Short-Term Stress Response of Juvenile Rainbow Trout Subjected to Two Different Rearing Densities ()
1. Introduction
Density is the total weight of fish per rearing unit volume during fish rearing [1]. Higher rearing densities maximize fish production, which is desirable, but once carrying capacity is exceeded, fish growth slows or stops [2]. Fish subjected to higher densities over longer time periods can also experience increased stress, impaired fish health, decreased post-stocking survival in natural environments, and exhibit undesirable behaviors [3] - [9].
Changes in rearing density usually occur slowly over time, with rearing unit biomass gradually increasing as fish grow. Prior studies have focused on longer-term density-related stress responses. Physiological stress responses in fish increase during periods of intense crowding associated with netting and movement during culture operations [10] [11] [12] [13]. However, fish may become habituated over time from repeated chronic handling stressors [14], and the short-term stress responses of fish subjected to different rearing densities have not been evaluated.
Starvation may also increase salmonid stress hormone levels [15]. Sumpter et al. [16] reported that cortisol levels in rainbow trout Oncorhynchus mykiss deprived of food initially increased but quickly returned to basal levels. Starvation also produces oxidative and other stress responses [17] [18] [19]. The effects of multiple stress inducers can lead to deleterious effects, including mortality [20]. Thus, rearing density and the presence or absence of food may interact to produce stress responses in fish.
The objective of this study was to document the stress response of fed and unfed juvenile rainbow trout subjected to a sudden change in rearing density using blood glucose, hematocrit, and other secondary stress indicators [10] [13] [21].
2. Methods
This experiment occurred at McNenny State Fish Hatchery, Spearfish, South Dakota, USA, over the course of two weeks, using de-gassed and aerated well water (11˚C; total hardness 360 mg/L CaCO3; alkalinity as CaCO3, 210 mg/L; pH 7.6, total dissolved solids 390 mg/L). Twelve 190-L semi-square tanks were used, with three tanks for each of the four treatments (n = 3). The experimental design was a 2 × 2 factorial, with two rearing densities (1.8 g/m3 or 30.1 g/m3) and two feeding regimes (fed or fasted). Thus, the treatments were: 1) Fed at a density of 1.8 g/m3, 2) Fasted at 1.8 g/m3, 3) Fed at 30.1 g/m3, and 4) Fasted at 30.1 g/m3 (Table 1). For the tanks that were fed, feed amounts used the hatchery constant method [22], with a projected feed conversion rate of 1.1 and a projected growth rate of 0.075 cm/day. These projected rates were at, or slightly above, satiation,
![]()
Table 1. Study design with 2 × 2 factorial of if fish were fed or unfed, or with low or high density.
based on prior hatchery experience. Feeding occurred every hour from 0900 to 1700 using 1.5-mm pellets (Protect FW, Skretting USA, Tooele, Utah, USA) in vibratory feeders (Pentair AES 0.5-L Vibratory Feeder, Pentair Aquatic Eco-Systems Inc., Apopka, Florida, USA).
Trout were euthanized in 200 mg/L tricaine methanesulfonate (MS-222; Tricaine-S, Syndel, Ferndale, Washington, USA), and blood was collected via caudal fin severance. Glucose (mg/dL) was obtained using a blood glucose monitor (Accu-Chek Aviva Plus; Roche Diabetic Care, Indianapolis, Indiana, USA). Heparinized microhematocrit capillary tubes (Fisher Scientific, Pittsburg, Pennsylvania, USA) were used to measure hematocrit. The tubes were centrifuged for 10 minutes at 11,500 revolutions per minute (RPM), after which red blood and total blood volume in the tube were measured, and the volume percentage of red blood cells (hematocrit) was calculated.
