Reviews in Fish Biology and Fisheries

, 18:143

Resource allocation in yolk-feeding fish

Authors

Research paper

DOI: 10.1007/s11160-007-9070-x

Cite this article as:
Kamler, E. Rev Fish Biol Fisheries (2008) 18: 143. doi:10.1007/s11160-007-9070-x

Abstract

An insight is made into the main processes that occur in fish during endogenous feeding period. The ways in which yolk absorption rate can be measured are evaluated. Essential amino acids and polyunsaturated fatty acids are preferentially retained for incorporation into body tissue. Profound physiological and anatomical changes in yolk and a sequence of slow, fast, and a second period of slow absorption occur during the endogenous feeding period. Attempts to quantify the ontogenetic sequence are reviewed. Various methods of body size assessment are compared, and sources of bias in individual and population growth estimates are discussed. Several calorimetric methods are compared of which direct calorimetry using an oxygen bomb is the reference method. An advanced elemental analysis (CHNS) is a reliable technique that is adequate for early stages. Indices of growth potential are reviewed including a comparison of different measures, models and approaches used to estimate growth. Changes in body hydration, caloric value, content of lipids, protein, free amino acids (FAA) and minerals, and in content of RNA and DNA occur in early ontogeny. Ways to quantify metabolic rate are identified. Mean relative respiration rate of initial egg before activation is very low, about 20 mm3 g−1 h−1. Ontogenetic sequence in absolute metabolic rate of fish embryos and yolk-feeding larvae involves an increase through hatching to a peak at the time of first feeding ability, and a decrease under starvation. Models predicting the relationship between oxygen consumption and age in yolk-feeding fish are reviewed. Sequence of metabolic fuels begins with use of small molecules as carbohydrates, soon switched to FAA. Later lipids are progressively used, they provide energy for swimming activity. After yolk depletion body protein-bound amino acids are mobilised. In this review I focused on the major environmental variables as temperature, oxygen, salinity, pH, toxic xenobiotics, light, UV radiation, magnetic field and substrate, along with intrinsic factors as egg or body size, sex and genetic factors. A question was posed on how the extrinsic and intrinsic factors determine yolk absorption, growth and metabolic rates in yolk-feeding fish. Special attention is devoted to fish body size attained exclusively on yolk. A considerable variety of body size responses to temperature was found, for which several explanations are forwarded. Methodological progress made recently is characterised and the most conspicious advances in understanding of fish early life history are highlighted. Information derived from these studies can be used in management of fish populations in the field and to optimise activities in aquaculture.

Keywords

Body growthEmbryosExtrinsic factorsMetabolismYolk-feeding larvaeYolk absorption

Introduction

This review focuses on the main processes during the earliest life period of fish during which they rely exclusively on yolk. Unlike externally feeding fish, the yolk-feeding fish are energetically closed systems in which energy consumed from yolk is allocated chiefly among growth and metabolism. During this endogenous feeding period energy allocation is not clouded by conflicting demands and constraints of food foraging, egestion of faeces, locomotion, and social interactions. Rapid changes occur in yolk-feeding fish. All these characteristics make them an attractive object for studies of factors affecting energy allocation patterns.

High mortality in the early life stages of fish is well established. It averages 96.40% and 99.98%, respectively, for freshwater and marine species over the whole larval period (Houde 2002). It is generally assumed that the major causes of fish early mortality are predation, starvation, and environmental and endogenous factors (reviews by: Miller et al. 1988; Fuiman and Magurran 1994; Houde 1987, 1994, 1996; Chambers and Trippel 1997; Bunn et al. 2000). Recruitment was defined as fish survival to a particular time in ontogeny (Chambers and Trippel 1997). Myers (1997) combined data sets across 14 populations of marine fishes, and concluded that the larval period rather than the juvenile period was the source of variability in year-class strength.

Thus, in wild fish, studies of early life stages are becoming increasingly important in forecasting recruitment. For example, pelagic fish eggs and larvae are used to estimate fish stock abundance (Pepin 2002). Availability of published data for estimating reproductive potential of Northwest Atlantic stocks was reviewed by Tomkiewicz et al. (2003). Among the vital information they considered egg and larval production and viability, and environmental influences, and concluded that recently such studies have became more frequent. In cultured fish, the supply of stocking material is one of the major constraints on aquaculture development. A great many works have been involved in optimising technologies for larval rearing.

Energetic performance of fish during yolk-feeding and early exogenous feeding was considered in my book (Kamler 1992). Similar topics were covered from different perspectives by Chambers and Trippel (1997), Fuiman and Werner (2002), and Rombough (2006). However, physiological and ecological aspects of the earliest life period in which yolk is the only source of energy and matter have not been summarised yet.

The present review synthesises data on vital rates in yolk-feeding period: yolk absorption, tissue formation and metabolism. The core topic will be how fish embryos and yolk-feeding larvae allocate their energy and matter as a response to intrinsic ontogenetic processes and extrinsic variation in the environment. An energetics approach will be adopted. Evaluation of selected methods will be given with new approaches highlighted. Attempts to quantify energy and matter transformations will be shown. Studies on endogenously feeding fish in the field will be briefly summarised, and application of such studies in aquaculture will be considered. I will focus mainly on recent works that were not available when I wrote my earlier (Kamler 1992), and will point at the progress made in recent decades.

Yolk absorption (CY)

General remarks

Energy absorbed from yolk (CY) is partitioned mainly between energy invested in newly formed tissue (P) and energy expended in respiration (R) (Fig. 1). No faeces are egested prior to the onset of external feeding, but small amount of energy is excreted as nitrogenous excretion (U):
$$ {\text{C}}_{{\text{Y}}} {\text{ = P + R + U}} $$
(1)
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig1_HTML.gif
Fig. 1

Schematic representation of fish egg energy partitioning in early ontogeny with no access of external food. Fe – egg fertilisation, H – hatching, Re – final yolk resorption. The compartment “metabolism” contains both, energy expended in respiration (R) and energy of nonfaecal excretion (U). E.c. + p.f. denotes energy of egg capsules and perivitelline fluid discarded at hatching (reprinted from Jaworski and Kamler 2002, with permission from Elsevier)

Yolk is the major component of freshly fertilised fish eggs. For example, yolk diameter was about 73% of the diameter of swollen eggs in common carp (Cyprinus carpio) (Baruš et al. 2002), 76% in tench (Tinca tinca) (Geldhauser 1989), 81% in Danube bleak (Chalcalburnus chalcoides) (Smirnova 1961) and 88% in northern pike (Esox lucius) (Tański et al. 2000). Yolk was 76% of total egg wet weight in Atlantic salmon (Salmo salar) (Hamor and Garside 1977) and 83% in chinook salmon (Oncorhynchus tshawytscha) (Heming 1982). In teleosts an extraembryonic yolk sac occurs. Yolk platelets, rich in lipoproteins and phosphoproteins, and oil globules, composed primarily of triglycerides, are the main components of yolk (review by Heming and Buddington 1988). Yolk serves as a store of nutrients for embryo or yolk-feeding larva; no metabolic activity is presumed to occur in it. Apart from provisioning energy, yolk is also the source of hormones and enzymes (review by Brooks et al. 1997).

Once spawned and fertilised the oviparous fish egg is a semi-closed system, in which heat, gases and water are exchanged freely with the environment. Low-molecule chemical compounds smaller than 500–700 Da such as salts, glucose and free amino acids can penetrate through egg capsules (chorion) into the egg, but very little nutrients are taken up (Lasker and Theilacker 1962; Potts and Rudy 1969; Eddy 1974; Riis-Vestergaard 1982; Thorsen et al. 2003; reviews by Buznikov 1961; Heming and Buddington 1988; Rønnestad and Fyhn 1993). Uptake of free amino acids from sea water was shown for herring embryos (Siebers and Rosenthal 1977). Free histidine supplemented to incubation water at 0.1 mg dm−3 prior to gastrulation enhanced by 26.5% survival of Cyprinus carpio during the first month post-hatching and supressed the percentage of deformities (Vladimirov 1973). Certain amino acids (10−5 M taurine, l-glutamic acid, l-tryptophan, l-proline and l-cystine) taken up from the surrounding water significantly (P < 0.05) shortened the time from fertilisation to hatching in red sea bream (Pagrus major), whereas branched amino acids (l-valine, l-isoleucine and l-leucine) prolonged the incubation time (Takii et al. 1997b). Uptake of 14C-labelled alanine was demonstrated for turbot (Scophthalmus maximus) yolk feeding larvae with an autoradiographic method (Korsgaard 1991). Uptake of free amino acids from sea water enriched with free amino acids (5, 50 and 200 μM FAA) by yolk-sac larvae of Atlantic halibut (Hippoglossus hippoglossus) was positively related with FAA concentration, but at the low concentration found in open sea water (0.3 μM FAA) only 0.6% of the metabolic demands of the larvae could be covered by external FAA (Rønnestad 1993). The uptake of the exogenous nutrients from the surrounding water results probably from drinking, but in the field their contribution to the total energy demands of fish embryos and yolk-sac larvae is negligible (Siebers and Rosenthal 1977; Fauconneau et al. 1986; Rønnestad 1993).

Thus, in most fish species yolk is the major source of energy during endogenous feeding period. Exceptions are viviparous fishes, in which yolk is supplemented or replaced by a direct food supply, e.g., ovarian fluid or nutritional substances provided from resorption of unfertilised ova (Grodziński 1961; Wourms 1981; Boehlert et al. 1986).

Methodological approaches

Yolk absorbed (mm3 indiv.−1, mg indiv.−1or J indiv.−1) is estimated as:
$$ {\text{C}}_{{\text{Y}}} {\text{ = Y}}_{{\text{0}}} - {\text{Y}}{\text{.r}} $$
(2)
The amount of matter or energy consumed daily from yolk by an embryo or larva (absolute rate of yolk absorption, aCY, mm3 indiv.−1 day−1, mg indiv.−1day or J indiv.−1 day−1) is equivalent to food ration in exogenously feeding organisms. It is taken from:
$$ {\text{aC}}_{{\text{Y}}} {\text{ = (Y}}{\text{.r}}_{{\text{1}}} - {\text{Y}}{\text{.r}}_{{\text{2}}} {\text{)/($\tau$}}_{2} -\uptau_{{\text{1}}} {\text{)}}{\text{.}} $$
(3)
Relative rate of yolk absorption (rCY, day−1 or h−1) is quantified using equation:
$$ {\text{rC}}_{{\text{Y}}} {\text{ = (lnY}}{\text{.r}}_{{\text{1}}} - {\text{lnY}}{\text{.r}}_{{\text{2}}} {\text{)/($\tau$}}_{2} - \uptau_{{\text{1}}} {\text{)}}{\text{.}} $$
(4A)
or can be obtained from an exponential equation:
$${\text{Y}}{\text{.r($\tau$) = Y}}_{{\text{0}}} {\text{e}}^{{{\text{$\tau$(}} - {\text{rC}}_{y}) }}.$$
(4B)

In the above equations Y0 denotes initial size (volume, weight or energy content) of the yolk sac at egg activation, Y.r is the size of the remaining yolk, and τ is time post-activation (days).

The ways in which yolk size can be assessed are summarised below.
  1. (1)

    Direct weight measurements on separated tissues and formalin-hardened yolk sacs were made, for example, in Salmo salar (Gunnes 1979; Ojanguren et al. 1999), striped bass (Morone saxatilis) (Rogers and Westin 1981), rainbow trout (Oncorhynchus mykiss) (Kamler and Kato 1983), Nile tilapia (Oreochromis niloticus) (De Silva et al. 1986), sea trout (Salmo trutta) (Raciborski 1987), and 14 species of Central Amazonian fish (Araujo-Lima 1994). In nase (Chondrostoma nasus) formalin-hardened yolk was crumbly, the weight of total embryo or larva was taken before separation, the separated embryonic or larval tissues were weighed, and yolk weight was calculated by subtraction (Kamler et al. 1998). Gruber and Wieser (1983) separated Alpine charr (Salvelinus alpinus) tissues from yolk hardened by freezing. Yolk-sac larvae of Hippoglossus hippoglossus were lyophilized prior to separation (Rønnestad et al. 1993, 1995). This separation technique was applied also by Trippel (1998) to Atlantic cod (Gadus morhua) and was recommended for marine fish by Thorsen et al. (2003). The advantages of that technique are direct weight measurements and supply of materials for calorimetric and—after hardening by freezing or lyophilisation—for chemical determinations. Disadvantages include the facts that the fish need to be sacrificed, and the technique is hardly applicable for fish species producing very small eggs.

     
  2. (2)

    Indirect methods. Computing wet weight from volume is a popular technique of estimation yolk size in small eggs, similarly as it is routinely done in small planktonic invertebrates (Edmondson and Winberg 1971). Wet weight/dry weight relationship and caloric value and or chemical composition are determined at the beginning of development when yolk mass is relatively large.

     
  3. (2A)
    Yolk sac volume can be calculated using several equations, depending on the shape: ellipsoidal spherical, pyriform and conical, or cylindrical (review by Heming and Buddington 1988). In the majority of works (for example Blaxter and Hempel 1966; Huuskonen et al. 2003; Firat et al. 2003; Bonisławska et al. 2004; Klimogianni et al. 2004; Tamada and Iwata 2005; Mendiola et al. 2007) an elongated ellipsoidal shape was assumed:
    $$ {\text{V}}_{{\text{Y}}} {\text{ = (}}\Pi {\text{/6) l*h}}^{{\text{2}}} {\text{,}} $$
    (5)
    where l is the major axis, and h is the minor axis (mm). In some species yolk sac absorption is accompanied by changes of its shape. For example, in the 6th embryonic stage of Chondrostoma nasus yolk sac was spherical with a thin caudal projection, during the 7th stage it changed to a pyriform shape, in the 8th stage yolk sac assumed a simple conical form (at the end of this stage hatching occurs), and in the 9th stage yolk sac became cylindrical (Peňáz 1974). Then a combination of appropriate equations needs to be used. Decrease of yolk volume is measured at time intervals (Lasker 1962 in Pacific sardine, Sardinops caerulea; Blaxter and Hempel 1963, Klinkhardt 1986 in herring, Clupea harengus; Laurence 1969 in largemouth bass, Micropterus salmoides; Howell 1980 in yellowtail flounder, Limanda ferruginea; Spannhof and Pavlov 1984 in Oncorhynchus mykiss; Quantz 1985 and Rønnestad et al. 1992a in Scophthalmus maximus; Rønnestad et al. 1992b in lemon sole, Microstomus kitt; Fyhn and Govoni 1995 in Atlantic menhaden, Brevoortia tyrannus and spot, Leiostomus xanthurus). The volume of oil globule after a coalescence of numerous globules into one by centrifugation, and the volume of total reserves (lipoprotein yolk plus oil) were measured in vendace (Coregonus lavaretus); the volume of yolk was calculated by subtraction (Escaffre et al. 1995). The advantages of technique (2A) are that it can be applied to small yolk sacs and does not require fish to be sacrificed. A weakness of this method is that repeated measurements of the same individuals at short time intervals (e.g., daily) are limited, because yolk-sac larvae of many species poorly survive handling involved with sub-microscope measurements.
     
  4. (2B)

    Recently, a promising technique of yolk size measurement in vivo has been developed (Sarnowski 2002, 2003; Jordaan and Kling 2003; Hardy and Litvak 2004; Martell et al. 2005). In Sarnowski’s experiments Cyprinus carpio larvae were reared individually. A live larva was placed on a concave-depression microscopic slide with a small amount of water in which the larva remained immobile. The image of the larva was transmitted from under a stereoscopic microscope to a computer. The computer image analysis system MultiScan 8.08 was used. Yolk sac perimeter was drawn with a computer mouse on the image displayed on the screen; the yolk sac area could be computed automatically even for a yolk of complicated shape. Larval mortality was reduced because the handling time was short. The same individuals were measured repeatedly during 12 days from hatching until the final yolk sac resorption: a procedure that generated very smooth yolk absorption curves and assessed among-individual variability. Jordaan and Kling (2003) sampled yolk-sac larvae of Gadus morhua reared at 2, 4, 8 and 12°C according to a sampling programme based on degree-days. Photographic images of anesthetised larvae were taken and yolk sac area was calculated from them. The square root was calculated from the yolk sac area to convert it to a linear measure. Hardy and Litvak (2004) collected yolk sac length, height and area for shortnose sturgeon (Acipenser brevirostrum) and Atlantic sturgeon (A. oxyrinchus) with an image analysis system, and computed yolk utilisation rate from the slope of yolk sac area against age regression. Martell et al. (2005) took digital images of anesthetised haddock (Melanogrammus aeglefinus) larvae, and analysed yolk sac profile area to assess the effect of incubation at 2, 4, 6, 8 and 10°C on the course of yolk absorption. The notable features of the method 2B include a possibility of repeated in vivo measurements of the same individuals (as in Sarnowski 2002, 2003), and an assessment of the area of yolk of complicated and changing shape. The method provides a relative measure of yolk absorption rate which is useful for evaluation effects of, for example, temperature (Jordaan and Kling 2003; Martell et al. 2005) or xenobiotics (Sarnowski 2003). The weakness is that the changes in yolk sac area are assumed to be representative of changes in amount of nutrient reserves.

     
However, the common limitations of the indirect methods (2A and 2B) are that yolk specific gravity is assumed to be 1, and hydration and chemical composition/caloric value of yolk dry matter are assumed to be constant. Heming and Buddington (1988) suggested that yolk hydration does not remain constant. In Oncorhynchus mykiss yolk hydration increased significantly between hatching and swimming-up from 49% to 53–59% (Escaffre and Bergot 1984), in Hippoglossus hippoglossus rate of water loss form yolk was faster than that of protein absorption (Rønnestad et al. 1993). Although no selective absorption of nutrients from yolk was reported for Sardinops caerulea (Lasker 1962), and only minor changes of yolk composition were observed in Oncorhynchus mykiss (Smith 1957), other works reported on changes of yolk chemical composition (salmonids—Hayes 1949; bluegill sunfish, Lepomis macrochirus—Toetz 1966; review by Heming and Buddington 1988). Absorption of energy-rich oil globules was retarded as compared to absorption of lipoprotein yolk in many species. Examples are Morone saxatilis (Eldridge et al. 1981; Rogers and Westin 1981); Pacific sandeel (Ammodytes personatus) (Yamashita and Aoyama 1985); Coregonus fera (Loewe and Eckmann 1988); walleye (Stizostedion vitreum) (Moodie et al. 1989); Scophthalmus maximus (Rønnestad et al. 1992a); Coregonus lavaretus (Escaffre et al. 1995); Leiostomus xanthurus (Fyhn and Govoni 1995). Larvae of European sea bass (Dicentrarchus labrax) still had about 30% of the initial oil globule at the onset of exogenous feeding (Rønnestad et al. 1998). In larvae of gilthead sea bream (Sparus aurata) 60% of the oil globule was still present at the time when body proteins were being used for energy purposes (Rønnestad et al. 1994). Minor changes in the relative composition of lipid classes (phosphatidylcholine, phosphatidylethanolamine, triacylglycerol, cholesterol and sterol ester, all expressed as % of total lipids) were reported by Rønnestad et al. (1995) during development of Hippoglossus hippoglossus yolk-sac larvae, but incorporation of lipid classes into larval body was selective.

Expression of size in terms of energy is considered in the chapter “body growth” below.

Description of yolk feeding

In mature females a phosphoglycolipoprotein, vitellogenin, is synthesized by hepatocytes, transported by the blood flow and incorporated in oocytes. Along with yolk protein derived from vitellogenin, a proteolytic enzyme, cathepsin D is included in the oocytic yolk (Sire et al. 1994).

In teleostean embryos yolk nutrients are mobilised through vitelline syncytium which reaches its maximum at completion of epiboly and decreases afterwards. Ultrastructural characteristics of syncytium and histological presentation of yolk absorption process were made for larvae of pike perch (Sander lucioperca) (Ostaszewska 2002) and plaice (Pleuronectes platessa) (Skjaerven et al. 2003). Two factors contributing to yolk resorption rate are activity of hydrolytic enzymes and surface area of the syncytium layer. Microvillar extensions increase the area of inner membrane of the syncytium (Skjaerven et al. 2003).

In freshly stripped Salmo trutta eggs nine protein types of 15–95 kDa were found to occur in concentrations >1 μg egg−1 (Lahnsteiner 2007). The major proteins were a 15 kDa phosphoprotein, two glycoproteins (62 kDa and 85 kDa), and a 95 kDa lipoprotein. Embryonic survival to the eyed stage was positively correlated with concentrations (μg egg−1) of each: 39, 77, 85, and 95 kDa proteins, a multiple regression model explained 96% of the survival variability (Lahnsteiner 2007). Several earlier studies have demonstrated contribution of maternal effects to embryonic viability via egg matter composition (review by Kamler 2005).

Eggs of marine pelagic fish contain large amounts of free amino acids (FAA) (Fyhn 1988), which form up to 50% of the total amino acid pool (Rønnestad et al. 1999). Depletion of a free amino acid pool, which is located mainly in yolk, begins after epiboly. In barfin flounder, Verasper moseri (Ohkubo and Matsubara 2002) and Hippoglossus hippoglossus (Zhu et al. 2003) sparing of essential amino acids (EAA) versus non-essential amino acids (NEAA) was reported, suggesting that EAA are retained for protein synthesis. However, in Pleuronectes platessa (Skjaerven et al. 2003) differential utilization of EAA and NEAA was not observed. In Scophthalmus maximus mobilisation of free amino acids from the FAA pool occurred mainly before and shortly after hatching, in Microstomus kitt—in embryos and early yolk-feeding larvae (Rønnestad and Fyhn 1993), while in Hippoglossus hippoglossus—mainly in yolk-sac larvae (Rønnestad and Fyhn 1993; Zhu et al. 2003). In early yolk-feeding larvae of H. hippoglossus (2–12 days post hatching) free amino acids from the yolk FAA pool were recruited to the body and incorporated into newly synthesised protein, while in older yolk-feeding larvae (12–32 days post hatching) amino acids entered the larval body from a yolk protein pool (Rønnestad et al. 1993). Protein stored in yolk is degraded into amino acids by cathepsins. Procathepsin L is located in the yolk syncytial layer and is transferred to yolk globules that have detached from a central yolk mass. Cathepsin D contained in yolk globules activates the proenzyme to cathepsin L which hydrolyses yolk protein (Sire et al. 1994). In summary, then, in yolk-feeding fish free amino acids from the yolk FAA pool are utilised earlier in ontogeny than amino acids mobilised from yolk protein.

A high assimilation efficiency (90%) was found in post-larval Hippoglossus hippoglossus fed liquid radiolabelled FAA diet, compared to peptide (PEPT, 32%) and protein (PROT, 12%) diets. The FAA diet was absorbed faster than the PEPT and PROT diets by factors of six and eight, respectively (Rojas-Garcia and Rønnestad 2003). Thus, FAAs are easily absorbed and highly assimilated nutrients, which makes them a superior amino acid source in earliest ontogeny.

Oil globules are generally regarded as an energy reserve. They are primarily neutral lipids with high proportion of monounsaturated fatty acids (MUFA) which are preferentially used as a fuel. In contrast, lipids in lipoprotein yolk contain primarily polar lipids rich in polyunsaturated fatty acids (PUFA) (see review by Wiegand 1996a). It is well known that marine fish receive an abundance of EPA (C20:5n-3) and DHA (C22:6n-3) in diets, thus their somatic tissues and eggs have relatively high levels of n-3 PUFA, while n-6 PUFA (linoleic acid C18:2n-6 and AA C20:4n-6) occur particularly in freshwater fish (reviews by Wiegand 1996a; Tocher 2003; Steffens 2005).

