseira was again found to be the dominant species, although Stephanodiscus, Surirella and zooplankton (mainly Diaptomus) made up 35–70% of the stomach contents in August and September.

In Lake Malombe, zooplankton (mainly Bosmina) and the large solitary diatom Surirella were found to the dominant items, zooplankton being especially important from August to December.

Comparison with plankton hauls, indicates that chambo positively select cladocerans (such as Bosmina and Diaphanosoma) and large diatoms, while taking fewer Mesocyclops (a large fast copepod) and Closterium (a small alga).

7. IMPLICATIONS OF BIOLOGY FOR MANAGEMENT OF FISHERIES

The synonymising of O.saka with O.karongae has meant that identification of adult male chambo is now relatively straightforward. However, in Lake Malombe, females and non-breeding males of O.karongae and O.squamipinnis could not be reliably distinguished. Thus separate stock assessments could not be performed on the two species in this lake (Seisay et al. 1992b). In Lake Malawi, identification of larger fish was sufficiently reliable that length-based methods could be employed to carry out an independent assessment of each species.

Although concern has been expressed that seining or trawling on nesting areas could be detrimental to chambo stocks, this seems unlikely. High mortality of males is unlikely to greatly affect population fecundity, since males play no part in parental care, and are able to mate with a large number of females during the course of a breeding season. Females spend very little time on the nesting areas, but are vulnerable to seining in shallow water nursery areas where fry are guarded and released. With both small fry and mature spent females in shallow water, chambo are particularly vulnerable to beach seining. Trawling and gillnetting are likely to be able to make substantially higher sustainable catches.

Chambo have longer breeding seasons than was previously believed. Thus there is no biological justification for the timing of a closed season. Should it be thought necessary to retain a closed season, the time of year should thus be determined by social considerations, for example, closure of fisheries immediately after the harvest (March-April) would cause less hardship than the present closed season during the rainy season, when there are often food shortages.

The size at maturity of chambo in Lake Malombe is not smaller than in Lake Malawi; Thus there is no justification for the present regulations in which smaller mesh sizes are permitted in Lake Malombe.

The size-related depth preferences of chambo mean that all species are exploited by artisanal and industrial and semi-industrial fisheries, and thus these fisheries must be considered to exploit a single stock.

Acknowledgements

Much of the sample analysis and data entry was carried out by Anderson L.Bowa, Ojinja K. Mhone, Lydia Mzunzu, Lameck M.Phiri, and Arthur Phiri. Collection and transport of artisanal fisheries samples was organised by H.Jobo Phiri and Mohammed B.D. Seisay.

References

Kornfield, I.L., Smith, D.C., Gagnon, P.S., & Taylor, J.N. (1982) The cichlid fish Cuatro Cienegas, Mexico: direct evidence of conspecificity among distinct morphs. Evolution, 36, 658–664.

Lowe, R.H. (1952) Report on the Tilapia and other fish and fisheries of Lake Nyasa. Fishery Publications of the Colonial Office 1:1–126.

Lowe, R.H. (1953) Notes on the ecology and evolution of Nyasa fishes of the genus Tilapia, with a description of T.saka Lowe. Proc. zool. Soc. Lond. 122:1041–1053.

Meyer, A. (1987) Phenotypic plasticity and heterochrony in Cichlasoma managuense (Pisces, Cichlidae) and their implications for speciation in cichlid fishes. Evolution 41, 1357–1369.

Meyer, A. (1990) Morphometrics and allometry in the trophically polymorphic cichlid fish, Cichlasoma citrinellum: alternative adaptations and ontogenic changes in shape. J. Zool. Lond. 221, 237–260.

Reid, D.M., Zaba, B.N. & Pitcher, T.J. (1990) Identification of Oreochromis species from Lake Malawi using mitochondrial DNA. Report to FAO, Rome.

Seisay, M.B.D., van Zalinge, N.P. & Turner, G.F. (1992a) Relationships between chambo stocks of Lake Malombe, the Upper Shire River and Lake Malawi. GOM/FAO/UNDP MLW/86/013 Field Document 18, July 1992.

Seisay, M.B.D., van Zalinge, N.P. & Turner, G.F. (1992b) Population dynamics and stock estimates of chambo (Oreochromis spp.) in the South-East Arm and Lake Malombe- length-based approach. GOM/FAO/UNDP MLW/86/013 Field Document 19.

Thorpe, R.S. (1976) Biometric analysis of geographic variation and racial affinities. Biol. Rev. 51: 407–452.

Trewavas, E. (1942) Nyasa fishes of the genus Tilapia and a new species from Portuguese East Africa. Ann. Mag. Nat. Hist. 11:294–306.

Trewavas, E. (1983) Tilapiine fishes of the genera Sarotherodon, Oreochromis and Danakilia. London: British Museum (Natural History). 583pp.

Turner, G.F. & Robinson, R.L. (1990) Ecology, morphology and taxonomy of the Lake Malawi Oreochromis (Nyasalapia) species flock. Ann. Mus. Roy. Afr. Centr. Sc. Zool. 262, 23–28.

Turner, G.F., Grimm, A.S., Mhone, O.K., Robinson, R.L. & Pitcher, T.J. (1991a) The diet of Oreochromis lidole (Trewavas) and other chambo species in Lakes Malawi and Malombe. J.Fish Biol. 39, 15–24.

Turner, G.F., Witimani, J., Robinson, R.L., Grimm, A.S. & Pitcher, T.J. (1991b) Reproductive isolation and the nest sites of Lake Malawi chambo, Oreochromis (Nyasalapia) spp. J.Fish Biol. 39, 775–782.

Figure 1. Breeding seasons of chambo, expressed as percentage of ripe adults taken from all gears sampled. High levels of breeding activity were recorded from all months, except April to June. Reproduction begins earlier in Lake Malombe than in Lake Malawi, but north of Boadzulu Island O.karongae has a longer, and possibly bimodal breeding season. The main breeding area for Oreochromis lidole is north of Boadzulu Island.

Figure 2. Oreochromis squamipinnis from Lake Malombe have higher fecundity than conspecifics from the south-east arm of Lake Malawi. Fecundities of Oreochromis karongae are not different in the two water bodies, and are lower than O.squamipinnis in Lake Malombe. Oreochromis lidole has lower fecundity at all sizes.

Figure 3. Experimental trawling indicated that chambo move to progressively deeper water as they grow, until, at around the size at maturity (25cm), they move back to shallower water. Data from 760 fish caught in July 1991 in the SE Arm of Lake Malawi.