Energy Allocation and Yield and Cost of Growth in Young Esox Lucius and CoregonuS Lavaretus (Teleostei): Influence of Species, Prey Type and Body Size

We are still quite ignorant about the rules by which metabolic energy is allocated to different physiological functions in small animals. Small body size implies a high rate of maintenance metabolism with little scope for additional physiological functions such as locomotion, stress responses, defence, repair and growth (Wieser, 1991). In consequence, the energy metabolism of small animals must respond sensitively to changes in internal or external conditions, so that metabolic energy can be allocated to different functions according to priority of demand. The rules by which priority is established are unknown, but the rapidity with which switching is likely to proceed makes it mandatory to measure simultaneously energy consumption (usually as oxygen consumption) and the intensity of the function to be investigated.

By means of a ‘feeding respirometer’ we have been able to identify some of the rules governing the conversion of food energy into body substance in small fish (Wieser and Medgyesy, 1990a,b, 1991). Since in these (and probably other) small metazoans the energy expenditures for the sum of all physiological functions may exceed the aerobic maximum, individual functions are believed to substitute for each other when metabolic energy becomes limiting. Thus, uncertainty exists as to the true costs of growth, particularly during periods of embryogenesis and larval development (Rombough, 1992). In such animals a precise accounting of energy fluxes will probably never be possible since this would require the simultaneous measurement of all energy-consuming functions. However, our knowledge of the rules of energy allocation could be furthered by comparing species with different life styles and by studying the responses of organisms to changes in external and internal factors, such as food, temperature, stress, body size, etc.

The present investigation extends previous analyses of flux and allocation of metabolic energy in fish larvae (Wieser and Medgyesy, 1990a,b, 1991). Specifically, we asked to what extent the rules of energy partitioning may be influenced by (1) the life style of the fish species, and (2) the type of food offered. In order to answer the first question we compared the larvae of the white fish Coregonus lavaretus (L.), a fast-swimming predator, with those of the pike, Esox lucius (L.), a sit-and-wait predator, feeding on the nauplii of Artemia sp. To answer the second question we compared the larvae of pike feeding on Artemia with juveniles of the same species feeding on the larvae of a cyprinid species, Rutilus rutilus. An additional aim of the present investigation was to compare the two forms, or subspecies, of C. lavaretus, i.e. C. l. wartmanni (Gangfisch) and C. l. macrophthalmus (Blaufelchen), which spawn at different depths in Lake Constance and are believed by aquaculturists and fishermen to differ in their rates of growth (Ruhlé, 1986; Roesch, 1987). In a feeding experiment with dry food, Roesch and Dabrowski (1986) observed Gangfisch to grow faster than Blaufelchen, but this difference may have been due to the latter hatching from smaller eggs. At any rate, we found no differences between the two forms and shall, for the purposes of this investigation, consider them as representing a single taxon.

In conjunction with the results of previous reports, the present investigation permits the formulation of some general rules of energy allocation and growth efficiency in small fish.

Larvae and juveniles of Coregonus lavaretus (L.) and pike, Esox lucius (L.), were raised from eggs fertilized in the laboratory, the parent fish originating from Bodensee (Lake Constance; Vorarlberg) and Mondsee (Upper Austria), respectively. After hatching, the fish were maintained in running water at 15°C and fed Artemia nauplii. When the pike had reached a wet body mass of about 200 mg, they were fed larvae of the cyprinid species Rutilus rutilus (L.). A group of C. lavaretus was also raised at 10°C.

Oxygen consumption was determined on groups of fish by means of the automatic feeding respirometer described previously (Wieser and Medgyesy, 1990a,b; Medgyesy and Wieser, 1991). The experiments comparing the two fish species were performed at 15°C, but a group of C. lavaretus was also studied at 10°C. The experiments consisted of four feeding periods in the light, each lasting 15 h, interrupted and terminated by four non-feeding periods in the dark, each lasting 7h. Between these periods, the fish were removed from the respirometer for lh, and the bacterial oxygen consumption of the system was determined. When Artemia nauplii were used as food, each feeding period was again subdivided into eight 1-h periods of actual feeding, and seven 1-h periods when the nauplii were flushed out of the system, the unconsumed prey being retained by a plankton net for later counting. The approximate number of prey presented to the fish was known from the concentration of Artemia in the holding cylinder and the volume of the solution passing through the magnetic valve into the experimental chamber (which had a volume of about 0.5 dm³). At the end of a feeding period the live nauplii retained by the faeces sampler were counted and added to the number remaining in the holding cylinder. Subtracting this figure from the initial number of nauplii in the holding cylinder yielded an estimate of the number of prey consumed by the fish.