Saturated fatty acids and PUFA (EPA C20:5n-3, DHA C22:6n-3 and AA C20:4n-6) initially present in the yolk were preferentially retained for incorporation into tissues while MUFA were preferentially utilised as fuel in larval energy metabolism. These preferences were stronger at a low temperature presumably to maintain membrane fluidity, and in larvae from high egg quality (goldfish Carassius auratus, Wiegand et al. 1991; Wiegand 1996b; Hippoglossus hippoglossus, Rønnestad et al. 1995; Esox lucius, Desvilettes et al. 1997).

Yolk amino acid and lipid utilisation in early fish ontogeny have been usually studied separately, but Zhu et al. (2003) suggest that in H. hippoglossus carbon skeleton of FAA can be used for lipid synthesis. A tri-phasic sequence of yolk nutrients utilization for Verasper moseri embryos (days 0–10th post fertilisation at 8°C) and yolk sac larvae (days 11–21) was proposed by Ohkubo and Matsubara (2002). During days 0–4th no FAA were utilised. The main depletion of FAA occurred after 4th day: only 13% of the initial FAA level remained by the 13th day. Lipovitellin (the major yolk protein) and phospholipids were utilised mostly on 16–21st day post fertilisation. Taken together, common features of yolk nutrient utilisation are that FAA are utilised early and are followed by lipid and protein, and that essential amino acids and PUFA are retained preferentially.

Ontogenetic sequence in yolk feeding rate

Yolk absorption rate is low at the beginning of ontogeny and slowly increases afterwards. After hatching yolk absorption is faster than in embryos (Figs. 1 and 2; Wiegand et al. 1991 for Carassius auratus; Kinnison et al. 1998 for Oncorhynchus tshawytscha; review by Kamler 1992). Larger larval size, higher motility and better developed blood vessels in yolk sac contribute to this difference. In Oncorhynchus mykiss the period of rapid protein decrease was associated with increase of cathepsin L activity (Sire et al. 1994). This was followed by a decrease in the absolute yolk absorption rate associated with yolk depletion shortly before the final yolk absorption in starved larvae of Lepomis macrochirus (Toetz 1966), Micropterus salmoides (Laurence 1969), Oncorhynchus tshawytscha (Heming 1982) and Hippoglossus hippoglossus (Rønnestad et al. 1995). Conceição et al. (1997b) measured the absolute yolk absorption rate in starved larvae of African catfish Clarias gariepinus (of which Clarias lazera is a junior synonym, Teugels 1984). They found an increase in aCY between hatching and beginning of exogenous feeding (recomputed data: 0.106, 0.171 and 0.190 mg dry yolk day−1 at 0.54, 1.54 and 2.08 days post hatching), but a decrease afterwards (0.119 mg dry yolk day−1 at 2.58 days post hatching). Fatty acid (FA) utilization rate in endogenously feeding Scophthalmus maximus larvae was studied by Cunha and Planas (1997) during three developmental phases: 1—from hatching (H) to the onset of exogenous feeding ability (S), 2—from S to the mean point between S and 50% starvation mortality (SM50%), and 3—from the mean point between S and SM50% to SM50%, at 14, 18 and 22°C. The absolute rate of FA utilisation per effective day-degree (expressed in ng FA larva−1 D°eff−1) was significantly different in the three phases, highest in phase 2 and lowest in phase 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig2_HTML.gif
Fig. 2

Combined effect of temperature and egg size on the absolute rate of yolk absorption. Comparison of mean values for embryos (from fertilisation to hatching, Fe-H - dashed lines), and yolk-feeding larvae (from hatching to final yolk resorption, H-Re - solid lines).:Cn - Chondrostoma nasus (egg dry weight 2 mg), Om - Oncorhynchus mykiss, a three-year-old female (24 mg), Ot - O. tshawytscha (254 mg). Recomputed from Kamler et al. (1998), Kamler and Kato (1983), and Heming (1982), respectively. Semi-logarithmic scale

Studies on physiological and anatomical aspects of yolk resorption in Pleuronectes platessa, a marine teleost that spawns pelagic eggs, were reported (Skjaerven et al. 2003). Changes in protein, free amino acids (FAA) and NH4+, associated with morphometric and histological events, were monitored at 7°C from fertilisation to final yolk absorption. Four distinct developmental phases of endogenous feeding were observed. The first one lasted from fertilisation to gastrulation and formation of yolk syncytial layer (0–4 days post fertilisation). No changes of total (yolk and embryonic tissue) wet and dry weights, water content, content of protein, FAA and NH4+ were observed during this phase, and yolk resorption was undetectable. The second phase lasted from gastrulation to hatching (4–13 days post fertilisation). Total wet and dry weights, and water content continued to remain at an about stable level, FAA decreased rapidly, total protein increased and NH+ accumulated. Low pinocytic activity of the yolk syncytial layer was observed in embryonic P. platessa (i.e., during the first two phases). The third phase commenced after hatching and lasted until maximum larval size on yolk was attained (13–24 days post fertilisation). Loss of wet and dry weights, and water, further decline of FAA level, and rapid NH4+ decrease were observed. Yolk resorption rate was high at that phase, resulting from a high pinocytic activity. The mouth opened two days after hatch. In the presence of food larvae would begin mixed feeding at this phase. The fourth phase (24–28 days post fertilisation) started when maximum larval size solely on yolk was attained. In the absence of food, yolk size, and FAA and NH4+ levels remained constant, while protein content and larval size decreased (Skjaerven et al. 2003).

There have been several attempts to quantify the decrease in yolk size during ontogeny. In yolk-feeding chum salmon (Oncorhynchus keta) larvae yolk weight was found to decrease linearly (Zhang et al. 1995). Decrease of lipoprotein yolk and oil globule volumes in Brevoortia tyrannus and Leiostomus xanthurus during the whole endogenous feeding period was represented by Fyhn and Govoni (1995) with linear regressions which explained 79% to 90% of the variance. An exponential model used to quantify the decrease of yolk sac volume in common pandora (Pagellus erythrinus) larvae between hatching and final yolk resorption explained 99, 96 and 98% of the variance at 16, 18 and 21°C (Klimogianni et al. 2004). Similarly, decrease of yolk sac volume in yolk-feeding larvae of Atlantic mackerel Scomber scombrus was fitted through exponential equations, yielding 97, 99, 95, 81 and 70% of variance explained at 8.6, 11.1, 13.2, 15.1 and 17.8°C, respectively (Mendiola et al. 2007). Rønnestad et al. (1993) described dry weight of yolk (Y.r, μg indiv.−1) or free amino acid content (Y.r, nmol indiv.−1) at age τ (days from hatch) in Hippoglossus hippoglossus larvae at 7°C using a polynomial regression:
$$ {\text{Y}}{\text{.r($\tau$)}} = a + b\uptau + c\uptau^{{\text{2}}} + d\uptau^{{\text{3}}} {\text{.}} $$
(6)

The parameters a, b, c and d were: 1181.4, −18.281, −0.19282 and −0.0030, R2 = 1.000 for dry weight, and 1775.4, −192.77, 7.644 and −0.10284, R2 = 0.999 for FAA, respectively. Simulation of the decrease of yolk amount in Clarias gariepinus between hatching and final yolk absorption was made with a logistic model by Conceição et al. (1993). In Oncorhynchus tshawytscha a trajectory of yolk sac weight between fertilisation and first feeding was shown as an output from a mechanistic model based on differential equations (Beer 1999).

For the entire yolk-feeding period a three-phase sigmoid relationship between yolk area and age in embryos and yolk-feeding larvae of Melanogrammus aeglefinus was described by Martell et al. (2005) with a logistic dose-response curve. Yolk dry mass depletion over time in Central Amazonian fish species was fitted by Araujo-Lima (1994) to the Gompertz model:

$$ {\text{Y}}{\text{.r}}(\tau ) = {\text{Y}}_{0} {\text{e}}^{{ - {\text{e}}^{{ -g(\tau _{0} - \tau )}} }} $$
(7)
where Y.r is remaining yolk dry weight at age τ, Y0 is the asymptotic yolk dry weight, and g is the instantenous growth rate at the inflection point (τ0). Jaworski and Kamler (2002) considered four models describing the decrease of yolk dry weight between egg activation and final yolk resorption of Oncorhynchus mykiss, Chondrostoma nasus and Clarias gariepinus: von Bertalanffy, Gompertz, logistic and the Richards models. Reduced mean absolute per cent error and limited number of parameters make the Gompertz model useful for predicting yolk weight or energy throughout endogenous feeding period.

In summary, the above models describe the sequence of slow, fast and again slow absolute yolk absorption rates that occurs during yolk feeding with no access of external food. The final decrease of the absorption rate has not been revealed in some studies, possibly because of inter-individual variability of absorption and developmental rates, or measurements not frequent enough or truncated.

Factors affecting the yolk absorption rate

The effect of both, the syncytium surface area and metabolic rate on the rate of yolk absorption is well established (Heming and Buddington 1988). Syncytium surface area depends on egg size, metabolic rate depends on temperature. These two major factors will be considered now.

Egg size

Faster absolute yolk absorption rate (aCY, J indiv.−1 day−1) was observed in species producing larger eggs (Fig. 2). From an expanded interspecific comparison (Table 1) mean aCY increases as a power function of the initial energy in the yolk (Y0, J indiv.−1):
$$ {\text{aC}}_{{\text{Y}}} {\text{ = 0}}{\text{.260Y}}_{{\text{0}}} ^{{{\text{0}}{\text{.544}}}} {\text{,}} $$
(8)
n = 12, R2 = 0.930, P < 0.001, 95% c.i. for intercept 0.147 to 0.457, 95% c.i. for slope 0.440 to 0.649. The slope was just below the value 0.66, thus aCy was nearly surface-dependent. That confirms the observation of Conceição et al. (1993) that in Clarias gariepinus the rate of yolk absorption depends on the syncytium surface area. Beer and Anderson (1997) included the limiting surface area between yolk sac surface and fish body surface in their model of Oncorhynchus tshawytscha growth model. The aCY value 0.937 (95% c.l. 0.641 to 1.368) predicted for Clarias gariepinus from Eq. 8 was lower than the observed one, 2.363 J indiv.−1 d−1 (Table 1) which seems to be associated with exceptionally fast growth rate of yolk-feeding C. gariepinus (Kamler et al. 1994; Conceição et al. 1997a). In the interspecific comparisons shown in Table 1 yolk size differences were of three orders of magnitude. Differences between small and large eggs of Sander vitreum were much smaller (ca. 15%); the effect of initial egg size on the yolk consumption rate was not detected (Moodie et al. 1989). Body size-dependence of maximum yolk absorption rate at given time (Cmaxi) was described (Jaworski and Kamler 2002) at an intraspecific level for Oncorhynchus mykiss, Chondrostoma nasus, Cyprinus carpio, Tinca tinca and C. gariepinus by a power function. Food ration size (the proportion of Cmaxi realised at time i) of these yolk-feeding fish declined with depletion of yolk reserves in a way similar to the Ivlev’s (1961) dependence of ration size on food concentration in externally feeding fish (Jaworski and Kamler 2002).
Table 1

Effect of initial yolk size (Y0) on mean absolute yolk consumption rate (aCY) from fertilization to final yolk resorption in unfed larvae at near-optimum temperatures. Recomputed data

Species

Temperature (°C)

Y0 (J indiv.−1)

aCY (J indiv.−1d−1)

Source

Oncorhynchus tshawytscha

10.0

4784.00

32.500

Heming (1982)

O. ketaa

4.8

3981.00

21.800

Beacham et al. (1985)

O. kisutcha

4.6

3238.00

20.500

Beacham et al. (1985)

O. mykissb

12.0

674.00

12.500

Kamler and Kato (1983)

Salmo truttac

3 to 11

625.00

3.720

Raciborski (1987)

O. mykissd

12.0

280.00

7.000

Kamler and Kato (1983)

Chondrostoma nasus

16.0

48.99

2.355

Kamler et al. (1998)

Clarias gariepinus

25.0

10.56

2.363

Kamler et al. (1994)

Micropterus salmoides

12 to 18

9.56

0.735

Laurence (1969)

Clupea harengus pallasie

12.5 to 13.5

5.44

0.453

Eldridge et al. (1977)

Tinca tinca

21.7

3.95

0.483

Kamler et al. (1995)

Lepomis macrochirus

23.5

3.90

0.459

Toetz (1966)

aFemale #2, small eggs, to yolk “button up”

bThree-year-old female

cSeries 1984

dTwo-year-old female

e Control larvae

The duration of ontogenetic intervals within endogenous feeding period is size-related (review by Kamler 2002). Miller et al. (1988) described the dependence of the number of days from hatching to final yolk absorption (τH-Re) on the total length of newly hatched larvae (TL, mm) of marine, freshwater and anadromous fish (data standarized to 15°C):
$$\tau_{{{\text{H}} - {\text{Re}}}} {\text{ = 4}}{\text{.76 + 1}}{\text{.3 TL, }}n{\text{ = 88, }}R^{{\text{2}}} {\text{ = 0}}{\text{.33, }}P < {\text{0}}{\text{.0001}} $$
(9)

The combined effect of egg size (positive) and length of incubation period (negative) on dry weight of yolk remained unabsorbed at hatching was quantified by Koho (2002) by a polynomial equation which explained 76% of variance in the remaining yolk size. Thus, the efficiency of using yolk energy (or matter) resources and time resources both contribute to production large weaned larvae.

Temperature

Many studies were dedicated to the temperature effect on yolk absorption rate. Earlier studies were reviewed by Blaxter (1988) and Kamler (1992). Coregonid larvae caught in deep (2–6 m) cold waters had more unabsorbed yolk than larvae from the near-surface (0–2 m) (Karjalainen and Viljanen 1992). Just prior to hatching in embryos of Salmo salar the amount of unresorbed yolk decreased linearly with increasing temperatures (R2 = 0.93), but a logarithmic function was suitable for yolk-sac larvae (R2 = 0.97) (Ojanguren et al. 1999). Age at which 50% of Gadus morhua larvae absorbed their yolk decreased with increasing temperature between −1°C and 7°C (Pepin et al. 1997). Yolk utilization by Dicentrarchus labrax (Pelosi et al. 1993), spotted wolffish Anarhichas minor (Falk-Petersen 2001), and two sturgeons Acipenser brevirostrum and A. oxyrinchus (Hardy and Litvak 2004) was faster at higher temperatures. In larvae of Pagellus erythrinus relative yolk consumption rates (h−1) were 0.029 ± 0.001, 0.041 ± 0.003 and 0.056 ± 0.004 (±SE) at 16, 18 and 21°C, respectively, the differences were significant at P < 0.05 (Klimogianni et al. 2004). Relative yolk consumption rates (day−1) derived from volume-at-age exponential models for larvae of Scomber scombrus were very strongly related to temperature: 0.39, 0.66, 0.81, 0.96 and 1.29 at 8.6, 11.1, 13.2, 15.1 and 17.8°C (R2 = 0.99) (Mendiola et al. 2007). In the relationships between log absolute yolk absorption rate and temperature (Fig. 2) R2 values were: Chondrostoma nasus Fe-H 0.99 and H-Re 0.99, Oncorhynchus mykiss Fe-H 0.84 and H-Re 0.98, O. tshawytscha Fe-H 0.58 and H-Re 0.99. Thus, the yolk absorption rate typically increases with temperature. However, in O. tshawytscha embryos above 10°C yolk absorption rate decreased at a temperature of 12°C which seems to be supra-optimal for them. In O. mykiss embryos below 10°C the yolk absorption rate decreased dramatically; temperature 9°C was probably below the optimum range. In addition, no yolk-feeding larvae survived at 9°C until the final yolk sac resorption (Fig. 2).

Heming and Budington (1988) summarised 29 values of Q10 for yolk absorption rate in 23 fish species. The overall mean value was 2.92 ± 0.17 (±SE).

Temperature dependence for yolk absorption was summarised in Jaworski and Kamler (2002). Within the tolerated range, at low temperatures yolk absorption increases with temperature, at optimum temperature it is the highest, while further increase of temperature is accompanied by decrease of the yolk absorption rate. At temperatures below the theoretical optimum a simple exponential function describes well the temperature dependence of yolk absorption rate (Jaworski and Kamler 2002, further examples in Fig. 2). Absorption rate of oil globules seems to be more affected by temperature than the rate of absorption of yolk platelets (review by Heming and Buddington 1988).

Light

Yolk absorption of Salmo salar (Ryzhkov 1976) was delayed in light, but there are data that suggest an acceleration of endogenous reserve absorption by light. After incubation in complete darkness newly hatched Clarias gariepinus had less unresorbed yolk (by 6%) than after incubation under illumination from 50 lx (night minima) to 400 lx (midday maxima) (Appelbaum and Kamler 2000). In common dentex (Dentex dentex) larvae the yolk sac volumes at the onset of external feeding were 0.0176 mm3 and 0.0044 mm3 at total darkness and under constant illumination of 450 lx, respectively, and oil globule volumes were 0.0020 mm3 and 0.0006 mm3, respectively (Firat et al. 2003). Effect of illumination on yolk resorption and development of digestive organs in Sparus aurata was also reported by Sarasquete et al. (1995).

Xenobiotics

Reduction of yolk absorption rate was one of the responses to toxic xenobiotics. Bleached kraft mill effluent slowed down yolk absorption in Salmo trutta (Vuorinen and Vuorinen 1987). Yolk absorption was retarded in Al concentrations up to 800 μg dm−3 in acidic water down to pH 5.25 during a 10-day exposure in Esox lucius larvae (Vuorinen et al. 1993; Keinanen et al. 2000) and down to pH 5.75 in roach (Rutilus rutilus) larvae (Vuorinen et al. 1993). Exposure to 0.2 mg Cd dm−3 for 4 days reduced yolk absorption in Mozambique tilapia (Oreochromis mossambicus) (Hwang et al. 1995). The same concentration (0.2 mg Cd dm−3) of Cd or Cu or a mixture of Cd + Cu (1:1) applied during embryogenesis resulted in a reduction of yolk absorption rate and delayed final yolk resorption in Cyprinus carpio larvae reared in pure water from hatching (Ługowska 2005). Reduction of yolk absorption rate was a response to cadmium-induced damages of embryonic vascular system (Salmo salar, Rombough and Garside 1982).

Two more extrinsic factors contribute to yolk absorption rate. In Salmo salar low oxygen (Hamor and Garside 1977) delayed yolk absorption. High salinity of seawater reduced the yolk absorption rate in Scophthalmus maximus compared to brackish water in which extra energy is required to keep from sinking (Quantz 1985).

Increase in energy expenditure during locomotor activity accelerates yolk absorption rate. This was observed in zebrafish (Danio rerio) yolk-sac larvae under chronic exercise by swimming (swim training) compared to control (untrained) larvae (Bagatto et al. 2001). Yolk absorption was delayed in Salmo salar larvae in which incubation on rugose substrate reduced motor activity (Hansen and Møller 1985).

Depletion of yolk reserves slowed down yolk absorption rate of O. mykiss, C. nasus, C. carpio, T. tinca and C. gariepinus (Jaworski and Kamler 2002) in a similar way as exogenous food ration decreased with reduction of concentration of external food in older fish (Ivlev 1961). Mixed feeding slowed the rate of yolk absorption in Theragra chalcogramma (Hamai et al. 1974) and Salmo trutta (Raciborski 1987), but in some other species an access to external food accelerated yolk absorption rate or had no effect (review by Hemming and Buddington 1988).

A genetic effect on yolk absorption rate was demonstrated by Wood and Foote (1990). Smaller eggs were produced by nonanadromous (kokanee) than anadromous (sockeye) Oncorhynchus nerka. Although yolk absorption rates were similar in pure kokanee and pure sockeye larvae, “hybrid” larvae obtained from fertilisation with sockeye sperm absorbed yolk faster than hybrids resulted from fertilisation with kokanee sperm, which suggests a mismatch of egg size with male genotype.

Body growth (P)

General remarks

Growth is an increase in size with age. Increase of body tissues will be considered here, thus models describing total weight changes (embryonic tissues plus yolk plus perivitelline fluid plus egg capsules in embryos, or larval tissues plus yolk in yolk-feeding larvae, for example Finn et al. 1995a) will not be discussed here. During the whole endogenous feeding period the fish body mass increases by a factor of several orders of magnitude. Negative growth can occur under starvation: shrinkage in length and/or weight loss. Early growth rates strongly affect survival rates in marine fish (recent review by Houde 2002). Consequently, strength of a year-class responds to early growth rates, therefore the selection pressure for the maximization of larval growth is strong. Growth rate is a key process in studies of fish early life.

Methodological approaches

The size of tissues in an embryo or a larva can be expressed as total length (Lt), standard length (body length) (Ls), wet weight (Ww), dry weight (Wd), as a content of energy in embryonic/larval tissues (C.e.), or else as a content of any chemical compound in the tissues. Measures based on length and wet weight are less reliable due to variability of body shape and hydration of tissues.

In fish early ontogeny weight-to-length relationship (condition) is very variable. Weight is more useful to describe growth pattern than length. For example, response of growth in weight to heavy metals is stronger than that in length (review by Jezierska and Witeska 2001). Body size determinations in yolk feeding fish require body tissues to be separated from yolk (see discussion in yolk size determination section). Recently works have been published in which matter and energy content and conversions between yolk and body were studied (for example see Rønnestad et al. 1993; Kamler et al. 1998, and several others).

In small aquatic organisms “individual fresh weight is not really useful to describe growth pattern” (Dawirs 1981), because water that enters the body tissues or is removed from them may contribute to wet weight gains or losses. For example, no significant temperature - induced wet weight changes were detected in lake minnow (Eupallasella percnurus) at final yolk resorption (1.26, 1.26, 1.24, 1.20 and 1.19 mg in groups kept from fertilisation at 13, 16, 19, 22 and 25°C, respectively, ANOVA P = 0.0632). In contrast, dry weight did increase with temperature (0.111a, 0.114ab, 0.118bc, 0.119bc and 0.122c respectively, P = 0.0001, results followed by the same letter are not significantly different, Tukey HSD test), because hydration of tissues decreased with increasing temperature (percent of dry weight in wet weight was 8.84a, 8.99ab, 9.47b,9.99c and 10.21c, respectively, P = 0.0000) (Kamiński et al. 2006). An important water absorption after hatching was found in the body of Oncorhynchus tshawytscha (Heming 1982). Water increase was included into growth model of yolk-feeding O. tshawytscha (Beer and Anderson 1997). Thus, it appears that dry weight provides more reliable information about body size than wet weight.

Dry weight of embryonic and larval tissues during yolk feeding is usually very small. For example, the maximum dry weight attained with exclusively endogenous feeding was about 0.04–0.06 mg in an Amazonian characiform Potamorhina latior (Araujo-Lima 1994). Therefore, dry weight measurements of body tissues call for high accuracy (minimalised difference between the amount actually present and the amount found by measurement, Sutclife 1979). Dry weight needs to be taken after desiccation to constant weight over a drying agent, for example silica gel or NaOH. The temperature of drying is important, because too high temperature may result in underestimated weight due to loss of fats. Temperatures of about 60°C is routinely used (Edmondson and Winberg 1971). Modern techniques include freeze drying. In yolk-feeding Scophthalmus maximus desiccation at 58°C and lyophilisation gave overlapping results (Finn et al. 1996).

