WO1997038590A1

WO1997038590A1 - Production of pet food and wet fodder for fur-bearing animals and fish by surface treatment

Info

Publication number: WO1997038590A1
Application number: PCT/NO1997/000093
Authority: WO
Inventors: Almar Ellingsen
Original assignee: Silver Feed A/S
Priority date: 1996-04-12
Filing date: 1997-04-10
Publication date: 1997-10-23
Also published as: EP0893954A1; NO961453L; NO304631B1; NO304631B2; NO961453D0; AU2578797A

Abstract

There is described a method for the production of surface-coagulated pet food and feed for farmed fish, fur-bearing animals, birds and other animals, which is characterised in that the paste is made of by-products from the meat and/or fish industry and/or whole fish, plus some energy-giving meal, is heat treated at 100-200 °C for 5-60 seconds to apply to the feed a surface film of coagulated protein, the total heat supplied (time x temperature) being controllable when the buoyancy is adjusted in fish feed, whereupon the feed is cooled quickly after coagulation.

Description

PRODUCTION OF PET FOOD AND WET FODDER FOR FUR-BEARING ANIMALS AND FISH BY SURFACE TREATMENT

Surface coagulation of ground meat and/or fish paste is a method which has been developed for the production of pet food and wet feed for fish and/or fur-bearing animals. The feed is produced from a paste made of meat and/or fish by-products and/or whole fish, plus some energy-giving meal. Pellets of ground fish paste are surface treated with sufficient heat to apply a film of coagulated protein to the feed particles, whilst the protein in the rest of the feed particles will have a decreasing degree of coagulation towards the particle core. The method which is described in this application comprises cooking in hot animal and/or vegetable oil. The buoyancy of feed intended for fish can be adjusted by controlling the amount of heat applied. Cooking in hot oil (animal and/or vegetable) at different temperatures and for different residence times provides in addition the possibility of controlling the water content, fat content and fatty acid content of the feed.

By-products from fisheries, animal husbandry and marine aquaculture are a source of protein and fat, and the amount of by-products from the fish processing and foodstuffs industry is vast. To use large amounts of fish by-products for our traditional domestic animals is practically impossible, as such amounts would give the products an off taste. However, this is not a problem if the waste is used for pet food and feed for fur-bearing animals and/or farmed fish. There is already a limited use of fish waste in dry feed (feed having < 10% water content) for farmed fish and domestic animals. The production of dry feed for fish is nevertheless primarily based on the use of fishmeal and fish oil. The greatest potential for the use of fish waste resides therefore in an extended use of wet feed (feed having > 10% water content) in salmon farming.

Production of feed for fish is used here as an example, but the method described can also be used for the production of pet food and feed for fur-bearing animals (and optionally for feed for other animals and birds).

When salmon farming started in Norway at the beginning of the 1970s, production was based on wet feed. The protein and fat contents of this feed varied greatly and its structure was poor. This led to a high level of feed waste. Dry feed therefore soon became a great competitor of wet feed. In the course of the last ten to 15 years the use of wet feed for fish has almost come to a halt. The technical quality of dry feed has improved steadily, and intensive research on salmon feed has resulted in a salmon feed which today contains about 40% protein and 30% fat. This has improved feed utilisation, and at the same time has resulted in lower feed production costs as raw protein material is more expensive than fat (oils). Combined with less feed wastage, the consumption of feed has been reduced from about 3.5 kg of dry matter to about 1.1 kg of dry matter per kg of fish produced.

The cost of salmon production has therefore seen a dramatic reduction since the start in about 1970. However, the cost must in all probability be reduced further, and since feed costs account for more than 40% of the production costs, the price of feed can contribute to the reduction. Fish by-products are very inexpensive (0-70 øre per kg) as to get rid of such waste products often means extra costs for the fish processing industry. Wet feed is therefore once more of interest to fish farming.

New wet feeds must satisfy the requirements that modern fish farming has with respect to the physical structure of feed, and must not stick together, be crushed or dissolve too quickly in water. The object of the method described in this application is to attain such stabilisation by heat treating the surface of wet feed pellets (which chiefly consist of meat and/or fish by-products and some energy-giving meal) in hot oil. Treatment with hot oil will cause the pellet surfaces to coagulate, so that they neither stick to each other and surrounding equipment nor dissolve in water. The degree of coagulation will decrease towards the core of the pellet, and the depth of the coagulation will therefore be controllable.

The aim is that central units with access to meat and/or fish waste will in the future be able to provide salmon farmers with quality feed which is considerably cheaper than dry feed. The concept will mean great savings for the salmon farmers who are able to make use of it, whilst ensuring an enhanced utilisation of Norwegian fish resources.

Pet food and fur-bearing animal feed of good quality can be produced using the same equipment, optionally in the same factories.

Today's production of farmed fish is dry feed based, feeding is largely automated and the amount of feed is easy to control. However, the feed is expensive. Several wet feed concepts are therefore being tested, including gelling by using alginates and coagulation using microwaves.

In Norway, the production of dry feed for farmed fish is in the hands of five firms: T. Skretting, Felleskjøpet, Ewos, BioMar and Nor Aqua. All these firms supply feed of good physical and nutritious quality. The major disadvantage of dry feed is however its high price due to production using dry and processed raw materials (fishmeal. animal and/or vegetable oil, soya and pure starch products).

Gelled wet feed has a good physical structure and can be fed to its recipients automatically. Alginates however are both expensive and indigestible and may also reduce feed utilisation. Through-coagulation of wet feed by means of microwaves is, in terms of energy, very expensive. This is due to (1): microwave heat can only be produced using electricity, and the generation of microwaves has a heat loss at a temperature which is poorly suited for recovery. (2): Heating the pellets with microwaves takes place from the inside. The residual energy in a through-heated feed must therefore inevitably be very high compared with a surface-treated feed. Wet feed produced using microwave heat can therefore not provide very much cheaper feed than today's dry feed.

In NO 177731 a method is described wherein the feed is first extruded. To make extrusion possible, about 25% water is added which must be removed again after extrusion. In reality this is a drying method and one purpose of the process is to ensure a fat content. Bumping is achieved since the hot bath is at more than 100°C, and cooling takes place in a hot bath at about 50-60°C. In the method of the present invention, the purpose of the fat bath is however to achieve surface coagulation. Shape stabilisation or shaping is also to be achieved. The product is not extruded beforehand, but the paste is cut into suitable lengths which are dropped into the hot bath, and which are thus given a stable shape.

Surface-coagulated wet feed has special advantages. The method is based on the use of inexpensive, local, wet primary products, mixed with some energy-giving meal. It does not include the addition of indigestible binders.

