© 2005 International Council for the Exploration of the Sea
Food ingestion in juvenile cod estimated by inert lanthanide markers – effects of food particle size
a Department of Molecular Biology, University of Bergen Thormøhlensgt. 55, N-5020 Bergen, Norway
b Institute of Marine Research PO Box 1870, Nordnes, N-5817 Bergen, Norway
c National Institute of Nutrition and Seafood Research PO Box 2029, Nordnes, N-5817 Bergen, Norway
d Institute of Aquaculture Research (AKVAFORSK) PO Box 5010, N-1432 Ås, Norway
*Correspondence to O. Garatun-Tjeldstø: tel: +47 5558 4379; fax: +47 5558 9683. e-mail: oddvard.garatun-tjeldsto{at}mbi.uib.no.
Development of formulated starter diets and improvement of diets for juvenile marine fish species are major challenges in aquaculture. The ingestion rate may be regarded as a parameter for evaluating whether a diet particle is available, recognized, of adequate size, palatable, and preferred by small fish. In this study, we evaluated the effect of food particle size (150–3425 µm) on ingestion rate in juvenile cod (36–826 mg wet weight). Lanthanide oxides were used as markers. Several mixtures of lanthanide marker-labelled diets were produced by combination of the mono-labelled size classes. Each combination was fed for one, four, or seven days to groups of juvenile cod in separate tanks. After termination by anaesthetic, the fish were collected, individually weighed, and frozen until analysis. The amount of lanthanides in the homogenized fish was measured by inductively coupled plasma-mass spectrometry (ICP-MS). There were no significant differences in ingestion of any of the markers when given in a mixture of mono-labelled particles. Thus, the markers Y2O3, La2O3, Nd2O3, Dy2O3, and Yb2O3 may be used as inert tools for evaluating diet ingestion. The total intake of dry diet was about 13 mg g–1 body weight. The particle size that caused the highest food intake was less than 1.2 mm in all experiments.
Keywords: cod, diet preference, Gadus morhua, inert lanthanide marker, particle size
Received 13 June 2004; accepted 16 November 2005.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
The optimal nutrient and physical properties of formulated start and weaning diets for marine fish larvae are not well established (NRC, 1993; Watanabe and Kiron, 1994). One of the first requirements in the chain of processes leading to growth is that the diet particle be available, recognized, of adequate size, palatable, preferred, and ingested by the fish. Several dietary characteristics, such as particle size, colour, taste, texture, initial water content, buoyancy, etc., may be of significance for the ingestion rate of marine fish larvae. Ingestion rates may be studied by visual observation of food intake, a time consuming and difficult method for larval and juvenile fish, or by analysis of gut content after a period of feeding. Markers like chromic oxide (Austreng, 1978) and radioactive isotopes (Kolkovski et al., 1993) have been added to the diets to quantify the food content in the gut, often in connection with digestibility studies. An approach based on sets of several markers used simultaneously makes it possible to differentiate between ingestion of several diets offered together to fish in one tank, and thereby to establish selection criteria based on the diet characteristics. A prerequisite for doing these types of experiments is finding suitable non-toxic inert markers that can be incorporated in the diet without adding biasing factors, and which can be quantified by examination of the individual after ingestion. It is of particular importance that the markers do not influence the ingestion of the diet by acting as food repellents or attractors. Minimally, a set of markers used within one experiment should be equally attractive or repellent to the fish.
The 15 trivalent oxides of lanthanides and yttrium make it possible to carry out work with sets of several markers simultaneously within one experiment. Trivalent oxides of yttrium and rare earth metals have been evaluated as inert markers in apparent digestibility studies with salmonids (Austreng et al., 2000). Such markers have been used to study the effects of several feeding stimulants on diet preference by gibel carp (Carassius auratus gibelio) (Xue and Cui, 2001), and used on juvenile cod (Gadus morhua) to study effects of dietary moisture content (Otterå et al., 2003). The lanthanide oxides are finely powdered and water insoluble. It was concluded that La, Dy, and Yb may be safely used as nutritional markers (Luckey et al., 1975), and lanthanide oxides have been used for nutrient studies in humans (Hutcheson et al., 1979). However, inhalation of dust containing lanthanide during food production should be avoided (McDonald et al., 1995; Hirano and Suzuki, 1996). The lanthanide oxides may be quantified in mixtures of individually labelled diets and in the stomach content of larvae and juveniles by inductively coupled plasma-mass spectroscopy (ICP-MS). In the present study, first, we aim to verify that lanthanide oxide markers are inert with respect to ingestion rate of juvenile cod, and second, to estimate food particle size preferences related to juvenile cod body size.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Two types of experiments were performed: a single experiment to test the marker inertness on cod juveniles, and then a series of six feeding experiments on juveniles of various sizes to test particle size preference. The basic content of the test diet (Table 1) is the same in all experiments. The basic test diet was divided into five batches, each batch labelled with a different lanthanide marker to get mono-labelled pellets. Each pellet type was crushed and sieved into eight mono-labelled size classes. From this group of labelled size grades we produced seven multi-labelled diets (A–G) with five markers in each (Table 2).
