Aquaculture Nutrition 2000 6; 73^76 .............................................................................................. The in¯uence of food particle size on gastric emptying and growth rates of ®ngerling African cat®sh, Clarias gariepinus Burchell, 1822 M.A.R. HOSSAIN, G.S. HAYLOR & M.C.M. BEVERIDGE Institute of Aquaculture, University of Stirling, Stirling, Scotland Abstract This study investigates the e€ect of feed particle size on gastric evacuation and growth of ®ngerling African cat®sh, Clarias gariepinus. Fish (0.97 ‹ 0.07 g) held in 40-cm diameter tanks at a stocking density of 5 ®sh L±1 (25 ®sh per tank) were presented with one of four sizes of food particles (1, 1.5, 2 and 3 mm diameter). Gastric evacuation rate could be described by an exponential function with small feed particles being evacuated more rapidly: values of 0.076, 0.054, 0.047 and 0.029 were observed for ®sh fed 1 mm, 1.5 mm, 2 mm and 3 mm food particles, respectively. Growth rates were highest for ®sh fed the 1.5 and 2 mm pellets. Based on these ®ndings, recommendations are made on the optimum food particle size for C. gariepinus ®ngerlings. KEY WORDS: African cat®sh, Clarias gariepinus, feed size, gastric evacuation, growth 1987; Bromley 1994), most with the purpose of determining daily ration and food consumption. The present work is part of a larger study aimed at developing general models for evacuation and feed intake in African cat®sh based on feed particle size and energy level. Talbot & Higgins (1983) developed a radiographic X-ray method that enabled quantitative determination of stomach contents of ®sh. In some ®sh species the rate of passage of X-ray dense markers appears to di€er from that of other food components (Jùrgensen & Jobling 1988; dos Santos & Jobling 1991), but Hossain et al. (1998b) observed that inclusion of marker in diets fed to C. gariepinus had no e€ect either on ingestion or on gastric evacuation rate. The study described in this paper was designed to examine the e€ect of feed particle size on food intake, growth and gastric evacuation rate of ®ngerling Clarias gariepinus using X-ray radiography. Materials and methods Received March 1998, accepted 14 May 1999 Fish Correspondence: Mostafa Ali Reza Hossain, Fisheries Research Station, Kyoto University, Naga-Hama, Maizuru City, Kyoto 625-0086, Japan. E-mail: marhossain@yahoo.com C. gariepinus ®ngerlings, weighing 0.97 ‹ 0.7 (SE) g (n ˆ 300) were obtained from broodstock maintained at the Institute of Aquaculture, University of Stirling, Scotland. Fish had been fed unhatched, de-cysted Artemia (Argent Chemical Laboratories, Redmond, WA, USA) for 4 days from 48 h after hatching, and were then weaned over a 4-day period by substituting Artemia with a commercial trout diet, ground and sieved to 250 lm (Trouw Aquaculture, Northwich, UK). The diet was analysed according to AOAC (1990) methods (Table 1). After weaning, the ®sh were fed on the ground trout diet for a further 16 days, with feed being distributed continuously using a battery operated belt feeder (Fiap Fish Technik, GmbH, Hohenburg, Papermell, Germany; supplied by Aquatic Service (International) Ltd, Hans, England). Introduction Although it has been suggested that food particle size is an important factor governing gastric evacuation in ®sh (Jobling 1987), few data are available on the in¯uence of food particle size on feed intake (Swenson & Smith 1973; Jobling 1986, 1987, 1988). Such knowledge is a prerequisite to optimizing production of a ®sh species because the role of feed size in determining food acceptance, growth and food eciency (Wankowski & Thorpe 1979; Tabachek 1988). In recent years, an extensive literature has appeared on gastric evacuation of ®sh (FaÈnge & Grove 1979; Jobling .............................................................................................. Ó 2000 Blackwell Science Ltd 73 74 M.A.R. Hossain et al. Table 1 Composition of the test diet (Trouw UK). According to the manufacturer the dietary ingredients included cereal