African Journal of Aquatic Science ISSN: 1608-5914 (Print) 1727-9364 (Online) Journal homepage: http://www.tandfonline.com/loi/taas20 Reproductive strategies of smooth-head catfish Clarias liocephalus (Boulenger, 1898), in the RwiziRufuha wetland system, south-western Uganda J Yatuha, J Rutaisire, L Chapman, J Kang’ombe & D Sikawa To cite this article: J Yatuha, J Rutaisire, L Chapman, J Kang’ombe & D Sikawa (2018) Reproductive strategies of smooth-head catfish Clarias�liocephalus (Boulenger, 1898), in the RwiziRufuha wetland system, south-western Uganda, African Journal of Aquatic Science, 43:2, 101-109, DOI: 10.2989/16085914.2018.1470082 To link to this article: https://doi.org/10.2989/16085914.2018.1470082 Published online: 17 Jul 2018. Submit your article to this journal View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=taas20 African Journal of Aquatic Science 2018, 43(2): 101–109 Printed in South Africa — All rights reserved Copyright © NISC (Pty) Ltd AFRICAN JOURNAL OF AQUATIC SCIENCE ISSN 1608-5914 EISSN 1727-9364 https://doi.org/10.2989/16085914.2018.1470082 Reproductive strategies of smooth-head catfish Clarias liocephalus (Boulenger, 1898), in the Rwizi-Rufuha wetland system, south-western Uganda J Yatuha1*, J Rutaisire2, L Chapman3, J Kang’ombe4 and D Sikawa4 1 Biology Department, Mbarara University of Science and Technology, Mbarara, Uganda National Agricultural Research Organization, Entebbe, Uganda 3 Department of Biology, McGill University, Montreal, QC, Canada 4 Bunda College of Agriculture, University of Malawi, Lilongwe, Malawi *Corresponding author, email: jyatuha@must.ac.ug 2 The reproduction of the smooth-head catfish (Clarias liocephalus), a heavily exploited wetland fish in Uganda, in a data-deficient fishery, was studied from January to December 2011. Analyses were based on a sample of 854 fish specimens obtained from a chain of wetlands that fringe the Rwizi River, a tributary of the Nile River. Samples were collected monthly from four sites. Fecundity, gonadosomatic index, size at sexual maturity, condition factor and growth patterns were used to describe the reproduction of the species. Mean female fecundity was 2 484.03 ± 1 289.90 (range 266–3 474). Females attained sexual maturity at a smaller size (12.0 cm TL) than males (13.79 cm TL), but were in better condition than males. Gonadosomatic index peaked in the wettest months of the study period, with highest proportion of mature ova occurring during July to November. Although C. liocephalus seems less fecund than other known clariids and is therefore prone to adverse effects from overexploitation, knowledge of its spawning periodicity and its size at sexual maturity could be useful in the management of the fishery in wetlands where it is still abundant. Keywords: condition factor, fecundity, gonadosomatic index, size at maturity Introduction The smooth-head catfish Clarias liocephalus is a small air-breathing wetland fish species that is exploited in rural environments in Uganda, for food, income generation and medicinal purposes. Increases in sales of live fish apparent from records in the office of the regional coordinator for environment and natural resource management, Mbarara, indicate that exploitation of C. liocephalus in Western Uganda is increasing. This may partly be attributed to the persistently high demand for the fish as live bait in the Nile perch fishery, as observed by Ajangale (2007). Despite such demand and exploitation, the species has not been widely studied and there is no clear management strategy to sustain its fishery. Fish utilise a variety of reproductive strategies to maximise offspring production and subsequent survival to adulthood (Murua and Saborido-Rey 2003). These strategies are species-specific and may differ among populations of a given species depending on prevailing environmental conditions (Melvin et al. 2009). The current study focused on the reproduction of C. liocephalus. It is known that information of a species’ reproductive potential, gonad development and natural spawning patterns help to define the survival and resilience of a population in space and time, and this is important for its management (Kohinoor et al. 2003; Murua et al. 2003; Rutaisire 2004). Generation of such information for C. liocephalus is therefore a requisite towards its management and conservation. Whereas a number of studies have been done on the reproductive strategies of related clariid species, especially C. gariepinus (De Graaf and Janssen 1996; Yalcin et al. 2001; El Naggar et al. 2006; Pouomogne 2008; Offem et al. 2010), little is known of the basic reproductive biology of C. liocephalus. Fecundity, gonadosomatic index (GSI), size of oocytes and size at sexual maturity were chosen as key parameters to study