A source of bias in size determination is shrinkage—a reduction in size resulting from distortions and damages caused by capture, and water and other compound loss at fixation (review by Ferron and Leggett 1994). Body shrinkage at death may be important. In Clupea harengus and smelt (Osmerus eperlanus) it amounted to 15% Ls (Fey 2002), in yolk-feeding larvae of freshwater tropical fish to 24% Wd (Araujo-Lima 1994), and in Gadus morhua to 40% L (Radtke 1989). Shrinkage mostly occurs early after preservation, is species- and preservative-specific and depends on larval size. Araujo-Lima (1994) proposed an equation:
$$ W_{{{\text{dU}}}} = W_{{{\text{dP}}}} ^{{{\text{0}}{\text{.903}}}} $$
(10)
for correction of preserved dry weight (WdP) to unpreserved one (WdU). Fey (2002) predicted unpreserved standard length of C. harengus and O. eperlanus (LsU, mm) from preserved one (LsP):
$$ \hbox{L}_{{{\text{sU}}}} {\text{ = 0}}{\text{.91 }}\hbox{L}_{{{\text{sP}}}} {\text{ + 2}}{\text{.695}} $$
(11)

The losses of weight and length were negligible in relatively large individuals (≥1 mg and ≥30 mm respectively, in Eqs. 10 and 11), but were important in the smallest specimens (0.06 mg Wd and 10 mm Ls, respectively). Thus, in studies of growth in fish early life, this source of bias needs to be taken into consideration.

More components of bias are in population growth estimates. Two of them act in contrasting ways. The first is a size-selective net avoidance. In early developmental stages of fish, an escape of larger larvae makes the mean size and growth rate increasingly underestimated. The size-selective mortality acts in an opposite direction. There is a strong negative trend between risk of predation and larval fish prey size (Fuiman 1994; Houde 1996). Greater mortality experienced by smaller eggs was demonstrated for Gadus morhua and Pleuronectes platessa in the field (Rijnsdorp and Jaworski 1990), but it was also observed in laboratory experiments on Salmo trutta larvae with no predation (Hansen 1985). Successive elimination of smaller individuals results in an overestimation of growth during fish early life (Kamler 1992; Houde 2002; Jones 2002). This speculative hypothesis has been confirmed with elephant seal, Mirounga leonina, an Antarctic mammal larger than fish larvae by a factor of 1 × 106. Burton (2001) followed individual weight trajectories of pups from weaning to the age of 200–300 weeks when their weight reached 300–1300 kg. In female pups he found a negative relationship between growth in weight and initial weight at weaning, thus the pups that had been lighter at weaning exhibited an apparently faster growth. According to Burton (2001) the upward bias was attributable to the over-representation of the faster growing survivors. Thus, an apparent population growth is the growth of survivors.

Size- and temperature-dependent migrations were identified as another factor contributing to apparent changes in growth of perch (Perca fluviatilis) and Sander lucioperca during first growing season (Kjellman et al. 2001).

Records of fish growth, age and key ontogenetic events are retained in otoliths, scales and other hard parts. Scales and fin rays do not develop enough before metamorphosis, while two pairs of otoliths are formed on the day of hatch. They are sagittae and lapilli, examples are given by Hoff et al. (1997) for shortnose sucker Chasmistes brevirostris and Lost River sucker Deltistes luxatus. Thus, for fish larvae the relevant records of growth are otoliths (review by Jones 2002) on which increments are deposited daily (Pannella 1971). Otolith daily marks result from alternate deposition of calcium carbonate and protein. During periods of fast growth more protein and less calcium is deposited in the otoliths. The records of sizes of individual fish at past ages can be reconstructed from the relation between otolith size and fish size. Growth trajectories are derived from repeated measurements on individuals, and they provide valuable information about the past fate of an individual fish, but require an adequate statistical treatment (Chambers and Miller 1994). The first daily increment in Gadus morhua otoliths was found the day after hatching (Radtke 1989). The author pointed out that distinct marks were formed at yolk sac absorption and first external feeding. However, in species in which daily marks begin at final yolk resorption (review by Jones 2002) otoliths cannot be applied for determination of growth during yolk-feeding period.

For energetics purposes body size needs to be expressed in terms of energy. Caloric equivalent is the energy content of an individual (J indiv.−1). Caloric value of a unit of mass (i.e., specific caloric value, for example J mg−1 or kJ g−1 dry matter) can be assessed (1) by direct calorimetry or (2) indirectly from chemical composition of the material.
  1. (1)

    Two direct calorimetric methods, wet (dichromate) oxidation (Maciolek 1962) and “dry combustion” using two forms of oxygen bomb calorimeters gave similar caloric values of Perca fluviatilis somatic tissues (Craig et al. 1978). Direct calorimetry using an oxygen bomb calorimeter is considered to be the most reliable method and is used as a reference (Prus 1993; Kamler et al. 1994; Finn et al. 1995b). In studies on fish early life small samples are available, thus microbomb calorimeters (Phillipson 1964; Klekowski and Bęczkowski 1973) are appropriate, in which size of combusted pellets is typically 5–30 mg dry matter. They were used for whole eggs with developing embryos and for whole yolk-sac larvae of Cyprinus carpio (Kamler 1972), Hippoglossus hippoglossus (Finn et al. 1991), Gadus morhua (Finn et al. 1995b) and Scophthalmus maximus (Finn et al. 1996). A microbomb was also used for initial eggs, larval tissue, remaining yolk, egg cases and liquid wastes discharged at hatch of Oncorhynchus mykiss (Kamler and Kato 1983).

     
  2. (2)

    Caloric values can also be computed indirectly from proximate analysis (2.1), from chemical determination of several constituents (2.2) or from the CHNS analysis (2.3).

     
  3. (2.1)

    Protein, lipids and carbohydrates derived from the proximate analysis are converted to joules. That procedure has been commonly reported in the literature. Energy conversion factors adopted by the International Biological Programme (Winberg and Collaborators 1971) for freshwater organisms were 23.0 (probable range 22.2–24.1), 39.8 (38.5–41.8) and 17.2 (15.7–17.6) J mg−1 for protein, lipids and carbohydrates, respectively. Slightly different (albeit remaining within the IBP range) factors, were reported by Brafield and Llewellyn (1982), Jobling (1983), and Gnaiger and Bitterlich (1984). Due to residues present in fish oils, and, probably, greater amounts of PUFA, a lower lipid conversion factor, 35.5 J J mg−1was recommended (Beamish et al. 1975; Craig et al. 1978; Finn et al. 1995b). A deficit in the tally is typical for the proximate analysis because of incomplete extraction and numerous minor constituents left unidentified (see Kamler 1992 and literature herein). Thus, the caloric values based on the proximate analysis are often underestimated. For example, caloric values of eggs of vendace (Coregonus albula), Cyprinus carpio and Sardinops caerulea computed from proximate analysis were 83%, 96% and 97% of the caloric values determined directly in a bomb calorimeter (review by Kamler 1992).

     
  4. (2.2)

    The most comprehensive data set of chemical composition for whole eggs with developing embryos (8 data points) and for whole yolk-sac larvae (1 data point) was provided for Gadus morhua by Finn et al. (1995b). They determined protein, free amino acids, lipids, carbohydrates, NH4 and lactate. The caloric values computed from the chemical composition ranged from 93.2% to 101.1% of the values derived from direct determinations in a bomb calorimeter when a conversion factor of 38.95 J per 1 mg of lipids was used, and from 84.9% to 98.9% when a factor of 35.56 J mg−1 was taken.

     
  5. (2.3)

    In the CHNS analysis percentages of main elements: carbon (C), hydrogen (H), nitrogen (N) and sulphur (S) in dry matter are directly determined, sulfanilamide or acetanilide serve as reference. Ash fraction (A) is determined separately, the remaining fraction is assumed to be oxygen (O). Caloric value is computed from the C, H, O and S fractions with a formula given by the Analyser’s programme. There are several advantages of that method. First, reliability of results was confirmed by a good agreement of within 5% between larval carp caloric values assessed with the CHNS method and measured directly with a bomb calorimeter (Kamler et al. 1994). Second, precision of the method (replicability of results) is typically very high. For example, coefficient of variation (standard deviation among replicate subsamples as a percentage of the sample mean) amounted to 0.6, 0.3, 0.1, 0.1 and 0.1%, respectively, in Clarias gariepinus initial eggs, dissected larval tissues at hatch and larval body at yolk resorption at 22, 25 and 28°C (Kamler et al. 1994). In Chondrostoma nasus initial intact eggs and in dissected yolk coefficient of variation was 0.6%, in dissected embryonic tissues, tissues of larvae at hatch and at final yolk resorption it amounted to 0, 0.5 and 0.3%, in perivitelline fluid and in egg cases it was 1.3 and 1.2%, respectively (Kamler et al. 1998). Similarly, low variability about the mean (coefficient of variation ≤1) of CHNS-derived caloric values was found for yolk-feeding Tinca tinca (Kamler et al. 1995), Eupallasella percnurus (Kamiński et al. 2006) and in another study on C. gariepinus (Appelbaum and Kamler 2000). Third, the sample size can be as small as 3–5 mg dry matter. Fourth, the measurements are highly automatised and done in a short time. Fifth, information about carbon, hydrogen, nitrogen and sulfur content is derived from the same analysis, and protein content can be computed using the nitrogen-protein conversion factor of 5.78 recommended by Gnaiger and Bitterlich (1984) for aquatic organisms. Thus, it appears that elemental analysis is a promising method that can be recommended for determination of main elements, energy and protein contents during fish early ontogeny.

     

Expression of growth rate and potential

Growth rate is a change in body size over a time interval. Body size can be expressed as any measure listed in the section Methodological approaches above.

Absolute growth rate (AGR) in weight can be approximated by:
$$ {\text{AGR = }}{{\left( {W_{{\text{2}}} - W_{{\text{1}}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {W_{{\text{2}}} - W_{{\text{1}}} } \right)}} {\Updelta \tau }}} \right. \kern-\nulldelimiterspace} {\Updelta \tau }{\text{,}} $$
(12)
where W1 and W2 are the body weight at times τ1 and τ2, respectively. For example Rombough (1988c) used this equation to provide estimates of absolute growth rates of Oncorhynchus mykiss.
There are several measures of growth rate that are expressed as proportion of body size. The most commonly used for exponential growth of young fish is the specific growth rate (G, day−1):
$$ {\text{AGR = }}{{\left( {W_{{\text{2}}} - W_{{\text{1}}} } \right)}}\mathord{\left/ {\vphantom {{{\left( {W_{{\text{2}}} -W_{{\text{1}}} } \right)}} {\Updelta \tau }}} \right.\kern-\nulldelimiterspace} {\Updelta \tau }{\text{,}} $$
(13)
Winberg (1956) proposed to calculate daily weight gains as per cent of initial weight (WGR, % day−1):
$$ G = {{\left( {\ln W_{{\text{2}}} - \ln W_{{\text{1}}} } \right)}}\mathord{\left/ {\vphantom {{{\left( {\ln W_{{\text{2}}} - \ln W_{{\text{1}}} } \right)}} {\Updelta \tau }}} \right. \kern-\nulldelimiterspace} {\Updelta \tau } $$
(14)

Later Winberg (1971) abandoned Eq. 14 in favour of the Eq. 13, but Eq. 14 had remained in use by earlier authors (for example Baranova 1974; Ryzhkov 1976; Buckley 1979; Ostroumova et al. 1980) for description of growth rate in fish embryos and larvae.

Ricker’s (1975) relative growth rate (RGR, % day−1):
$$ {\text{WGR}} = {\text{100}}{\left[ {10^{(1/\Updelta \tau {\text{) (log }}W_{{\text{2}}} - {\text{ log }}W_{{\text{1}}} {\text{)}}} - {\text{1}}} \right]} $$
(15)
has been considered as appropriate for fish larvae (Myszkowski 1997), and used, for example, for larval Gadus morhua (Folkvord 1997) and for yolk-feeding Clarias gariepinus (Conceição et al. 1997b).

It is important to remember that growth rates computed with Eqs. 14 and 15 are identical (that is, WGR = RGR), while Eq. 13 gives G values that are lower than WGR/100 and RGR/100, especially when growth is fast. Therefore, the method of the growth rate computation needs to be always defined, and the G values should not be compared directly with the RGR = WGR values.

A weak point in embryonic growth rate calculation is uncertainty of the determination of the initial size (size of germinal disc, G.d.), because G.d. is small and its separation from yolk is uncertain. Dry weight of Oncorhynchus mykiss body measured in 25 replicates 0–2 h after fertilisation was 0.03 mg (Ryzhkov 1976). Volume of germinal disc in Salmo sp. and Coregonus sp. was 0.6–0.7 mm3 (Korovina 1978), that is about 0.06 mg dry weight. It should be noted, however, that values of the specific growth rate (G) are not very sensitive to variability in the initial weight. Assuming O. mykiss tissue dry weight at hatching 2 mg, incubation period 30 days (Kamler and Kato 1983), and germinal disc dry weight 0.03 or 0.06 mg (factor 2.0×), one can expect G values 0.140 or 0.117, respectively (factor 1.2×). This agrees with the findings of Jaworski and Kamler (2002) who conducted a sensitivity analysis of various parameters in a bioenergetics model for yolk-feeding larvae. They found that changes in embryo size at the beginning of development (i.e., weight of germinal disc) had a negligible effect on the generated maximum body size. The latter was mainly dependent on temperature and yolk size, and caloric value. Thus, it appears that the uncertain estimation of the germinal disc size has a minor effect on final results.

In Acipenserbrevirostrum and A. oxyrinchus the rate of body area growth was expressed as a slope of the regression of body area against age (Hardy and Litvak 2004). These authors collected body area from anaesthetised yolk sac larvae using an image analysis system by drawing a line around larval body excluding fins and yolk sac.

Growth potential that differed among 30 larvae of Brevoortia tyrannus can be suggested from very rare crossing of their individual growth trajectories: the individuals that had a higher growth rate during the first week post hatching maintained it during the two successive weeks (Chambers and Miller 1994). An indirect measure of the potential for growth is the RNA/DNA ratio, and lipid and protein content. Quantity of RNA reflects the amount of recent protein synthesis, while the amount of DNA is believed to be related to the number of cells, basing on an assumption that DNA is conserved on a relatively constant level. For example, for first-feeding Gadus morhua larvae a positive, highly significant (P < 0.01) correlation of the RNA/DNA ratio with growth rate in terms of protein was reported by Buckley (1979). The RNA/DNA ratio has been since accepted as an index of short-term growth potential in several aquatic organisms, from phytoplankton to larval, juvenile and adult fish. A longer-term growth potential is mirrored by total protein content, while total lipids are commonly used as a long-term growth potential index. Detailed reviews are provided by Ferron and Leggett (1994), Jones (2002) and Weber et al. (2003), among others.

Ontogenetic changes in body size during yolk feeding

Although growth patterns differ between fish species, typically a piecewise continuum is observed (Fig. 1, Table 2, more examples in the section “Factors affecting growth rate” below):
  • Fe to H. Embryonic growth from fertilisation (Fe) to hatching (H) is slow, but accelerating.

  • H to S. Shortly after hatching (H) growth of yolk-sac larvae further accelerates, then slows down until the larvae reach the maximum weight (S) that their yolk reserves can support. At S growth rate approaches zero, an equilibrium is reached between nutrient absorption and nutrient utilisation.

  • S to Re. Small remaining energy reserves (yolk and/or oil globule) do not cover metabolic expenditures. Unfed larvae show an energy deficit and tissue absorption (negative growth) until final yolk resorption (Re).

For example, in Gobio albipinnatus maximum growth rate, 11% wet weight day−1 was found few days after hatch (Wanzenböck and Wanzenböck 1993). In two New Zealand Oncorhynchus tshawytscha populations specific growth rate in terms of dry weight at 12°C decreased within the period from H to Re: 0.087, 0.042 and 0.026 day−1, and 0.083, 0.042 and 0.024 day−1 during 40–52, 52–64 and 64–73 days post-fertilisation, respectively (Kinnison et al. 1998). Specific rate of protein synthesis in coregonid yolk-sac larvae was higher than in embryos and starved larvae (Fauconneau et al. 1986).
Table 2

Examples of ontogenetic sequence (Fe – egg activation, H – hatching, S – initiation of external feeding, Re – completion of yolk resorption) on absolute growth rates (J indiv.1d1) Recomputed data

Species

Source

Lepomis macrochirus

Toetz (1966)

Micropterus salmoides

Laurence (1969)

Oncorhynchus tshawytscha

Heming (1982)

Initial yolk size (J egg−1)

3.9

9.6

4784

Fe to H

0.28

0.26

8.2

H to S

0.33

0.40

28.9

S to Re

−0.11

−0.13

−26.4

Left (initial) part of the growth curve for yolk feeding fish was described by several simple models (earlier reviews by Hayes 1949 and Kamler 1992). Linear properties (length, height) are usually assumed to grow linearly with time. For example, in Hippoglossus hippoglossus larvae between H and S body length and myotome height increased linearly with age (Finn et al. 1995c), see also McGurk (1984) for Clupea harengus pallasi. Martell et al. (2005), however successfully (R2 0.69–0.87) applied exponential functions to describe increase in length over time in yolk-feeding Melanogrammus aeglefinus. The exponential model was extensively used to predict weight of endogenously feeding fish at time τ(W(τ)) during an accelerating phase of growth before the inflection of growth curve:
$$ W(\tau) = W_{0} {\rm e}^{{{\rm G}}\tau} $$
(16)
In the exponential model W0 is the size at time 0, and the specific growth rate, G, is constant. For example, in Coregonus sp. embryos tissue dry weight increased exponentially from 0.1 mg at eyed stage to 0.6 mg at hatching (Ortner et al. 1988). In H. hippoglossus larvae between H and S at 6°C dry weight (μg indiv.−1) was well predicted (R2 0.999) by the model (15) where W0 and G were 106.37 and 0.0587, respectively (Finn et al. 1995c). A cube root model:
$$ {W(\tau) = W}_{{\text{0}}} {\text{e}}^{{{\text{G}}}\tau} $$
(17)
(Hogendoorn 1980; Verreth and Den Bieman 1987) is not a valid alternative for the exponential model (Haylor 1992). Growth as depicted in (16) and (17) is observed for short time intervals, only.
Growth rate usually decreases with increasing age, and yolk-feeding fish larvae are not an exception (Fig. 3). Thus, a log-log relationship:
$$ {W(\tau)}^{{{\text{1/3}}}} {= W}_{{\text{0}}} ^{{{\text{1/3}}}}{\text{ + G}\tau} $$
(18)
(where W0 is the initial weight when time τ = 1, and b is a constant) was used, for example, by Eldridge et al., (1982) for Morone saxatilis embryos and Rombough (1988c) for yolk-feeding Oncorhynchus mykiss after blastopore closure. In the MacDowell’s modification of the equation (18) (see, for example Ricker 1979) τ is the time from the establishment of the embryonic axis. In Scomber scombrus yolk-feeding larvae solely on yolk standard length-at-age (Ls(τ)) between hatching and maximum Ls was successfully described for five temperatures by a Laird-Gompertz model (Mendiola et al. 2007). However, the above models do not represent the S-shaped growth curve that occurs in yolk-feeding larvae without access to external food.
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig3_HTML.gif
Fig. 3

Age effect on relative growth rate in terms of dry matter (RGR, % day−1) in yolk feeding larvae within the period from hatching to final yolk resorption. C.g. 28 – Clarias gariepinus reared at 28°C, initial egg dry weight 0.4 mg (from Conceição et al. 1997b); O.n. 24 and O.n. 28 – Oreochromis niloticus at 24 and 28°C, respectively, initial egg dry weight 4 mg (from Rana 1990)

There were attempts to describe growth of yolk-feeding larvae between hatching and a point in time after reaching maximum body weight. Escaffre and Bergot (1984) proposed a model (19) for dry weight of Oncorhynchus mykiss larvae:
$$ {W(\tau) = W}_{{\text{0}}} \tau^{{\text{b}}} $$
(19)
where: Wmax is maximum dry weight, Z = 1/[1 + expλ(τmax – τ)], τmax is time needed to reach Wmax, and λ is a growth coefficient. Rønnestad et al. (1993) described growth of Hippoglossus hippoglossus yolk-feeding larvae using a 3rd order polynomial:
$$ {W(}\tau {\text{) = a + b}}\tau {\text{ + c}}\tau ^{{\text{2}}}-{\text{d}}\tau ^{{\text{3}}} $$
(20)
where τ is time post-hatch (days), and a, b, c and d are constants. For dry weight of body tissues (μg indiv.−1) the constants were 67.814, 15.246, 0.1306 and 0.0003, respectively, R2 = 1.00. Klimogianni et al. (2004) used the Gompertz model to describe growth in total length of Pagellus erythrinus larvae between hatching and yolk sac resorption at 16, 18 and 21°C.
There are models that fit observed data of embryonic and larval body size-at-age to a single model over the entire yolk-feeding period from egg activation to final yolk absorption. Araujo-Lima (1994) fitted dry body weight at age in 14 species of Central Amazonian fishes to:
$$ {W(}\tau {\text{) = ae}}^{{{\text{(b}}\tau {\text{)}}}} {\text{e}}^{{{\text{(c}}\tau ^{{\text{2}}} )}} $$
(21)
where τ is age from egg activation (h), and a, b and c are constants. Beer and Anderson (1997) developed a mechanistic model to describe the growth of Oncorhynchus tshawytscha during the whole endogenous feeding period using physiological and geometric parameters. Bertalanffy’s conception assuming growth resulting from the difference between anabolism (absorption) and catabolism (loss) was followed. The growth of fish body as the yolk mass is declining was described by two coupled differential equations. A limiting surface area between yolk and fish controls the nutrient absorption, which is a function of temperature, and water absorption was also taken into account. Four phases were considered. (1) In embryo yolk is unlimited, the limiting absorptive area is fish surface area, no water absorption occurs. (2) At hatching water absorption begins. (3) The yolk sac surface defines the limiting area when yolk is completely surrounded by vitelline circulatory system and is smaller than fish surface. (4) Starvation with the exponential loss of weight occur after complete yolk resorption. Thus, the output of the model are times to and weight at different developmental events. In Oncorhynchus mykiss embryonic and larval weight between egg activation and beginning of feeding was also derived from a mechanistic model based on differential equations (Beer 1999). Fiksen and Folkvord (1999) proposed a simulation model for the entire larval period (endogenous and external feeding included) of Gadus morhua. In that model growth was taken from the difference between assimilation and metabolism. Jaworski and Kamler (2002) expressed body dry weight of Oncorhynchus mykiss, Chondrostoma nasus and Clarias gariepinus between egg activation and final yolk resorption as a difference between two sigmoid curves, one describing total embryonic or larval dry weight (tissue plus unresorbed yolk) and another representing yolk dry weight. Four models were tested, the Gompertz model (equation 7), von Bertalanffy, Richards and logistic models. In Oncorhynchus keta total weight (tissue plus yolk) and body weight were fitted well with the Gompertz growth model (Zhang et al. 1995). Both, accurate representation of size and limited number of parameters make the Gompertz model useful for predicting the changes in weight of fish embryos and larvae during complete yolk feeding period (Jaworski and Kamler 2002).

Ontogenetic changes in body composition during yolk feeding

Changes in composition of body matter are much less conspicuous than those of body size.

Hydration of fish eggs is relatively low (review by Kamler 1992), but, unlike reptiles and birds, fish yolk-feeding larvae have no water limitation. Water absorption is high after hatching (Heming 1982; Beer and Anderson 1997). For example, in freshly spawned (unswollen) Cyprinus carpio eggs 70.6% (Semenov et al. 1974) and 67.0% (Kamler 1976) water was reported, while in unfed C. carpio larvae at the end of yolk feeding 86.4% (Peňáz et al.,1976) and 85.0% (Kamler 1976) water was found. In unswollen Chondrostoma nasus eggs 67.9 ± 0.5% water ( ± SD) was observed (raw numerical data from Keckeis et al. 2000). Higher hydration levels, 86.2 ± 0.2%, 86.4 ± 0.6%, 85.9 ± 0.1% and 86.2 ± 0.4% water were found in larvae at the final yolk resorption after incubation at 10, 13, 16 and 19°C, respectively (raw numerical data from Kamler et al. 1998). Thus, percentage of dry matter in wet matter of initial eggs was higher than that of larval body at the end of endogenous feeding. In C. nasus the contribution of dry matter in larval body at yolk resorption did not vary with temperature, thus did not respond to differences in developmental time that ranged from 13 days at 19°C to 46 days at 10°C (Kamler et al. 1998).

