Juvenile olive flounder Paralichthys olivaceus achieved full compensatory growth when they were fed daily for 6 weeks after a 2-week feed deprivation period in the earlier studies (Cho, 2005; Cho et al., 2006). Therefore, applying a feeding strategy to achieve full compensatory growth is effica-cious in fish farming, particularly when fish should be starved under unfavorable conditions such as a coldwater mass, red tide or handling stress resulting from grading, transporting, and medicating. Compensatory growth has been reported in several fish species including coldwater fish (Jobling, 1988; Riley et al., 1993; Nanton et al., 2003), temperate water fish (Cho, 2005; Cho et al., 2006; Oh et al., 2008; Cho and Heo, 2011), and tropical water fish (Xie et al., 2001; Tian and Qin, 2004).

Biological indices of fish such as the hepatosomatic index (HSI) and condition factor (CF) haven been used to identify compensatory growth. HSI is a good compensatory growth index in rainbow trout Oncorhynchus mykiss, channel catfish Ictalurus punctatus) and olive flounder (Farbridge and Leath-erland, 1992; Gaylord and Gatlin, 2000; Cho, 2005). Further-more, Bavevi et al. (2010) proposed that body length (or some other measure that incorporates length, such as condition) should always be assessed when characterizing the compensa-tory growth of gilthead sea bream Sparus aurata.

A relationship between hormonal changes in fish after fast-ing or refeeding has been reported (Eales et al., 1992; MacK-enzie et al., 1998; Gaylord and Gatlin, 2000). Additionally, MacKenzie et al. (1993) and Gaylord et al. (2001) reported that increase of plasma thyroid hormone is associated with an increase in growth rate of red drum Sciaenops ocellatus and channel catfish that achieved compensatory growth. Because the quality and/or quantity of the diet largely affect the proxi-mate composition of fish, manipulation of the diet could affect thyroid hormones of fish undergoing compensatory growth.

In the present we investigated the effects of dietary nutri-ents on the biological indices and serum chemistry of olive flounder achieving compensatory growth.

Juvenile olive flounder were purchased from a private hatchery and transferred to the laboratory. Before initiation of the feeding trial, the fish were acclimated to the experimen-tal conditions for 2 weeks. In total, 450 fish averaging 16.0 g were randomly chosen and distributed into 18 180-L flow-through tanks (25 fish/tank). The flow rate in each tank was 6.5 L/min, and the water temperature was 16.0-25.5℃.

Five experimental diets were prepared: control (C), high protein (HP), high carbohydrate (HC), high lipid (HL), and intermediate protein, carbohydrate and lipid (IPCL) diets. The ingredient and nutrient composition of the diets are given in Table 1. Crude protein (48.6%) and lipid (7.3%) levels in the C diet satisfied the requirements for juvenile olive flounder (Lee et al., 2000, 2002). Protein, carbohydrate (nitrogen-free

extract), and lipid levels in the HP, HC, and HL diets were increased to 56.3%, 31.6% and 12.4%, respectively, at the ex-pense of cellulose and/or wheat flour. Finally, protein and lipid levels of the IPCL diet were increased to 53.1% and 9.2%, respectively, at the expense of cellulose and wheat flour.

Six treatments were prepared in triplicate. Fish were hand-fed with the C diet to apparent satiation twice daily (09:30 and 17:00), 6 days per week, for 8 weeks (8W-C), or fish were starved for 2 weeks and then hand-fed with the C, HP, HC, HL, or IPCL diets to apparent satiation twice per day for 6 weeks, referred to as 6W-C, 6W-HP, 6W-HC, 6W-HL, and 6W-IPCL, respectively.

Total length and body weight were measured in five ran-domly chosen fish from each tank at the end of the 8-week trial. Fish were dissected to measure liver weight to determine the biological indices: condition factor (CF) = fish weight (g)/total length (cm)³, and hepatosomatic index (HSI) = liver weight (g) × 100/fish weight (g).

Blood samples were obtained from the caudal vein of 5 ran-domly chosen 24-h starved fish from each tank by syringes. Serum was collected after centrifugation (3,000 rpm for 10 min) and stored at -70℃ in separate aliquots to analyze total protein, glucose, glutamic oxaloacetic transaminase (GOT), glutamic pyruvic transaminase (GPT), and triglyceride (TG) using automatic chemistry system (Vitros DT60 II, Vitros DTE II, DTSC II Chemistry System; Johnson and Johnson Clinical Diagnostics Inc., New York, NY, USA). Additionally, triiodothyronine (T₃) and thyroxine (T₄) hormones were also analyzed by radioimmunoassay and a gamma counter (Cobra II; Packard, Dallas, TX, USA).

A one-way analysis of variance and Duncan's multiple range test (Duncan, 1955) were performed to analyze differ-ences among the means of treatments using SAS program version 9.1 (SAS Institute, Cary, NC, USA). A P < 0.05 was considered significant.