C. liocephalus based on their use in the reproductive potential of other fishes (Hislop et al. 1978; Morgan and Hoening 1997; Thorsen and Kjesbu 2001). Information on size and age at sexual maturity have been reported as useful in determining the appropriate size of fish to be harvested, such that an adequate proportion of a population survives to contribute to the gene pool (Law 2000). Lengthweight relationships and condition have also been reported to influence fish reproduction (Froese 2006). The current study was aimed at determining the reproductive strategies of C. liocephalus using the above key reproductive parameters for the population of C. liocephalus inhabiting the Rwizi-Rufuha wetland system in southwestern Uganda. Materials and methods Study area and sampling sites The study was conducted in the Rwizi-Rufuha wetland system in the southwestern region of Uganda (Figure 1). African Journal of Aquatic Science is co-published by NISC (Pty) Ltd and Informa UK Limited (trading as Taylor & Francis Group) Published online 17 Jul 2018 Yatuha, Rutaisire, Chapman, Kang’ombe and Sikawa 102 0 20 4° N 40 km AFRICA 0° 0° Uganda 2° N Kiruhura Greater Bushenyi UUG GAANNDDAA See enlarged area Mbarara Bush 0° Rucece LMCA But 0 32° E 30° E Isingiro 350 km 34° E 2° S Legend Ntungamo 1° S 30° E 175 31° E Sampling Point River Lake Wetland District Figure 1: Location of sampling sites () along the Rwizi-Rufuha wetland system in south-western Uganda. (Inset: map of Uganda showing the location of the study region) The wetlands that form the Rwizi-Rufuha wetland system fringe the Rwizi River, a tributary of the Nile River, whose catchment stretches over a large part of south-western Uganda. The choice of sampling sites was largely based on the availability of C. liocephalus, and this was guided by the presence of a landing and/or selling point of this fish in an area. The four selected sampling sites were: Site 1, Bush (00°32' S, 30°23' E), located in the Bushenyi area and close to the source of the river system, Site 2, But (00°43' S, 30°22' E,) in the Ntungamo area, Site 3, Rucece (00°38' S, 30°34' E,) in the Mbarara area, and Site 4, the Lake Mburo Conservation Area (LMCA) (00°41' S, 30°56' E,), located in the Lake Mburo Conservation Area where C. liocephalus fishing is regulated by the park management. Fish sampling and preparation Clarias liocephalus was identified by general visual inspection based on specific morphological features, as described by Greenwood (1966). Samples were collected once every month from each of the four study sites for a period of 12 months from January to December 2011. Local cane basket traps were exclusively used to capture the fish. The traps, similar to the ones used by all the C. liocephalus fishers in the study area, were made out of cane with two openings; a narrow entry and a wider rear opening. The size of the entry openings ranged from 4.0 to 6.5 cm diameter, the rear opening was covered with fresh grass prior to setting the trap. The size range of the trap entrance was meant to cater for the various sizes of the target species in each site. After retrieving the traps, the trapped fish were carefully removed from the trap though the wider rear entrance. The target species was sorted out, counted and immediately euthanised in lethal solutions of clove oil (2.5 ml l−1), placed on ice in an ice box, and transported to Mbarara University for analysis. Laboratory analysis In the laboratory, the specimens were individually measured and weighed. A top-loading electronic scale (Ohaus Scout SPU 202), with a 200 g capacity and 0.01 g accuracy, was used to measure body mass, whereas the total length was measured using a 30 cm graduated fish ruler. Length and weight values were recorded to the nearest 0.1 cm and 0.1 g, respectively. The specimens were subsequently dissected to expose the gonads, which were examined to ascertain the sex of the specimens and determine their sexual maturity. Specimens whose sex was not accurately determined were recorded as undifferentiated and used only in the length-weight relationship (LWR) analyses. After determining and recording the sex and stage of sexual maturity for each specimen, the gonads were carefully excised, blotted dry and weighed to the nearest 0.01 g. The rest of the viscera were removed and the eviscerated body mass measured to obtain the eviscerated weight (WE). The gonads of each specimen were individually preserved in Gilson’s fluid in labelled containers for additional analyses. The following data were recorded for each specimens: site and date of capture, total length (TL), total body mass (WT), sex, and stage of sexual maturity, gonad mass and eviscerated weight (WE). This information was used to compute the subsequent indices. Computation of length-weight