grains, ®sh products, oil seed products and byproducts, land-animal products, oils and fats, and minerals Table 2 Sampling schedule on day 41±43 (after feeding at 09.00 h on day 41) for African cat®sh ®ngerlings fed three di€erent diets Sampling time (h) Proximate composition Sample no. Deprivation period (h) Ingredient Manufacturer stated (g kg)1) Laboratory analysed (g kg)1) Crude oil Crude protein Crude ash Crude fibre N-free extract (by subtraction) Moisture Vitamin A Vitamin D3 Vitamin E 70 400 100 25 ö ö 10 000 IU kg)1 1000 IU kg)1 100 IU kg)1 77 426 89 30 288 90 ö ö ö 1 2 3 4 5 6 7 0 4 8 16 24 32 48 Preparation of feed marked with Ballotini The pelleted trout diet (same diet used as in Table 1) was ground to a ®ne powder in a hammer mill and Ballotini (0.16± 0.25 mm; Jencons Scienti®c, Leighton Buzzard, UK) added at a concentration of 1% w/w. After several hours mixing in a food mixer (Hobart A200; Hobart Electronic Company, Troy, OH, USA) the feed was re-pelleted in four di€erent sizes: 1, 1.5, 2 and 3 mm (California Pellet Mill, Model CL2; CPM Europe Ltd., West March, Daventry, Northants, UK), freeze-dried and stored in sealed containers at 5 °C until used. Twenty-eight samples of marked feed of known weights (0.05±1.0 g) were X-rayed to establish a relationship between feed weight and numbers of Ballotini (feed weight (g) ˆ 0.00419 + 0.00209 Ballotini; r2 ˆ 0.99). Experimental procedure Three-hundred 25-day-old ®ngerlings were randomly allocated to 12, 40-cm diameter plastic tanks with a diameter± depth ratio of 10. The stocking density was 5 ®sh L±1 or 25 ®sh per tank in a recirculation system (Hossain et al. 1998a) with a water ¯ow rate of 0.4 L min±1. Tanks were covered by black polythene to reduce light levels and a 12 h light/12 dark photoperiod (light period 0830 h±2030 h) was imposed. Water temperature was maintained at 30 ‹ 1 °C. From day 26 (from the day ®sh started feeding), ®sh were fed the marked feed to apparent satiation three times each day (at 0900, 1300 and 1700 h). Every ®fth day, following the morning feed, the weights (precision 0.01 g) of 15 ®sh taken from each treatment was determined. On day 41 (at 0900 h) the ®sh in all 12 tanks were fed to satiation with marked pellet as usual. After various depriva- Day 41 Day 42 Day 43 09.00 13.00 17.00 01.00 09.00 17.00 09.00 tion periods between 0 and 48 h (0, 4, 8, 16, 24, 32 and 48 h) (see Table 2 for detailed sampling schedule), 10 ®sh from each treatment were selected at random, anaesthetized, weighed and X-rayed. All procedures were performed on ®sh anaesthetized with a 100 mg L±1 benzocaine solution. No losses of ingested feed were observed in any ®sh before or during the X-ray operation. The stomach contents were calculated in terms of per cent body weight following the relationship between feed weight and numbers of Ballotini. The changes in the amount of feed present in the stomach over time were used to estimate gastric evacuation rate (GER). Since no X-rayed ®sh were returned to the tanks (based on the assumption that the feeding and other behavioural patterns of ®sh would be changed following anaesthesia and X-raying), on the last day of the experiment, day 45, the weights of the ®ve remaining ®sh were determined. The ®sh weight data collected every ®fth day over the experimental period were described by the exponential relationship Wt ˆ W0eGwt, where W0 is the initial ®sh weight and Wt the weight at time t and speci®c growth rate is Gw. The stomach contents of pelleted feed of di€erent pellet size from the stomach of 41-day-old ®sh after various deprivation periods were described by the equation St ˆ S0 eÿRt 1† where S0 ˆ stomach contents after ®rst feeding to satiation, St ˆ stomach contents after time t, R is the rate constant, gastric evacuation rate and t is the time in hours. X-ray protocol The method followed that described by Hossain et al. (1998b), using a Machlett Aeromax 2 X-ray apparatus and Kodak Industrex ®lm; exposure time was 2 s at 2 kV. The numbers of Ballotini on X-ray