Sustained high temperature at a low moisture level can cause heat damage to the protein and thus reduce both digestibility and growth. In surface coagulation however, the heat treatment takes place at a high temperature (100-200°C), whilst the feed particles are wet (<50% dry matter), and the period of heat transfer is very short (5-60 seconds). The surface of the feed particles is subjected to heat sufficient to apply a film of coagulated protein to the feed particles, whilst the protein in the rest of the feed particles will have a decreasing degree of coagulation towards the particle core. The method enhances the digestibility of the protein in comparison with raw ground meat and/or fish paste, probably due to a protein structure which is more readily available for enzymatic cleavage, and thus is also more readily available for absorption. A protein structure of this kind may also bring about a "protein mix effect" with different rates of solubility and digestibility of the protein in the different parts of the feed particle. In this case, this may lead to a more even absoφtion of amino acids from the intestinal tract and better utilisation of the amino acids for deposits in the muscle.

The economics of energy in the production of wet feed pellets are extremely important as the amounts of raw materials that are to be treated are large. In surface coagulation the consumption of energy for the heat coagulation process will be low initially as only the pellet surface is to be heated. Oil baths (animal and/or vegetable oil) can be heated wholly or partly using steam. More than 90% of the energy in a steam circuit can normally be recovered in such processes. The main amount of the energy consumption in the production will therefore be in the form of residual heat in the finished feed. This energy is difficult to recover.

Cooking in animal and/or vegetable oil and subsequent cooling gives solid pellets having a "meatball consistency" which do not stick, even after freezing. Cooking in oil reduces the water content in the feed and adds extra fat to the feed. This allows the possibility of adjusting the fat content in the feed and also of adding desired fatty acids. The higher fat content, probably together with the formation of gas bubbles in the feed particles during heat treatment, gives the oil-cooked feed entirely different buoyancy than raw ground fish paste. This creates the possibility of controlling the buoyancy in oil-cooked fish feed by controlling the temperature (degree of bumping) and residence time (degree of fat penetration) in the oil bath.

The fat content in the feed particles can also be increased by spraying on oil prior to cooling.

In order to obtain a wet feed pellet having the said physical properties, the pellets of ground meat and/or fish paste must be surface treated with heat, so that they are given a surface film of coagulated protein. The methods described in this application include as stated above: (1): cooking in hot animal and/or vegetable oil (deep-frying).

Examples

On cooking in hot oil, the ground fish paste is pelleted directly in the oil bath (animal and/or vegetable oil, 100-200°C), and is lifted out of the bath and passed into a cooling tunnel by means of a conveyor belt having buckets. The residence time in the oil bath (5-60 seconds) is controlled using the speed of the conveyor belt. The oil is heated using steam pipes or electric elements. The cooled feed is then put into suitable packaging, e.g., large sacks or containers.

I. Preliminary tests on digestibility of wet feed for salmon produced by surface coagulation in hot oil.

MATERIALS AND METHODS

Production of test feed

Mixing of premix

Two premixes were mixed in a food processor, both consisting of 2 kg gelatinised wheat-oat mixture (70% wheat + 30% oats), with 5 g of indicator added to it for determination of digestibility. Five grams of ytterbium oxide (Yb₂O₃) were added to Premix J, whilst five grams of yttrium oxide (Y₂O₃) were added to Premix 2. To ensure an even distribution of the indicator in the premixes, each indicator was first mixed with a small amount of the meal mixture. The rest of the meal was added gradually during the course of 15 minutes, and each premix was subsequently kneaded for 20 minutes.

Mixing of ground fish paste

Three parts whole herring (27 kg) and one part whole cod (9 kg) were ground twice. Half of the fish mass was frozen in blocks of 750 g, whilst the rest was kneaded as Premix 1 was gradually added. The finished mixture contained 100 g of wheat-oat mixture per kg and was kneaded for a further 20 minutes. The paste was then left in a cool place over night. Production of oil-cooked feed

Pellets of suitable size were produced with the aid of a food mixer with a sausage attachment and manual cutting of the string of paste. They were subsequently cooked in sunflower oil for 15 seconds at 180°C in a deep fryer, shaken dry, cooled on a wire tray, and frozen singly.

Production of untreated feed

The block of 750 g of ground fish paste was defrosted in a cold-storage room over night, kneaded in a food mixer as 83 g of Premix 2 was gradually added. The finished mixture contained 100 g of wheat-oat mixture per kg, and was kneaded for a further 20 minutes. The paste was then rolled out into strings and cut up into suitable pellets.

Digestibility tests

Fish and facilities

The test was carried out using salmon in seawater, in eight tanks (lm x lm), with 20 fish of 500 g in each tank Light was on continuously, and the feed was distributed 24 hours a day with the aid ofa belt auto feeder.

Dredging out of dung

Dung was dredged out for digestibility measurement on day 8, day 15 and day 18. All the samples of dung from each tank were combined. Due to small amounts of dung, the samples from pairs of tanks within each treatment were combined. Digestibility was further determined following the method described by Austreng (1978).

Chemical analyses

The determination of dry matter (105°C, 16-18 hours), ash (combustion at 550°C to constant weight), nitrogen (Kjeldahl N), fat (by hydrolysis) and gross energy (Bomb calorimeter) was carried out at AKVAFORSKs (Norwegian Institute of Aquaculture Research) laboratory at Sunndalsøra. The concentration of yttrium and ytterbium was analysed spectrophotometrically at the Agricultural analysis centre in As (Agricultural University of Norway).

Statistical analysis The results were analysed using one-way analysis of variance (n=2), with a level of significance selected at P<0.05 (5% probability of false rejection of the null hypothesis). The results are given as mean ± standard error of the mean. RESULTS AND DISCUSSION

Effects of heat treatment on the structure of the feed

The composition of the ground fish paste and the feeds is shown in Table 1. The finished paste had a viscous and fine consistency, but with only a 10% content of dry raw materials was still too sticky for pellet production. However, cooking in oil and subsequent cooling gave solid pellets having a "meatball consistency", which did not stick even after freezing. The cooking in oil reduced the water content in the feed from 60 to 51% and added extra fat to the feed.

Table 1. Composition of feed

Feed 1 Untreated Oil-cooked

Formulation, g/kg (prior to cooking in oil)

Whole herring 675 675 Whole cod 225 225

Premix l ¹ 100

Premix 2 100

Chemical composition (after cooking in oil) Dry matter (DM), g/kg 389.5 490.9

Protein, g/kg DM³ 427.9 384.3

Fat, g/kg DM 310.2 405.2

Ash, g/kg DM 51.5 42.1

Energy, MJ/kg DM 25.2 27.5

70% gelatinised wheat + 30% gelatinised oats, with 2.5 g/kg Y₂0₃ added.