relationships, growth patterns and fish condition The length-weight relationship (LWR) was determined by fitting the length weight data to a parabolic least-squares equation (i) and its logarithmic form (ii) (Wooton 1990; Offem et al. 2010) using the equations: i) WT = aTLb ii) logWT = loga + blogTL, African Journal of Aquatic Science 2018, 43(2): 101–109 where WT = fish weight (gram), TL = total fish length (cm), a = proportionality constant or intercept and b = the allometric coefficient. Both constants, a and b, were estimated by the least squares regression analysis. The Relative Condition Factor (Le Cren 1951) was used to describe the general condition of the fish. The Relative Condition Factor and was calculated using the equation Kn = WT /aTLb; where: Kn = Relative Condition Factor, WT = total body mass, a and b are constants derived from the LWR described earlier and TL = total length. A linear regression analysis was also used to describe relationships between fecundity and three variables of total length, total body mass and gonad mass. Determination of sexual maturity, fecundity, oocyte size, size at maturity and spawning season The stage of sexual maturity was based on the macroscopic morphological appearance of the gonads and classified according to a five-stage scale (Table 1) following a modification of the methods of Murua and Saborido-Rey (2003) and Yin et al. (2012). Fecundity was estimated using the gravimetric method, which is based on the relationship between ovary mass and number of oocytes in the ovary, as described by Murua and Saborido-Rey (2003), and Klibansky and Juanes (2008). To estimate the number of oocytes in the ovary, three sub-samples of ovarian tissue were taken from the anterior, middle and posterior regions of 55 mature ovaries and weighed to the nearest 0.01 g. Each subsample was spread in a petri dish to identify and count the oocytes under a dissecting microscope. Procedures for subsample extraction, preparation and ova counting followed Murua and Saborido-Rey (2003). Fecundity (F) was determined as the product of gonad (ovary) mass and oocyte density (number of oocytes per unit measure of ovarian tissue). Fecundity = WO × (Oi /Wi), where: WO = ovary weight; Oi = number of oocytes; Wi = weight of ovarian tissue sample. Relative fecundity (RF) = number of ova per unit of total length (cm) or unit of body mass (gram). The formula for relative fecundity is logF = loga + blogx i, where: F = fecundity, x i = independent variables (body mass, total length, ovary mass), a = constant, b = allometric coefficient, both of which were used to determine relative fecundity (Offem et al. 2010). Length at 50% sexual maturity (L 50) was determined by fitting a logistic curve using a logistic equation: p = 1/(1 + exp(a−b × TL), where p = proportion of mature fish in each length range (0.5 cm) 103 and a and b are constants. A logistic curve was fitted by plotting the proportion of mature fish against their lengths. The length at which 50% of the fish in a size class were mature was considered length at first maturity. Both males and females whose gonads were in macroscopic stages 3 to 5 were considered mature. The peak spawning season was estimated by analysing changes in the gonadosomatic index (GSI). The GSI was calculated by expressing the gonad mass as percentage of total body weight (WT). GSI = (WO /WT) × 100 where: WO = wet weight of ovary, WT = total wet body weight. The monthly variations in the GSI were used to estimate the spawning seasonality of C. liocephalus populations in the study area. Rainfall data was obtained from Mbarara regional Meteorological Centre, Kakoba. Statistical analysis Spearman’s correlation coefficient and linear regressions were used to quantify associations between measured variables. The chi-square goodness-of-fit test was used to determine whether there was a significant departure from the 1:1 male to female ratio in the populations of C. liocephalus. Data were entered in Excel spreadsheets and the analyses run using SPSS Inc. 17 (IBM Corp. Chicago, USA), Excel and Minitab Inc. 14 statistical software. The significance threshold was set at 0.05. Results Length-weight relationships and condition Table 2 and Figures 2, 3 and 4 summarise the findings on the length-weight relationships for C. liocephalus. Based on total body mass and total length, males were generally longer and heavier than females (Figure 2). There was a strong positive correlation between total length and total body mass (Figure 3) for the entire sample set (r = 0 .99, n = 854, p < 0.01). The b-values from regression analysis were 2.79, Table 2: Length-weight regression coefficients for male and female Clarias liocephalus collected from the Rwizi-Rufuha wetland system in January to December 