Inter-specific differences between caloric values of initial egg dry matter were mirrored by differences in caloric values of larval tissues at final resorption of yolk (Table 3), R2 = 0.586, d.f. 8, P < 0.01. The caloric values of larval body at Re were lower by about 12% than those of initial eggs, but in E. percnurus that difference was only 2%, whereas in C. carpio it was as much as 23%. Caloric value of tissues declined with age of yolk-feeding larvae (Table 3). Changes of chemical composition of dry matter and formation of skeletal structures contribute to that decrease.
Table 3

Comparison of caloric values (J mg1 dry matter, mean ± SD) in initial eggs and in larval tissues separated from yolk at hatching (H) and beginning of external feeding (S), and in larval body at final resorption of yolk (Re). Based on 12 data sets of 9 species

Species

Eggs

Larval tissue

Source

H

S

Re

Chondrostoma nasus

28.4 ± 0.2

27.1 ± 01

25.9 ± 0.1

Kamler et al.(1998)

Tinca tinca

28.1 ± 0.1

25.1 ± 0.0

Kamler et al. (1995)

Clarias gariepinus

27.9 ± 0.2

27.2 ± 0.1

25.4 ± 0.4

Kamler et al. (1994)

C. gariepinus

27.9 ± 0.1

26.6 ± 0.3

Appelbaum and Kamler (2000)

Oncorhynchus mykiss, 3ya

27.9 ± 0.5

24.8 ± 0.8

23.4 ± 1.7

Kamler and Kato (1983)

O. mykiss, 2ya

27.4 ± 2.5

24.3 ± 0.6

23.8 ± 0.7

Kamler and Kato (1983)

Eupallasella percnurus

26.8 ± 0.1

26.3 ± 0.3

Kamiński et al. (2006)

Cyprinus carpio

26.1

23.3

Peňàz et al. (1976)

C. carpio

25.3 ± 1.7

20.0

Kamler (1972)

Micropterus salmoides

25.1

20.8

Laurence (1969)

Lepomis macrochirus

24.4 ± 4.3

21.3

Toetz (1966)

Hemiramphus sajori

23.0

21.8

Kimata (1982)

aOffspring of 3 or 2 years old females

Total lipids contributed to 11% of dry matter in tissue of newly hatched Hippoglossus hippoglossus larvae (Rønnestad et al. 1995). That value is low compared to egg lipids which usually range between 10 and 35% of dry matter (mean 19.3%, Kamler 1992). A minor increase in the level of total fatty acids (sum of FA in phosphatidylcholine, phosphatidylethanolamine and neutral lipids) occurred in tissue of developing yolk-sac larvae of Hippoglossus hippoglossus between hatching and maximum larval weight: 8.55 and 9.38 g/100 g dry matter, respectively (recomputed from Rønnestad et al. 1995). At the same time the levels of DHA (22:6n−3) and EPA (20:5n−3) increased more than twice, from 1.45 to 3.26 and from 0.51 to 1.40 g/100 g dry tissue, respectively. These compounds are of high biological value. For example a high dependence of Sparus aurata growth on DHA was demonstrated by Koven et al. (1993). During development of yolk-feeding fish n-3 PUFA tend to be retained selectively (review by Tocher 2003).

Protein content in tissues of Hippoglossus hippoglossus yolk-feeding larvae between hatching and maximum tissue mass increased exponentially with age (τ, days post hatch) (Finn et al. 1995c):
$$ {\text{Protein (}}\mu {\text{g indiv}}{\text{.}}^{{{\text{ - 1}}}} {\text{) = 42}}{\text{.67 e}}^{{{\text{0}}{\text{.0677 }}\tau }} $$
(22)

Percentage of protein in larval dry body matter was about 54%; it did not vary with age. Protein deposition rate was 6.8% day−1, thus it was a little higher than the dry matter growth rate (6.0% day−1) (Finn et al. 1995c). Protein synthesis rate in yolk-feeding larvae of H. hippoglossus was 30% day−1 while in Clarias gariepinus it was as much as 300% day−1 (Conceição et al. 1995).

In general, a high concentration of free amino acids (FAA) is typical of marine pelagic eggs, well above what is found in tissues of adult fish, marine demersal eggs and freshwater eggs (Fyhn 1989, review by Rønnestad and Fyhn 1993). Newly spawned eggs of Scophthalmus maximus contained 50–90 nmol egg−1 of free amino acids (FAA) (Rønnestad et al. 1992a), in eggs of Clupea harengus amount of FAA increased from 54 nmol egg−1 on days 1–3 post fertilisation to 79 nmol egg−1 prior to hatching (days 18–20) (Hølleland and Fyhn 1986). Free amino acids play a role in osmoregulation, they are important osmolytes and as such make a contribution to buoyancy (Fyhn 1993). They are also used as a fuel in energy metabolism (see chapter Metabolism below). However, the primary function of FAA in marine fish early life is to serve as substrate for protein synthesis: rapid growth of fish embryos and early yolk-feeding larvae is supported by a FAA pool (Houlihan 1991; Rønnestad et al. 1992a; Rønnestad et al. 1993; review by Rønnestad and Fyhn 1993). In yolk-feeding Scophthalmus maximus larvae almost all assimilated 14C – labelled alanine was incorporated into molecules insoluble in TCA, while only a negligible amount was released as labelled CO2 (Korsgaard 1991). Tissues of Oncorhynchus mykiss at final yolk resorption contained 22 times more FAA than the tissues at hatching (Zeitoun et al. 1977). In tissues of H. hippoglossus yolk-feeding larvae the increase of the FAA content with age (τ, days post hatch) was linear from shortly after hatching to shortly before final yolk sac resorption (Rønnestad et al. 1993):
$$ {\text{FAA (nmol indiv}}{\text{.}}^{{{\text{ - 1}}}} {\text{) = 27}}{\text{.54 + 6}}{\text{.66 }}\tau $$
(23)

However, Dabrowski et al. (2003) disagreed with the conclusion that in early fish growth is supported by FAA. They conducted experiments involving first-feeding Oncorhynchys mykiss larvae (initial weight 114 mg) to investigate growth response to different sources of amino acids, and pointed out on a retarded growth in fish fed a FAA-based diet compared to dipeptide-based diet as a protein source.

In embryos or yolk-feeding larvae ash is rarely determined, because large samples, ≥25 mg dry matter of tissue separated from yolk are required. Few data existing suggest a low content of minerals in larval tissue at hatching as compared with larval body at final yolk resorption (Table 4), which may reflect skeleton formation as ontogeny progresses. Larvae with high ash content emerge from eggs rich in minerals: as much as 64.7% of variance in the ash content in larval body at the end of endogenous feeding can be attributed to the ash content in initial egg, P < 0.05 (Table 4).
Table 4

Comparison of ash content (% dry matter, mean ± SD) in initial eggs and larval tissues separated from yolk at hatching (H), and in larval body at final resorption of yolk (Re). Based on 7 data sets of 7 species

Species

Eggs

Larval tissue

Source

H

Re

Chondrostoma nasus

10.2 ± 0.1

9.1 ± 0.7

15.1 ± 0.4

Kamler et al. (1998)

Scophthalmus maximus

9.3

20.1

Finn et al. (1996)

Tinca tinca

8.3

9.9

Kamler et al. (1995)

Clarias gariepinus

5.8

2.3

9.5 ± 0.2

Kamler et al. (1994)

Cyprinus carpio

5.2 ± 1.0

9.7 ± 2.4

Kamler (1972)

Eupallasella percnurus

4.8

5.2

Kamiński et al. (2006)

Oncorhynchus mykiss

3.9

8.2

Suyama and Ogino (1958)

The carbon/nitrogen ratio (the C/N ratio) has been used as one of measures of condition in externally feeding animals (review by Ferron and Leggett 1994). It is believed to reflect the ratio of fats to protein, with C/N for protein near 3, and increasing above 3 when lipids, rich in carbon, are present. The C/N ratio was reported to decrease with age in yolk-feeding Scophthalmus maximus (Korsgaard 1991). In yolk-feeding larvae of Clupea harengus (tissues plus yolk) the C/N ratio was 4.54, while in exogenously feeding larvae it was 3.8 (Kiørboe et al. 1987). In Danio rerio whole eggs 24 h post fertilisation 3.11 ± 0.090 μmol C egg−1 was found, while just before hatch (age 75 h) the carbon content was reduced to 2.81 ± 0.003 μmol C egg−1 (P < 0.05). In contrast, nitrogen contents did not differ, 0.572 ± 0.014 and 0.538 ± 0.06 μmol N egg−1, respectively, (P > 0.05) (Bang et al. 2004). In Chondrostoma nasus carbon contents (% dry matter) decreased in sequence yolk, embryonic tissue and larval body at final yolk resorption: 56.0 ± 0.33c, 51.6 ± 0.09b and 48.9 ± 0.13a, respectively, while nitrogen contents increased: 11.8 ± 0.06a, 11.7 ± 0.02a and 12.0 ± 0.03b, respectively (non-significant results are followed by the same letter). Thus, C/N ratio decreased in this sequence: 4.75, 4.41 and 4.08, respectively (Kamler et al. 1998). All these reflect the mobilisation pattern of yolk energy stores: carbon-rich lipids are preferentially catabolised for energy, whereas protein is retained for growth.

Both, RNA and DNA content in whole egg or larval body plus remaining yolk increased between early cleavage and the onset of external feeding (Zeitoun et al. 1977 for Oncorhynchus mykiss; Buckley 1981 for winter flounder Pseudopleuronectes americanus; Seoka et al. 1997 for Pagrus major). From that increase, an intense protein synthesis and cell proliferation can be deduced. In starved larvae of P. americanus, and just prior to the onset of external feeding in P. major, a reduced RNA/DNA ratio was found, indicating a suppressed growth.

Factors affecting growth rate

Retarded ontogeny and growth prolong embryonic and early larval developmental periods, in which mortality rate reaches maximum. Thus, effective use of time resources is of vital importance in early fish life. Temperature, oxygen, substrate, pH, salinity, xenobiotics, light/ultraviolet radiation, and magnetic field are probably the major environmental variables that control growth, along with three intrinsic factors, egg size, sex and genetic factors.

It is well known that in poikilotherms growth rate is strongly accelerated by temperature. Salvelinus alpinus embryos and yolk-feeding larvae grew at a faster rate at 8°C (temperature close to the upper limit of the tolerance zone) than at 4°C which is close to the lower limit (Gruber and Wieser 1983). In exponential functions describing embryonic growth in length of Melanogrammus aeglefinus over time post-fertilisation regression exponents (slopes) increased with incubation temperature, 0.054, 0.057, 0.075, 0.082 and 0.091, respectively at 2, 4, 6, 8 and 10°C (Martell et al. 2005). Further examples of temperature-accelerated growth rate in yolk-feeding fish are growth in length of Cyprinus carpio (Peňáz et al. 1983), orangethroat darter Etheostoma spectabile (Marsh 1986), Acipenser brevirostrum and A. oxyrinchus (Hardy and Litvak 2004), tissue dry weight of Salmo salar morpha sebago (Ryzhkov 1976), and whole body protein synthesis in coregonids (Fauconneau et al. 1986). During the first growing season in three bays in the Gulf of Finland (Baltic Sea) growth in length of Perca fluviatilis and Sander lucioperca was positively correlated with temperature, expressed as effective day-degrees over 10°C (Kjellman et al. 2001).

Response of growth rate in yolk-feeding fish to temperature is summarised in terms of tissue dry matter in Fig. 4 for five species of contrasting temperature requirements, from a cold water salmonid Oncorhynchus tshawytscha to warm water fishes such as African catfish Clarias gariepinus and Nile tilapia Oreochromis niloticus. Typically, the growth rate over the whole viable temperature range is dome-shaped, increasing from near zero at low temperature to a maximum value at optimum temperature and decreasing back at sub-lethal temperature (post-yolk fishes: Gadus morhua larvae, Jordaan and Kling 2003; review by Jobling 1997). The whole course of the response of growth rate to temperature is not shown in Fig. 4. The response at high temperatures is often truncated. It may be hardly detectable and clouded by high mortality.
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig4_HTML.gif
Fig. 4

Temperature effect on relative growth rate in terms of dry matter (RGR, % day−1) in yolk feeding larvae within the period from hatching to final yolk resorption. O.t. – Oncorhynchus tshawytscha, initial egg dry weight 490 mg (from Heming 1982); O.m. – Oncorhynchus mykiss, initial egg dry weight 10 mg (from Kamler and Kato 1983); C.n. – Chondrostoma nasus, initial egg dry weight 2 mg (from Kamler et al. 1998); O.n. – Oreochromis niloticus, initial egg dry weight 4 mg (from Rana 1990); C.g. – Clarias gariepinus, initial egg dry weight 0.4 mg (from Kamler et al. 1994)

Incubation temperatures may have a delayed effect in post-embryonic ontogeny. In Melanogrammus aeglefinus an effect of incubation temperatures of 4, 6 and 8°C on larval growth at a common post-hatch temperature 6°C was significant (P < 0.0001) and was observed well beyond hatch (>35 days) (Martell et al. 2005). Incubation temperatures of 8.0 and 10.2°C displayed a very long (8 months) persistence on Oncorhynchus tshawytscha growth rate (Heath et al. 1993).

Oxygen content influences growth rate much less than temperature. It is well established that the size of newly-hatched larvae is reduced under hypoxia (see reviews by Zhukinskij 1986; Rombough 1988b; Kamler 1992, 2002). For example, dry weight of tissues of newly-hatched Salvelinus alpinus larvae decreased with the decrease of oxygen concentration: from 4.44 mg at 100% air saturation to 3.40–4.07 mg at 50% and to 3.02–3.13 mg at 20–30% saturation (Gruber and Wieser 1983). In Oncorhynchus mykiss embryonic growth did not differ between groups incubated at 10°C in hyperoxia (15 mg O2 dm−3) and normoxia (10 mg O2 dm−3), but in hypoxia (5 mg O2 dm−3) it was significantly less (Ciuhandu et al. 2005). Three probable explanations for this effect can be proposed. Firstly, hypoxia can slow down the rate of tissue differentiation compared with normoxia. Secondly, premature hatching of smaller larvae can result from low oxygen concentration through stimulation of chorionase secretion and embryo mobility (review by Kamler 2002). Thirdly, in yolk feeding fish growth and metabolism compete for yolk energy resources. With a decrease in oxygen supply metabolic processes can be partly shifted towards less efficient anaerobic processes.

In O. mykiss embryos the same pattern (depressed growth in hypoxia compared to normoxia) was observed by Ciuhandu et al. (2005) in both, intact and dechorionated eggs, but in the latter embryos grew faster in all three oxygen treatments. The positive response of O. mykiss growth to chorion removal was also reported by Ninness et al. (2006). This response in fish embryos apparently involves three mechanisms. First, chorion and perivitelline fluid that isolated developing embryo from the surrounding water, were eliminated from the dechorionated eggs, thus supply with oxygen was improved. Constraints on oxygen supply to eggs by chorion and perivitelline fluid will be discussed in more details in the section Factors affecting metabolic rate. Second, the factors that inhibit growth, hormones and growth factors may be lost, and ammonia level drastically reduced in dechorionated eggs (review by Ciuhandu et al. 2005). Third, spontaneous movements of the embryos are no more restricted in dechorionated eggs: number of movements of dechorionated embryos increased by a factor of 36, but respiration did not differ (Ninness et al. 2006). Significantly greater body mass and body protein content in dechorionated larvae compared with intact ones after hatching was explained by exercise training effect, similar to that observed in adult salmonids.

Yolk-feeding larvae of Salmo salar were reared on two contrasting substrates, a flat one and on gravel. The rugose substrate accelerated growth rate at all five temperatures tested (Peterson and Martin-Robichaud 1995).

No differences were observed between the size of newly hatched Perca fluviatilis larvae from two lakes of different pH: an acid (pH 4.8–5.0) and a circum-neutral one (pH 6.6–6.9) (Laitinen 1994).

Growth of Oncorhynchus mykiss embryos in brackish water (salinity 6‰) manifested a small depression, but later on (between hatching and onset of external feeding) growth of larvae was normal (Spannhof and Pavlov 1984). In desert pupfish (Cyprinodon macularius) total length at hatching was 4.5 mm at salinity 0.02‰ (half freshwater), 4.4 mm at 0.04‰ (freshwater), 4.2 mm at 35‰, 3.9 mm at 55 ‰ and 3.7 mm at 70‰ (Kinne and Kinne 1962). Total length of ruffe (Gymnocephalus cernuus) larvae increased, and then decreased with increasing salinity. The maximum length in a freshwater population remained at lower salinity than the maximum in brackish water G. cernuus (Vetemaa and Saat 1996). With increasing salinity sizes of larvae at hatching have been reported to decrease, to increase, to remain constant, or an initial increase of larval size with increased salinity was followed by a decrease at higher salinity level (review by Chambers 1997). Thus, the response of growth to salinity is dose- and species-specific, and tolerance to salinity may be adaptive.

Several studies have shown impaired growth in yolk-feeding fish exposed to xenobiotics. Heavy metals received most attention. Exposure to 50 μg dm−3 Cd for 4 days had no significant effect on total length of yolk-feeding larvae in Oreochromis mossambicus, while concentration of 200 μg dm−3 reduced growth in length (Hwang et al. 1995). Retarded growth of yolk-feeding Salmo trutta larvae subjected to 80 nmol Cu dm−3 or 50 nmol Pb dm−3 in soft, acid water was reported by Sayer et al. (1991). A 10-days exposure to Al slowed down growth of yolk-feeding Esox lucius at pH 5.25, and that of Rutilus rutilus at pH 5.75 (Vuorinen et al. 1993). Yolk-feeding larvae of rudd, Scardinius erythrophthalmus aged 4 days post-hatch were exposed to 0.0, 0.1, 0.2 and 0.3 mg Cd dm−3 for 24 h, then they were transferred to pure water and fed for 10 days. A delayed effect of the short-term exposure to Cd on larval growth in pure water was found (Sikorska and Wolnicki 2006). Egg incubation in solutions of Ag, Cd, Pb, Cu or Zn reduced the size of newly hatched larvae in several fish species (review by Ługowska 2005). Reduced body growth under heavy metals may result from energy-consuming detoxication processes (Marr et al. 1998) and/or from metal-induced decrease of thyroid hormone (T4 and T3) levels (review by Ługowska 2005). In Salmo trutta embryos growth in length was retarded by bleached kraft mill effluents of 0.5, 1 and 2% (v/v), the concentrations of 0.5 and 1% impaired growth of yolk-feeding larvae (Vuorinen and Vuorinen 1987).

Precocious hatching at smaller size was reported by Leshchinskaya (1954) for European anchovy (Engraulis encrasicolus) larvae under strong light. After incubation in complete darkness newly hatched Clarias gariepinus had slightly higher dry weight of body tissues (by 2%) as compared to their siblings exposed to light (diurnal fluctuation from 50 lx to 400 lx) (Appelbaum and Kamler 2000). Later on (at the onset of external feeding) dry weight of tissue of dark-reared C. gariepinus was higher by 7% than that of light-reared individuals. Light slightly reduced growth in length of yolk-feeding larvae of Dentex dentex: total length of larvae at the onset of external feeding was 3.52 ± 0.08, and 3.36 ± 0.06 mm (±SD) after rearing from hatching in complete darkness or under constant illumination of 450 lx, respectively (Firat et al. 2003). In contrast, higher weight of newly hatched starry sturgeon (Acipenser stellatus) larvae after incubation in daylight as compared to darkness was demonstrated by Detlaf et al. (1981). In Sevan trouts, Salmo ischchantypicus, S. ischchan aestivalis and S. ischchan gegarkuni largest size (body length, wet weight and dry weight) at hatching and at the onset of external feeding, and highest embryonic and larval relative growth rates (computed as WGR, Eq. 14) were found in constant light. Total darkness resulted in smallest larvae and lowest growth rates. Exposure to natural photoperiod, to 16 h L:16 h D and to 8 h L:8 h D regimes resulted in intermediate growth (Ryzhkov 1976). No effects of different light regimes (natural photoperiod, or continuous light, or else total darkness) on Rutilus rutilus heckeli and Sander lucioperca incubation period and size of newly hatched larvae were observed by Belyj (1961). In the above examples growth response to light in yolk-feeding fish was species-specific: negative, positive, or no effect was observed. Zhukinskij (1986) suggests that the positions of optimal or tolerated zones within an illumination gradient depend on environmental adaptations.

The most injurious component of solar radiation is ultraviolet radiation, specifically UV-B (Bunn et al. 2000). In embryos of northern anchovy (Engraulis mordax) and Pacific mackerel (Scomber japonicus) that survived UV radiation retarded growth was found (Hunter et al. 1979). In Brachydanio rerio UV radiation impaired epiboly, thus preventing blastopore closure, but gastrulation was less affected (Straehle and Jesuthasan 1993). There are mechanisms to minimise effects of radiation. Transparency of egg and embryonic structures in some marine pelagic eggs that float near sea surface is developed as a response to solar radiation which passes through them without absorption (Bunn et al. 2000). In embryos and newly hatched larvae of Eurasian minnow (Phoxinus phoxinus) and Rutilus rutilus a substance acting as a sun screen was identified by Hofer and Kaweewat (1998).

Increase of the constant magnetic field from natural geomagnetic field (the control value) to 5 mT was paralleled by the increase of length and weight of Oncorhynchus mykiss newly hatched larvae (up to 119 and 128% of the control, respectively) (Formicki and Winnicki 1998). In Salmo trutta larvae they found an increase of the size at full yolk resorption (up to 114% and 127%, respectively).

Initial yolk size had a significant (P < 0.001) effect on embryonic and larval absolute growth rates (AGR) and explained 75% and 96% of the variance in AGR values, respectively, in interspecific comparisons (Table 5). However, no clear response was obtained for the yolk-size-dependence of the specific growth rate (G, Table 5). From a model for Gadus morhua the advantage of large yolk supply for growth of externally fed larvae was particularly manifest if yolk-size-dependent development of perceptive ability was assumed (Fiksen and Folkvord 1999).
Table 5

Effect of initial yolk size (Y0) on mean absolute growth rate (AGR, equation (12)) and specific growth rate (G, equation (13)) in fish embryos (from egg activation to hatch) and yolk-feeding larvae (from hatch to first feeding (S) or from hatching to full yolk sac resorption (R)). Recomputed data

Species

Temp. (°C)

Y0 (J indiv.−1)

AGR (J indiv.−1day−1)

G (day−1)

Sourcea

Embryos

Larvae

Embryos

Larvae

Oncorhynchus tshawytscha

10

4784.00

8.20

28.90S

0.05

0.02S

1

Oncorhynchus keta

4.8

3981.00

2.90

0.06

2

Oncorhynchus kisutch

4.6

3238.00

3.30

0.06

2

Salmo ischchan aestivalis

7.8–15

1112.00

0.70

7.90S

0.07

0.10S

3

Oncorhynchus mykiss

12

861.00

1.18

3.40R

0.14

0.03R

4

Oncorhynchus mykiss

12

664.00

1.76

5.49R

0.16

0.04R

5

Salmo trutta

3–11

625.00

0.60

3.90S

0.04

0.03S

6

Oncorhynchus mykiss

7.5–11.3

592.00

0.76

8.10R

0.08

0.06R

7

Oncorhynchus mykiss

12

278.00

1.00

3.52R

0.14

0.05R

8

Chondrostoma nasus

16

49.00

1.33

1.39R

0.72

0.07R

9

Micropterus salmoides

18–21

9.56

0.26

0.40S

0.01

0.20S

10

Clarias gariepinus

28.1

9.22

0.97

2.54R

4.87

0.79R

11

Clarias gariepinus

28.0

9.22

1.29

2.93R

3.60

0.70R

12

Morone saxatilis

18.0?