plates were counted using a binocular microscope (´40 magni®cation). .............................................................................................. Ó 2000 Blackwell Science Ltd Aquaculture Nutrition 6; 73^76 Feed and gastric evacuation Figure 1 Speci®c growth rate in ®sh fed pellets of four di€erent sizes. Error bars represent 95% con®dence limit. Figure 2 Gastric evacuation rate in Claris gariepinus ®ngerlings fed pellets of four di€erent sizes. The y-axis represents the gastric evacuation rate (±R). Error bars represent 95% con®dence limit. Statistical analyses These data are summarized in Fig. 2, evacuation rate was highest in ®sh fed 1 mm pellet diet and lowest in ®sh fed 3 mm pellet diet; however, there was no signi®cant di€erence between the ®sh fed 1.5 and 2 mm pellet diets. Ninety-®ve percent con®dence limits (CL) were calculated as, CL ˆ X ‹ t0.05 (n±1) (S/Ön); where X ˆ mean, t0.05 (n)1) =value from a two-tailed t table where 0.05 is the proportion expressing con®dence and n ± 1 is the degree of freedom and S ˆ standard deviation. The percentage body weight data were arcsine transformed and a Bartlett's test used to con®rm homogeneous variance (Sokal & Rohlf 1981). A single classi®cation ANOVA was carried out to investigate di€erence in stomach content at various deprivation periods between 0 and 48 h. Exponential regression between the deprivation time and stomach content were carried out using computer program Excel (Version 4 Windows NT, from Microsoft). Results Weights changes over time (measured every 5 day) were not signi®cantly di€erent (P < 0.05) in ®sh fed 1.5 and 2 mm pellet but were signi®cantly higher than those of ®sh fed 1 and 3 mm diets: 1 mm: Wt ˆ 1.04 ´ e0.087t, R2 ˆ 0.95, n ˆ 6, P < 0.01; 1.5 mm: Wt ˆ 1.04 ´ e0.099t, R2 ˆ 0.98, n ˆ 6, P < 0.01; 2 mm: Wt ˆ 1.12 ´ e0.099t, R2 ˆ 0.97, n ˆ 6, P < 0.01; 3 mm: Wt ˆ 0.95 ´ e0.077t, R2 ˆ 0.98, n ˆ 6, P < 0.01. Speci®c growth rates (Gw) did not di€er between the ®sh fed 1.5 and 2 mm pellets but were higher than those of ®sh fed with 1 and 3 mm pellets (Fig. 1). The relationships between stomach content and time for the four pellet sizes were: 1 mm: St ˆ 4.67 ´ e±0.077t, r2 ˆ ±0.96, n ˆ 6, P < 0.01; 1.5 mm: St ˆ 6.47 ´ e ±0.054t, r2 ˆ ±0.97, n ˆ 6, P < 0.01; 2 mm: St ˆ 6.54 ´ e±0.046t, r2 ˆ ±0.97, n ˆ 6, P < 0.01; 3 mm: St ˆ 3.89 ´ e±0.029t, r2 ˆ ±0.92, n ˆ 6, P < 0.01. .............................................................................................. Ó 2000 Blackwell Science Ltd Aquaculture Nutrition 6; 73^76 Discussion Growth rate was found to be closely related to food particle size, and the fact that the highest growth rate occurred among ®sh fed 1.5 and 2 mm pellets indicates that there is an optimum, intermediate particle size range (Fig. 1). The largest food items that ®sh can manipulate and ingest are not necessarily the most pro®table (Wanzenboeck 1995) and although large ®sh may be able to consume small particles, the net energy may be low (Pandian & Vivekanandan 1985). Thus, intermediate sized feed particles may result in greater net energy gain and promote best growth, as found here and in studies on other species, such as young Atlantic salmon (Salmo salar) (Wankowski & Thorpe 1979), Arctic charr (Salvelinus alpinus) (Tabachek 1988) and common carp (Wang et al. 1994). There is wide agreement that an exponential model can describe the evacuation of small easily digestible feed particles from the stomach (Persson 1986; Jobling 1987; Macpherson et al. 1989; Haylor 1993). In the present experiment, small feed particles were evacuated more rapidly than large Ð ®ndings that are similar to those reported previously (Swenson & Smith 1973; dos Santos & Jobling 1991). The present trials demonstrated that feeding ®sh small food particles results in faster stomach evacuation rates, and this may be expected to lead to an increase in meal frequency, even though growth may be poorer. In contrast, when the cat®sh were fed larger particles, both feed intake and growth rates were observed to be lower. 