2011 Category All specimens Males Females a −2.010 (0.00977237) −1.943 (0.0114025) −2.099 (0.00796159) b 2.86 2.79 2.95 r2 0.99 0.985 0.982 n 853 353 389 Table 1: Macroscopic staging of female ovary development in Clarias liocephalus from the Rwizi-Rufuha wetland system in January to December 2011 Ovary stage Immature/resting stage Maturing Ripe Ripe and running Spent Macroscopic appearance and description Ovary thin, pale pink, small, shrunken, with no macroscopically visible oocytes. Occupies less than ¼ of the abdominal cavity Ovaries in the early stage of this phase small, pear-shaped and appeared pale red with smooth walls. Small oocytes were macroscopically visible. Later in this stage the ovaries enlarge, become thick-walled and darker in colour, filling more than half of the abdominal cavity Thin ovary membrane, large pale yellow ova clearly visible, but not released on pressing the abdomen Ovaries large and fully distended to fill the abdominal cavity by up to 90%. Yellowish ova easily seen through the thin, greenish ovary membrane (tunica albuginea). Ova released on applying slight pressing to abdomen Ovary appears dark red (haemorrhagic appearance), deflated, floppy, shrunken and with very few eggs Yatuha, Rutaisire, Chapman, Kang’ombe and Sikawa 104 2.95 and 2.86 for males, females and the total sample, respectively; tending towards a negative growth form for this species. The logarithmic equations derived were as follows: i) males logW = −1.937 + 2.795 logTL (r = 0.984), ii) females logW = −2.099 + 2.949 logTL (r = 0.940) iii) males and females logW = −2.01 + 2.861 logTL (r = 0.906) and a general predictive equation of WT = 0.0098 × TL2.861. The mean relative condition factor (Kn) for the total sample (n = 854), including those whose sex was not accurately (a) 25 20 15 Reproductive strategy The key findings to describe the reproductive strategies of the species based on the fecundity, size at maturity, peak spawning season and sex ratio are summarised in Table 1 and in Figures 5–10. The size range of the spawning females was between 12 and 24 cm TL and the highest number of spawning individuals fell between 12 and 22 cm TL. The mean fecundity was 2 484.03 ± 1 289, with a minimum of 266 and maximum of 3 474 ova, and mean relative fecundity was 51.08 ova per gram of body mass. Ovary mass was positively related to the total length (Figure 5). Similarly, fecundity was positively related to total length, total body mass and ovary mass (Figure 6). 10 150 135 120 105 90 75 60 45 30 15 1.6 (b) CONDITION FACTOR (K n) BODY MASS (g) TOTAL LENGTH (cm) 30 determined, was 1.01 ± 0.12 (SD), with a minimum of 0.60 and maximum of 1.92. The mean Kn for females n = 384 and males n = 352 was 1.04 ± 0.11 (SD) and 0.97 ± 0.12 (SD), respectively, and, though males were larger than females, the females were in better condition than males (n1 = 384, n2 = 352, t = 7.14, p < 0.01; Figure 4). Female 1.4 1.2 1.0 0.8 Male Bush SEX But LMCA Rucece SITE Figure 2: Box-and-whisker plot representation of length and weight distribution of male and female Clarias liocephalus samples from Rwizi-Rufuha wetland system, January–December 2011 (n = 736) Figure 4: Mean condition factor of male and female Clarias liocephalus samples collected from the four study sites in the Rwizi-Rufuha wetland system in January–December 2011 120 10 OVARY WEIGHT (g) BODY MASS (g) SEX F M 100 80 60 40 y = 0.5643x − 5.4059 R 2 = 0.7339 8 6 4 2 20 5 10 15 20 25 TOTAL LENGTH (cm) Figure 3: Relationship between total length and body mass of Clarias liocephalus males and females obtained from the Rwizi-Rufuha wetland system from January to December 2011 5 10 15 20 TOTAL LENGTH (cm) 25 Figure 5: Relationship between ovary mass and total length of female Clarias liocephalus (n = 55) from the Rwizi-Rufuha wetland system, January–December 2011 African Journal of Aquatic Science 2018, 43(2): 101–109 105 The length at 50% maturity was calculated based on 701 specimens macroscopically staged as mature or immature (329 males and 372 females) and the estimated length at 50% maturity across the four sampling sites was 12.00 cm for females and 13.57 for males (Figure 7). Gonadosomatic index (GSI) scores ranged from 2.69% to 10.24% in females and 0.31% to 0.8% in males. The mean GSI values were highest in August and September and lowest in March for both males and females. The months (a) 3.6 LOG FECUNDITY 3.4 of February and November also registered high values of GSI (Figure 8). GSI peaked in the wettest month of the study period and was low during the dry months (Figure 9). Although all maturity stages from 1 to 5 appeared in all the samples collected throughout the year, the highest percentages of fish with mature oocytes were recorded in the months of July, August and November, (64.5%, 