|
|
Food production
As shown in Table 1, the test diet was formulated to be rich in protein, with herring (Clupea harengus) meal as the main component. Based on earlier experiments, the diet should be suitable for feeding juvenile cod (Lie et al., 1989). The vitamins included were twice the level recommended by NRC (1993) in order to compensate for manufacturing losses and leakage. Squid mantle (Gonatus fabricii) and krill (Euphasia superba) were dehydrated by evacuation and ground together with the main mixture powder using a Retsch Ultra Centrifugal Mill, type ZM1. A base mixture of 20 kg was made and portioned into five batches of 2250 g. 4.5 g of one of the marker oxides was stirred vigorously into each batch followed by the addition of 320 g of water. Each batch of 2575 g moist mixture, with one marker included, was run through a 7-mm matrix on a pellet machine in the following marker sequence: La, Y, Yb, Dy, and Nd. The water content was measured to 19.4% before pelleting.
The markers Y2O3, La2O3, Nd2O3, Dy2O3, and Yb2O3 of 99.9% purity were purchased from Sigma Chemical Company. To reduce contamination of marker from the preceding batch, portions of 1.4 kg base mixture without marker were used to clean the pellet machine between each batch. This procedure resulted in approximately 2.2 kg of each batch of marked diet collected as pellets. The pellets were dried in air stream overnight to a dry weight of 95 g/100 g and crushed. A Fritsch Analysette rotatory shaker equipped with a staple of stainless steel sieves of decreasing apertures was used for grading the granulated mixture into eight particle classes, ranging from 0.0–0.3 to 3.35–3.5 mm. Then each particle size class was separately stored and analysed by ICP-MS to determine the marker element content.
The mono-labelled particle class 1.4–2.0 mm was used in a feeding experiment to verify that the markers' influence on the ingestion rate of a specific diet was not relative. Five fractions of this size class, with one marker in each fraction, were blended by equal weight to produce the test diet that was used for the assessment of possible marker taste preferences by the fish (Table 2, diet G).
To evaluate preferences for particle size, sets of five differently labelled size classes were blended by equal weight and offered to juvenile cod of various weight classes (Table 2, diets A–F).
Elemental analysis
For the analysis of 39Y, 57La, 60Nd, 66Dy, and 70Yb concentrations in samples of the food and fish, the samples were digested by adding 2 ml concentrated nitric acid (HNO3) and 0.5 ml hydrogen peroxide (H2O2). 1 g of diet was homogenized to measure marker concentration in the food samples, while two, ten-cod specimens were pooled in each analysis in order to have enough material for analysis of marker content in the fish. The digestions were run in tetrafluorine methoxil digestion vessels through a standardized temperature programme (S1, 250 W, 5.00 min; S2, 0 W, 1.00 min; S3, 250 W, 5.00 min; S4, 400 W, 5.00 min; and S5, 650 W, 5.00 min) in a microwave oven (Milestone MLS-1200, Italy) (Julshamn et al., 1999). After digestion, the digests were diluted to 25 ml, and the elements were analysed by inductively coupled plasma-mass spectrometry (ICP-MS) (Perkin–Elmer SCIEX Elan 5000A). In ICP-MS, the ICP is established as an effective ion source for MS as detector for the lanthanide elements. The blank samples were prepared and run together with the sample solutions. The method of standard addition calibration was used for quantification of the element in question. The lanthanide content was calculated as milligramme elemental lanthanide per gramme total wet weight of fish in the vial.