Weight gain (%) of olive flounder in the 8W-C, 6W-HP and 6W-IPCL treatments was significantly (P < 0.05) higher than that of fish in the 6W-C treatment, but not significantly (P > 0.05) different from that of fish in the 6W-HC and 6W-HL treatments (Table 2). Poorer growth of olive flounder in the 6W-C treatment compared to that of fish in the 8W-C treat-ment contradicted results from earlier studies (Cho, 2005; Cho et al., 2006).

However, no significant difference in weight gain of all fish groups experiencing 2-week feed deprivation (6W-HP, 6W-HC, 6W-HL, and 6W-IPCL treatments), except for fish in the 6W-C treatment compared to that of fish in the 8W-C treatment indicated that supplementing diets with nutrient such as protein, carbohydrate, and lipids and their combi-nation resulted in compensatory growth. Moreover, a slight overcompensation of fish in the 6W-HP and 6W-IPCL treat-ments was observed suggesting that supplementing diets with protein and a combination of protein, carbohydrate, and lipids was effective to accelerate compensatory growth. Similarly, Cho and Heo (2011) showed that a combined HP (54.8%) and HL (14.0%) diet achieved overcompensation in juvenile olive flounder subjected to 1-week feed deprivation. Additionally, Gaylord and Gatlin (2001) showed that compensatory growth of channel catfish is primarily affected by dietary protein level rather than dietary energy level.

The CF of olive flounder in the 6W-HP, 6W-HC, and 6W-IPCL treatments was significantly (P < 0.05) higher than that

of fish in the 8W-C and 6W-C treatments, but not significantly (P > 0.05) different from that of fish in the 6W-HL treatment. This result suggests that CF could be a compensatory growth index and that fish fed the HP, HC, and IPCL diets after a 2-week food deprivation became fatter when compensatory growth was achieved. Similarly, Bavevi et al. (2010) demon-strated that gilthead sea bream Sparus aurata did not compen-sate length, but showed increased in condition. Those authors proposed that length (or some other measure that incorporates length, such as condition) should always be analyzed when characterizing compensatory growth.

The HSI of olive flounder in the 6W-HC, 6W-HL and 6W-IPCL treatments was significantly (P < 0.05) higher than that of fish in the 8W-C, 6W-C and 6W-HP treatments. In addition, the HSI of fish in the 6W-HP treatment was significantly (P < 0.05) higher than that of fish in the 8W-C and 6W-C treat-ments. The higher HSI of fish in the 6W-HP, 6W-HC, 6W-HL and 6W-IPCL treatments compared to that of fish in the 6W-C and 8W-C treatments indicated that HSI could be a compensa-tory growth index. Similarly, a high HSI is observed in fish that achieve compensatory growth (Farbridge and Leather-land, 1992; Gaylord and Gatlin, 2000; Cho, 2005).

Another reason for the high HSI in the 6W-HP, 6W-HC, 6W-HL, and 6W-IPCL treatments was probably due to the higher energy content of the diets (Table 1). This agreed with other studies showing that fish receiving higher energy diets have a higher HSI (Jobling, 1988; Nanton et al., 2003; Kjær et al., 2009). Kjær et al. (2009) demonstrated that a higher HSI in Atlantic cod (Gadus morhua) fed a high-fat diet was primarily due to a increase in liver cell size when fish were fed either a high-fat (30.5%) or low-fat (11.4%) diet for 112 days and then starved for 3 weeks.

Total protein, glucose, GOT, GPT, and T₄ were not significantly different among the treatments (Table 3). However, TG of olive flounder in the 6W-HL treatment was significantly (P < 0.05) higher than that of fish in all other treatments. Ad-ditionally, TG of fish in the 6W-IPCL treatment was signifi-cantly (P < 0.05) higher than that of fish in the 8W-C, but not significantly (P < 0.05) different from that of fish in the 6W-HP and 6W-HC treatments. The high TG in fish in the 6W-HL and 6W-IPCL treatments probably resulted from the HL (energy) content in the diets, which agreeed with Lee and Kim (2005) who showed that increased dietary lipid levels increase serum TG and total cholesterol in fish. Additionally, Kjær et al. (2009) reported that increased dietary lipid levels increase serum triacylglycerol levels in Atlantic cod.

The T₃ levels of olive flounder in the 6W-HL and 6W-IPCL treatments were significantly (P < 0.05) higher than those of fish in the 8W-C and 6W-C treatments but not significantly different from those of fish in the 6W-HP and 6W-HC treat-ments. This result partially agreed with Riley et al. (1993) and MacKenzie et al. (1998) who showed that diet quality, par-ticularly dietary protein levels largely affect thyroid hormone levels in fish, which seem to play a major role in achieving compensatory growth (Gaylord et al., 2001; Cho, 2009).

The results of this study demonstrated that CF and HSI of fish could be compensatory growth indices and that T₃ seemed to play a partial role in achieving compensatory growth in ol-ive flounder.