76.3% and 58.8%, respectively), and the lowest numbers were in January and March when 4.8% and 3.7%, respectively, were recorded (Figure 10). A significant departure from the expected 1:1 male:female sex ratio was observed only in June (p = 0.001), with a preponderance of females over males. But generally there was no significant difference between the observed and expected sex ratio (1:1) 3.2 1 (a) 3 0.9 2.8 y = 3.3085x − 0.859 R 2 = 0.7052 2.6 General 0.8 Expected Observed 0.7 0.6 p 2.4 2.2 0.5 0.4 1.1 1.2 1.3 1.4 0.3 LOG TL (cm) 0.1 (b) 3.8 0 1 (b) 0.9 Males 3.6 3.4 3.2 0.8 3 Expected Observed 0.7 0.6 2.8 y = 1.1505x + 1.4248 R 2 = 0.6654 2.6 2.4 p LOG FECUNDITY a = 8.0; b = 0.58; Size@50% = 13.79 0.2 0.5 0.4 0.3 2.2 a = 7.6; b = 0.56; Size@50% = 13.57 0.2 1 1.2 1.4 1.6 1.8 2 0.1 LOG BODY WT (g) (c) 3.6 3.4 0.8 3.2 0.7 Expected Observed 0.6 3 p LOG FECUNDITY 0 1 (c) 0.9 Females 2.8 y = 1.0211x + 2.6426 R 2 = 0.537 2.6 0.5 0.4 0.3 2.4 0.2 2.2 0.1 a = 6.6; b = 0.55; Size@50% = 12.0 0 0 0.2 0.4 0.6 0.8 1 LOG OVARY WT (g) Figure 6: Log-transformed data, depicting relationship between a) fecundity and total length, b) fecundity and total weight c) fecundity and ovary weight for Clarias liocephalus from the Rwizi-Rufuha wetland system in January–December 2011 3 6 9 12 15 18 21 24 27 TL (cm) Figure 7: Estimated size at 50% sexual maturity for a) male and female, b) male, c) female Clarias liocephalus samples collected from four study sites of the Rwizi-Rufuha wetland system in January–December 2011 Yatuha, Rutaisire, Chapman, Kang’ombe and Sikawa 106 14 Female Male 0.80 0.60 10 8 0.40 MALE GSI FEMALE GSI 12 6 0.20 4 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH Figure 8: Female (bars) and male (lines) gonadosomatic index for Clarias liocephalus from the Rwizi-Rufuha wetland system in January–December 2011 200 Female GSI Rainfall 12 150 10 8 100 6 4 50 MEAN RAINFALL (mm) MEAN FEMALE GSI 14 2 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH Figure 9: Relationship between Clarias liocephalus mean female gonadosomatic index (GSI; bars) and mean rainfall (lines) from January–December 2011 (Source of rainfall data: Mbarara Region Meteorological Centre) for the overall sample of C. liocephalus and for each of the four sites. Discussion Length-weight relationships form an important basis in the assessment and management of any given fishery (Bolger and Connolly 1989; Garcia et al. 1989; Waly et al. 2015). The allometric coefficient b value for C. liocephalus falls well within the range for finfish of 2.5 to 3.5 (Pauly and Gayinlo 1997), but because b was lower than 3 for both males and females, the species tends towards a negative allometric growth. This probably reflects the body shape of C. liocephalus, because elongate fishes reportedly exhibit negative allometric growth patterns (Froese 2006). Fish are considered to have negative allometric growth when their b-values are lower than 3, positive allometric growth if b is greater than 3, and isometric growth when b = 3 (Oso et al. 2011). These growth patterns indicate that a group of the fish is heavy, light or isometric, respectively; C. liocephalus could then be categorised as a light fish. Other Clariid species with negative allometric growth include: Clarias batrachus (2.19), Clarias macrocephalus (2.79) and C. gariepinus (2.76) (Das et al. 1997; Yusof et al. 2011). Because allometric coefficients were similar between males and females, the generated predictive equation will be useful in estimating the average mass at a given length group in any C. liocephalus population of interest. The condition factor observed in the current study presents a reference point in assessing the relative wellbeing of C. liocephalus populations in space and time. These values have been used to predict the sexual maturity, the availability of food resources, and sex of some species (Anibeze 2000) and so can also apply for C. liocephalus. The significant difference in size between males and females, with males predominantly larger than females, may be useful in guiding the managers of the fishery on the appropriate size of the fish traps, particularly during the peak spawning seasons. In so doing, the chances of capturing gravid females would be minimised. Disparity in male/female clariid catfish has also been reported by Bruton (1979), who found that male Clarias gariepinus presented higher biomass than females. However, although males were significantly larger than females, females were in better condition. Higher condition factor in females may be attributed to the physiological changes occurring in sexually mature females in preparation for spawning. These include accumulation of fat deposits and increased gonad mass in spawning females (Offem et al. 2010). Better condition