70% gelatinised wheat + 30% gelatinised oats, with 2.5 g/kg Yb₂0₃ added.

N x 6.25.

The higher fat content, probably in conjunction with the formation of gas bubbles in the feed particles during the heat treatment, gave the feed cooked in oil a buoyancy entirely different from that of the raw ground fish paste. The ground fish paste sank immediately, whilst the oil-cooked feed could remain floating for about 24 hours. As a part of the test, a preference study was planned to investigate whether the oil cooking had any effect on the palatability of the paste. This was found impossible to carry out since the salmon took the sinking raw ground fish paste relatively quickly , but took longer to become accustomed to the floating, oil-cooked feed. A palatability study would therefore be greatly affected by the fishes' preference for floating or sinking feed. That all fish groups took feed over time indicates, however, that taste was not crucial to the reluctance of the fish to take the oil-cooked feed in the starting phase of the test.

Effects of heat treatment on digestibility

As can be seen from Table 2, cooking in oil had a positive effect on the digestibility of the feed. Of particular interest was the increased protein digestibility. The digestibility of both protein and energy was improved by 7%.

Table 2. Digestibility of the feeds

Feed 1 2 Untreated Oil-cooked

Protein¹ 84±0.2^b 90±0.2^a

Fat 96±0.4 97±0.3 Energy 83±l.l^b 89±0.2^a

^a indicate statistically certain differences ¹ N x 6.25

Sustained high temperature at a low level of moisture can cause the protein heat damage, and so reduce both digestibility and deposits in the muscle (Opstvedt et al., 1984; Pike et al., 1990). The heat treatment in this test was carried out at a high temperature (180°C), but the feed particles were wet (39% DM), and the duration of the heat transfer was short (15 seconds). Probably only the surface of the feed particles has been subjected to sufficient heat potentially to destroy protein. The protein in the rest of the feed particles seemed to have a decreasing degree of coagulation towards the core of the particle. All together, this may have given a protein structure which is better available for enzymatic cleavage, and is thus also better available to the fish.

Such protein structure may also bring about a "protein mix effect", with different rates of solubility and digestibility of the protein in the different parts of the feed particle. In this case, this may result in a more even absoφtion of amino acids from the intestinal tract, and better utilisation of the amino acids for deposits in the muscle. However, to examine such effects would call for further tests.

The fat digestibility was high in the case of both feeds, and was virtually unaffected by the heat and the extra addition of fat by cooking in oil. However, the extra fat supply reduced the ratio of digestible protein to digestible energy (g digestible protein/digestible MJ) from 17.2 to 14.1 g/MJ. The optimal ratio of digestible protein to digestible energy in salmon feed in the seawater phase is between 16 and 20 g/MJ (Einen and Roem, 1996). When cooking in oil, the paste should therefore be mixed with more protein than fat than in this test, either by using leaner raw fish material or by adding extra fishmeal. If one chooses to add fishmeal, one will at the same time obtain a drier and better consistency of the paste which is to be pelleted.

Conclusions

This test has shown that cooking ground fish paste in oil results in:

• A stable, coagulated particle structure

• Less water and more fat in the feed particles (more dry matter) • Better digestibility of the protein in the feed

Consequently, cooking in oil is a promising form of production for surface coagulated wet feed for salmon with respect to feed utilisation.

Proposed composition

Alternative 1 , Use pf fatty fish only,

Raw material, g/kg Part Drv matter Protein Fat Carbohydrate

Herring, 15% fat 850 314 153 128

Wheatmeal

(gelatinised) 100 92 12 3 70

Fishmeal 50 46 35 5

Total 452 200 136 70

Digestible energy.

MJ/kg Total From. Protein Fat Carbohydrate

9 8 4.27 4 84 0 72

Percentage of digestible energy 43% 49% 7%

Alternative 2, Use of lean fish onl\

Raw material, g/kg Part Drv matter Protein Fat Carbohydrate

Cod, 1 5% fat 750 365 135 15

Wheatmeal

(gelatinised) 100 92 12 3 70

Fishmeal 50 46 35 5

T otal 403 182 123 70

Digestible energy,

MJ/kg Total From. Protein Fat Carbohvdrate

9,0 3 88 4 37 0 72

Percentage of digestible energy 43% 49% 8% The calculations of digestible energy have been made given the following premises:

Protein: 90% digestible

Fat: 90% digestible

Carbohydrate: 60% digestible

Both the compositions are well-balanced with respect to the fishes' need for protein, fat and carbohydrates.

The compositions include the extreme situations where only fatty fish or lean fish are used. However, the content of wheatmeal and fishmeal are kept constant. Therefore by inteφolation one may easily arrive at the correct ratio of raw fish material to oil in new recipes which include both fatty and lean fish.

It will necessary to add vitamins, macrominerals and astaxanthin (colouring) to the feed.

II. Test carried out on digestibility of surface coagulated wet feed for salmon after different frying times in oil at 180°C.

MATERIALS AND METHODS

Production of test feed

The grinding of raw fish material, mixing the paste and deep frying the pellets was carried out by the Norwegian company Alamar AS, Svolvaer.

Mixing the ground fish paste The formulation of the ground fish paste is presented in Table 1. The paste was produced in two batches, but using the same composition: 82.9% raw fish material (frozen whole herring and frozen cod scraps), 1 1.7% dry meal (LT fishmeal and sieved wheatmeal), 4.9% fish oil and 0.5% water.

Before the fishmeal was mixed into the paste, lg of the indigestible marker yttrium oxide (Y₂O₃) per kg was added to it to give 50 mg of Y₂O₃ per kg of the ready mixed paste. To ensure an even distribution of the marker in the fishmeal, the Y₂O₃ was first mixed into a small amount of the fishmeal in a food processor. The rest of the fishmeal was then gradually added in a larger kneading machine.

Butylated hydroxytoluene (BHT) was added to the paste as antioxidant. Crystalline BHT was first dissolved in alcohol in the ratio of 60g BHT to one litre of alcohol, and 16 ml of this solution was then added per kg of fish oil before being mixed into the paste (50mg BHT per kg paste).

When mixing the pastes the defrosted raw fish material was first cut up and mixed in a large high-speed chopping machine. Then fishmeal, wheatmeal, oil and water were added in turn and mixed into the fish mass in the high-speed chopping machine. Finally the mixtures were passed through a microcutter with a cutting rate of 3000 - 4000 revs/min. The finished pastes had a temperature of about 30°C.