Fish stocking, experimental conditions, and sampling
The juvenile cod used in the feeding experiments were collected from a population of pond-reared juveniles (Blom et al., 1991). Newly hatched larvae had been released into the pond and preyed on naturally occurring zooplankton until they were collected for experiments. The fish were distributed randomly among black 180-l experimental tanks, and the feeding was started. The multi-labelled food was supplied with automatic disc feeders at 1-min intervals for 24 h d–1. Feeding was in excess, approximately ten times the estimated requirement. Surplus food and dead fish sedimented in the conical bottom of the tanks were tapped out once a day. The water temperature during the experiments was approximately 10°C.
Three tanks were used for each feeding experiment, of which one tank was terminated at day 1, one at day 4, and the last at day 7 after start. One experiment (experiment G) was performed to verify that the marker was inert. Six experiments (A–F) were conducted to investigate the preference for diet particle size.
The fish in the tank were euthanized with an anaesthetic (metomediate) and gently tapped out through the outflow. The fish were rinsed in seawater, dried on a blotting paper, individually weighed (wet weight), put in scintillation vials, and frozen until analysis. Individuals of similar size (Table 2) were pooled in each vial. The experimental conditions, such as stocking of fish, labelling diets, feeding, and sampling, are specified in Table 2. Standard statistical tests (Zar, 1999) were used as specified in the appropriate figures and tables, specifically the Shapiro–Wilks W test of normality, two-way ANOVA, and chi-square for distribution fitting.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Food production
The marker level (Table 3) and the corresponding level of contaminating markers (Table 4) in each mono-labelled food and size fraction were assayed. A contaminating marker is referred to as the remaining portion of the marker used to produce the preceding batch. The average contamination accompanying each main marker (Table 4) was used to calculate contamination correction factors (Table 5) to be applied to the ICP-MS measurements of lanthanide levels in the juveniles. An example of correction is that the level of observed Yb (mg kg–1) in the fish from ingestion experiments should be reduced by 14.2% of the Dy value (mg kg–1) (Table 5). Even though an identical amount of marker was used in each batch during diet production, the levels measured by ICP-MS after production were not the same. Most of the variation could be explained by marker (Table 3), while size fraction had less impact on the variability (Table 3). Only one ICP-MS analysis for each of the 40 feed batches was performed, as these analyses are very costly. Therefore, we chose to normalize the lanthanide level in each individual food particle batch for segregation (different lanthanide level for each size grade) and marker profile (different level of marker for each lanthanide used) in addition to corrections for contamination level when ingestion in cod was calculated. The normalizing procedure and resulting recalculated values are shown in Table 6.
|
|
|
|
In the case of feeding experiment G (Table 2), the size class 1.4–2.0 mm of mono-labelled particles was blended on an equal weight basis and fed to a group of juveniles with initial body weight of 826 mg.
The apparent food content associated with the cod juveniles was estimated from the ICP-MS analysis of marker oxides, corrected for cross-contamination, and normalized as described in the previous section. The corrected marker weight per cent distribution of each sample is shown in Figure 1. The hypothesis of marker inertness was not rejected (Table 7; chi-square = 0.6239, 4 d.f., p = 0.96 based on 41 samples). Therefore, we could use the investigated lanthanides and yttrium as inert markers, equally repellent or attractive if used together in the same experiment.
|
|
Preferences for particle size (experiments A–F)
The fish in experiments A–F had the opportunity to ingest from a mixture of five particle size classes. The estimated amount of dry diet in the fish originating from each of the five particle size classes is reported in Figure 2. Ideally, the particles presented in each experiment should include the entire size range from which fish of a certain size are capable of ingesting. The experiments A, B, C, and E seem to fit this criterion. The smallest fish (experiment A; 33–41 mg) ingested particles from nominal size 0.15 to 1.7 mm, but most from size 0.15 mm. Fish weighing 146–178 mg (experiment B) had highest ingestion of particle sizes from 0.85 to 1.2 mm. In experiments C (fish; 400–506 mg) and E (fish; 848–1037 mg), the maximum (peak) ingestion came from particles near 1.2 mm. That was also the case for experiments D and F, where particle sizes from below 1.2 mm and up to 3.35 and 3.5 mm were offered to fish from 375 to 1180 mg. We conclude that small larvae ingested optimally small particles, and the particle size of "preference" increased up to about 1.2 mm at a body weight of approximately 600 mg. In none of the experiments did we observe maximum (peak) ingestion from particles larger than 1.2 mm.
|
Small fish (below 500 mg bw) had not started eating on the first day (Figure 2, experiments A–D), but larger fish (800–1200 mg) had started. The ingested material associated with the fish from all experiments at days 4 and 7 was approximately the same (12.7 ± 2.6 and 13.4 ± 2.0 mg g–1 bw ± s.d., respectively) regardless of body weight, so it is not necessary to carry out feeding experiments of this type for more than four days.