in females has also been observed in other clariids of similar size to C. liocephalus. For example, in C. macromystax the condition was 1.05 for males and 1.19 for females, in C. anguillaris it was 0.64 and 1.35 and in C. nigrodigitatus, 0.718 and 1.26 for males and females, respectively (Offem et al. 2008; Offem et al. 2010). Fish collected at the LMCA site were in better condition than those from the other sites (Figure 4). This difference may partly be attributed to differences in food composition at these sites. Results of stomach contents analyses confirm that the LMCA site had a richer and more diverse prey selection than the other sites (Yatuha et al. 2012). The commonly encountered size of C. liocephalus is 20 cm (Greenwood 1966), and this is considered the optimum size at which C. liocephalus may be harvested. The fact that the majority of the collected specimens were below this optimum could be is an indication that the stock is indiscriminately exploited. This may, in the long term, negatively affect the spawning stock (Agger et al. 1974). Findings on mean fecundity seem to indicate that C. liocephalus is less fecund than other clariid catfishes, such as C. gariepinus (Gaigher 1977) and C. anguillaris (Offem et al. 2010). Although low fecundity in C. liocephalus could be a result of factors like food supply, fish size, water temperature and stress conditions, among others (Zamidi et al. 2012), the species could be inherently less fecund, and this may partly be attributed to the relatively large oocytes, compared with those of other clariids, as observed by Yatuha (2015). Low fecundity implies less offspring from an individual, but African Journal of Aquatic Science 2018, 43(2): 101–109 Spent Ripe and running 107 Ripe Ripening Mature 100 90 80 MATURITY (%) 70 60 50 40 30 20 10 Jan Mar Apr May Jun Jul Aug Sep Oct Nov Dec TIME (months) Figure 10: Monthly distribution of sexual maturity stages of Clarias liocephalus fish samples from the Rwizi-Rufuha wetland system in 2011 this can be compensated for by the large egg size, which confers adequate reserves for the survival of the young. Because a small sample size (55) of females was used to determine fecundity, results from a larger sample could give more accurate fecundity estimates for C. liocephalus. A positive correlation between fecundity, body mass and length is apparent for most clariids, such as Clarias ebriensis, C. gariepinus and Heterobranchus bidorsalis (Offem et al. 2008), and the results of the current study show that this is also the case in C. liocephalus. Surviving to sexual maturity, and the ability to contribute to the gene pool, define fitness of an individual, and such surviving individuals determine the survival of the population. For a management regime to ensure, in the face of exploitation, that a sufficient number of juveniles reach maturity; information on the size and age at first maturation is critical. Based on the findings of the current study, the mean size at maturity for male and female C. liocephalus is 12.00 and 13.57 cm TL, respectively. This provides a management clue on whether the harvested stock has had a chance to contribute to the gene pool. Clarias liocephalus females appeared to mature at a slightly smaller size than males. This is in contrast to some other small catfishes, where males mature at a smaller size. In the case of C. anguillaris, males mature at 14.8 and females at 15.7 cm, respectively (Offem et al. 2010), whereas in Chrysichthys nigrodigitatus the males and females mature at 11.5 and 16.7 cm, respectively (Offem et al. 2008). This finding may not be conclusive, because only macroscopic characterisation was used to determine the maturity of the male gonads. However, the estimate of size at which C. liocephalus attains sexual maturity is very important as an indicator of the minimum permissible capture size for sustainable utilisation of the fishery. Based on the macroscopic categorisation of the female gonads into the different developmental stages, and on the variation in GSI values throughout the study period, it appears that C. liocephalus has one major spawning event, in July to September, and two minor ones in November and February. The months with the maximum spawning activity, as shown by high GSI and large number of ripe ovaries, also happened to be the wettest, whereas the lowest were the driest months of the year of study. This implies that the species has more than one spawning event per year and also suggests that rainfall is a major factor in the breeding patterns of C. liocephalus. The trend of having the highest spawning activity during the wettest period has been observed for other tropical catfish species (see Ezenwagi 1992; Offem 2010) and for other catfishes (see Khan et al.1990; Jons and