A different approach, had to be used when juvenile pike were fed larvae of R. rutilus. During each of the four 15-h feeding periods, two groups of 15–20 prey larvae were introduced into the respirometer chamber, one at 1 and one at 6h after the beginning of the feeding period. The pike preyed upon the larvae rapidly and efficiently. After about 15 min they seemed to be satiated and stopped feeding, at which time the respirometer chamber was opened, the remaining prey were removed by suction and the chamber was closed again. In this way, a nearly uninterrupted record of oxygen consumption of the pike was obtained prior to, during and after consumption of two meals of R. rutilus larvae during four feeding periods of 15h each. The mass of the prey eaten was estimated by weighing samples of R. rutilus from the same group as those used in the experiments.

The fish fed ad libitum were deprived of food for about 12 h and were then introduced into the respirometer chamber, where they were left for another 6h without food to ensure that the lowest rate of oxygen consumption measured during this period represented what, in the usual terminology, has been called the ‘basal’, ‘standard’ or ‘resting’ rate of metabolism (Brett, 1972). The fish were then removed, bacterial oxygen consumption was determined and the fish were returned to the experimental chamber.

To monitor growth, wet body mass, dry body mass and protein concentrations (measured by the method of Lowry) were determined at the beginning and at the end of each experiment. The initial values were determined on samples of fish from the same group from which the experimental animals had been taken. There was no significant difference in body water and protein concentration between fish at the beginning and at the end of the experiments (Wieser and Medgyesy, 1990b). The method of sample weighing was employed for fish up to a wet body mass of about 70 mg, which were too sensitive to be subjected to the stress of weighing. In larger fish, the initial wet body mass was determined on the experimental animals themselves by lightly anaesthetising them with benzocaine (stock solution: 1 g dm⁻³ 70% ethanol, 4 ml added to 150 ml tap water) before weighing.

Assimilation efficiency was expressed as (P+R)/C×100; P, R, and C being the energy equivalents of growth, total oxygen consumption and food consumption, respectively, during a 4-day experiment. The following energy equivalents were used for the construction of energy budgets: l μmol O₂=0.45 J; 1 mg dry body mass fish=22.7 J; 1 mg dry body mass Artemia nauplii=20 J; 1000 nauplii=42 J (Wieser and Medgyesy, 1990a).

In each experiment the following rates were distinguished: R_min, the average of the three lowest preprandial rates of oxygen consumption; R_max, the average of the three highest postprandial rates of oxygen consumption; R_rout, the average rate of oxygen consumption during an experiment, determined by dividing the total amount of oxygen consumed by the number of experimental hours; and R_f (called FIT in previous pubheations), the rate of oxygen consumption connected with feeding, i.e. R_rout—R_min. This definition assumes that all oxygen consumed above the preprandial minimum is connected with the uptake and conversion of food energy. In agreement with Jobling (1981,1983,1985), we assume that most of R_f has to do with fuelling growth but, as will be shown below, this is not the whole story.

All metabolic rates are given in μmol O₂ g⁻¹h⁻¹, based on either wet or dry body mass.

We also determined g, the relative or specific rate of growth, calculated from (ln dbm_end–ln dbm_begin)/ Δt (where dbm is dry body mass), and expressed as either% per day or mg dry body mass g⁻¹h⁻¹, where t is time.

Simple regressions were calculated for the relationship between metabolic or growth rate and body mass. The relationship between metabolic rate and growth rate, which are both subject to error, was defined by calculating Bartlett’s best fitting lines. Differences in physiological variables between groups were estimated by means of Student’s t-test.

In Table 1 the results of the present investigation are summarized and compared with relevant data from previous investigations. Although the pike larvae contained significantly more body water than the white fish larvae (columns A,C; P<0.001), the water content of the two species was fairly similar up to a wet body mass of about 1.5g (mean 85±1.5%). However, the water content of the larger juvenile pike described in Wieser and Medgyesy (1991) was very much lower (76%) than that of smaller fish. Protein content was significantly higher in pike than in white fish larvae and early juveniles (60±5.5% as against 44.4±2.2% of dry body mass).

In fish of identical size, the routine, minimum and maximum rates of oxygen consumption were 1.3-1.7 times higher in C. lavaretus than in E. lucius. However, Rf was higher in the latter species by a factor of 1.15 (12.3±1.7 as against 10.8±1.7μmol g⁻¹ h⁻¹; P=0.0166) (Table 1, columns A and C). The ratio of the highest to the lowest metabolic rates measured (R_max/R_min) ranged from 1.9 to 2.1 in C. lavaretus but from 2.4 to 3.1 in E. lucius, which may reflect the characteristic ‘sit-and-wait’ life style of the pike, already fully developed in the larvae.