8.52

1.82

0.08S

13

Clupea harengus pallasi

12.5–13.5

5.44

0.37

0.14

14

Lepomis macrochirus

23.5

3.90

0.28

0.33S

>1.00?

0.30 S

15

Scophthalmus maximus

15.0

0.85

0.06S

0.19S

16

n

14

13

16

14

r

0.868

0.979

–0.356

−0.337

P

<0.001

<0.001

>0.05

>0.05

aSources: 1, Heming (1982) converted from dry matter to energy basing on Kamler and Kato (1983); 2, Beacham et al. (1985) converted from wet to dry matter basing on Ryzhkov (1973), converted from dry matter to energy basing on Kamler and Kato (1983); 3, Ryzhkov (1976) embryos at 7.8–8.3°C, larvae at 15°C; 4, Kamler and Kato (1983) 4-years old female; 5, Kamler and Kato (1983) 3-years old female; 6, Raciborski (1987); 7, Ryzhkov (1976) embryos at 7.5–8.9°C, larvae at 11.3°C; 8, Kamler and Kato (1983) offspring of 2-years old female; 9, Kamler et al. (1998); 10, Laurence (1969); 11, Conceição et al. (1997b) Y0 and caloric values of larval tissues at hatch and at yolk resorption from Kamler et al. (1994); 12, Kamler et al. (1994); 13, Eldridge et al. (1982) the value for embryos is given by the authors, that for yolk-feeding larvae is recomputed; 14, Eldridge et al. (1977) control larvae; 15, Toetz (1966); 16, Quantz (1985)

A sex-specific growth pattern was demonstrated with cherry salmon (Oncorhynchus masou): in a 60-d experiment yolk-feeding males grew faster than females, and the difference was greater in half-siblings hatched from large eggs (Yamamoto 2004).

Diploid Oncorhynchus mykiss larvae were significantly heavier at swimming up than their triploid full-sibs, and responded to subsequent starvation with less wet weight reduction and hydration increase (Happe et al. 1988).

At the end of the section of growth in yolk-feeding fish the case of a fast-growing species, Clarias gariepinus, will be considered. Data presented in Figs. 3 and 4, and Table 5 demonstrate that during yolk feeding C. gariepinus is a very fast growing fish. Conceição et al. (1997a) measured protein synthesis (Ks), protein growth (Kg) and protein degradation (Kd) (all in terms of % protein weight d−1) in C. gariepinus yolk-feeding larvae aged from 34 h to 40–42 h post activation at 28°C: Ks = Kg + Kd. They found Ks 137.8, Kg 95.9 and Kd 41.9, a fast growth, confirming earlier data for C. gariepinus yolk-feeding larvae (Conceição et al. 1993: Kamler et al. 1994). Protein retention efficiency, Kg 100 Ks−1 = 69.6% was similar to that in older fish. However, the costs of protein synthesis, 55.3 mmol ATP g−1 protein synthesised were low, approaching the theoretical minimum level, 50 mmol ATP g−1. Therefore, the fast growth of C. gariepinus yolk-feeding larvae is attributable to the high rates of protein synthesis at minimum costs, rather than to the maximization of the protein retention efficiency. Fast and efficient growth is typical also of C. gariepinus older developmental stages, externally feeding larvae and juveniles (Hogendoorn et al. 1980; Hecht and Appelbaum 1987; Conceição et al. 1997b). The fish is native to tropical and subtropical fresh waters, and is a successful newcomer to European aquaculture in heated waters.

Metabolism (R)

General remarks

Metabolism is the sum of the reactions that release the energy useful for physical and chemical work. Total metabolism (Rt) is partitioned into the costs of different functions. In externally feeding, growing animals cost of maintenance (Rm) represents energy expenditure by a unfed, resting individual under optimal conditions. Energy required for growth (Rg) is associated with the cost of synthetic processes. Feeding metabolism (Rf) involves an amount of energy used for foraging, ingestion and food conversion. Scope for activity (Ra) is the amount of energy used for locomotion and associated with response to non-optimal conditions.

In yolk feeding fish the partitioning of the maintenance component from the feeding and growth components has been difficult. Two sub-components of the total metabolism in yolk-feeding fish had been proposed by Rombough (1988b): routine metabolism (defined as “the average rate of aerobic metabolism under normal rearing conditions”, see also Rombough 1988c) and active metabolism (metabolism during burst activity). Recently Rombough (2006) reviewed the ways to estimate maintenance metabolism in endogenously feeding fish. In yolk-feeding larvae of Danio rerio locomotor activity was blocked by anesthesis, and growth was inhibited by blocked protein synthesis. Metabolic rate measured in that experiment was probably devoted to maintenance, it amounted to about 45% of the routine metabolic rate in untreated individuals. Similar values for maintenance metabolism (about 46% and 50% of the routine metabolism) were arrived at from theoretical calculations and from measurements of metabolic rate at the incipient lethal oxygen level, respectively. Ability of fish embryos to reduce maintenance metabolism during dormant state was reported (review by Rombough 2006).

Growth, and especially protein synthesis is energy-demanding (Brafield and Llewellyn 1982). Increase of respiratory rate in proportion to the rate of growth was summarised for postlarval fish by Jobling (1985). The net cost of deposition of 1 g dry matter in externally feeding fishes and an aquatic mollusc Mytilus edulis amounted to 15 mmol O2, assuming caloric value 22 J mg−1 dry matter, thus, the efficiency of utilization of metabolisable energy (P + Rg) for growth (P), P/(P + Rg) is about 0.75 (Wieser 1994). This author indicated that costs of growth are not fixed, and at high growth rates they may be reduced. However, Rombough (2006) demonstrated with fish larvae that this relationship is not ubiquitous.

Ammonia is an endproduct of utilisation of amino acids as a fuel in energy catabolism. The level of ammonia excretion is several-fold lower than that of oxygen consumption (both in terms of nmoles indiv.−1 h−1). For example, routine rate of ammonia excretion in Hippoglossus hippoglossus yolk-feeding larvae ranged from about 2.5 to 7 nmol indiv.−1 h−1 while routine rate of oxygen consumption ranged from about 10 to 52 nmol indiv.−1 h−1 (Finn et al. 1995c).The molar ratio of NH3 produced to O2 consumed defines the apparent nitrogen quotient (NQ). The average NQ value for protein oxidation was 0.27 (Gnaiger 1983), the NQ for aerobic catabolism of an “average amino acid” in Microstomus kitt embryos was 0.26 (Rønnestad et al. 1992b).

Teleost fishes are predominantly ammonotelic, urea usually constitutes no more than 10–20% of total nitrogen (ammonia-N plus urea-N) excreted (review by Wood 1993).

Quantifying of metabolic rate

Indirect calorimetry determines the amount of energy released from catabolic substrates: carbohydrates, protein and lipids. The complete procedure requires the determination of oxygen consumption, carbon dioxide production and excretion of non-faecal nitrogen (Brafield 1985). However, for juvenile grass carp Ctenopharyngodon idella and a beetle, Tribolium castaneum small differences (<2%) were found between energy output computed from complete indirect calorimetry and from oxygen consumption alone converted to energy with a composite oxycalorific coefficient (Kamler 1970). In fish embryos, carbon dioxide is buffered in perivitelline fluid (Smith 1957), and nitrogen metabolites are retained in it (Blaxter 1969; Kaushik et al. 1982). Thus, in many studies metabolic rate of endogenously feeding fish is estimated from oxygen consumption measured alone (review by Finn et al. 1995d).

Oxygen consumption can be expressed in mg O2, in cm3 (1 cm3 O2 = 1.429 mg O2) or in mmol O2 (1 mmol O2 = 22.4 cm3 O2) (0°C, 760 mm Hg, Schmidt-Nielsen 1983). Absolute rate of oxygen consumption is expressed in terms of an amount of oxygen consumed per an individual per unit time (for example, mm3 O2 individual−1 h−1). Mass-specific (relative) oxygen consumption rate is expressed per unit of weight (for example, mm3 O2 mg−1 dry weight h−1). Absolute oxygen consumption rate is used in an instantaneous form (for example mm3 O2 day−1 larva−1) or as a cumulative form (Klekowski et al. 1967). In the latter the daily metabolism is cumulated from the beginning of the life cycle to successive days of development (e.g., mm3 O2 larva−1 from egg activation to hatching).

Oxygen consumption is converted to energy using an oxycalorific coefficient. For fully aerobic metabolism a composite oxycalorific coefficient, 13.6 J per mg oxygen consumed is an estimate for fish catabolizing chiefly protein and fats, while 14.1 J per mg O2 is recommended when carbohydrates play a significant role as a metabolic fuel (Elliott and Davison 1975; Brett and Groves 1979; Brafield 1985; Gnaiger and Kemp 1990). One mol of O2 consumed is assumed to be equivalent to the synthesis of 6 mol ATP (Reeds et al. 1985).

Koho (2002) estimated daily metabolism of Coregonus albula embryos as loss of egg dry weight.

Ontogenetic sequence in metabolism

Oxygen consumption rates of freshly stripped (non-activated) fish spermatozoans ranged from 20 to 164 mm3O2 cm−3sperm h−1, after activation the rates increased two- to three-fold (review by Linhart et al. 1991).

Oxygen consumption rates of fish eggs before and shortly after activation are listed in Table 6. Mean mass-specific oxygen consumption rate in fish initial eggs at circum-optimal temperatures amounted to 20.5 (95% C.L. 7.7 to 33.3) mm3 g−1 h−1 (Table 6). Winberg (1956) described absolute oxygen consumption rates for juveniles and adults of several fish species basing on 364 pairs of data:
$$ R{\text{ = 0}}{\text{.3 }}W^{{{\text{0}}{\text{.8}}}} $$
(24A)
thus mass-specific oxygen consumption rate is:
$$ R{\text{/}}W{\text{ = 0}}{\text{.3 }}W^{{ - {\text{0}}{\text{.2}}}} $$
(24B)
where R is in cm3O2 indiv.−1 h−1 at 20°C and W in g wet weight. The R/W values extrapolated from the eqn (24B) for the extremal egg weights (840 mg and 0.18 mg, Table 6) are higher than the initial egg mean respiration rate (20.5 mmg−1 h−1) by factors of 15 and 80, respectively. Thus, the oxygen consumption rates of initial fish eggs are very low. In these eggs the amount of metabolising tissue (germinal disc) is very low compared to large amount of yolk and other inert egg structures.
Table 6

Effect of egg size on oxygen consumption in unfertilised or freshly fertilised fish eggs

Species

Temp. (°C)

Egg size (mg wet wt)

Oxygen consumption rate

Source

Absolute (mm3egg−1h−1)

Relative (mm3g−1h−1)

Trematomus borchgrevinki

−1

840.00

c.13.0000

c.15.5

Rakusa-Suszczewski (1972)

Salmo salar

229.00

2.7900

12.2

Ozernyuk (1985)

Oncorhynchus mykiss

72.40

1.0100

14.0

Ozernyuk (1985)

O. mykiss

10

0.1750

Yarzhombek (1986)

O. mykiss

7.5–8.9

c.60.00

0.1205

c.2.0

Ryzhkov (1976)

O. mykiss

12

29.60

0.1000

3.4

Kamler and Kato (1983)a

Huso huso

25.80

0.3020

11.7

Ozernyuk (1985)

Chondrostoma nasus

16

20.35

0.1193b

5.9

Kamler et al. (1998)

C. nasus

13

20.35

0.0995b

4.9

Kamler et al. (1998)

Esox luciusc

12

10.80

0.0600

5.6

Lindroth (1946)

Hippoglossus hippoglossus

5.7

6.27d

0.0075b

2.2

Finn, et al. (1991)

Cyclopterus lumpus

5

4.60d

0.0120

2.6

Lønning et al. (1988)

Misgurnus fossilis

21

0.0420

Yarzhombek (1986)

M. fossilis

2.38

0.0350

14.7

Ozernyuk (1985)

Clarias gariepinus

28

1.46

0.0508

34.9

Kamler et al. (1994)

C. gariepinus

25

1.46

0.0298

20.5

Kamler et al. (1994)

Cyprinus carpio

23

1.00

0.0510

51.0

Kamler (1976)

Tinca tinca

22

0.79

0.0163

20.5

Kamler et al. (1995)

Scophthalmus maximus

15

0.62e

0.0155

25.0

Finn et al. (1996)

Brachydanio rerio

0.48

0.0074

15.4

Ozernyuk (1985)

Gadus morhua

5

0.29d

0.0070

24.1

Lønning et al. (1988)

Solea senegalensis

19.5

0.18d

0.0224

124.4

Parra et al. (1999)

n

20

20

r

0.997

0.095

P

<0.001

>0.05

aRaw numerical data

bCalculated from an equation

cFour-cell stage

dEgg weight recomputed from dry to wet using 74% water in eggs of sea spawners (Kamler 1992)

eEgg wet weight from Finn and Rønnestad (2003)

In eggs of mummichog (Fundulus heteroclitus) an increase of respiration just after activation was reported from an early study (Boyd 1928). However, an absence of abrupt changes of respiration on egg activation was reported from several studies (Philips 1940 for Fundulus heteroclitus; Nakano 1953 for Japanese rice fish, Oryzias latipes; Kamler 1976 for Cyprinus carpio, and Ozernyuk 1985 for sterlet, Acipenser ruthenus; review by Zhukinskij 1986). High ammonia excretion just after egg activation is unrelated to metabolism, but ammonia is eliminated during the process of chorion hardening (review by Finn et al. 1991). The absence of changes of oxygen consumption rate and ATP levels at activation is typical of fish and amphibian eggs (review by Ozernyuk 1985).

In early ontogenesis (cleavage, morula) absolute rate of oxygen consumption per fish embryo increases slowly (examples in Figs. 57). Short-term fluctuations of relative metabolic rate per unit weight at the transition form one developmental step to another were broadly reviewed for embryogenesis of several fish species (Ryzhkov 1976). Two peaks (during cleavage and epiboly) or only one (during epiboly) were indicated in Rombough’s (1988b) review. The apparent controversies may be of methodological nature because of fast changes and minute amounts of respiring tissue in early embryogenesis. Oxygen uptake in early embryos is cutaneous, a progressive increase of gill relative importance in terms of respiratory function was shown by Rombough (1999) for Oncorhynchus mykiss (earlier reviews by Fry 1957 and Rombough 1988b). After onset of heart beat oxygen uptake is aided by blood circulation. A short-term acceleration of the oxygen consumption in Cyprinus carpio at the age of 36–44 D° coincided with haemoglobin formation in the blood (Kamler et al. 1974). Embryonic motility breaks down oxygen gradients in perivitelline fluid.
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig5_HTML.gif
Fig. 5

Ontogenetic changes of absolute rate of oxygen consumption in Clarias gariepinus embryos and larvae solely on yolk at three temperatures. Initial egg dry weight 0.41 mg. Based on raw numerical data from Kamler et al. (1994)

A conspicuous temporary acceleration of the absolute oxygen consumption rate was manifest at hatching in Clarias gariepinus (see Fig. 5) at 22°C (50% of larvae hatched at the age of 2.05 days, Kamler et al. 1994), at 25°C (hatched at 1.26 days), and at 28°C (0.90 days). In Oncorhynchus mykiss (see Fig. 7) short-term increases in oxygen consumption rates were observed at hatching, which peaked 35th, 31st, 25th and 21st days post-activation at 9, 10, 12 and 14°C, respectively (Kamler and Kato 1983). Increase of oxygen consumption rates at hatching is well established. It was reported for three species of Acipenser (Korzhuev et al. 1960), Clupea harengus (Holliday et al. 1964; Eldridge et al. 1977), Micropterus salmoides (Laurence 1969), Cyprinus carpio (Kaushik et al. 1982), Pleuronectes platessa and lumpsucker (Cyclopterus lumpus, Lønning et al. 1988), Tinca tinca (Kamler et al. 1995), and Dicentrarchus labrax (Rønnestad et al. 1998) among others. In a tropical demersal fish, rabbitfish (Siganus randalli) the factorial scope for post-hatching activity was expressed as a ratio between oxygen consumption values after and before hatching (0.67 and 0.08 μg O2 indiv.−1 h−1, respectively) (Nelson and Wilkins 1994); an increase by a factor 8.4 occurred over a short time (less than 1 h). The authors conclude that increased activity of larvae newly hatched from marine demersal eggs could facilitate their dispersal. In a cold stenothermic species, Salvelinus alpinus, the factorial scope for post-hatching activity was the lowest (1.6) at a low temperature (4°C) and low oxygen content (20% air saturation), and the highest (about 5) at high temperature (8°C) and 50–100% air saturation, which stimulated locomotor activity (Gruber and Wieser 1983). However, in Gadus morhua (Davenport and Lønning 1980; Lønning et al. 1988; Finn et al. 1995a), in three other marine species with pelagic eggs (Hippoglossus hippoglossus, Scophthalmus maximus, and Microstomus kitt) (review by Finn et al. 1995a), and in Chondrostoma nasus (Fig. 6), the temporary accelerations of respiratory rate at hatching were not found. Oxygen consumption may accelerate at hatching or may remain constant. On the other hand, burst of respiration at hatching is a short-lived event, which can be underestimated, or even missed when the frequency of measurements is not high enough. The respiratory acceleration at hatching is attributable to, both, the increased mobility of embryo, and discarded respiratory barriers: egg capsule (chorion) and perivitelline fluid. In Perca fluviatilis embryonic mobility reached a peak just before hatching (Korzelecka et al. 1998). An abrupt decrease in glycogen content in embryos of Cyprinus carpio and Carassius carassius prior to hatching was reported (review by Chepurnov 1989). Another factor contributing to the increase of oxygen uptake just after hatching may be repayement of oxygen debt. Measurements of oxygen consumption of Salmo salar embryos in intact eggs and of embryos freed from egg capsules demonstrated that egg capsule suppresses oxygen consumption, especially just before hatching (Buznikov 1961). A decrease of respiratory quotient (RQ) below 0.6 and a depression of lactate in Cyprinus carpio at hatching were shown by Kamler et al. (1974). In Gadus morhua a temporary decrease of lactate was also observed at hatching (Finn et al. 1995a).
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig6_HTML.gif
Fig. 6

Ontogenetic changes of absolute rate of oxygen consumption in Chondrostoma nasus embryos and larvae solely on yolk at four temperatures. Initial egg dry weight 2.14 mg. Based on raw numerical data from Kamler et al. (1998)

In further ontogeny of yolk-feeding larvae a low level and lack of change in lactic acid were indicative of fully aerobic metabolism (Finn and Fyhn 1995: Gadus morhua and Scophthalmus maximus; Finn et al. 1995a: Gadus morhua; Finn et al. 1995c: Hippoglossus hippoglossus; Finn et al. 1995d: Scophthalmus maximus). Yolk-feeding larvae of Salmo salar have reduced anaerobic performance as compared to parr, their swimming is fuelled aerobically (Wakefield et al. 2004).

Oxygen consumption of yolk-feeding larvae continues to increase until a peak, which coincides with food searching activity. For example, in unfed larvae of Chondrostoma nasus this peak (Fig. 6) was observed shortly after first food was ingested by 50% of their fed siblings in a parallel experiment (days 41, 23, 15 and 10 post activation at 10, 13, 16 and 19°C, respectively, Kamler et al. 1998). Afterwards oxygen consumption rate decreases when endogenous sources of energy are depleted in larvae receiving no external food (Figs. 57). At the same time in fed larvae oxygen consumption increases after a slight depression during transition from yolk to external feeding. The depression was demonstrated for fed larvae of salmonid fishes (Ryzhkov 1976), Morone saxatilis (Eldridge et al. 1982), Russian sturgeon Acipenser gueldenstaedtii (Korzhuev’s data in Ozernyuk 1985), Oreochromis niloticus (De Silva et al. 1986) and Salmo trutta (Raciborski 1987).
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig7_HTML.gif
Fig. 7

Ontogenetic changes of absolute rate of oxygen consumption in Oncorhynchus mykiss embryos and larvae solely on yolk at four temperatures. Initial egg dry weight 10.4 mg. Based on raw numerical data from Kamler and Kato (1983)

Thus, the ontogenetic sequence in the absolute oxygen consumption of fish solely on yolk follows a general pattern. A low level occurs at initiation of embryogenesis, then an increase through hatching to reach peak at the time when larvae are ready for first feeding, and, finally, a decrease is observed under absence of external food. This pattern is followed irrespective of taxonomic position and temperature, and is preserved in fish that differ in oxygen consumption level by several orders of magnitude. For example, in yolk-feeding Oncorhynchus mykiss the maximum oxygen consumption was about 30 mm3 O2 individual−1 h−1 (Fig. 7), while in Scophthalmus maximus it peaked at about 0.13 mm3 O2 individual−1 h−1 (Finn et al. 1996). Detailed descriptions of the ontogenetic sequence are provided for Cyprinus carpio (Belyaeva 1959; Kamler 1976; Kaushik et al. 1982), Oncorhynchus mykiss (Winnicki 1968), Micropterus salmoides (Laurence 1969), Acipenser gueldenstaedtii (Korzhuev’s data in Ozernyuk 1985), Salmo trutta (Winnicki 1968; Raciborski 1987), Tinca tinca (Kamler et al. 1995), Gadus morhua (Finn and Fyhn 1995; Finn et al. 1995a), and Scophthalmus maximus (Rønnestad et al. 1992a; Finn and Fyhn 1995; Finn et al. 1995d; Finn et al. 1996; Finn and Rønnestad 2003), among others, see also Figs. 57. Usually only minor differences occur among different species in the shape of their respiration-on-age curve, but in Microstomus kitt (Rønnestad et al. 1992b) and Sparus aurata (Rønnestad et al. 1994) oxygen consumption nearly levelled after hatching, and a small hump was observed at the onset of feeding ability.