Miranda 1997; Coward and Bromage 1998; Murua and Saborido-Ray 2003). Throughout the year of study, there were almost always some specimens with mature gonads, an indication that C. liocephalus has the capacity to spawn throughout the 108 year, which may influence the management, conservation and culture of the species. Conclusions This study is the first quantitative description of the reproductive strategies of C. liocephalus. Its findings define the spawning period, fecundity and size at sexual maturity in the C. liocephalus population inhabiting the Rwizi-Rufuha wetland system. We recommend the regulation of fishing activities in the wetlands where the species is dominant, and management measures should be in place to prevent indiscriminate harvesting of this species, especially during the spawning season. Acknowledgments — We acknowledge the Regional Universities Forum for Capacity Building in Agriculture (RUFORUM) and partners for funding this research, the Bunda College of Agriculture and the Mbarara University of Science and Technology, for providing space, expertise and technical assistance for the study to succeed, the Lake Mburo National Park management and the Mbarara District Environment office for clearance and field guidance to access the wetlands and their fishers; the team involved in the C. liocephalus fishery at all the landing sites used in the study, for their cooperation. References Agger P, Bagg O, Hansen O, Hoffman E, Kesteven GL, Williams T. 1974. Manual of fisheries science. Part 2– Methods of resource investigation and their application. FAO Fisheries Technical Paper T115Rev1. FAO Fisheries and Aquaculture Management Division, Rome, Italy. Ajangale, NI. 2007. The aquaculture potential of indigenous catfish (Clarias gariepinus) in the Lake Victoria Basin, Uganda. PhD thesis, University of Stirling, UK. Anibeze CLP. 2000. Length-weight relationship and relative condition of Heterobranchus longifilis (Valencienness) from ldodo River, Nigeria. The ICLARM Quarterly 23: 34–35. Bolger T, Connolly PL. 1989. The selection of suitable indices for the measurement and analysis of fish condition. Journal of Fish Biology 34: 171–182. Bruton MN. 1979. The food and feeding behaviour of Clarias gariepinus (Pisces: Clariidae) in Lake Sibaya, South Africa, with emphasis on its role as a predator of cichlids. Transactions of the Zoological Society of London 35: 47–114. Chapman LJ. 1995. Seasonal dynamics of habitat use by an air-breathing catfish (Clarias liocephalus) in a papyrus swamp. Ecology of Freshwater Fish 4: 113–123. Coward K, Bromage NR. 1998. Histological classification of oocyte growth and the dynamics of ovarian recrudescence in Tilapia zillii. Journal of Fish Biology 53: 285–302. Das NG, Majumder AA, Sarwar SMM. 1997. Length-weight relationships and condition factor of catfish Arius tenuispinis Day, 1877. Indian Journal of Fisheries 44: 81–85. De Graaf JG, Janssen J. 1996. Handbook on the artificial reproduction and pond-rearing of the African catfish Clarias gariepinus in sub-Saharan Africa FAO Fisheries Technical Paper No. 362. Rome, Italy: Food and Agriculture Organization of the United Nations. El Naggar GO, John G, Rezk MA, Elwan W, Yehia M. 2006. Effect of varying density and water level on the spawning response of African catfish Clarias gariepinus: Implications for seed production. Aquaculture 261: 904–907. Ezenwaji H M G. 2002. The biology of Clarias ebriensis Pellegrin, Yatuha, Rutaisire, Chapman, Kang’ombe and Sikawa 1920 (Osteichthyes: Clariidae) in an African rainforest river basin. Fisheries Research 54: 235–252. Froese R. 2006. Cube law, condition factor and weight–length relationships: history, meta-analysis and recommendations. Journal of Applied Ichthyology 22: 241–253. Gaigher IG. 1977. Reproduction in the catfish (Clarias gariepinus) in the Hardap dam, South West Africa. Madoqua 10: 55–59. Garcia CB, Buarte JO, Sandovol N, Von-Schiller D, Najavas, P. 1989. Length-weight relationships of demersal fishes from the Gulf of Salamanca, Colombia. Fishbyte 21: 30–32. Greenwood PH. 1966. The fishes of Uganda. Kampala, Uganda: The Uganda Society. Hislop JRG, Robb AP, Gauld JA. 1978. Observations on the effects of feeding level on growth and reproduction in haddock, Melagogrammus aeglefinus (L.), in captivity. Journal of Fish Biology 13: 85–98. Hutchings JA, Reynolds JD. 2004. Marine fish population collapses: consequences for recovery and extinction risk. Bioscience 54: 297–309. Jons GD, Miranda LE. 1997. Ovarian weight as an index of fecundity, maturity, and spawning periodicity. Journal of Fish Biology 50: 150–156. Khan MS, Ambak MA, Ang KJ, Mohsin AKM. 1990. Reproductive biology of a tropical catfish Mystus nemurus Cuvier and Valenciennes in Chendoroh reservoir, Malaysia. Aquaculture and