As has been pointed out before (Giguere et al. 1988; Rombough, 1988; Wieser, 1991), the mass-specific rates of metabolism of fish pass through a short mass-independent phase after hatching. In C. lavaretus and E. lucius this phase lasted until the larvae had reached a wet body mass of about 10 mg (from a hatching mass of 1.5–2mg). Above 10mg, R_rout at 15°C is defined by the following scaling relationships (m in mg wet body mass; R_rout in μmol O₂g⁻¹ wet body mass h⁻¹; Fig. 1):

A comparison of these results with those of previous studies (Wieser and Medgyesy, 1990b, 1991) revealed some general rules of energy partitioning in small fish. From the basic data summarized in Table 1, we extracted the values for C, P and R and converted them to the common unit, J g⁻¹ dry body mass h⁻¹. Table 2 suggests a general pattern of energy partitioning for a wide range of experimental conditions, such as species used (C. lavaretus and E. lucius), prey (Artemia and R. rutilus), experimental temperature (10–20°C) and body size (7–3600 mg). Although E. lucius allocated a larger proportion of energy to somatic production than did C. lavaretus, the difference was small. Across all experimental groups 62±3.7% of energy was allocated to P and 38±3.7% to R.

Table 1 (columns D,E) also shows that the reduced consumption rates of the larger pike feeding on R. rutilus were partly offset by high assimilation efficiencies and that the Q₁₀ values of physiological variables in the larvae of C. lavaretus feeding on Artemia (columns A,B) ranged from 1.6 to 2.7.

Fig. 3 combines the data of the present investigation with previous data on C. lavaretus (C. wartmanni in Wieser and Medgyesy, 1990b), thus including an experiment with reduced rations. By distinguishing between two sets of experiments according to the type of prey used it can be shown that the relationship between g and R_rout (x) is defined by the following equations (based on Bartlett’s best fitting lines):

However, this generalisation has to be qualified in one important respect: in calculating the best fitting lines through the data of the Artemia experiments, the two experiments with the smallest pike larvae (1.01 mg) were excluded. The specific growth rates of these fish were nearly twice as high (9.4 and 10 mg g⁻¹ h⁻¹) as those of the next largest fish, whereas R_rout was in the range of the highest values recorded in other experiments (Fig. 3). Since two samples are not sufficient to determine the shape of the curve in the upper range, we take these samples as merely suggesting that in E. lucius R_TOut may reach an aerobic maximum at about 200–220 μmol O₂g⁻¹ dry body mass h⁻¹. In our previous investigations, an aerobic maximum for growth has been defined with regard to R_f only (Wieser and Medgyesy, 1990a,b).

By using the scaling exponents given in equations 1–5 as well as that of a lowration group of C. lavaretus studied previously (Wieser and Medgyesy, 1990b), we have also calculated R_rout, g and the cost of growth, for two standard fish weighing 10 and 100 mg dry body mass. The values summarized in Table 3 allow the following conclusions. (1) When comparing groups of similar ranges of body mass subjected to identical experimental conditions (groups A and D), the cost of growth was found to be lower in E. lucius than in C. lavaretus, savings ranging from 17 to 7% in standard fish weighing 10 and 100 mg, respectively. (2) The cost of growth was lower by 17–18% in the larger fish-fed pike than in the smaller Artemia-fed pike. (3) Between 10 and 15°C the cost of growth was more or less independent of experimental temperature in C. lavaretus. (4) In the experiments with reduced rations the lower growth rate was not accompanied by an equivalent reduction in total oxygen consumption, leading to exaggerated gross costs of growth. This is due to the increased swimming activity of small fish maintained on low rations, a phenomenon observed previously (Wieser et al. 1988). According to the rules of classic nutrition physiology, a proportional decrease in growth rate and metabolic rate would have been expected under these circumstances (Wieser, 1991).

Arecent subject of discussion has been whether in small poikilothermic animals the allocation of metabolic energy follows the rules of addition or of compensation (Wieser, 1989, 1991; Rombough, 1992). As far as the metabolic costs of growth and production are concerned, the classic approach consists of determining the relationship between growth and oxygen consumption following ingestion of food in excess of the maintenance ration (Brody, 1945; Jobling, 1985). The amount of oxygen consumed after ingestion of a meal has been called the specific dynamic action (SDH: Rubner, 1902), diet-induced thermogenesis (DIT: Rothwell and Stock, 1989), feeding-induced thermogenesis (FIT: Wieser and Medgesy, 1990a) or simply - as in the present paper - oxygen uptake rates associated with feeding (R_f: Jobling, 1985). Despite the difficulty of defining a maintenance level in most poikilothermic animals, a proportional relationship between metabolic intensity above maintenance (R_f) and growth rate (g) has been reported in several studies with fish (Hogendoom, 1983; Jobling, 1981, 1985). However, in the larvae and small juveniles of several species of freshwater fish R_f was found to become independent of g above a critical threshold (Wieser and Medgyesy, 1990a,b).