Several attempts were made to quantify the relationship between oxygen consumption and age in yolk-feeding fish taking no external food. In weatherfish, Misgurnus fossilis (Nejfakh 1960) and Gadus morhua (Davenport and Lønning 1980) oxygen consumption between egg activation and hatching was reported to increase nearly linearly. During 205 days that elapsed between egg activation and hatching of Coregonus albula metabolic rate (R, mJ  indiv.−1 d−1, measured by direct calorimetry at 3–5°C) increased curvilinearly:
$$ R{\text{(}}\tau {\text{) = 20}}{\text{.332}} + {\text{0}}{\text{.001 }}\tau ^{{\text{2}}} {\text{ (}}r^{{\text{2}}} {\text{ 0}}{\text{.984, }}P = {\text{0}}{\text{.000)}} $$
(25)
from 20 mJ indiv.−1 d−1 at egg activation to 70 mJ indiv.−1 d−1 at hatching (Koho et al. 2002).
An exponential model:
$$ R{\text{(}}\tau {\text{)}} = R_{{\text{0}}} {\text{ e}}^{{{\text{b}}\tau }} $$
(26)
has been widely used to describe age (τ, days)-induced changes in oxygen consumption (R, mm3 O2 individual−1 h−1) during different sectors of yolk feeding between egg activation and peak of respiration. Earlier works were reviewed by Devillers (1965). Further, successful applications of the model (26) were reported by Laurence (1973) for tautog (Tautoga onitis), Kamler (1976) for Cyprinus carpio, Kamler and Kato (1983) for Oncorhynchus mykiss, Ozernyuk and Lelyanova (1985) for Brachydanio rerio, and Kamler et al. (1998) for Chondrostoma nasus. Usually, the model provided a high statistical fit. For example, in O. mykiss determination coefficient was 0.90 for the whole sector from the beginning of ontogeny to the onset of feeding ability (Kamler and Kato 1983), in Hippoglossus hippoglossus embryos (from egg activation to hatching) the determination coefficient amounted to 0.98 (Finn et al. 1991), in yolk and mixed feeding period of C. nasus at four temperatures it ranged from 0.96 to 0.99 (Kamler et al. 1998).
A third order polynomial:
$$ {R(}\tau {\hbox{) = a + b}}\tau {\hbox{ + c}}\tau ^{{\hbox{2}}} + {\hbox{ d}}\tau ^{{\hbox{3}}} $$
(27)
with parameters a, b, c and d amounting to 0.086, 0.526, 0.089 and −0.008, respectively, well described (determination coefficient = 0.98) oxygen consumption (nmol O2 individual−1 h−1) of Scophthalmus maximus throughout the entire yolk-feeding period at 15°C (Finn et al. 1996). In an earlier study on Hippoglossus hippoglossus yolk-feeding larvae Finn et al. (1995c) used separate 3rd order polynomials to describe an ascending and a descending sector of the respiration-on-age curve. Very high determination coefficients (0.99–1.00) were provided by two pairs of polynomials measured in H. hippoglossus exposed to light or to darkness. Another polynomial well described routine oxygen consumption increase between blastopore closure and maximum tissue weight in Oncorhynchus tshawytscha; combined action of age and temperature explained 96.4% of variance (Rombough 1988a).
In a bioenergetic model for the entire endogenous feeding period of Oncorhynchus mykiss, Chondrostoma nasus, Clarias gariepinus, Cyprinus carpio and Tinca tinca Jaworski and Kamler (2002) computed the maximum absolute metabolic rate at the optimum temperature and at age τ (Rmaxτ) from the relation to body weight:
$$ R_{{{\text{max}}\tau }} = {\text{a}}W_{\tau } ^{{\text{b}}} $$
(28)
where Wτ is body weight without yolk (tissue weight) at age τ, and a and b are constants. Age-induced changes in body weight were computed as shown in the section “Body growth” above.
Mass-specific oxygen consumption rate of Salvelinus alpinus embryos before hatching was below the size-corrected basal rate of juvenile S. alpinus; newly hatched larvae consumed less oxygen than slowly swimming adults (Gruber and Wieser 1983). In yolk-feeding Gadus morhua mass-specific oxygen consumption rates (mg O2 h−1g dry weight−1) were lower than in exogenously feeding larvae (Hunt von Herbing and Boutilier 1996). The actual absolute oxygen consumption rates measured in larvae of four freshwater fish species at the end of endogenous feeding period were lower than the size- and temperature-corrected oxygen consumption rates for postlarval fishes by factors 1.1–2.2 (Table 7). Humphreys (1979) analysed 235 published energy budgets for natural animal populations. Metabolic costs per unit production in these populations were about 10 times higher than in yolk-feeding Oncorhynchus mykiss (Kamler 1992). Summing up, dissipation of energy for metabolic processes is limited in yolk-feeding fish.
Table 7

Comparison between the actual absolute oxygen consumption rates measured in unfed fish larvae at the end of endogenous feeding period (at the final yolk resorption, Re) and the oxygen consumption rates predicted from Winberg’s (1956) equations for juvenile and adult fishes (R = aWb, where R is in cm3O2 indiv.−1 h1 at 20°C and W is wet weight of larvae at Re, g). The predicted values were temperature-corrected according to Backiel (1977)

Species (and sourcea)

Temp. (°C)

Wet wt at Re(mg)

Winberg’s equation

R (mm3indiv.−1h−1)

a

b

Actual

Predicted

Tinca tinca (1)

22

0.723

0.230

0.79

0.52

0.90

Cyprinus carpio (2)

22

1.500

0.343

0.85

0.90

1.62

Chondrostoma nasus (3)

16

7.267b

0.336c

0.80c

4.10

4.54

Oncorhynchus mykiss (4)

12

95.533b

0.498d

0.76d

18.00

38.73

aSources: (1)—Kamler et al. (1995); (2)—Kamler (1976); (3)—Kamler et al. (1998); (4)—Kamler and Kato (1983, offspring of a 3-years old female).

bComputed from dry weight, assuming 15% of dry weight in wet weight.

cEquation for cyprinids (Carassius spp., Tinca tinca and Cyprinus carpio excluded)

dEquation for salmonids

Several mechanisms can be considered to explain reduced metabolic expenditures in fish early life. Cost of growth (Rg) and energy required for maintenance (Rm) were postulated to be the main components of total energy expenditure in yolk-feeding larvae of Clarias gariepinus (Conceição et al. 1993). Earlier postulations that the metabolic rate increases with increasing rate of growth (review for postlarval fish by Jobling 1985) have been challenged for fish embryos and early larvae, growing at very high rates. In these fish the metabolic rate not always increases with increasing rate of growth. The rates of metabolism decoupled from the rates of growth were shown for Oncorhynchus tshawytscha from fertilisation to complete yolk resorption by Rombough (1994), and reviewed by Pedersen (1997) for young externally fed fish larvae. In yolk-feeding larvae of Clarias gariepinus costs of growth were low (63.5 mmol ATP g−1 dry weight), approaching the minimum cost of protein synthesis (50 mmol ATP g−1 protein sythesised) (Conceição et al. 1997a). A hypothesis of reduced metabolic costs of protein synthesis in young fish larvae was considered (Pedersen 1997).

Endogenously feeding fish experience no feeding stimuli, their costs of feeding (Rf) are minimised by the lack of foraging and ingestion of external food, and, probably, by minimised costs of absorption of endogenous food. The scope for activity (Ra) is strongly reduced. A prevalent mode of locomotion of yolk-feeding Gadus morhua was burst-swimming, which is a way to dissipate boundary layer and reduce oxygen gradient around the body (Hunt von Herbing and Boutilier 1996).

An exponential increase of ammonia excretion was observed during the few final days of embryogenesis of H. hippoglossus (Terjesen et al. 1998). In embryos of Gadus morhua ammonia was gradually retained in yolk, which, because of a pH of about 5, is probably the major site of ammonia ion entrapment (Fyhn and Serigstad 1987). Exchange of heavy potassium ions for lighter ammonia ions probably plays a role in buoyancy in marine fish eggs (see reviews by Rønnestad and Fyhn 1993; Terjesen et al. 1998). At hatching, the heavy egg capsule is discarded. A post-hatch period of elevated ammonia excretion was observed, for example, in Gadus morhua (Fyhn and Serigstad 1987; Finn et al. 1995a), Scophthalmus maximus (Rønnestad et al. 1992a), Microstomus kitt (Rønnestad et al. 1992b) and Hippoglossus hippoglossus (Finn et al. 1995c; Terjesen et al. 1998). Afterwards ammonia reaches a low, stable value. A delayed burst of ammonia excretion was reported for H. hippoglossus which spawn at great depths and larvae need more time to float to near-surface layer where they can start external feeding (Terjesen et al. 1998). Summing up, the general pattern, ammonia accumulation during embryogenesis and release after hatch, is similar among species. However, the timing and rate of release are species-specific. Reviews for marine fish are given by Rønnestad and Fyhn (1993) and Finn et al. (1995c).

In an air-breathing fish, Clarias gariepinus, higher percentages of urea in total nitrogen excreted were observed in embryos and larvae than in adult fish, a probable explanation is detoxification of ammonia as a response to uncertain water availability at their spawning grounds (Terjesen et al. 2001).

Metabolic fuels and their sequence in yolk-feeding fish

In earlier studies carbohydrates were reported to be the main energy source for catabolism during early embryogenesis of fishes prior to gastrulation, a shift towards lipids and protein was suggested to occur in further ontogeny (Milman and Yurowitzky 1973; Terner 1979; Boulekbache 1981; Kaushik et al. 1982; Gosh 1985). A burst of interest in metabolic substrates of yolk-feeding marine fish was seen in late 1980’s and early 1990’s to improve the understanding of nutritional requirements in larvae of commercially important fish species.

There is by now extensive evidence that carbohydrates are, indeed, the primary energy substrate during a very short period of cellular divisions of the blastodisc in several freshwater and marine fishes (reviews by Finn et al. 1995d and Finn et al. 1996). Only a minor part of total aerobic energy dissipation derives from carbohydrates (see examples for Gadus morhua and Scophthalmus maximus in Fig. 8 and Table 8). In several species carbohydrates were not found to serve as an initial energy substrate (examples in Fig. 8 and Table 8). It remains to be seen whether carbohydrates really are not used as fuels in these species, or they were undetectable during that short period of very low metabolic rate (Figs. 57).
https://static-content.springer.com/image/art%3A10.1007%2Fs11160-007-9070-x/MediaObjects/11160_2007_9070_Fig8_HTML.gif
Fig. 8

Schematic representation of metabolic fuels sequence in three species of marine fish embryos and larvae. Leiostomus xanthurus has a relatively large oil globule, volume ratio of the oil globule to yolk was 1:40 in newly spawned eggs, yolk was resorbed completely on about 5th day post fertilisation, oil globule on about 8th day (Fyhn and Govoni 1995). Brevoortia tyrannus has a smaller oil globule and larger yolk, the ratio was 1:400, both were resorbed on about 5th day, no data for protein were available for that species (Fyhn and Govoni 1995). Gadus morhua has no visible oil globule, complete yolk resorption occurred on 23 to 25th day post fertilisation (Finn et al. 1995a). H – hatching time, S – time of first feeding, larvae were not fed during experiments

Table 8

Relative contribution (% of total oxygen consumed) of substrates to catabolism during endogenous feeding period of marine fish with pelagic eggs: Scophthalmus maximus, Solea senegalensis, Gadus morhua and Hippoglossus hippoglossus. Species with (+) or without (−) oil globule and with different levels of total lipids (% dry matter) in initial eggs are compared. Summary for embryos, yolk-feeding larvae and whole yolk-feeding period

Species

Eggs

Substrates

Substrate contribution (% of O2)

Oil glob.

Lipids

Embryos

Larvae

Whole period

S. maximusa

+

25–32a

Glycogen

2.0

0.0

0.4

FAA

84.0

10.0

23.8

∑Lipids

14.0

58.0

50.2

Protein

0.0

32.0

25.7

S. senegalensisb

+

>15b

Glycogen

FAA

86.8

41.4

61.5

∑Lipids

21.3

47.4

35.9

Protein

0

13.1

7.3

G. morhuac

10–12d

Glycogen

2.0

0.0

1.0

FAA

75.0

32.0

47.0

∑Lipids

23.0

37.0

32.0

Protein

0.0

31.0

20.0

H. hippoglossuse

11–15d

Glycogen

FAA

21.0

∑Lipids

26.0

Protein

53.0

– No data

aFinn et al. (1996)

bParra et al. (1999)

cFinn et al. (1995a)

dLønning et al. (1988)

eFinn et al. (1995c)

Free amino acids enter the embryonic/larval body from the yolk and are used for body protein synthesis or as catabolic fuel. The importance of free amino acids (FAA) as an energy substrate in metabolism was suggested from many studies. Probably amino acids were the major energy fuels during embryogenesis (from early cleavage to hatching) of bastard halibut, Paralichthys olivaceus (Seoka et al. 1998). In Microstomus kitt FAA were not used as a metabolic fuel during first three days post-activation, afterwards their use increased to 60% of total aerobic energy dissipation 9 days after activation and to 100% at hatching. However, in yolk-feeding larvae the percentage of energy dissipated derived from FAA decreased to about 50–75% (Rønnestad et al. 1992b). Free amino acids were the major fuel in embryogenesis of Scophthalmus maximus, except for the first day post activation when they accounted for only 10% of energy. During the second day 100% of energy dissipated derived from FAA, and 63% at hatching (Rønnestad et al. 1992a).

In contrast, after hatching of S. maximus lipids were the dominant fuel (Rønnestad et al. 1992a). The sequence of energy sources for catabolism in S.maximus embryos and yolk-feeding larvae was confirmed by Finn and Fyhn (1995). They pointed out that during the whole yolk-feeding period the contribution of total amino acids (free + protein−bound) to total energy dissipation was 40% and that of total lipids—60%. Another study on S. maximus (Finn et al. 1996) reports 49.5% and 50.2% for total amino acids and total lipids, respectively (Table 8). Free amino acids were a significant (60–70%) endogenous energy source during embryonic development of Sparus aurata, whereas neutral lipids derived from the oil globule were the dominant metabolic energy substrate (80–90%) in yolk-feeding larvae (Rønnestad et al. 1994). Free amino acids were the main energy source for embryos and early yolk-feeding larvae of Dicentrarchus labrax, while neutral lipids derived from oil globule were an important energy substrate afterwards (Rønnestad et al. 1994). Similarly two main periods for endogenous nutrient utilization were observed in Theragra chalcogramma (Ohkubo et al. 2006). The first period coincided with embryogenesis, from egg activation to hatching (day 18th post fertilisation at 5°C): during the first 6 days no substantial changes of the FAA content was measured, while they decreased to 28% of their initial level at hatching. During the second period the major utilisation of protein (lipovitellin) and of lipids (phospholipids and triacylglycerols) was found in yolk-feeding larvae.

Lipids were the main metabolic fuel during yolk feeding of Clarias gariepinus (Conceição et al. 1993), caloric value of C. gariepinus eggs was 27.9 J mg−1 dry weight (Kamler et al. 1994), a high value compared to remaining 24 species listed in Kamler (2005). In embryos of northern bluefin tuna (Thunnus thynnus) the main energy source was triacylglycerol (Takii et al. 1997a), caloric value of T. thynnus eggs, 30.1 J mg−1 dry weight (Takii et al. 1997a), was one of two highest values among 25 species listed in Kamler (2005). In Leiostomus xanthurus the ratio of oil globule to yolk was 10 times greater than in Brevoortia tyrannus (data by Fyhn and Govoni 1995 in Fig. 8). Prior to tissue autolysis the main substrate oxidised in L. xanthurus were lipids (70%) while 30% of energy derived from FAA, whereas in B. tyrannus the corresponding values were 35% lipids and 65% FAA (Fig. 8). In contrast, Gadus morhua is a species whose eggs have a low caloric value of 19.5 J mg−1 (Finn et al. 1995b), the lowest caloric value among 25 species listed in Kamler (2005). In G. morhua embryos and yolk-feeding larvae the preferred metabolic substrates were FAA (73%), whereas lipids contributed to only 27% of aerobic energy dissipation (Finn and Fyhn 1995) or to 32% (data by Finn et al. 1995a in Table 8). Protein is mobilised early, and is a significant metabolic substrate (Fig. 8). Free amino acids, mainly alanine, leucine, serine, isoleucine, lysine and valine were used as a major energy substrate in Gadus morhua embryos and yolk-feeding larvae (Fyhn and Serigstad 1987). Thus, a clear reliance on amino acids (both, free and protein-bound) is seen during endogenous feeding of G. morhua (Table 8, Fig. 8 and Finn et al. 2002). In newly hatched Hippoglossus hippoglossus larvae FAA were the sole metabolic fuel, but, at the same time, FAA were polymerised into protein. Hereafter metabolism became fuelled predominantly by protein-bound AA, joined initially by polar lipids (mainly phosphatidyl choline) and then by neutral lipids (mainly triacylglycerol). Amino acids (FAA + protein-bound AA) were the dominant energy substrate (relative contribution of 74% of oxygen consumed). Hippoglossus hippoglossus eggs have low lipid content and no oil globule (data by Finn et al. 1995c in Table 8).

Later protein becomes progressively used for catabolic purposes. That was demonstrated for Clarias gariepinus (Conceição et al. 1993), Dicentrarchus labrax (Rønnestad et al. 1998) and Senegal sole, Solea senegalensis (Parra et al. 1999). Amino acids were recruited from protein stored in yolk of Hippoglossus hippoglossus larvae 12–32 days post-hatch when about 20% of yolk remained unresorbed (Rønnestad et al. 1993). After attainement of first feeding ability under absence of external food body protein are increasingly important as a catabolic fuel, this was suggested for Scophthalmus maximus (Finn and Fyhn 1995 and Finn et al. 1996).

Finn and Rønnestad (2003) quantified the mass fraction of amino acids catabolised when carbohydrate is the co-substrate:
$$ {\text{K}}_{{{\text{AA/CHO}}}} {\text{ = 3}}{\text{.221 (NQ) + 2}}{\text{.571 (NQ)}}^{{\text{2}}} {\text{ + 4}}{\text{.064 (NQ)}}^{{\text{3}}} $$
(29)
and the mass fraction of amino acids catabolised when lipid (neutral or polar) is the co-substrate:
$$ {\text{K}}_{{{\text{AA/LIPID}}}} {\text{ = 8}}{\text{.069 (NQ)}} - {\text{25}}{\text{.267 (NQ)}}^{{\text{2}}} + {\text{36}}{\text{.732 (NQ)}}^{{\text{3}}} $$
(30)
where NQ is the apparent nitrogen quotient (the molar ratio of NH3 produced to O2 consumed).

In conclusion, a general pattern of ontogenetic sequence of catabolic substrates for fish embryos and unfed endogenously feeding larvae may be summarised as follows. In contrast to other vertebrates, fish obtain a high proportion of metabolic energy from nitrogen compounds (FAA and protein). Typically, fish ontogeny begins with a short period of carbohydrate use, soon switched to FAA as an important fuel. Thus, in early ontogeny small molecules are the energy substrates that are catabolised quickly and readily. Lipid mobilisation follows after hatching in response of increased energy demand. At the onset of first feeding lipids provide energy for swimming activity. The second peak of amino acid catabolism occurs after depletion of endogenous reserves, when body protein-bound amino acids are mobilised. Timing and extent of these shifts is species-specific, modified by egg properties. Lipids are important metabolic fuels in species whose eggs have an oil globule, high lipid level and high caloric value, whereas amino acids are the major energy substrates in species having low lipid contents, low caloric value of dry matter and no oil globule. A reliance on amino acids (both, FAA and protein-bound) was extended on first few weeks of external feeding of Gadus morhua (Finn et al. 2002), deamination of dietary FAA was also reported for first-feeding Oncorhynchus mykiss (Dabrowski et al. 2003). Knowledge of metabolic substrates helps to optimise formulated diets for first-feeding larvae.

Factors affecting metabolic rate

Three major factors governing metabolic rate, size, temperature and oxygen content, will be considered here, as well as evidence indicating response of metabolic rate to light and magnetic field.

Two aspects of the size effect on oxygen consumption rate will be considered. An effect of the whole egg size on the oxygen consumption rate of eggs will be shown, and an effect of tissue size on metabolic rate in developing embryos and yolk feeding larvae will be discussed.

In an inter-specific analysis egg size explained 99% of variance in absolute oxygen consumption rates of initial eggs at activation, whereas mass-specific oxygen consumption rate (mmg−1 h−1) was unrelated to egg size (Table 6). At an intra-specific study Bang et al. (2004) used an oximetric microtechnique to monitor oxygen consumption of individual Danio rerio embryos. From individual trajectories of oxygen consumption cumulated values were computed for the age from 24 to 75 h (just before hatching). The cumulated values ranged 0.261–0.462 μmol O2 individual−1, egg volume explained 26% of the among-individual variation.

Within an animal species the relationship between absolute oxygen consumption rate and body size is described by the Eq. (28). In juvenile and adult fish oxygen consumption rate typically scales to body mass ≈ 0.8 (Winberg 1956). However, there is by now extensive evidence that high b exponents (≥1) can occur in early life of several (but not all) fish species (reviews by Kamler 1992; Rombough 1988b and c, 2006). During yolk-feeding period the weight exponent b ranged from 0.75 to 1.83 (examples in Table 9). Several values of b were higher than 0.8, and approached or exceeded 1.0, showing a nearly direct dependence of metabolism on body mass. Similarly, in equations describing combined effect of body size and temperature on respiration of Oncorhynchus mykiss between final epiboly and maximum metabolic rate on yolk the weight exponents were 1.05 and 1.25 for wet and dry weights, respectively (Rombough 1988c).
Table 9

The b exponent in the relationship between absolute oxygen consumption rate (R) and tissue weight (W) during fish endogenous feeding period, R = aWb

Species

Period

b

Source

Oncorhynchus mykiss

Entire yolk-feeding period

0.75

Jaworski and Kamler (2002)

Hippoglossus hippoglossus

Dark-reared yolk-feeding larvae, 3–33a days post hatch

0.77

Finn et al. (1995c)

Clarias gariepinus

Entire yolk-feeding period

0.79

Jaworski and Kamler (2002)

Cyprinus carpio

Entire yolk-feeding period

0.83

Jaworski and Kamler (2002)

Tinca tinca

Entire yolk-feeding period

0.90

Jaworski and Kamler (2002)

H. hippoglossus

Dark-reared yolk-feeding larvae, 10–33a days post hatch

0.98

Finn et al. (1995c)

Oncorhynchus tshawytscha

From blastopore closure to maximum tissue weight

1.05

Rombough (1988a)

H. hippoglossus

Light-reared yolk-feeding larvae, 3–33a days post hatch

1.08

Finn et al. (1995c)

Chondrostoma nasus

Entire yolk feeding period

1.09

Jaworski and Kamler (2002)

Clupea harengus

Yolk-feeding larvae

1.10

Holliday et al. (1964)

Blaxter and Hempel (1966)

H. hippoglossus

Light-reared yolk-feeding larvae, 10–33a days post hatch

1.25

Finn et al. (1995c)

Salmo trutta

Yolk-feeding larvae

1.83

Raciborski (1987)

aAge of maximum tissue weight

Several reasons have been advanced to explain the ontogenetic shift in metabolic scaling. Von Bertalanffy (1964) suggested that metabolic types (i.e., size-dependencies of metabolism) are associated with growth types. A direct relation of the absolute respiration rate on body mass (b ≈ 1) is associated with an exponential type of growth, while 0.66 < b < 1 is associated with S-shaped growth. The association of b ≈ 1 and exponential growth was found in externally feeding larvae (review by Kamler 1992). In yolk-feeding period of some fish absolute respiration rate nearly directly related to body mass (b ≈ 1) is not surprising, because the exponential growth type is common prior to the inflection of growth curve (see chapter “Body growth” above). Exponential growth of body dry weight in Hippoglossus hippoglossus yolk-feeding larvae up to the time of maximum tissue mass was accompanied by a proportional increase of respiration with weight (Finn et al. 1995c). The authors think that probably mitochondria concentration could be maintained proportionally. Other hypotheses aimed to explain the ontogenetic shift in metabolic scaling were also considered (reviews by Calow 1984; Rombough 2006).