Fisheries Management 21: 173–179. Klibansky N, Juanes F. 2008. Procedures for efficiently producing high-quality fecundity data on a small budget. Fisheries Research 89: 84–89. Kohinoor AHM, Islam MS, Thilsted SH, Wahab MA. 2003. Reproductive biology of three indigenous small fish, mola (Amblypharyngodon mola), chela (Chela cachius) and punti (Puntius sophore). In: Wahab MA, Thilsted SH, Hoq ME (eds), Small indigenous species of fish in Bangladesh. Proceedings of BAU-ENRECA/DANIDA workshop on potentials of small indigenous species of fish in Bangladesh, 30–31 October 2002, Bangladesh Agricultural University, Mymensing, Bangladesh pp. 3–22. http://www.npajournals.com/ijab/index.php/ijab/article/ view/35. [Accessed 19 February 2010]. Law R. 2000. Fishing, selection, and phenotypic evolution. ICES Journal of Marine Science 57: 659–668. Le Cren ED. 1951. The length-weight relationship and seasonal cycle in gonad weight and condition in the perch (Perca fluviatilis). Journal of Animal Ecology 20: 201–219. Melvin GD, Stephenson RL, Power MJ. 2009. Oscillating reproductive strategies of herring in the western Atlantic in response to changing environmental conditions. ICES Journal of Marine Science 66: 1784–1792. Morgan MJ, Hoenig JM. 1997. Estimating maturity-at-age from length stratified sampling. Journal of Northwest Atlantic Fisheries Science 21: 51–63. Murua H, Saborido-Rey F. 2003. Female reproductive strategies of marine fish species of the North Atlantic. Journal of Northwest Atlantic Fishery Science 33: 23–31. Offem B, Akegbejo-Samsons Y, Omoniyi IT. 2008. Diet, size and reproductive biology of the silver catfish, Chrysichthys nigrodigitatus (Siluformes: Bagridae) in the Cross River, Nigeria. International Journal of Tropical Biology 56: 1785–1799. Offem BO, Akegbejo-Samsons Y, Omoniyi IT. 2010. Aspects of Ecology of Clarias angularis (Teleostei: Clariidae) in the Cross River, Nigeria. Turkish Journal of Fisheries and Aquatic Sciences 10: 101–110. Oso JA, Idowu EO, Fagbuaro O, Olaniran TS, Ayorinde BE. 2011. Fecundity, condition factor and gonado-somatic index of Hepsetus odoe (African pike) in a tropical reservoir, south Nigeria. World Journal of Fish and Marine Sciences 3: 112–116. Pauly D, Gayanilo Jr. FC. 1997. An alternative approach to African Journal of Aquatic Science 2018, 43(2): 101–109 estimating the parameters of a length-weight relationship from length-frequency samples and their bulk weights. Manila, Philippines: NAGA ICLARM. Pouomogne V. 2008. Capture-based aquaculture of Clarias catfish: case study of the Santchou fishers in western Cameroon. In: Lovatelli A, Holthus PF. (eds), Capture-based aquaculture. Global overview. FAO Fisheries Technical Paper. No. 508. Rome, Italy: Food and Agriculture Organization of the United Nations. Rutaisire J. 2004. The reproductive biology and artificial breeding of ningu, Labeo victorianus (Pisces: Cyprinidae). PhD thesis, Rhodes University, South Africa. Thorsen A, Kjesbu OS. 2001. A rapid method for the estimation of oocyte size and potential fecundity in Atlantic cod using computer-aided particle analysis system. Journal of Sea Research 46: 295–308. Waly N, Khady D, Ousseynou S, Papa N, Jacques P. 2015. The length-weight relationship and condition factor of white grouper (Epinephelus aeneus, Geoffroy Saint Hilaire, 1817) at the south-west coast of Senegal, West Africa. International Journal of Advanced Research 3: 145–153. Wooton JR. 1990. Ecology of teleost fishes. Chapman and Hall, London, England. 109 Yalcin S, Akyurt I, Solak K. 2001. Certain reproductive characteristics of the catfish (Clarias gariepinus Burchell, 1822) in the River Asi, Turkey. Turkish Journal of Zoology 25: 453–460. Yatuha J. 2015. Feed, feeding habits and aspects of reproductive biology of Clarias liocephalus, a potential culture species for smallholder fish farmers in southwestern Uganda. PhD thesis, University of Malawi, Malawi. Yatuha J, Kang'ombe J, Chapman L. 2013. Diet and feeding habits of the small catfish, Clarias liocephalus in wetlands of western Uganda. African Journal of Ecology 51: 385 -392. Yin J-X, Racey P, Li J, Zhang Y-G. 2012. The ovarian cycle of the fish Leptobotia elongata Bleeker, endemic to China. Pakistan Journal of Zoology 44: 997–1005. Yusof MF, Siraj SS, Daud SK. 2011. Length-weight relationships of seven catfish species in peninsular Malaysia. Journal of Fisheries and Aquatic Sciences 6: 828–833. Zamidi I, Samat A, Zaidi CC, Mazlan AG, Alam GM, Al-Amin AQ, Simon KD. 2012. Fecundity and Temporal Reproductive Cycle of Four Finger Threadfin (Eleutheronema tetradactylum) in Malaysian Coastal Water. Asian Journal of Animal and Veterinary Advances 7: 1100–1109. Manuscript received: 13 October 2017, revised 16 April 2018, accepted 23 April 2018 Associate Editor: MJ Genner