Table 1. Formulation and chemical composition of the paste

Paste, Paste,

1 st production 2nd production

Formulation, g/kg

Frozen whole herring¹ 585 585

Frozen cod scraps 244 244

Fishmeal 49 49 Wheatmeal⁴ 68 68

Fish oil⁵ 49 49

Water 5 5 Chemical composition

Dry matter (DM), g/kg 422 425 Protein, g/kg DM⁶ 421 439

Fat, g/kg DM 397 353

Ash, g/kg DM 62 66

Supplied by K.J. Ellingsen AS, Skrova, Norway Supplied by M. Johansen AS, Stamsund, Norway

Norse LT-94 (Nordsildmel, Bergen, Norway), with the addition of 1.0g Y₂0₃ (Sigma Chemical

Company, St. Louis, Mo, USA), per kg

Sieved wheatmeal (Statkom, Oslo Norway)

Red fish oil (BioMar AS, Myre, Norway) ), with the addition of 1.0g BHT (Sigma Chemical Company, St. Louis, Mo, USA) per kg ⁶ N x 6.25

Production of deep-fried feed

The test feed was produced in a modified fish ball cooker, where the mould diameter in the fish ball cutter was 1 1 mm. The cooker was six metres long and was filled with 1200 litres of oil, which was heated by 28 electric elements of 3.0 - 3.2 kW (a total of 86 - 90 kW). In the cooker there was mounted an inclined conveyor belt which emerged from the oil bath after three metres. By adjusting the speed of this conveyor belt the residence time of the paste balls in the oil bath could be controlled. From the deep-fryer the ready fried feed moved on a seven-metre open conveyor belt before passing onto a six-metre conveyor belt through an air-cooled tunnel.

Three different feeds (Table 2) were produced from each paste mixture:

Feed 1 , fried for 15 seconds Feed 2, fried for 25 seconds Feed 3, fried for 35 seconds

Consequently, each feed was produced from two separate paste mixtures in two separate frying batches. The production of each test feed started when the temperature in the oil bath reached 184°C. In each production the temperature then fell gradually to about 179°C. The core temperature in the feed particles was measured in 55±7.9 (mean± standard deviation) particles from each production using an electronic thermometer at the end of the conveyor belt in the deep fryer, i.e., three metres after the feed had exited the oil bath.

Table 2. Chemical composition of oil-cooked feed

lst productior L 2nd production

Seconds in oil bath 15 25 35 15 25 -25

Chemical composition

Dry matter(DM), g/kg 502 507 51 1 497 523 536

Protein, g/kg DM¹ 354 362 372 364 356 357

Fat, g/kg DM 461 492 471 497 517 505

Starch, g/kg DM 48 49 49 49 44 58

Ash, g/kg DM 65 57 52 56 49 56

Energy, MJ/kg DM 28.3 28.4 28.5 28.4 29.1 28.5

N x 6.25

The finished feed was collected in plastic trays, cooled in a cold-storage room (2°C) and frozen in a freezing room (-30°C) on the same afternoon. Digestibility tests

Fish, facilities and dredging out of dung

The test was carried out using salmon in seawater in 12 test pens (3x7x5 m) with 40 fish of 2.5 kg in each pen. Fish were put out 11 days prior to the start of the test to allow them to become accustomed to the test pens. In the test period the fish were fed according to appetite twice a day.

Samples of dung inside the pens were dredged out for digestibility measurement on day 19. Digestibility was further determined following the method described by Austreng (1978) using yttrium oxide as an indigestible marker (Austreng et al., 1996).

Chemical analyses

The determination of dry matter (105°C, 16-18 hours), ash (combustion at 550°C to constant weight), nitrogen (Semi-micro-Kjeldahl, Kjeltec-Auto System, Tecator,

Hoganas, Sweden), fat (petroleum ether extraction in a Fosstec (Tecator) analyser after HCl hydrolysis, Stoldt (1952)), starch (as described by Storebakken et al. (1991)) and gross energy (Parr 1271 Bomb calorimeter, Parr, Moline, IL, USA) was carried out at AKVAFORSK's laboratories. Yttrium in feed and dung was measured in ash (dissolved in aqua regia (HCl:HNO₃ 2:1 (v/v)) at AKVAFORSK) using an ICP spectrometer (Model 1 100, Thermo Jarrel/Ash, Franklin, MA, USA) at the Agricultural analysis centre in As.

Statistical analysis The results were analysed using one-way (core temperature in feed particles, n=2; digestibility, n=4) and two-way (digestibility, feed * cooking batch) analysis of variance by using the General Linear Models (GLM) procedure in the SAS software program (SAS 1985). For core temperature in feed particles the level of significance was selected as PO.10 (10% probability of false rejection of the null hypothesis). For digestibility the level of significance was selected as PO.05 (5% probability of false rejection of the null hypothesis). Significant differences were ranked using Duncan's Multiple Range Test. The results are given as mean±standard error of the mean. RESULTS AND DISCUSSION

Physical and chemical changes infeed. The measurement of the core temperature in the feed particles (Table 3) was somewhat unreliable as the speed of the conveyor belt exiting the oil bath was high. It was therefore difficult to position the measuring probe exactly in the centre of the feed particles. At the same time the cooling of the feed particles was already in progress, as the measurements were taken three metres after the feed particles had emerged from the oil bath. Together, this resulted in a relatively large variation between individual measurements.

Table 3. Core temperature in deep-fried feed particles after different periods in oil bath (Mean±standard error, n= 2)

Seconds in oil bath 15 25 35

Core temperature(°C) 86.5±4.9^b 95.4±2.8^a 96.0±l.l^a

Footnotes ^a indicate significant differences within column at 10% level.

The core temperature in the feed particles which were fried for 15 seconds was significantly lower than the core temperature in the particles which were fried for longer. There was no difference in temperature between the particles which were cooked for 25 seconds and those cooked for 35 seconds. This shows that the core temperature in the feed particles had reached boiling point after 25 seconds frying time at 180°C. That the temperatures measured at 25 and 35 seconds frying time are somewhat lower than 100°C is probably due to the fact that the temperatures were measured very quickly, that the feed particles had already cooled a little, and that the size of the feed particles was uneven, where the core temperature was still lower in the very largest particles. This last-mentioned is also illustrated by the fact that the variation in core temperature decreased when the frying time of the feed particles was increased from 25 to 35 seconds. The deep frying resulted in an increase in dry matter content in the feed and an increase in the fat content in the feed dry matter (see Table 4). In each production batch the water loss in the feed particles increased as the frying time increased. However, fat absoφtion was highest after the middlemost frying time, i.e., after 25 seconds. These calculations have been carried out on the basis of analyses ofa start sample of the paste from before each production batch, and would consequently be affected by non- homogeneous paste. Differences in particle sizes between the samples for analysis of the different feeds will also affect the calculations, as larger particles have a smaller surface area for fat penetration during deep frying (Orthoefer et al., 1996), and for absoφtion of fat on cooling (Banks, 1996). It is nevertheless probable that the reduction in fat content from 25 seconds frying time to 35 seconds frying time was an effect of the steam pressure in the frying feed particles reversing some of the oil absoφtion (Banks, 1996). This is supported by the fact that the fat content in the feed particles increased whilst the core temperature rose, whereas it fell somewhat after the core temperature had stabilised at boiling point.