Besides size, temperature is a factor strongly affecting metabolic rate, and, as such, it received much attention. Generally, in fully acclimated poikilothermic animals metabolic rate (R) is exponentially related to temperature (t) (Winberg 1983, 1987; Cossins and Bowler 1987):
$$ R{\text{ = }}a{\text{e}}^{{bt}} $$
(31)
The acceleration of the metabolic rate by temperature is usually described by the van’t Hoff temperature coefficient Q10met for a 10°C difference in temperature:
$$ Q_{{{\text{10met}}}} = {\text{(}}R_{{\text{2}}} {\text{/}}R_{{\text{1}}} {\text{)}}^{{{\text{10/(}}t_{{\text{2}}} - t_{{\text{1}}} {\text{)}}}} $$
(32)
or, rarely, by the Q1met coefficient for a 1°C difference in temperature (Winberg 1987):
$$ Q_{{{\text{1met}}}} = Q^{{{\text{0}}{\text{.1}}}}_{{{\text{10met}}}} $$
(33)
In the case of exponential temperature dependence (Eq. 31) the temperature coefficients remain constant over a temperature range:
$$ {\text{ln }}Q_{{{\text{10met}}}} {\text{ = 10}}b $$
(34)
$$ {\text{ln }}Q_{{{\text{1met}}}} = b $$
(35)
The single temperature coefficient, Q10met = 2.25, is similar to the value 2.3 proposed for fish by Brett and Groves (1979).
However, the relationship deviates from exponential when metabolic rate becomes independent of temperature, forming a plateau over the zone of near-optimum temperatures (reviews by Stroganov 1956; Duncan and Klekowski 1975; Cossins and Bowler 1987; examples for larvae of three cyprinids in Wieser and Forstner 1986). Another popular deviation of metabolic rate from the exponential dependence on temperature is depicted by the empirical Krogh’s “normal curve” (Ege and Krogh 1914; Krogh 1916). It was based on measurements of standard (resting) metabolism in poikilotherms from temperate zone and anesthesised homoiotherms. According the Krogh’s “normal curve” the temperature coefficients Q10met are not constant but decline with temperature:

t range (°C)

0–5

5–10

10–15

15–20

20–25

25–30

Q10met

10.9

3.5

2.9

2.5

2.3

2.2

A quantification of the Krogh’s “normal curve” was proposed by Backiel (1977):
$$ R_{{\text{t}}} = R_{{{\text{20}}}} {\left( {{\text{0}}{\text{.3e}}^{{{\text{0}}{\text{.071}}t}} - {\text{0}}{\text{.24}}} \right)} $$
(36)
However, the Krogh’s “normal curve” is not ubiquitous. Scholander et al. (1953) measured resting metabolic rate in tropical and arctic exogenously feeding fish and crustaceans at their natural temperatures (high and low, respectively). Experimental Q10met remained within 2–3 irrespective of temperature range. In animals inhabiting steady warm environments (e.g., tropical seas), cold but fluctuating environments (e.g., arctic terrestrial environments) and temperate zone the Q10metExp values were consistent with those expected from the Krogh’s “normal curve” (Q10metKrogh). In contrast, in animals living in steady cold environments (arctic fish and crustaceans) the Q10metExp values were lower than the respective Q10metKrogh values (Scholander et al. 1953). But also in temperate climate there are biotopes inhabited by cold stenotherms in which Q10metExp can be depressed compared to Q10metKrogh values (e.g., Plecoptera larvae in mountain waters, Kamler 1971).

All these general relationships were described for exogenously feeding animals. How do they apply to fish embryos and yolk-feeding larvae? Temperature-induced changes in ontogenetic course of oxygen consumption rates are exemplified in Figs. 57 for early developmental stages of three fish species. Similar shift in age/respiration curves in Oncorhynchus mykiss at 6, 9, 12 and 15°C were reported by Rombough (1988c). It is obvious that fish of equal “calendar” age incubated at higher temperature consume more oxygen. However, the observed response is a combined result of, both, direct effect of temperature on metabolic rate, and an indirect one through ontogenetic rate, which is fast during early development and is accelerated by high temperature (review by Kamler 2002).

The acceleration of ammonia excretion rate by temperature was described by Q10metExp = 2.36 in Cyprinus carpio embryos for the range of 13–28°C (Kaushik et al. 1982), i.e., for a temperature range remaining within the viable range for embryogenesis (10–33°C, review by Kamler 2002). Thus Q10metExp did not deviate much from Q10metKrogh (2.42 for 13–28°C). In yolk-feeding Scophthalmus maximus Q10metExp values for both, oxygen consumption and ammonia excretion amounted to 2.60 ± 0.23 (±95% C.L., Finn and Rønnestad 2003) in response to short-term temperature fluctuations between 13 and 17°C. This temperature range is within the viable range for S. maximus embryos (10–23°C, Weltzien et al. 1999). The Q10metExp was similar to the Q10metKrogh = 2.67 for the same temperature range. In embryos of beluga, Huso huso (Gershanovich 1983) and Misgurnus fossilis (Ozernyuk 1985) oxygen consumption rates were related to temperature according the Krogh’s curve. In Chondrostoma nasus during three developmental intervals (from-to: Fe-6E, Fe-H and Fe-Re) at circum-optimal ranges 13–16°C and 16–19°C the Q10metExp values did not deviate much from the Q10metKrogh ones, while elevated Q10metExp values 4.05, 4.45 and 7.86, respectively, were found at suboptimal temperatures 10–13°C (Kamler et al. 1998).

However, in some cold-adapted yolk-feeding fish the actual Q10met values were depressed as compared with those derived from the Krogh’s curve. In Coregonus albula embryos (viable range 0.5–12°C, Luczynski and Kirklewska 1984) Q10metExp for heart rate was 2.73 for 1–9°C temperature interval (Koho et al. 2002), well below Q10metKrogh = 5.66 at respective temperature range. In newly hatched larvae of another cold-adapted fish, Salvelinus alpinus (embryonic viable temperatures range 0.5–8°C, Humpesch 1985) mean Q10metExp for mass-specific standard oxygen consumption rate was 3.4 for temperature range 2–7°C (Huuskonen et al. 2003). Thus, the Q10metExp was markedly lower than the Q10metKrogh, 5.72 for respective temperatures.

A summary of Q10met values for fish embryos and larvae (mostly freshwater species) yielded an average of 3.0, range 1.5–4.9 (Rombough 1988b). He concluded that fish embryos and larvae are more stenothermal than juveniles and adults in which Q10met typically is about 2. In yolk-feeding marine fish the Q10met value was 2.5 ± 0.78 (n = 9) (two upper extreme values of 4.77 and 6.40 excluded) (Finn et al. 1995c).

Thus, it is well established that metabolic rates of yolk-feeding fish are temperature-related similarly as they are in exogenously feeding animals. At circum-optimal temperatures Q10met values remain within the range of 2–3. While temperate species are related to temperature more or less according to the Krogh’s curve, cold water stenotherms have thermal optimum shifted to the lower range. Their Q10met values are below those predicted from Krogh” “normal curve” for respective temperatures, but remain at the level of 2–3 within respective optimum temperatures.

In Scophthalmus maximus the temperature coefficient Q10metExp did not change with ontogenetic stage (embryos or yolk-sac larvae), light (total darkness or 25–30 μE s−1 m−2), nor with temperature change direction (increase or decrease) (Finn and Rønnestad 2003).

We demonstrated that metabolic rate varies with size in addition to being positively related to temperature. A combined effect of initial egg dry weight (Wd0, mg) and temperature (t, °C) on absolute rate of oxygen consumption at hatch (R, mm3 O2 indiv.−1 h−1) was quantified by Dabrowski et al. (1984) for 11 fish species:
$$ R = {\text{0}}{\text{.1334 }}W^{{{\text{0}}{\text{.6634}}}} t^{{{\text{0}}{\text{.7143}}}} $$
(37)
Similar equations were used by Rombough (1988c) to estimate effects of body size (alternatively: wet or dry weight) and temperature on absolute rate of oxygen consumption (routine metabolism, R, μg O2 indiv.−1 h−1) in Oncorhynchus mykiss embryos and larvae between completion of epiboly and maximum metabolic rate on yolk. Dependence of routine metabolism of Clarias gariepinus embryos (R, mol ATP fish−1 h−1) on wet weight of embryonic tissue (Ww, g) and temperature (°C, where t is water temperature, tR is reference temperature) was quantified by Conceição et al. (1993):
$$ R = a{\text{ }}W_{{\text{w}}} ^{b} {\text{ }}Q_{{{\text{10met}}}} ^{{{{\left( {t - {t}_{R} } \right)}} \mathord{\left/ {\vphantom {{{\left( {t - {\text{t}}_{R} } \right)}} {{\text{10}}}}} \right. \kern-\nulldelimiterspace} {{\text{10}}}}} $$
(38)
The combined effect of initial egg size and temperature on daily metabolic dry matter loss of Coregonus albula eggs over incubation period, modelled with a polynomial equation by Koho (2002), accounted for 81% of variance in metabolism.

Oxygen depletion is not considered to have significance to pelagic eggs, but is a limiting factor in embryonic development of demersal species; freshwater species are more tolerant than marine species (reviews by Kryzhanovskij 1949; Balon 1975; Rombough 1988b). Eggs situated in deeper layers of demersal eggs thick masses may suffer oxygen depletion (reviews by Rombough 1988b; Bunn et al. 2000). In Clupea harengus pallasi typically four to six egg layers are deposited, but sometimes more than 12 (Alderdice and Hourston 1985). For spawning intensities of 1, 2, 4–6, 8–10 and 12–16 layers the larval production per unit area was 1.0, 1.8, 3.0, 1.8 and 1.4, respectively, with the decreasing hatching success of 100%, 90%, 60%, 20% and 10%, respectively, due to oxygen depletion. To hatch successfully C. harengus requires at least 20% of air saturation (Alderdice and Hourston 1985).

Constraints on oxygen supply to eggs were reviewed, among others, by Rombough (1988b) and Kamler (1992). It is well established that blood circulation is poorly developed in fish embryos with no haemoglobin in some species. Gas diffusion is considered to play a role, but partial pressure of oxygen in surrounding water is much lower than in air. In addition, oxygen deficits are not uncommon to spawning grounds. At the very beginning of their development embryos rely on oxygen diffusion. Egg capsule (chorion) and perivitelline fluid create barriers. Thus, partly anaerobic conditions prevail in eggs incubated in water. Anaerobic glycolysis is less efficient than aerobic metabolism by a factor of nearly 20 (Hochachka and Somero 1973). In unripe oocytes and mature non-activated eggs of Rutilus rutilus heckeli and bream, Abramis brama (Gosh and Zhukinskij 1979), and during embryogenesis of Salvelinus alpinus (Gnaiger et al. 1981) energy is partly channeled through glycolytic pathways.

Sensitivity to low oxygen level increases as embryogenesis proceeds (Ortner et al. 1988Coregonus sp.; Rombough 1988cOncorhynchus mykiss; reviews by Devillers 1965; Doudoroff and Shumway 1970; Zhukinskij 1986; Rombough 1988b). For example, both, control Chondrostoma nasus embryos incubated in air-equilibred water and those exposed to a strongly reduced oxygen content (about 10% of air saturation) during earliest ontogenesis (from fertilisation to gastrula) revealed high hatching success and post-hatching viability. In the groups exposed from gastrula to eyed stage and from eyed stage to hatching a high proportion of larvae hatched but all were deformed and died during following 1–5 days (Keckeis et al. 1996). In Misgurnus fossilis the rate of anaerobic glycolysis increased prior to gastrulation, during gastrulation it slowed down, and no further anaerobic glycolysis was found (Milman and Yurowitzky 1973).

Several environmental factors, as well as behavioural, morphological, and biochemical adaptations were identified to contribute to oxygen transport into eggs and metabolite removal. Movements of surrounding water (wave action and convection currents) are the environmental factors that mitigate oxygen gradients formed in boundary layers (Ambühl 1959) of deoxygenated water around eggs (Klyashtorin 1982; Rombough 1988b). Ventilation of eggs by parents guarding them is a behavioural adaptation that improves oxygen conditions within the eggs. Oxygen gradient accross the perivitelline fluid is attenuated by embryonic respiratory movements: contractile movements of the blastoderm in early embryos, and neuromuscular movements of the embryo in older ones. The latter were observed after tail bud formation (Reznichenko et al. 1967) or shortly after blastopore closure (Rombough 1988c). Tański et al. (2000) described movements in Esox lucius early embryos: circular movements of blastodisc resulted from contractions of a thin layer of amorphous cytoplasm, whereas later, after blastopore closure, quasi-peristaltic contractions along the yolk sphere originated from endo- and meso-dermal cell structures. Trunk flexures in Salmo salar embryos produced movement of perivitelline fluid; in hypoxia, frequency of these flexures increased which confirms their respiratory nature (Peterson and Martin-Robichaud 1983). Shortly prior to hatching embryos of Scardinius erythrophthalmus became very motile, and rotated around their axis (Korzelecka and Winnicki 1998).

In very early embryos no morphological adaptations that improve oxygen availability were observed. However, salmonid eggs have oil globules concentrated at the animal pole during blastodisk formation. Aggregation of oil globules is probably an adaptation to place the embryo in the upper part of the egg at better oxygen conditions (Korovina 1978). In early embryos gas exchange occurs through body surface. Several respiratory organs are formed further in ontogenesis: ducti Cuvieri, hepatic vein, subintestinal vein, and respiratory vessels in fin folds, gills and gill covers (reviews by Kryzhanovskij 1949; Soin 1968). For example, several morphological adaptations to improve respiratory conditions were found in a Cuban endemic species incubating in warm water, Cichlasoma tetracanthum. They are: a narrow perivitelline space (mean diameter ratio of swollen to unswollen eggs 1.59), a rich vascular system of the yolk sac, branching caudal vein in the postanal lobe of the embryonic finfold, and the subintestinal vein in the preanal finfold (Prokeš et al. 1985).

Ontogenetic changes of the nature and role of respiratory pigments were reviewed by Rombough (1988b). He concluded that embryonic and/or larval hemoglobin could differ from adult hemoglobin in greater oxygen affinity and pH independence. Carotenoids were suggested to provide an oxygen reserve for embryos developing in eggs (reviews by Dabrowski et al. 1987; Balon 1991). Large salmonid eggs and/or eggs developing in poor oxygen conditions (e.g., Oncorhynchus masu) have more carotenoids than small eggs of Plecoglossidae and Osmeridae which are better supplied with oxygen (Korovina 1978). However, some controversy remains about the respiratory role of carotenoid pigments in fish early ontogeny (Rombough 1988b).

Limited access of oxygen to fish embryos makes their respiration an oxygen-dependent process. In progressively decreasing ambient oxygen fish embryos and larvae (similarly as older fish and other aquatic animals) are able to maintain their routine oxygen consumption at a level down to a critical partial pressure of oxygen (Ocrit) below which oxygen consumption decreases (Rombough 1988b). An increase of Ocrit during embryogenesis (data for Oncorhynchus mykiss by Rombough 1988c; review by Rombough 1988b) confirms once again that tolerance to low oxygen decreases during embryogenesis. But in Salvelinus alpinus the total amount of oxygen consumed between eyed stage and hatching was directly related to the ambient oxygen concentration (Gruber and Wieser 1983). A linear relationship (P < 0.001) between oxygen consumption rate (R, mm3 O2 indiv.−1 O2 d−1) and oxygen concentration in water (OC, 4–14 mg O2 dm−3) was found at 1.4 and 2.7°C in Coregonus albula (R = 0.173 + 0,055 OC) and C. lavaretus (R = 0.447 + 0.187 OC) embryos aged 119–145 days post-activation. Oxygen concentration explained 75–85% of the variability in the oxygen consumption rate of these coregonid embryos (Koho et al. 2002). Thus, embryos and larvae of some fish species can be classified as metabolic regulators, some other as conformers.

Premature hatching is a common response to hypoxia (review by Rombough 1988b). At hatching oxygen supply improves markedly. Respiratory barriers, egg capsule and perivitelline fluid are discarded at hatching. Temporary increase of oxygen consumption rate at hatching (see section “Ontogenetic sequence in metabolism” above) can be partly attributed to higher oxygen content in the immediate surroundings of the body. Another response is an abrupt decrease in Ocrit.during hatching. For example, the critical levels dropped to 2–3 mg O2 dm−1 at hatch of Oncorhynchus mykiss, then they declined slowly, and during the second half of larval yolk-feeding stage they reached stabilised levels of 2.27, 3.35, 3.39 and 4.75 mg O2 dm−1 at 6, 9, 12 and 15°C, respectively (Rombough 1988c). Fish early larvae were found to be powered aerobically (Cyprinidae—El-Fiky et al. 1987; Hippoglossus hippoglossus—Finn et al. 1995c). Initially fish have thin and strongly vascularised skin, which makes cutaneous diffusion a major site of gas exchange, and motor activity has a respiratory significance; later fish undergo a transition from cutaneous to gill respiration. Ontogenetic changes in sensitivity to oxygen during larval life are the combined response to increase of oxygen demand, decrease in surface-to-volume ratio, gill formation and hemoglobin first appearance (review by Rombough 1988b). Elevated activity and temperature both result in an increase oxygen demand, which shifts Ocrit towards higher oxygen values. Interestingly, Bagatto et al. (2001) demonstrated that yolk-feeding larvae of Danio rerio exposed to chronic exercise by swimming (swim trained larvae) had a very low Ocrit value of 27 mm Hg, while the control (untrained) larvae had a Pcrit of 48 mm Hg.

Light-reared yolk-feeding larvae of Hippoglossus hippoglossus consumed more oxygen than dark-reared ones. Respiration cumulated for the complete period from hatching (H) to final yolk sac resorption (Re) amounted to 30.1 (light) and 24.5 (dark) μmol O2 indiv.−1. Within that period no significant difference in oxygen consumption between light- and dark-reared larvae were found before the onset of external feeding (S), while later (between S and Re) the larvae in the light consumed more oxygen (Finn et al. 1995c). In embryos and early yolk-feeding larvae of Scophthalmus maximus no differences between oxygen consumption levels in the presence of light (25–30 μE s−1 m−2) and in darkness were found. However, more oxygen was consumed by light-reared older larvae at final steps of yolk resorption (Finn and Rønnestad 2003). Similarly, in Gadus morhua embryos and early yolk-feeding larvae, light did not affect oxygen consumption prior to S, while between S and Re oxygen consumption rate was stimulated by light (Finn et al. 1995a). Appelbaum et al. (1995) reported mortality of yolk-feeding larvae of lingcod (Ophiodon elongatus) to be significantly higher in light tanks than that in dark tanks in which the light intensity was reduced by a factor of 75. The authors concluded that the lower mortality in darkness was due to reduced energy expenditure, because under darkness a limited swimming activity was observed. Reduction of energy expenditure in yolk-feeding Dentex dentex larvae reared in dark, compared to light conditions, was suggested by Firat et al. (2003): reduced locomotor activity was accompanied by reduced yolk absorption and better growth in the dark-reared larvae. Thus, there are ontogenetic changes in the response of metabolic rate to light. In visual predators the response may be explained by adaptation to feeding.

Application of a constant magnetic field of 50 and 70 mT increased heart beat frequency in yolk-feeding Salmo trutta, Esox lucius and Cyprinus carpio as compared to that in a natural geomagnetic field, and the frequency of pectoral fin movements increased in larvae under magnetic field (Formicki and Winnicki 1998). Laboratory experiments with Esox lucius embryos to investigate an effect of 30-min exposure to magnetic fields (constant 4 mT or alternating 15 mT, 7°C) were conducted by Winnicki et al. (2004). In blastula (age 3 D° post-fertilisation) and gastrula (1/2 epiboly, 20 D°) an increased angular velocity and amplitude of ectoplasmic waves were observed under both forms of magnetic field. In embryos just prior to hatching (110 D°) heart rate was accelerated, but somatic motorics was not affected by magnetic fields. The authors hypothesised about increased oxygen demand under intensified activity. A constant artificial magnetic field of 5, 10, 15, 150 and 300 mT enhanced oxygen consumption of Oncorhynchus mykiss embryos (Perkowski and Formicki 1997). In a further experiment (Formicki and Perkowski 1998) an increase of oxygen consumption was observed in O. mykiss embryos exposed to a constant magnetic field of 5 or 10 mT at a constant temperature of 10°C. The effect was more pronounced in periods of intensive morphogenesis.

Fish body size attained on yolk

There is ample data to show that large size of larval prey supresses the risk of predation (Miller et al. 1988; Fuiman 1994). That makes body size a key factor in early life of fish. Growth rate was considered in an earlier section (“Body growth”). Here I will focus on temperature-induced differences in the state, i.e., in the larval size reached at comparable ontogenetic events that are crucial for recruitment.

There exist a variety of responses of larval body size at comparable ontogenetic advancement to incubation temperature. The responses do not seem to differ between cold-water and warm-water species, but different responses were observed between species, within species between ontogenetic states, as well as between various temperature levels for the same species and ontogenetic advancement (Table 10). Four common types of response are listed below.
  1. I.

    Decrease of larval body size at hatching with increasing temperature along the whole temperature range tested was reported from many studies (↓ in Table 10). This was observed in all the six data sets of hatching length compiled by Blaxter (1969), in 10 out of 12 sets of hatching sizes shown by Blaxter (1992), and in 14 out of 33 data sets at H (42%) listed in Table 10.

     
  2. II.

    No effect of temperature on the larval size was detected in several studies over a whole range of temperatures (↔ in Table 10). Size of Clupea harengus larvae at final yolk resorption over a wide range of incubation temperatures (3.5, 5, 8, 12, 15 and 17°C) “showed a remarkable constancy” (Overnell 1977). Data compiled in Table 10 show that this response was more common in larvae at yolk resorption (35% of sets at Re) than in larvae at hatching (9% of sets at H). Under different temperatures petrale sole (Eopsetta jordani) had different body sizes at hatching, but they achieved similar sizes at the end of endogenous feeding, the same was found in Oncorhynchus keta, Chondrostoma nasus and Melanogrammus aeglefinus (Table 10). Apparently this adaptation minimises differences in first feeding success between larvae incubated at different temperatures.

     
  3. III.

    Increase of larval size with increasing temperature over the whole range of temperatures tested (↑ in Table 10) was observed at hatching in Gadus morhua, Eopsetta jordani, Oncorhynchus keta and O. mykiss and/or at the end of endogenous feeding period in O. mykiss and Eupallasella percnurus (Table 10).

     
  4. IV.

    A dome-shaped relationship between larval size with an ascending limb at the lower part of viable temperatures and an descending limb at higher temperatures(↑↓ or ↑↔↓) was found in 10 out of 60 data sets presented in Table 10. Similarly, standard length (Ls) of newly hatched Perca fluviatilis measured at 16 temperature levels over a wide range of 6–22°C showed the maximum length at intermediate temperatures of 12–16°C (Kokurewicz 1969). A similar pattern was also observed earlier by Lasker (1964) in Sardinops caerulea, and recently by Mendiola et al. (2007) in Scomber scombrus over the whole range of temperatures experienced in the field (8.6–17.8°C). An assymetric dome-shaped curve, with temperature of maximum larval size shifted towards colder temperatures, was observed in five data sets in Table 10 and in Scomber scombrus.

     
What are the mechanisms behind these variable results of experimental studies? Several explanations have been proposed.
  1. 1.

    A negative effect of high temperature on size of newly hatched larvae can be routed through additive action of depleted oxygen availability, and increased chorionase and locomotor activities. Increased embryonic oxygen consumption at elevated temperatures in addition to depressed supply with oxygen from surrounding water, both contribute to reduced oxygen content in perivitelline fluid. Retarded embryogenesis under low oxygen has been found in several fish species (reviews by Zhukinskij 1986 and Kamler 1992). A temperature-induced acceleration of hatching process is also involved. Embryonic motility increases in response to low oxygen. Intensive movements of embryos, accompanied by earlier formation of hatching gland cells, temperature-stimulated chorionase secretion and higher enzyme activity at elevated temperature, all may contribute to earlier disruption of egg capsule and premature hatching of smaller, less developed larvae at higher temperature (Kokurewicz 1969; Yamagami 1988; Rechulicz 2001). In Coregonus albula, Oncorhynchus mykiss and Chondrostoma nasus decreased tissue size at hatching with increasing incubation temperature (Table 10) was parallelled by an increase of the amount of unresorbed yolk (Viljanen and Koho 1991; Rombough 1988c; Kamler et al. 1998, respectively).

     
  2. 2.

    Alternatively, the decrease of larval size with incubation temperature was explained by Blaxter (1992) “high temperature...causes differentiation relatively early ... when there is less tissue to divide up”. Thus, in such cases developmental rate would be more accelerated by temperature than growth rate. Independent acceleration of differentiation rate and growth rate by temperature was assumed in an atroposic model proposed to explain contradictory results of temperature effect on the number of vertebrae in fish (Lindsey and Arnasson 1981). However recently pieces of evidence emerge suggesting that differentiation and growth are accelerated by temperature in a similar way.(see Q10dev and Q10gr in Table 11).

     
  3. 3.

    In contrast, a small size of newly hatched larvae at the lowest end of viable temperatures was assumed to be an effect of premature hatching resulting from a partial loss of fluid from perivitelline space in cold environment (Alderdice and Forrester 1974 for flathead sole, Hippoglossoides elassodon).