At the start of the experiment, glucose and hematocrit were sampled from ten rainbow trout from a common pool. An additional ten fish from the pool were euthanized, measured (total length) to the nearest 0.01 mm, and weighed to the nearest 0.01 g. Their dorsal, pectoral, and pelvic fin lengths were recorded to the nearest 0.01 mm, and viscera, spleen, and livers were weighed to the nearest 0.001 g. Immediately after this initial data collection, 300 g (approximately 25 fish) from the common pool were placed into each of the six tanks to achieve an initial density of 1.8 g/m3. Then 5000 g (approximately 418 fish) were placed into another six tanks to achieve an initial density of 30.1 g/m3.
At 4 hours, 6 hours, 48 hours (2 days), 168 hours (7 days), and 336 hours (14 days) after the start of the experiment, one fish from each tank was removed to obtain blood glucose and hematocrit. In addition, at the end of the two-week experiment (336 hours), total length, weight, dorsal, pectoral, and pelvic fin lengths, and viscera, spleen, and liver weights were obtained from one fish per tank.
The following formulas were used [23] [24]:
SPSS statistical analysis computer program (Version 24.0; IBM; Armonk, New York, USA) was used for data analysis, with significance predetermined at p < 0.05. Hematocrit and glucose data were analyzed using Repeated Measures Analysis and a Bonferroni post-hoc test. Glucose and hematocrit at each sampling point were analyzed using two-way Analysis of Variance (ANOVA). Two-way ANOVA was also used for length, weight, organosomatic indices, and relative fin lengths.
3. Results
Individual fish glucose, hematocrit, total length, weight, hepatosomatic index (HSI), viscerosomatic index (VSI), splenosomatic index (SSI), and relative fin lengths (dorsal, pectoral, pelvic) at the beginning of the experiment are listed in Table 2. Glucose levels over time were not significantly different among the density and feeding treatments (Figure 1). Hematocrit levels over time were also not significantly different among the treatments (Figure 2). Total lengths, weight, HSI, VSI, and any of the relative fin lengths were not significantly different between the low and high densities (Table 3). HSI was significantly greater in the fish that were fed compared to the unfed treatment. SSI was significantly greater in the higher density treatment. Dorsal fin relative length was also significantly shorter for the fish that were fed compared to the fish that were fasted.
![]()
Figure 1. Mean (±SE) glucosevalues over time of rainbow trout at either low (1.8 g/m3) or high (30.1 g/m3) densities and either fed or unfed.
![]()
Figure 2. Mean (±SE) hematocrita values over time of rainbow trout at either low (1.8 g/m3) or high (30.1 g/m3) densities and either fed or unfed. aHematocrit (%) = 100 × (red blood/whole blood).
![]()
Table 2. Individual fish mean (±SE) glucose, hematocrita, total length, weight, organosomatic indice [viscerosomatic index (VSI)b, splenosomatic index (SSI)c, hepatosomatic index (HSI)d], and relative fin lengths (% dorsal, pectoral, pelvic)e at the beginning of the experiment (n = 10).
aHematocrit (%) = 100 × (red blood/whole blood); bHSI = 100 × (liver weight/body weight); cVSI = 100 × (visceral weight/body weight); dSSI = 100 × (spleen weight/body weight); eDorsal = 100 × (dorsal fin length/fish length); fPectoral = 100 × (pectoral fin length/fish length); gPelvic = 100 × (pelvic fin length/fish length).
![]()
Table 3. Individual fish mean (±SE) total length, weight, organosomatic indices [hepatosomatic index (HSI)a, viscerosomatic index (VSI)b, splenosomatic index (SSI)c], and fin indices (dorsald, pectorale, pelvicf) after 336 hours of rearing at either at one of two densities and either fed or starved (unfed) (n = 3). Overall means with different letters in same row or column differ significantly (p < 0.05).
aHSI = 100 × (liver weight/body weight); bVSI = 100 × (visceral weight/body weight); cSSI = 100 × (spleen weight/body weight); dDorsal = 100 × (dorsal fin length/fish length); ePectoral = 100 × (pectoral fin length/fish length); fPelvic = 100 × (pelvic fin length/fish length).