Table 4 Changes in composition of feed on cooking in oil (Mean±standard deviation, n=2)

Seconds in oil Increase in Increase Water loss Fat absoφtion bath dry matter(%) i nn ffaatt ((gg//lkg paste)¹ (g/kg paste)'

% in dry matter)

15 7.6±0.6 10.4±5.7 163.5±6.0 74.5±29.0

25 9.2±0.9 13.0±6.0 205.0±33.9 86.5±29.0

35 10.011.611.3+5.5 213.5+41.7 77.5±37.5

' In relation to the concentration of yttrium in paste and feed.

Open deep frying gives both thermolytic and oxidative degradation of the frying oil, so that the oil will have an increasing content of virtually unavailable and/or toxic products the longer it is used (Marquez-Ruiz and Dobarganes, 1996; Orthoefer and Cooper, 1996; Orthoefer et al., 1996). Tests with rats have shown that mixing in more than 15% of used frying oil in the feed can result in weight reduction, and can in some cases also be toxic (Marquez and Dobarganes, 1996). The absoφtion of frying oil in the deep-fried wet feed is well below 15%. Nevertheless it will be important to ensure that the frying oil is changed adequately with a view to both turnover (i.e., the time it takes before all the oil has been changed on refilling) and change of oil in the deep fryer.

Digestibility

There was no difference in digestibility of nutrients between feed fried from 15 to 35 seconds, and nor was there any effect of production batch (Table 5). Compared with normal digestibility values of extruded fish feed found by using indigestible markers and dredging out dung (Refstie et al., 1996a, Refstie et al., 1996b; Storebakken et al., 1996), the values in this test were high. Deep frying of wet feed is therefore found to be a gentle heat treatment process, where the residence time of up to 35 seconds in oil at 180°C causes the protein in the fresh raw material very little heat damage, if any.

20 different DP/DE ratios (Table 5). Since the digestibilities of both fat and energy were the same for all the feeds, this was only an effect of different absoφtion of deep-frying oil at the different frying times (Table 4).

For dry feed the optimal DP/DE ratio in feed for salmon in the seawater phase is between 16 and 20 g/MJ (Einem and Roem, 1996). Higher water content and possibly better protein quality may perhaps cause the right DP/DE ratio to be lower in wet feed than in dry feed. In the wet feeds in this test the DP/DE ratio was however as low as 12.7±0.3 g/MJ (mean±standard deviation). When deep frying wet feed the paste should therefore contain more lean raw fish material and/or protein meal (fishmeal, soyameal, etc.) If more protein meal is added, the dry matter content in the finished feed will increase simultaneously.

Conclusions Deep frying of wet feed pellets (diameter =11 mm) at 180°C for 15, 25 or 35 seconds gives the following results:

• The wet feed pellets are heated through only after 25 seconds frying time.

• The dry matter content in the wet feed increases as frying time increases, due to vaporisation.

• The absoφtion of frying oil increases when the frying time increases from 15 to 25 seconds, but the absoφtion may be reversed somewhat when the frying time is further increased to 35 seconds. • The digestibility of nutrients is unaffected by frying times within the range of 15 to 35 seconds.

We therefore conclude that deep frying of wet feed is a gentle heat treatment process, where frying in oil at 180°C for up to 35 seconds does not cause the protein in the feed any heat damage.

III. Tests on digestibility of astaxanthin in surface coagulated wet feed for salmon after different frying times in oil at 180°C. 19

Table 5. Digestibility of oil cooked feed and the relationship between digestible protein and digestible energy in the feed after different periods in oil bath (Mean±standard error, n=4)

One-wav analvsis of variance in feed

Digestibility (%) DP/DE¹

Seconds Drv matter Protein Fat Energv (g/MJ) in oil bath

15 94.8±0.2 89.0±0.1 93.8±0.5 88.8±0.4 12.7±0.1

25 94.9±0.2 88.9±0.2 93.2±0.7 88.9±0.4 12.510.1

35 95.1+0.1 89.3±0.3 93.4±0.7 88.9±0.3 12.810.1

Two-wav analvsis of variance with interaction

Effect of

Seconds in oil bath N.S N.S. N.S. N.S. * Cooking batch N.S. N.S. N.S. N.S. N.S. Seconds x batch N.S. N.S. N.S. N.S. *

Footnotes ^{a c} indicate significant differences in the column at the 5% level * indicates significant differences at 5% level N.S. indicates no significant differences Grams digestible protein per digestible MJ

Preliminary studies of deep frying of wet feed for 15 seconds at 180°C in fact showed a positive effect on the availability of the feed protein (Refstie and Austreng, 1996). When compared, the results from this and the preliminary test may suggest that deep frying for up to 35 seconds has the same positive effect on feed protein.

As regards the ratio of digestible protein to digestible energy (DP/DE, g/MJ) in the various feeds, the differences were small. Nevertheless, different frying times gave MATERIALS AND METHODS

Production of test feed

Mixing of ground fish paste

The formulation of the ground fish paste is presented in Table 1. The paste was produced in two batches, but following the same recipe: 82.9% raw fish material (frozen whole herring and frozen cod scraps), 1 1.7% dry meal (LT fishmeal and sieved wheatmeal), 4.9% fish oil and 0.5% water.

Carophyll Pink, corresponding to 30 mg/kg wet weight, was added to the two different productions of paste. In the case of the first production of paste Carophyll Pink was dispersed in hot water (50°C) by stirring for about 30 minutes and then stirred into the oil phase. In the case of the second production of paste dry Carophyll Pink was added together with yttrium oxide (Y₂O₃). Before the fishmeal was mixed into the paste 1 g of the indigestible marker Y₂O₃ per kg was added to it to give 50 mg Y₂O₃ per kg of the ready mixed paste.