Based on a dry weight composition of cod of 18% (Van Pelt et al., 1997; Finn et al., 2002), we can calculate the dry gut content as percentage of the larval dry weight as approximately (12.7 + 13.4) x 0.5 x 100 x 180–1 = 7.63%. Correspondingly, based on dissection and weighing, 6% gut content in cod larvae feeding on zooplankton has been reported (Gamble and Houde, 1984). The low feeding ratio for the smallest fish is further visualized when the entire data set from each day is plotted (Figure 3) with a linear trend forced through zero. The average magnitudes of ingestion calculated from the trend line at days 4 and 7 are 14.3 and 11.4 mg g–1 body weight, respectively (r2 = 0.78 and 0.73).
|
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Even if no significant differences in marker concentrations among the particle size classes were observed, one should be cautious and aware of the possibility for marker segregation, which may arise as a result of incomplete blending or internal particle structures and breaking zones in the production of granulates from pellets. The use of normalization allows one to conduct a single measurement of the lanthanide concentration in each mono-labelled particle size fraction, which is time- and cost effective. With the exception of Yb as a contaminant of Dy, the contaminant correction was a minor adjustment. We recommend that cross-contamination during particle production be minimized or preferably avoided, which is feasible if thoroughly cleaned or separate equipment is used (Otterå et al., 2003).
We have assumed that hypothetical taste preferences are associated with the intended main marker in each diet particle, and that the effect of a minor contaminating marker on preference for ingestion is negligible. A rationale for this assumption might be that all the main marker sources may contribute to a low level lanthanide flavour in the tank, thus masking the presence of a minor contaminant in a certain particle. The variation in marker percentage distribution of experiment G seems to be largest in the tank terminated after one day as indicated by extreme outliers above 40% or under 5% in Figure 1, and also if percentage standard deviations are calculated from Table 7. The reason may be that only a few particles are ingested the first day, causing statistical fluctuations. In retrospect, the particle range (1.4–2 mm) used in experiment G lies above the optimal particle size suited for ingestion (Figure 3, experiments E and F).
We did not find the maximum (peak) ingestion from particle classes of nominal size larger than 1.2 mm. The particles presented to the fish in experiments D and F were too large, and the ingestion maximum lies to the left in the histograms; 1.2 mm is less than 30% of the estimated mouth height of the fish (Otterå and Folkvord, 1993). However, from our experiments we cannot conclude that particles smaller than 1.2 mm were preferred. Larger particles may be less available owing to higher sinking rates than small particles. Although larger particles may be more easily seen and have more nutrition per particle, they are distributed in lesser concentration, and may be less available. More experiments with diets blended in unequal weight ratios should therefore be designed to evaluate preference, availability, and optimal ingestion of food.
It is likely that we overestimated the percentage gut content. The value 7.63% is based on the level of the marker in the diet at the time of ingestion, but the diet is digested, and the marker is concentrated and finally evacuated after some time (Austreng et al., 2000). Since the amounts of ingested materials associated with the fish at days 4 and 7 were about the same, it seems that a steady state equilibrium between ingestion and evacuation had been reached.
The large variation in the amount ingested among individual samples suggests a heterogeneous eating behaviour within each experiment. Such heterogeneity is of particular interest, since it may influence individual growth rates and lead to cannibalism. The heterogeneity is probably underestimated owing to the pooling of several individuals within a sample. An increase in the sensitivity of the detection method would significantly increase the usefulness of this method for larval and juvenile fish. This could be done by increasing the marker concentration and modifying the dilution factor, or maybe by improving ICP-MS technology. The lanthanide marker methodology could then be used to generate ingestion data of single larva.
We conclude that yttrium and lanthanide labelling offers powerful tools for quantifying diet ingestion in fish and for comparing ingestion of differently labelled diets by the same group of fish. We also have used the technique to study effects of dietary water content (Otterå et al., 2003) and future possibilities includes the study of the effects of food type, colour, taste, hardness, or other diet properties.
|
Acknowledgements |
|---|
We thank Fiskeriforskning (Bergen, Norway) for use of facilities for food production. The technical assistance of Anne E. Madsen (Department of Molecular Biology, University of Bergen) and Jan Brenna at the National Institute of Nutrition and Seafood Research is gratefully acknowledged.
|
Footnotes |
|---|
Deceased 2003. 