     
  4. 4.

    Fish eggs contain a predetermined amount of yolk. Thus, temperature-induced variability of larval size at the end of yolk-feeding reflects the response of efficiency of yolk utilization for growth to temperature.

     
  5. 5.

    I think that larval size at a given ontogenetic state within the yolk-feeding period results from interplay between the response of three vital rates to temperature: ontogenetic rate (Q10dev, review by Kamler 2002), growth rate (Q10gr, this work, section. “Body growth”), and metabolic rate (Q10met, section “Metabolism”). One of the most obvious trade-offs in yolk-feeding fish is that between energy stored in tissues and energy expended in metabolism (Fig. 1). This trade-off is modified by temperature. Developmental rate, growth rate and metabolic rate, all increase with increasing temperature, but the extent of their acceleration may vary (Table 11). Usually development and growth rates are accelerated similarly, while Q10met sometimes deviates from Q10dev and Q10gr. Low acceleration of development and growth rates by temperature, accompanied by high acceleration of metabolic rate, results in decrease of larval size with temperature. An opposite combination of high Q10dev and Q10gr with low Q10met produces a size increase of larvae on yolk with increasing temperature. Larvae from both temperatures achieve similar size when the three temperature coefficients do not differ much (Table 11).

     
  6. 6.

    An interpretation of relationship between body size and incubation temperature can be correct only when the full range of viable temperatures was considered. For Gadus morhua Pepin et al. (1997) reported a significant increase in the larval Ls at hatching with incubation temperature increase within the range from −1°C to 7°C (Table 10). But in the same species Chambers (1997) observed a dome-shaped response of larval Ls at hatching to temperature within the full temperature range experienced by that fish (−0.8°C to 10.1°C), with an ascending part at lower and descending part at higher temperatures (Table 10). In Oncorhynchus tshawytscha only the descending limb of the relationship was revealed at temperatures ≥6°C, while in the data set MTWM), which included the lowest temperature of 5°C, the full dome-shaped response came out (Table 10). Generally, the dome-shaped pattern was revealed in the data sets that used many temperature levels (for example five for Hippoglossoides elassodon and Oreochromis niloticus, seven for Gadus morhua and eight for Morone saxatilis, but only three in Clarias gariepinus, Table 10), while remaining patterns were found at 2–6 (mostly two and three) temperature levels. Thus, an insufficient range of temperatures used and/or too wide temperature intervals between adjacent measurements may be insufficient to demonstrate the effect in full and may lead to a truncated shape of relationship between larval size and incubation temperature (Chambers 1997; Lambert et al. 2003). Assymetry in the dome-shaped curve may promote the truncated response.

     
Summing up, a considerable variety of body size responses to temperature was observed in yolk-feeding fish larvae. Mechanisms explaining different responses were tentatively proposed, but no unifying hypothesis has been advanced yet.
Table 10

Response of larval length (Lt – total length, Ls—standard length), tissue weight (Ww—wet weight, Wd—dry weight) or energy content (J) to incubation temperature. Compared are 60 data sets for 27 fish species at developmental stages: 6E and 7E—6th and 7th embryonic stages, respectively (Peňáz1974); H—hatching; MTW—maximum tissue weight on yolk; MLL—maximum larval length on yolk; EME—emergence; OFC—onset of feeding capability; Re—complete yolk resorption

Species

Stage

Unit

(Temperature °C)

Larval size

Summary

Source

Gadus morhua

  

−1°

    

1

 

HA1

mm Ls

2.9

3.7

3.6

4.0

4.2

   

 

G. morhua

  

−0.8°

0.8°

2.4°

3.9°

5.4°

7.0°

8.6°

10.1°

 

2

 

HA2

mm Ls

4.60

5.01

5.09

5.13

5.02

4.90

4.78

4.59

↑↓

 

Coregonus albula

  

1.0°

2.0°

3.0°

4.1°

     

3

 

HB

mm Lt

7.98

7.70

7.42

7.12

    

 

Salmo salar

  

1.8°

3.0°

4.1°

5.7°

     

4

 

H

mg Wd

5.0

5.1

5.2

5.1

    

 

S. salar

  

10°

       

5

 

HC

mg Ww

22.1

14.5

      

 
 

HD

mg Ww

26.6

12.5

      

 
 

HE

mg Ww

11.9

10.0

      

 

Salvelinus alpinus

  

       

6

 

H F

mm Lt

26.2b

23.6a

      

 

S. alpinus

  

       

7

 

H

mg Wd

4.44

4.44

      

 

Melanogrammus aeglefinus

  

10°

    

8

 

H

mm Lt

4.21

4.18

4.23

4.16

3.98

   

↔↓

 
 

Re

mm Lt

4.5

4.5

4.5

4.5

4.5

   

 

Hippoglossoides elassodon

  

2.4°

3.4°

6.0°

9.7°

10.7°

    

9

 

HG

mm Lt

3.32

3.80

5.78

5.57

5.11

   

↑↓

 
   

3.4°

4.0°

7.0°

8.5°

10.0°

10.7°

    
 

MLLH

mm Lt

6.30

6.96

6.97

6.97

7.04

7.20

  

↑↔

 

Gadus macrocephalus

  

2.7°

9.3°

       

10

 

HI

mm Lt

4.23

4.12

      

 
 

HJ

mm Lt

4.29

4.09

      

 

Anoploma fimbria

  

3.9°

4.0°

5.0°

6.0°

     

11

 

H

mm Lt

4.66

5.29

5.29

4.75

    

↑↔↓

 

Eopsetta jordani

  

4.1°

6.3°

8.5°

      

12

 

HK

mm Lt

2.51

2.77

3.07

     

 
   

6.3°

7.2°

8.1°

       
 

Re

mm Lt

5.70

5.65

5.50

     

 

Oncorhynchus keta

  

12°

      

13

 

H

mg Ww

105a

112b

117c

     

 
 

EME

mg Ww

340b

324a

333b

     

↓↑

 

Oncorhynchus tshawytscha

  

5.0°

7.3°

10.2°

12.5°

     

14

 

MTWL

mg Ww

375

372

343

    

 
 

MTWM

mg Ww

514

542

482

408

    

↑↓

 
 

MTWN

mg Ww

552

493

463

    

 

Oncorhynchus tshawytscha

  

6.0°

8.0°

10.0°

12.0°

     

15

 

H

mg Wd

20

20

19

17

    

↔↓

 
 

MTW

mg Wd

112

104

98

80

    

 
 

Re

mg Wd

75

71

68

56

    

 

Coregonus lavaretus

  

5.5°

10.0°

       

16

 

HO

mm Lt

10.4

9.9

      

 
 

HP

mm Lt

11.5

10.7

      

 

Oncorhynchus mykiss

  

6.0°

9.0°

12.0°

15.0°

     

17

 

ReR

mg Ww

205

196

196

193

    

↓↔

 
 

ReR

mg Wd

28.1

27.1

27.3

26.4

    

↓↔↓

 

Oncorhynchus mykiss

  

9.0°

10.0°

12.0°

14.0°

     

18

 

HS

mg Wd

0.99

1.01

1.08

1.17

    

↔↑

 
 

HT

mg Wd

1.04

0.99

1.08

1.28

    

↓↑

 
 

HU

mg Wd

1.10

1.51

1.85

2.40

    

 
 

HW

mg Wd

1.12

1.24

2.90

    

 
 

ReS

mg Wd

7.48

7.01

7.14

8.41

    

↓↑

 
 

ReT

mg Wd

7.43

7.37

7.32

8.56

    

↔↑

 
 

ReU

mg Wd

7.82

14.33

15.61

    

 

Leuciscus idus

  

9.0°

13.4°

22.0°

      

19

 

H

mm Lt

6.6

7.3

7.4

     

↑↔

 

Chondrostoma nasus

  

10°

13°

16°

19°

     

20

 

6E

mg Wd

0.11a

0.11a

0.11a

0.11a

    

 
 

7E

mg Wd

0.32b

0.44c

0.35b

0.24a

    

↑↓

 
 

H

mg Wd

0.79c

0.64b

0.49a

0.45a

    

 
 

Re

mg Wd

1.05a

1.09a

1.09a

1.08a

    

 

Morone saxatilis

  

12°

14°

16°

18°

20°

22°

24°

26°

 

21

 

HX

mm Lt

3.60

3.60

3.48

4.70

4.89

5.06

5.04

4.31

↑↔↓

 

Acipenser brevirostrum

  

13°

15°

18°

21°

     

22

 

Re

mm Ls

18.6a

17.5a

17.8a

17.8 a

    

 

Acipenser oxyrinchus

  

13°

15°

18°

21°

     

22

 

Re

mm Ls

14.5a

14.4a

15.2a

15.4a

    

 

Eupallasella percnurus

  

13°

16°

19°

22°

25°

    

23

 

Re

mg Ww

1.26a

1.26a

1.24a

1.20a

1.19a

   

 
 

Re

mg Wd

0.111a

0.114ab

0.118bc

0.119bc

0.122c

   

 

Gymnocephalus cernuus

  

14°

16°

18°

20°

     

24

 

H

mm Lt

3.70b

3.85c

3.67b

3.52a

    

↑↓

 

Scophthalmus maximus

  

15°

18°

21°

      

25

 

HY

μg Wd

17.1

17.1

17.1

     

 
 

ReY

μg Wd

18.4

17.3

20.2

     

↓↑

 

Pagellus erythrinus

  

16°

18°

21°

      

26

 

Re

mm Lt

3.03

2.93

2.80

     

 

Tautoga onitis

  

16°

19°

22°

      

27

 

H

mg Wd

3.1

6.4

4.2

     

↑↓

 
 

OFC

mg Wd

10.9

11.0

8.1

     

↔↓

 

Cyprinus carpio

  

21.3°

22.9°

       

28

 

Re

J

4.18

3.35

      

 

Cyprinus carpio

  

20.5°

22.5°

24.3°

26.7°

30.4°

    

29

 

H

mm Lt

5.05

4.95

4.71

4.67

4.34

   

 
 

H

mm Ls

4.80

4.71

4.49

4.47

4.10

   

 

Clarias gariepinus

  

22.1°

25.0°

28.1°

      

30

 

H

J

1.29

1.23

0.89

     

 
 

Re

J

7.36

8.18

7.86

     

↑↓

 

Oreochromis niloticus

  

20.0°

24.0°

28.0°

30.0°

34.5°

    

31

 

HZ

mg Wd

0.14a

0.29b

0.34c

0.38c

0.18

   

↑↔↓

 

Cyprinodon macularius

  

28.0°

30.0°

32.0°

34.0°

35.0°

    

32

 

HAA

mm Lt

5.25

5.00

4.30

3.90

3.65

   

 

a, b, c...nwithin rows non-significant results (P > 0.05) are followed by the same letter.

A1Significant increase in the size with increasing temperature (F4,75 = 2.53, P < 0.05). A1 and A2Approximate values read from the graph. BComputed from: Lt = 8.253 – 0.277t, P < 0.001, n = 278. C, D and EOxygen: 100, 50 and 30% air saturation, respectively. FOne example (offspring of the family ♀1 x ♂1) is shown, similar significant decrease of Lt with temperature was shown for 11 remaining families studied. GNear-optimum salinity of 25.0‰. HSalinity 25.0–39.6‰ required for normal development. I and JLow and high oxygen level (2.7 and 9.3 ppm O2, respectively), near-optimum salinity of 14‰. KNear-optimum salinity of 27.5‰. L, M and NInitial eggs 235, 341 and 384 mg wet weight, respectively. O and POffspring of hatchery families 1 and 2, respectively. RAt 90% yolk utilisation. S, T, U and WInitial egg 10.0, 10.4, 24.4 and 33.8 mg dry weight, respectively. X0‰ salinity. Within 0–10‰ salinity and 12–26°C Lt (mm) was related to temperature (t, °C): Lt = −0.013t2 + 0.62t – 2.22, R2 = 0.70. YAsh-free weight. ZAll groups were incubated at 28.0°C to retinal pigmentation before placing them at the test temperatures. AA100% air saturation, 35‰ salinity

Summary: ↑ - larger larvae tended to be produced with increased temperature, ↓ - size was reduced with increased temperature, ↔ - larvae from different temperatures achieved a similar size.

Source: 1—Pepin et al. (1997); 2—Chambers (1997); 3—Viljanen and Koho (1991); 4—Ryzhkov (1979); 5—Hamor and Garside (1977); 6 - Huuskonen et al. (2003); 7—Gruber and Wieser (1983); 8—Martell et al. (2005); 9—Alderdice and Forrester (1974); 10—Alderdice and Forrester (1971a); 11—Alderdice et al. (1988); 12—Alderdice and Forrester (1971b); 13—Beacham and Murray (1985); 14 - Rombough (1985); 15—Heming (1982); 16—Escaffre et al. (1995); 17—Rombough (1988c); 18—Kamler and Kato (1983); 19 - Witkowski et al. (1997); 20—Kamler et al. (1998); 21—Morgan et al. (1981); 22—Hardy and Litvak (2004); 23—Kamiński et al. (2006); 24—Bonisławska et al. (2004); 25—Quantz (1985); 26—Klimogianni et al. (2004); 27—Laurence (1973); 28—Kamler (1976); 29 - Peňáz et al. (1983); 30—Kamler et al. (1994); 31 - Rana (1990); 32 - Kinne and Kinne (1962)

Table 11

Body size of larvae attained at final yolk resorption after rearing at two contrasting temperatures. Temperature coefficients for developmental, growth and metabolic rates (Q10dev, Q10gr and Q10met, respectively) were computed as in Kamler et al.(1994)

Species

Temperature (°C)

Size of larvae

Acceleration of rates

Comparisonsa

Source

Q10dev

Q10gr

Q10met

Size (%)

Acceleration

Cyprinus carpio

21.3

4.18(1)

1.56

1.24

6.32

−22

0.22

1

22.9

3.35(1)

Oncorhynchus mykissb

6.0

28.1(2)

2.87

2.81

4.17

−6

0.68

2

15.0

26.4(2)

Clarias gariepinus

25.0

8.18(1)

1.93

1.89

4.09

-4

0.47

3

28.1

7.86(1)

O. mykissb

9.0

27.1(2)

2.26

2.25

3.83

−3

0.59

2

15.0

26.4(2)

O. mykissb

6.0

28.1(2)

3.92

3.82

3.81

−3

1.02

2

12.0

27.3(2)

Chondrostoma nasus

10.0

1.05(2)

3.92

3.94

2.88

3

1.36

4

19.0

1.08(2)

C. gariepinus

22.1

7.36(1)

5.85

6.17

1.67

11

3.60

3

25.0

8.18(1)

O. mykissc

10.0

7.37(2)

2.82

3.02

0.85

15

3.44

5

14.0

8.56(2)

O. mykissd

10.0

7.01(2)

2.82

3.06

0.81

18

3.63

5

14.0

8.41(2)

O. mykisse

10.0

7.82(2)

6.37

8.53

0.58

66

12.84

5

14.0

15.61(2)

(1)Larval size in terms of J indiv.−1; (2)larval size in terms of mg Wd indiv.−1

aComparison between larval sizes taken from: 200 (W2 − W1)/(W2 + W1), where W1 and W2 are body sizes at the lower and the higher temperatures, respectively. Acceleration rates compared as: (Q10dev + Q10gr)/2 Q10met

bAt 90% yolk utilisation

c, d and eInitial egg 10.4, 10.0 and 24.4 mg dry weight, respectively

Source: recomputed from 1—Kamler (1976); 2—Rombough (1988c); 3—Kamler et al. (1994); 4—Kamler et al. (1998); 5—Kamler and Kato (1983)

Concluding remarks

Progress made during last two decades

Recent technological advances, especially advances in electronics, resulted in methodological progress in eco-physiological studies. Selected methods have been discussed in more detail earlier in this work. Few examples of promising techniques and methods will be shortly summarised here.

Miniaturisation and an individual-based approach are noteworthy. Marine fish eggs were inspected individually under a binocular microscope during incubation of single eggs at controlled temperature in small cell culture multidishes containing 24 wells of 2 cm3 water each (Thorsen et al. 2003). The recent technique of individual yolk size monitoring in vivo coupled with a computer image analysis system was developed by Sarnowski (2002, 2003), Jordaan and Kling (2003), Hardy and Litvak (2004) and Martell et al. (2005). The same technique was applied to monitor body area increase (Hardy and Litvak 2004). Individual growth trajectories were constructed for larvae of Brevoortia tyrannus (Chambers and Miller 1994).

Reasons and extent of body shrinkage – reduction in length and/or weight at preservation - were better understood, and corrections for bias resulting from shrinkage were proposed (Araujo-Lima 1994; Fey 2002). Content of main elements (carbon, hydrogen, nitrogen and sulphur) in biological materials can be determined with an automatic CHNS analyser, and protein and energy content can be recomputed. The determinations are accurate and precise, and analyses are done in short time on small samples, which makes the CHNS method suitable for yolk-feeding fish (Kamler et al. 1994, 1998). An oximetric microtechnique was used in a device constructed to monitor oxygen consumption of individual eggs of Danio rerio (Bang et al. 2004). Thus, the methodological advance promotes the work with the yolk-feeding fish that are small, delicate and rapidly changing.

Conceptual aspects benefited from the methodological progress. A yolk-feeding fish can be regarded as a complex system that dynamically changes with ontogeny (Fig. 1). This review demonstrates recent progress in understanding the distribution of free amino acids, protein, fatty acids and lipid classes in the different compartments of yolk-feeding fish. Ontogenetic changes in the compartmentalisation have been highlighted. Advanced studies on amino acid metabolism have been performed. Few dozens of works by R.N. Finn, H.J. Fyhn, I. Rønnestad and co-workers contributed to this progress. Other studies (M.D. Wiegand, D.R. Tocher and co-authors) have expanded knowledge of the utilisation of various lipid components of yolk and their accumulation in teleost embryos and larvae.

Quantification is a noteworthy feature of recent studies on fish early life. Development is not synchronous, especially at low temperatures and later in ontogeny, thus a method of exact timing for any ontogenetic state (for example 16 cells, hatching, first feeding) was applied with appropriate statistics (Kamler 1992, 2002; Kamler et al. 1994, 1998; Fuiman et al. 1998; Kamiński et al. 2006). Several aspects of the main processes such as yolk feeding, growth and/or metabolism were described with equations. Some of them were presented earlier in this work. Summarising models have been developed by Conceição et al. (1993), Zhang et al. (1995), Beer and Anderson (1997), Beer (1999), and Jaworski and Kamler (2002).

Significance of studies on yolk-feeding fish in the field and aplication in aquaculture

Studies on fish early life history are of importance in forecasting recruitment (Pepin 2002; Tomkiewicz et al. 2003). These studies also play a part in the understanding of anthropogenic disturbances of environmental conditions in water bodies: mechanical destruction of habitat and pollution (Holt 2002). For example, riverine fish are good indicators of integrity of river systems. Discrepancy between state of river resulting from river engineering (river regulation, reduction of shoreline structure, and loss of connectivity with sidearms) and fish requirements during early ontogeny resulted in the threatened state of several rheophilic cyprinids in European rivers (Schiemer and Spidler 1989; Schiemer et al. 2003). In 1989 early life stages of Pacific herring (Clupea pallasi) and pink salmon (Oncorhynchus gorbuscha) were exposed to the “Exxon Valdez” spill of oil into Prince William Sound. Long-term complex studies demonstrated severe damages to both populations, recovery from the spill lasted 10 years (summary in Holt 2002).

Knowledge of critical oxygen level (Ocrit) may be useful to define dissolved oxygen content criteria in natural waters (Rombough 1988c). He found that Ocrit in Oncorhynchus mykiss was maximal just before hatch (Ocrit.max) and was positively related to temperature. He proposed to predict dissolved oxygen criteria from the equation:

$$\hbox{O}_{\rm crit.max}=2.07+3.06 \hbox{ ln}\,t$$
(39)
where t is temperature (°C) within the range of 6–12°C.

How aquaculture can profit from the studies on endogenously feeding fish? Incubation conditions may be manipulated in hatcheries to maximise fish recruitment to free-feeding stages. Optimal temperatures for growth and survival were determined in several species. Examples are Oncorhynchus mykiss (Ryzhkov 1976), Oncorhynchus tshawytscha (Heming 1982), Cyprinus carpio (Peňáz et al. 1983), Chondrostoma nasus (Kamler et al. 1998), Salmo salar (Ryzhkov 1976; Ojanguren et al. 1999), Acipenser oxyrinchus (Hardy and Litvak 2004), and many others.

Energy allocation during fish yolk-feeding period may also be optimised based on the knowledge of other environmental conditions required by fish. Substrate and light that imitate the natural conditions have been explored. Experiments on the role of substrate have been performed with two gravel hiders, developing from eggs burried in gravel (lithophils according to the classification by Kryzhanovskij 1949; Balon 1975, 1990). Yolk-feeding larvae of Salmo salar were reared at 4, 6, 8, 10 and 12°C with and without gravel, the rugose substrate improved gross yolk utilization efficiencies at all temperatures tested (Peterson and Martin-Robichaud 1995). Hansen and Møller (1985) introduced astro-turf as an artificial rugose substrate for yolk-feeding S. salar. The fish incubated with astro-turf exhibited higher survival, better growth of body tissues, and improved efficiency of yolk conversion compared to fish incubated on traditional flat screen. Another lithophilous species, Salmo trutta, survived better and converted yolk into body tissue more efficiently when reared with astro-turf (Hansen 1985). Larvae of Tinca tinca (a phytophil, Kryzhanovskij 1949; Balon 1975; 1990) hang on plants by cement glands during a long yolk-feeding period. Exogenously feeding stages of that fish grow slowly and non-efficiently, but endogenously feeding tench converted their yolk efficiently in a hatchery experiment in which metabolic expenditures were minimised in larvae attached to Carex stems placed in rearing tanks (Kamler et al. 1995). Clarias gariepinus is a photophobic fish which grew better and had a depressed ratio of total metabolism to body growth when reared under continuous darkness (Appelbaum and Kamler 2000). Taken together, the “fish-oriented approach” can provide simple and cheap alternative methods to be recommended in aquaculture.

Results of studies on yolk-feeding fish may be used to determine time at which first external food needs to be supplied. There are several criteria that potentially can be used for estimating times of first feeding, such as behavioural changes, maximum body weight, and use of thermal sums (which requires corrections for actual thermal regimes) (review by Peterson and Martin-Robichaud 1995). Other criteria are the ratio of embryonic tissue weight to the combined weight of tissue and remaining yolk (Marr 1966), developmental index KD (Bams 1970), ontogenetic index OL, (Fuiman 1994), hydration of whole body plus yolk (Peterson and Martin-Robichaud 1995) and effective day-degrees D°eff (Kamler 2002). In the latter work most of these indices have been critically reviewed.

First-feeding larvae of many fish species cannot be successively reared on formulated diets (review for cyprinid larvae by Kamler and Wolnicki 2006). That makes difficult evaluation of nutritional requirements of early larvae by manipulation of food composition. However, knowledge of nutrient composition in yolk and in fish larvae themselves, as well as studies on utilisation of endogenous nutrients from yolk has been suggested to be useful in elucidation larval nutritional requirements (Heming and Buddington 1988; Fyhn 1989; Rønnestad and Fyhn 1993; Rønnestad et al. 1995; Sargent et al. 1999). Summing up, studies on yolk-feeding fish can show ways of intensification in aquaculture.

Acknowledgements

Helpful comments on the manuscript provided by Andrzej Jaworski, Michał Korwin-Kossakowski and Jacek Wolnicki are gratefully acknowledged. The major improvement was made by Jennifer Nielsen and two anonymous reviewers. The permissions from Andrzej and Elsevier to re-use published material in the Fig. 1 is greatly appreciated.

Copyright information

© Springer Science+Business Media B.V. 2007