To ensure an even distribution of the marker in the fishmeal, Y₂O₃ was first mixed with a small amount of the fishmeal mixture in a food processor. The rest of the fishmeal was then gradually added in a larger kneading machine.

Butylated hydroxytoluene (BHT; 2,6-di-tert-butyl-p-cresol) was added to the paste as an antioxidant. Crystalline BHT was dissolved in ethanol (60g BHT/1). The solution (16 ml) was added to the fish oil before being mixed into the paste (50 mg per kg paste).

When mixing the pastes, the thawed raw fish material was first cut up and mixed in a high-speed chopper (Schneidmischer 325, Kramer-Grebe, Wallau/Lahn, Germany). Then fishmeal, wheatmeal, oil and water were added in turn and mixed into the fish mass.

The mixtures were then comminuted by means ofa microcutter with a cutting speed of 3000 to 4000 revs/min. The finished pastes had a temperature of about 30°C. Table 1. Formulation and chemical composition of the paste

Paste Paste

1 st production 2nd production

Formulation, g/kg

Frozen whole herring 585 585

Frozen cod scraps 244 244 Fishmeal³ 49 49

Wheatmeal 68 68

Fish oil 49 49

Water 5 5 Formulation, mg/kg Astaxanthin 30 30

Chemical composition, analysed

Dry matter (DM), g/kg 422 425

Protein, g/kg DM⁶ 421 439

Fat, g/kg DM 397 353 Carbohydrate, g/kg DM 120 142

Ash, g/kg DM 62 66

Supplied by K,J. Ellingsen AS, Skrova, Norway

Supplied by M. Johansen AS, Stamsund, Norway Norse LT-94 (Nordsildmel, Bergen, Norway) with an addition of l.Og Y₂0₃ (Sigma Chemical

Company, St Louis, Mo, USA) per kg

Sieved wheatmeal (Statkorn, Oslo, Norway) ⁵ Red fish oil (BioMar AS, Myre, Norway) with the addition of lg BHT (Sigma Chemical Company, St.

Louis,Mo. USA) per kg ⁶ N x 6.25

Production of deep-fried feed

The test feed was produced in a modified fish ball cooker, where the mould diameter in the ball cutter was 11 mm. The cooker was six metres long, and was filled with 12 litres of oil, which was heated by 28 electric elements of 3.0 - 3.2 kW (in total 86-90 kW). In the cooker there was mounted an inclined conveyor belt which emerged from the oil bath after three metres. By adjusting the speed of the conveyor belt the residence time of the paste balls in the oil bath was controlled. The ready-fried feed was conveyed further on an open conveyor belt (seven metres) and then through an air-cooled tunnel (six metres).

Three different feeds (Table 2) were produced from each paste mixture:

Feeds 1 and 4 cooked for 15 seconds Feeds 2 and 5 cooked for 25 seconds Feeds 3 and 6 cooked for 35 seconds

Each feed was consequently produced from two separate paste mixtures in two separate frying batches. The production of each test feed started when the temperature in the oil bath reached 184°C. The temperature then fell gradually to 179°C. The core temperature in the feed particles was measured in 5517.9 (meanlstandard deviation) particles for each production using an electronic thermometer at the end of the conveyor belt in the deep fryer, i.e., three metres after the feed had exited from the oil bath.

Table 2. Chemical composition of oil-cooked feed

1st production 2nd production

Seconds in oil bath 15 25 35 15 25 35

Chemical composition

Dry matter (DM), g/kg 502 507 51 1 497 523 536

Protein, g/kg DM¹ 354 362 372 364 356 357

, Fat, g/kg DM 461 492 471 497 517 505

Starch, g/kg DM 48 49 49 49 44 58

Ash, g/kg DM 65 57 52 56 49 56

Energy, MJ/kg DM 28.5 28.4 28.5 28.4 29.1 28.5

N x 6.25

The finished feed was collected in plastic trays, cooled in a cold-storage room (2°C) and frozen in a freezing room (-30°C) the same afternoon. Digestibility tests

Fish, facilities and dredging out of dung

The test was carried out using salmon in seawater, in 12 test pens (3x7x5 m), with 40 fish of 2.5 kg in each pen. The fish were set out 1 1 days prior to the start of the test to allow them to become accustomed to the test pens. In the test period the fish were fed according to appetite twice a day.

Samples of dung inside the pens were dredged out for digestibility tests on day 19.

Chemical analyses

Hόganas, Sweden), fat (petroleum ether extraction in a Fosstec (Tecator) analyser after HCl hydrolysis, Stoldt (1952)), starch (as described by Storebakken et al. (1991)) and gross energy (Parr 1271 Bomb calorimeter, Parr, Moline, IL, USA) was carried out at AKVAFORSK's laboratories. Yttrium in feed and dung was measured in ash (dissolved in aqua regia (HCl:HNO₃ 2:1 (v/v)) at AKVAFORSK) using an ICP spectrometer (Model 1100, Thermo Jarrel/Ash, Franklin, MA, USA) at the Agricultural analysis centre in As.

Analysis of astaxanthin in dung and feed. The dung was frozen on dry ice immediately after being dredged out in order to avoid decomposition of carotenoid. The samples were stored at -80°C until analysis. The sample was defrosted at room temperature and homogenised by stirring with glass rods. About 10.0 g of wet homogenised dung was weighed in and treated enzymatically with protease (Maxatase® P, 30mg; International Biosynthetics, Rijswijk, the Netherlands) in distilled water (10 ml) in an ultrasonic water bath (30 mins, 50°C) in order to break down the gelatine in Carophyll Pink™. Ethanol (100 ml) was added in a volumetric flask and dichloromethane was filled to 250 ml. An aliquot (10 ml) of the extract was purified using column chromatography with silica gel 60 (Merck, Darmstadt, Germany, No. 7733), and the eluate was steamed off. The sample was dissolved in an accurately determined volume (typically 1 ml) of n-hexane: acetone (86:14) and filtered (0.45 μm; Minisart SRP15, Sartorius, Germany) directly into a test tube for HPLC and stored at - 80°C until HPLC analysis. HPLC procedure for analysis of astaxanthin

The samples were analysed isocratically using HPLC on a silica gel column (Merck Hibar, LiChrosorb Si 60.5 μm particle size; inner diameter 4.6 mm, length 125 mm; flow rate 1.5 ml/min; (35 bar); detection wave length 470 nm) with 14% acetone in hexane as mobile phase, as described by Vecchi et al. (1987). The concentration was calculated with the aid of an external standard. The HPCL used was a Shimadzu LC- 10AS liquid chromatograph coupled to a Shimadzu SPD-M6A Photodiode array UV- VIS detector and Shimadzu CBM-10A Communications Bus module. The samples were injected by means ofa Shimadzu SIL-10 autoinjector. The chromatograms were reintegrated (Class LC10 software, Shimadzu, Japan) for base line correction. The retention time (R_τ) for astaxanthin was about 10.0 mins.