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Austreng E. (1978) Digestibility determination in fish using chromic oxide marking and analysis of content from different segments of the gastrointestinal tract. Aquaculture 13:265–272.[CrossRef][Web of Science]
Austreng E., Storbakken T., Thomassen M.S., Refstie S., Thomassen Y. (2000) Evaluation of selected trivalent metal oxides as inert markers used to estimate apparent digestibility in salmonids. Aquaculture 188:65–78.[CrossRef][Web of Science]
Blom G., Otterå H., Svåsand T., Kristiansen T.S., Serigstad B. (1991) The relationship between feeding conditions and production of cod fry (Gadus morhua L.) in a semi-enclosed marine ecosystem in western Norway, illustrated by use of a consumption model. ICES Marine Science Symposia 192:176–189.
Finn R.N., Rønnestad I., van der Meeren T., Fyhn H.J. (2002) Fuel and metabolic scaling during the early life stages of Atlantic cod Gadus morhua. Marine Ecology Progress Series 243:217–234.[Web of Science]
Gamble J.C. and Houde E.D. (1984) Growth, mortality and feeding of cod (Gadus morhua L.) larvae in enclosed water columns and in laboratory tanks. In Dahl E., Danielssen D.S., Moksness E., Solemdal P. (Eds.). The Propagation of Cod (Gadus morhua L.), Part 1 pp. 123–143 Flødevigen Rapportser, 439 pp.
Hirano S. and Suzuki K.T. (1996) Exposure, metabolism, and toxicity of rare earths and related compounds. Environmental Health Perspectives 104:Suppl. 1, 85–95.[CrossRef][Web of Science][Medline]
Hutcheson D.P., Venugopal B., Gray D.H., Luckey T. (1979) Lanthanide markers in a single sample for nutrient studies in humans. The Journal of Nutrition 109:702–707.
Julshamn K., Maage A., Wallin H. (1999) Determination of magnesium and calcium in foods by atomic absorption spectrometry after microwave digestion: NMKL collaborative study. Journal of AOAC International 81:1202–1208.[Web of Science]
Kolkovski S., Tandler A., Kissil G.W., Gertler A. (1993) The effect of dietary exogenous digestive enzymes on ingestion, assimilation, growth and survival of gilthead sea bream (Spares aurata, Sparidae, Linnaeus) larvae. Fish Physiology and Biochemistry 12:203–209.[CrossRef][Web of Science]
Lie Ø., Lied E., Lambertsen G. (1989) Feed attractants for cod (Gadus morhua). Fiskeridirektoratets Skrifter Serie. Ernæring 2:227–233.
Luckey T.D., Venugopal B., Hutcheson D. (1975) Heavy metal toxicity, safety and hormonology. Environmental Quality and Safety Supplement 1:75–80.
McDonald J.W., Andrew J.G., Sheehan C.E., Bernhardt P.F., Roggli V.L. (1995) Rare earth (cerium oxide) pneumoconiosis: analytical scanning electron microscopy and literature review. Modern Pathology 8:859–865.[Web of Science][Medline]
NRC. (1993) Nutrient Requirements of Fish(National Academy Press, Washington) 115 pp.
Otterå H. and Folkvord A. (1993) Allometric growth in juvenile cod (Gadus morhua) and possible effects of cannibalism. Journal of Fish Biology 43:643–645.[Web of Science]
Otterå H., Garatun-Tjeldstø O., Julshamn K., Austreng E. (2003) Feed preferences in juvenile cod estimated by inert lanthanide markers – effects of moisture content in the feed. Aquaculture International 11:217–224.[CrossRef][Web of Science]
Van Pelt T.I., Piatt J.F., Lance B.K., Roby D.D. (1997) Comparative composition and energy density of some North Pacific forage fishes. Comparative Biochemistry and Physiology 118A:1393–1398.
Watanabe T. and Kiron V. (1994) Prospects in larval fish dietetics. Aquaculture 124:223–251.[CrossRef][Web of Science]
Xue M. and Cui Y. (2001) Effect of several feeding stimulants on diet preference by juvenile gibel carp (Carassius auratus gibelio), fed diets with or without partial replacement of fish meal by meat and bone meal. Aquaculture 198:281–292.[CrossRef][Web of Science]
Zar J.H. (1999) Biostatistical Analysis 4th edn (Prentice Hall, New Jersey) 663 pp.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||