If the metabolic potential above what in this investigation has been called R_min (the lowest preprandial rate) is not sufficient to fuel the rates of growth observed, the most likely explanation is that R_min contains components of energy expenditure that can be replaced by some of the costs of fast growth. This is supported by the finding that extrapolation of the best fitting lines for Artemia-fed fish to zero growth resulted in a metabolic rate of 9.5 μmol O₂ g⁻¹ wet body mass h⁻¹, which is only half the minimum metabolic rate actually measured in C. lavaretus (Table 1). In E. lucius the difference between the extrapolated and the empirical maintenance level is much smaller (9.5 as against 11.0μmol g⁻¹h⁻¹), underscoring the conclusion that, in this species, as a result of its larger scope for growth, compensatory measures of energy allocation may have come into play only at the highest growth rates. Growth rates as high as 22–24% per day were measured in the two experiments with the smallest pike larvae (8.7 mg wet body mass), practically the same relative growth rates as those observed in pike larvae growing in small ponds at 15°C (Bry et al. 1991).

Whilst there can be no doubt that compensatory strategies of energy allocation do occur in the larvae of fish, the question remains whether these strategies affect only a fraction or all of the energy budget. Total lack of correlation between growth rate and oxygen consumption has recently been described in the embryos and yolk larvae of chinook salmon, Oncorhynchus tshawytscha, by Rombough (1992). This has also been claimed for the nematode species Caenorhabditis briggsae and for the feeding larvae of Rutilus rutilus by Wieser (1989). However, in the latter species oxygen consumption and growth were not measured simultaneously. The results of the present investigation in combination with those of a previous study (Wieser and Medgyesy, 1990b) provide only marginal support for the occurrence of compensatory strategies in the larvae and young juveniles of C. lavaretus and E. lucius if the total metabolic rate, R_rout, is taken into account. There is a hint that in E. lucius a metabolic ceiling is reached at growth rates of 6–7 mg dry body mass g⁻¹ h⁻¹ (=14–17% per day), but this suggestion is based on two 4-day experiments only (Fig. 3).

Another way of comparing the use of metabolic energy is by calculating partitioning into somatic production (P) and metabolic expenditure (R). As summarized in Table 2, E. lucius allocated slightly more energy into P than did C. lavaretus (62–67% as against 57–60%) but across all the experimental groups studied about 62% of metabolizable energy was allocated to P and 38% to R. This may be compared with the generalized energy budget for larger fish constructed by Brett and Groves (1979), in which the allocation of metabolizable energy to growth and respiration was almost exactly the opposite of our results for the larvae and young juveniles, about 40% being allocated to P and about 60% to R (see also Brafield, 1985).

The metabolic cost of production has been intensively studied in microorganisms and is usually expressed as the yield per ATP (Y_ATP) i-e-grams dry microbial mass per mole of ATP. According to Stouthamer (1973) the theoretical Y_ATP.max is about 32 g dry body mass mol⁻¹ ATP, provided that all the monomers for the synthesis of macromolecules are supplied externally. However, the yield found empirically is only about one-third of the theoretical maximum, and this discrepancy has been the subject of numerous investigations by microbiologists (Stouthamer, 1979; Tempest and Neijssel, 1984).

Since protein synthesis accounts for about 75% of the total cost of synthesis in microbial cells, and since the protein content of these cells is, on average, 50–60% of the dry mass, the values for microorganisms can be directly compared with those for the fish larvae in which the average protein content was 55% of dry body mass.

A net yield of 43 μg dry body mass μmol⁻¹ O₂ calculated for the data assembled in Fig. 3 converts to about 7 g dry body mass mol⁻¹ ATP (assuming 1mol O₂= 6 mol ATP), which is approximately 70% of the average empirical yield in microorganisms (Stouthamer and Bettenhaussen, 1973). The similarity of these values is, perhaps, not particularly surprising since it has been shown on several occasions that the cost of producing 1 g of organism is the same irrespective of the latter’s systematic position (Wieser, 1985). However, the fact that the empirical cost of growth is 3–4 times higher than the theoretical cost, in microorganisms as well as in metazoans, should puzzle zoologists as much as it has been puzzling microbiologists for some time (Stouthamer 1973, 1979).

This investigation has been supported by the ‘Fonds zur Forderung der wissenschaftlichen Forschung’ of Austria, project no. P 7804. We thank Joy Wieser for assistance with the language.

In a recent publication Kamler et al. (1992) reported net conversion efficiencies (K₂) of 55–80% in carp larvae fed zooplankton and different types of formulated diets.