Standards

External standards were used to quantify the carotenoid content in the samples. Standards having a known concentration were made and run each time samples were analysed. About 3 mg of crystalline all-trαnj-astaxanthin (Hoffmann-La Roche Ltd, Switzerland) was weighed in a 100 ml volumetric flask and dissolved in 4.5 ml chloroform Then 95.5 ml n-hexane was added. An aliquot of this astaxanthin solution (10 ml) was pipetted into a second 100 ml volumetric flask, and 90 ml of 4.5% chloroform in n-hexane was added. The absorbance of this 10% solution was measured in a glass cuvette in a spectrophotometer (UV-260, Shimadzu, Japan) against a reference cell containing 4.5% chloroform in n-hexane.

The concentration was calculated according to the following:

Concentration = 10*Vol*Abs E ι_%, 1 cm

wherein

Vol = the volume of solvent used to make the standard Abs = the absorbency when measuring at 470 mn for astaxanthin E ι%, _{I c}m - 2100 for astaxanthin, in hexane at 1^^ = 472 nm. Determination of digestibilities

Apparent digestibility coefficients (ADC) for astaxanthin in the different feeds were determined using an indirect method after sampling dung as described by Austreng (1978). Yttrium oxide was used as an indigestible marker (Austreng et al., 1996). ADCs were calculated according to Maynard and Loosli ( 1969).

Statistical analysis

The results were analysed by means of analysis of variance using the General Linear Models (GLM) procedure in the SAS software program (SAS, 1985). For digestibility the level of significance was selected at PO.05 (5% probability for false rejection of the null hypothesis). The results are given as meanlstandard error of mean.

RESULTS AND DISCUSSION

Process stability of pretreated and untreated Carophyll Pink

In Table 3 the astaxanthin content in feeds having pretreated or untreated astaxanthin is given for the different treatment times in oil. There was a significant difference between the content of astaxanthin based on dry matter content in the feeds (pO.OOl). There was also a significant difference (p<0.05) between feeds to which untreated astaxanthin (45.411.7) was added and those to which heat-treated astaxanthin (38.210.9) was added. This may suggest that there h ; greater loss of astaxanthin through the process when Carophyll Pink is dispersed in water before being added to the feed paste. There were no significant effects of the treatment time on the content of astaxanthin based on dry matter content. However, this will be elucidated in more detail in a later report.

Table 3. Content of astaxanthin in oil-cooked feed immediately after production (mg/kg DM)

Seconds in oil bath 15 25 35

Astaxanthin. mg/kg DM

Pretreated 38.711.0^bc 36.211.3^C 39.7+1.8^bc

Untreated 45.4l2.7^ab 41.3l3.3^bc 49.311.5^a

Footnotes ^{a c} indicate significant differences (p<0.05).

Digestibility

Apparent digestibility coefficients for astaxanthin (ADC) for the different feeds are given in Table 4. There were no significant differences in digestibility of astaxanthin between feed fried from 15 to 35 seconds, although there was a tendency for the feed pretreated with astaxanthin to have a different digestibility after treatment for respectively 15 and 35 seconds (Table 4). Compared with digestibility values for canthaxanthin found when using indigestible markers and dredging out dung (Torrissen et al., 1990), the values in this test were very high. Comparable results were obtained for astaxanthin in rainbow trout (Choubert and Storebakken, 1996; Bjerkeng et al., 1997). Deep frying of wet feed is therefore found to be a relatively gentle heat treatment process, where a residence time of up to 35 seconds in oil at 180°C does not cause any significant damage to the fishes' ability to utilise astaxanthin.

Table 4. Digestibility coefficients (ADC) for untreated and water-dispersed astaxanthin in oil-cooked feed treated for respectively 15, 25 and 35 seconds in an oil bath (180°C)

ADC (%^•) Seconds in oil bath Heat-treated Untreated

15 85.510.7' 82.211.7

25 84.911.0 80.111.1

35 77.814. l^r 81-011.0

Footnotes ^tr indicate trends (0.05<p≤0.1).

There was a mild tendency for the digestibility of astaxanthin to diminish as the treatment time in the oil bath increased, Table 5.

There were no significant differences in digestibility of astaxanthin as a result of any pretreatment of astaxanthin, Table 6. One would in fact expect a somewhat higher digestibility of astaxanthin in fish which were given pretreated astaxanthin because the concentration of astaxanthin in this feed was lower. Lower carotenoid concentration gives higher digestibility (Torrissen et al., 1990; Smith et al., 1992; Choubert and Storebakken 1996).

Conclusions Deep frying wet feed pellets (diameter = 11 mm) at about 180°C for 15, 25 or 35 seconds results in:

• The digestibility of untreated astaxanthin was very high, probably as a result of adequate dissolution of gelatine during heat treatment in oil. • Water dispersion of Carophyll Pink leads to a more rapid degradation of astaxanthin.

• By prolonging the heat treatment time one may expect significant loss and reduced digestibility of astaxanthin. Table 5. The effect of heat treatment time on digestibility (ADC) of astaxanthin in oil- cooked feed.

Seconds in oil bath ADCΓ%)

15 83.9

25 82.5

35 79.4

Effect p=0.15

Table 6. The effect of pretreating astaxanthin on digestibility (ADC) of astaxanthin in oil-cooked feed.

ADC (%)

Heat-treated Untreated

82.3 81.3 N.S.

N.S. indicates no significant differences

Consequently, the conclusion may be drawn that deep frying of wet feed is a gentle heat treatment process, where frying in oil at 180°C for up to 35 seconds does not result in reduced digestibility of astaxanthin. The heat process per se seems to be sufficient to obtain a satisfactory digestibility of astaxanthin from Carophyll Pink without any apparent need for pretreatment.

SOURCE REFERENCES

Austreng, E., 1978. Digestibility determination in fish using chromic oxide marking and analysis of contents from different segments of the gastrointestinal tract.

Aquaculture, 13: 265-272 Austreng, E., 1994. Rubin-fόret, vatfδr til oppdrettsfisk, Utprøving av teknikk og fόringsforsøk, RUBIN-rapport nr. 302/36, Trondheim, 20s. Austreng, E. and Draget, K., 1996. Rubin-fόret: Fra ide til kommersiell produksjon. Norsk fiskeoppdrett, 21 (20): 38-39.