4. Discussion
The results of this study generally indicate little short-term effects of density on stress in rainbow trout, if the fish are fed or not. Glucose and hematocrit levels typically increase following a stressful event [21], and such a response was not observed in this study. However, SSI did increase, likely indicating either an increase in hematopoietic capacity or antibody production [25]. In contrast, Iguchi et al. [5] observed a density-dependent increase in cortisol and a decrease in ayu Plecoglossus altirelis, a non-salmonid fish species.
The impacts of rearing density on SSI are uncertain. Wagner et al. [26] and Vijayan and Leatherland [27] reported no significant differences in SSI among rearing densities. However, another experiment by Wagner et al. [26] indicated a possible relationship between SSI and rearing density. It is possible that all fish in the current study were responding to a general immunological or environmental factor because all of the SSI values were higher than that previously reported for juvenile rainbow trout [25].
Similar to the current study, Papoutsoglou et al. [28] also reported no effect of density on blood glucose. However, Leatherland and Cho [29], Trenzado et al. [30], Yarahmadi et al. [31], and Wedemeyer [32] all observed increases in blood glucose with increasing rearing density. Contrarily, Vijayan and Leatherland [27] observed a decrease in blood glucose with increasing density. The difference in fish species, strains, sizes, densities, water chemistry, and rearing history may explain different results among the studies.
Hematocrit, the volume percentage of red blood cells, tends to increase when fish are under stress [21]. The lack of influence of density on hematocrit in this study supports the results reported by Leatherland and Cho [29] and Miller et al. [33]. In contrast, North et al. [34], Trenzado et al. [30], Mazur and Iwama [25], and Yarahmadi et al. [31] reported increases in hematocrit at higher fish rearing densities. Wagner et al. [26] observed inconsistent results with one experiment resulting in a density effect on hematocrit and another indicating no effect. Just as with blood glucose, differences in species, strains, sizes, densities used, and prior rearing experiences may all impact the response of hematocrit to elevated densities [10] [35].
The short-term stress results of blood glucose and hematocrit from this study were similar whether the fish were fed or not. However, feeding did influence HSI and VSI. This is not surprising. HSI is an indicator of nutritional status, indirectly measuring glycogen and carbohydrates, and VSI is an indicator of lipid shortage, being positively correlated with lipid levels [25]. Starvation, as experienced by the fasted fish in this study, will result in decreased HSI and VSI. Wagner et al. [26] and Vijayan and Leatherland [27] also reported no effect of rearing density on HSI. However, Leatherland and Cho [30] observed an increase in HSI at lower rearing densities. Similar to this study, Zahedi et al. [36] reported no effect of density on VSI.
Relative fin lengths may be long-term indicators of fish stress, and are influenced by numerous behavioral, environmental, or physiological factors [11] [26] [37]. Thus, it is not surprising that no differences, other than between fed and unfed fish and the dorsal fin, among the treatments were observed in this two-week study. Soderberg et al. [38] also reported no impact of rearing density on fin condition. The significantly shorter dorsal fin for the fish fed could be due to aggressive nipping commonly observed on the dorsal fin [39], and fish that were fed are aggressively going after the food available or diet [23] [26].
The relatively short duration of this study is obviously a limitation. In addition, the densities used in this study were generally lower than those used in other density experiments with rainbow trout [30] [31] [32] [34] [40] [41] [42] [43]. Lastly, it would have been beneficial to measure cortisol directly as recommended by Martinez-Porchas et al. [44] and Sopinka et al. [21].
5. Conclusion
In conclusion, blood glucose and hematocrit levels indicated no short-term stress response at either of the densities used in this experiment. The splenosomatic index was elevated in fish in the higher density treatments, however. Future studies should examine the stress response at higher densities, possibly identifying the limits of fish crowding during hatchery rearing.
Acknowledgements
We would like to thank Edgar Meza and Dante Bryce for their assistance in this study.