75 76 M.A.R. Hossain et al. Highest feed eciency and growth rates occurred when ®sh were fed the intermediate pellet sizes (1.5±2 mm). Acknowledgements The ®rst author thanks the British Council for funding this research and K. Ranson, W. Hamilton, A. Porter and I. Elliott for technical support. References AOAC (1990) Ocial Methods of Analysis, 15th edn. Association of Ocial Analytical Chemists, Arlington, VA. Bromley, P.J. (1994) The role of gastric evacuation experiments in quantifying the feeding rates of predatory ®sh. Rev. Fish Biol. Fish, 4, 36±66. Elliott, J.M. & Persson, L. (1978) The estimation of daily rates of food consumption for ®sh. J. Anim. Ecol.., 47, 977±991. FaÈnge, R. & Grove, D. (1979) Digestion. In: Fish Physiology, Vol. VIII (Hoar, W.S., Randall, D.J. & Brett, J.R. eds). Academic Press, London, UK. Haylor, G.S. (1993) Controlled hatchery production of Clarias gariepinus (Burchell, 1822): an estimate of maximum daily feed intake of Clarias gariepinus larvae. Aquacult. Fish. Manage., 24, 473±482. Hossain, M.A.R., Beveridge, M.C.M. & Haylor, G.S. (1998a) The e€ects of density, light and shelter on the growth and survival of African cat®sh (Clarias gariepinus Burchell, 1822) ®ngerlings. Aquaculture, 160, 251±258. Hossain, M.A.R., Haylor, G.S. & Beveridge, M.C.M. (1998b) Quantitative estimation of maximum daily feed intake of African cat®sh (Clarias gariepinus Burchell) ®ngerlings using radiography. Aquacult. Nutr., 4, 175±182. Jobling, M. (1986) Mythical models of gastric emptying and implications for food consumption studies. Env. Biol. Fish., 16, 35±50. Jobling, M. (1987) In¯uences of food particle size and dietary energy content on patterns of gastric evacuation of ®sh: test of a physiological model of gastric emptying. J. Fish Biol., 30, 299±314. Jobling, M. (1988) A review of the physiological and nutritional energetics of cod, Gadus morhua L., with particular reference to growth under farmed conditions. Aquaculture, 70, 1±19. Jùrgensen, E. & Jobling, M. (1988) Use of radiographic method in feeding studies: a cautionary note. J. Fish Biol., 32, 487±488. Macpherson, E., Henart, J. & Sanchez, P. (1989) Gastric emptying in Scyliorhynus canicula (L.): a comparison of surface-dependent and non-surface-dependent models. J. Fish Biol., 35, 37±48. Persson, L. (1986) Patterns of food evacuation in ®shes: a critical review. Env. Biol. Fish., 16, 51±58. Pandian, T.J. & Vivekanandan, E. (1985) Energetics of feeding and digestion. In: Fish Energetics: New Perspective (Tytler, P. & Calow, P. eds), pp. 99±124. Croom-Helm, Sydney, Australia. dos Santos, J. & Jobling, M. (1991) Factors a€ecting gastric evacuation in cod, Gadus morhua L., fed single meals of natural prey. J. Fish Biol., 38, 697±713. Sokal, R.R. & Rohlf, F.J. (1981) Biometry: the Principle and Practices of Statistics in Biological Research, 2nd edn. Freeman, New York, NY. Swenson, W.A. & Smith, L.L. (1973) Gastric digestion, food consumption, feeding periodicity and food conversion eciency in walleye (Stizostedion vitreum vitreum). J. Fish. Res. Brd. Can., 30, 1327±1336. Tabachek, J.L. (1988) The e€ect of feed particle size on growth and feed eciency of Arctic charr Salvelinus alpinus (L.). Aquaculture, 71, 319±330. Talbot, C. & Higgins, P.J. (1983) A radiographic method for feeding studies using metallic iron powder as a marker. J. Fish Biol., 23, 211±220. Wang, J., Pu, H., Tian, X., Zhao, D. & Zong, J. (1994) The role of food particle size in the growth of juvenile carp (Cyprinus carpio L.). J. Dalian. Fish., 9, 72±77. Wankowski, J.W.J. & Thorpe, J.E. (1979) The role of food particle size in the growth of juvenile Atlantic salmon (Salmo salar L.). J. Fish Biol., 14, 351±370. Wanzenboeck, J. (1995) Changing handling times during feeding and consequences for prey size selection of 0+ zooplanktivorous ®sh. Oecologia, 104, 372±378. .............................................................................................. Ó 2000 Blackwell Science Ltd Aquaculture Nutrition 6; 73^76