Austreng, E., Thomassen, M.S., Thomassen, Y., Refstie, S. and Storebakken T., 1996. Oxides of rare earth metals as new inert markers in digestibility studies with fish. Manuscript. Banks, D., 1996. Industrial frying. In: E.G. Perkins and M.D. Ericson (Editors), Deep Frying: Chemistry, Nutrition and Practical Applications, AOAC Press,

Champaign, USA, pages 258-270. Bjerkeng, B., Foiling, M., Lagocki, S., Storebakken, T., Olli, J.J. and Alsted, N., 1997. Different bioavailability of all E-astaxanthin and Z-isomers of astaxanthin in rainbow trout (Oncorhynchus mykiss). Manuscript. Choubert, G. and Storebakken, T., 1996. Digestibility of astaxanthin and canthaxanthin in rainbow trout as affected by dietary concentration, feeding rate and water salinity. Ann. Zootechn., 45: 445-453. Einem, O. and Roem, A.J., 1996. Growth, feed utilisation and body composition in Atlantic salmon (Salmo salar) fed extruded high energy diets of varying protein/energy ratios.

In course of preparation. Einem, O. and Roem, A.J., 1996. Dietary protein/energy ratios for Atlantic salmon in relation to fish size: Growth, feed utilisation and slaughter quality. Aquaculture Nutrition. In press. Marquez-Ruiz, G. and Dobarganes, M.C, 1996. Nutritional and physiological effects of used frying fats. In: E.G. Perkins and M.D. Ericson (Editors), Deep Frying: Chemistry, Nutrition and Practical Applications. AOAC Press, Champaign, USA, pages 160-182.

Opstvedt, J., Miller, R., Hardy, R.W. and Spinelli, J., 1984. Heat induced changes in sulfhydryl groups and disulfide bonds in fish protein and their effect on protein and amino acid digestibility in rainbow trout (Salmo gairdnerϊ). J.Agric. Food

Chem., 32(4): 929-935. Orthoefer; F.T., Gurkin, S. and Liu, K., 1996. Dynamics of frying. In: E.G. Perkins and

M.D. Ericson (Editors), Deep Frying: Chemistry, Nutrition and Practical Applications. AOAC Press, Champaign, USA, pages 223-244.

Orthoefer; F.T. and Cooper, D.S., 1996. Evaluation of used frying oil. In: E.G. Perkins and M.D. Ericson (Editors), Deep Frying: Chemistry, Nutrition and Practical

Applications. AOAC Press, Champaign, USA, pages 285-296. Pike, I.H., Andorsdόttir, G. and Mundheim, H., 1990. The role of fish meal in diets for salmonids. Int. Assoc. Fish Meal Manufacturers, Techn. Bull., No. 24, 35pp.

Refstie, S. and Austreng E., 1996, Fordøyelighet av vatfδr til laks produsert ved overflatekoagulering i varm olje. AKVAFORSK-rapport nr. 5/96, 1 1 pages.

CONFIDENTIAL. Refstie, S., Korsøen, 0.J., Storebakken, T. and Roem, A.J., 1996a. Soya to Atlantic salmon and rainbow trout: A comparative study. Manuscript.

Refstie, S., Storebakken, T. and Roem, A.J., 1996b. Adaptation to hulled and dehulled enzyme treated soybean meal in diets for Atlantic salmon. Manuscript. SAS, 1985. SAS/STAT® Guide for personal computers, Version 6 Ed. Cary, NC SAS

Institute Inc., 378 pages. Stoldt, W., 1952. Vorschlag zur Verienheitlichung der Fettbestimmung in Lebensmitteln.

Fette, Seiffen, Anstrichm., 54: 206-207. Smith, B.E., Hardy, R.W. and Torrissen, O.J., 1992. Synthetic astaxanthin deposition in pan-size coho salmon (Oncorhynchus kisutch). Aquaculture, 104: 105-119. Storebakken, T., 1985. Binders in fish feeds I. Effect of alginate and guar gum on growth, digestibility, feed intake and passage through the gastrointestinal tract of rainbow trout. Aquaculture, 47: 11-26. Storebakken, T. and Austreng, E., 1987. Binders in fish feeds II. Effect of different alginates on the digestibility of macronutrients in rainbow trout. Aquaculture, 60:

121-131. Storebakken, T., Hung, S.S.O., Calvert, C.C. and Plisetskaya, E., 1991. Nutrient partitioning in rainbow trout at different feeding rates. Aquaculture, 96: 191-

203. Storebakken, T., Refstie, S., Basverfjord, G. and Roem, A.J., 1996. Long-term effects of high energy soy diets for Atlantic salmon in seawater. Manuscript. Torrissen, O.J., Hardy, R.W., Shearer, K.D., Scott, T.M. and Stone, F.E., 1990. Effects of dietary canthaxanthin level and lipid level on the apparent digestibility coefficients for canthaxanthin in rainbow trout (Oncorhynchus mykiss). Aquaculture, 88:315-362. Vecchi, M., Glinz, E., Meduna, V. and Schiedt, K., 1987. HPLC separation and determination of astacene, semiastacene, astaxanthin and other keto-carotenoids. J. High Resolution Chromatogr. Chromatogr. Commun. 10: 348-351.

Claims

P a t e n t c l a i m s

1.

A method for the production of surface-coagulated pet food and feed for farmed fish, fur-bearing animals, birds and other animals, characterised in that the paste is made of by-products from the meat and/or fish industry and/or whole fish, plus some energy- giving meal, is heat-treated at 100-200°C for 5-60 seconds to apply to the feed a surface film of coagulated protein, the total heat supplied (time x temperature) being controllable when buoyancy is adjusted in fish feed, whereupon the feed is quickly cooled after coagulation.

2.

A method for the production of surface-coagulated pet food and feed for farmed fish and fur-bearing animals according to Claim 1 , characterised in that the ground fish paste is pelleted, surface coagulated in a hot oil bath (animal and/or vegetable oil) and the water content, fat content and specific fatty acids content of the feed are adjusted by controlling the residence time in the oil bath and/or the temperature in the bath, and subsequently the ground fish paste is lifted out of the oil bath for cooling.

3.

A method for the production of surface-coagulated pet food and feed for farmed fish and fur-bearing animals according to Claim 1, characterised in that surface-coagulated pellets are sprayed with oil (animal and/or vegetable) prior to cooling.

4.

A method for the production of surface-coagulated pet food and feed for farmed fish and fur-bearing animals according to Claims 1 to 3, characterised in that the surface coagulation is carried out to improve the digestibility ofa ground meat and/or fish paste.