Co-products are used in the feed industry worldwide, especially those that come from the cereal and flour industry or from rendering at meat and poultry processing plants. Co-products such as wheat middlings, meat and bone meal and feather meal are widely used and have been extensively tested and described. In contrast, in developing countries emerging human food industries related to fruits and plants are generating increasing amounts of co-products that are largely unexplored as potential feed ingredients. Knowledge about their management, processing, and use in animal feeds increases as their availability increases. However, this research rarely follows a planned strategy that encompasses issues beginning with co-product generation and ending with use in animal diets. Either due to scarce funding and/or limited facilities most research efforts focus on only one or two isolated issues related to coproduct use. The purpose of this paper is to present a holistic protocol for the evaluation of new feed ingredients, using the working example of passion fruit seed meal (PFSM).Why use unique local ingredients?
The search for alternative feed ingredients is driven directly by the economics of the industry to find nutrient sources at lower costs than those currently in use. Several countries depend heavily on imports of commodities such as corn and soybean meal for animal feeding. The search for alternative feed ingredients can lower dependency on imports. Alternative feed ingredients are usually co-products of local agricultural industries. Thus, finding a niche in animal feeding for co-products adds value to crops and promotes wealth in rural areas. Because most unexplored feed ingredients come from non-traditional crops, such uses support agricultural diversification and diminish the vulnerability of narrowbased economies to environmental and market fluctuations (Prasad and Chandra, 1980). Diversification usually involves crops that require a high labor input, and are otherwise uneconomical in developed countries due to high costs of labor.
Some examples of non-traditional materials that are available for use as feed ingredients have been presented by FAO (2005). Some of these ingredients require processing (e.g., grinding, heating, washing). However, the main concern in deciding whether to invest money and time in developing a feed ingredient is its availability.
Examples of available ingredients, ranked by suitability as feed or food, and limitations for their use, are shown in Table 1. Celery leaves, which are rich in protein and carotene, can be used to substitute other sources of pigments. Coffee pulp, although mainly used for ruminants, can be safely used at up to 5% in broiler diets if well processed. Mango seeds are not for use in nonruminants because of tannin contents, but can be used at levels up to 50% in rations for cattle. Tagua, a seed used in Ecuador for manufacturing buttons and other ornamental articles, has not been adequately evaluated; however because of the carbohydrate (mannans) content it might have potential use as an energy source.Feed ingredient evaluation protocol
A holistic evaluation of any new feed ingredient must not only include price and availability assessments and studies of its nutrient content, but also look at issues such as desirable physical characteristics, transportation and storage issues, performance at the process mill, and secondary effects on animals and the environment (e.g., quality of feces). Steps to evaluate a potential feed ingredient are described in Table 2. Not all steps may be needed for the evaluation of all ingredients since some information may already be available or not apply.
Likewise, the order in which steps are listed may not necessarily be ideal. However, it has been in the authors’ experience that problems in the daily use of nonconventional or new feed ingredients arise because steps were not followed and detailed information was not generated to properly evaluate the usefulness of a new ingredient. This protocol is presented using passion fruit seed meal as an example.
Table 1. Uses and ratings1 for selected, non-traditional ingredients.
To enlarge the image, click here 1Ratings indicate relative frequency of use in feed and food, 5 is high and 0 is low.Table 2. Protocol for evaluating potential feed ingredients.To enlarge the image, click here
STEP 1: ASSESS AVAILABILITY
Passion fruit (maracuya, maracuja, couzou, or parchita) is cultivated in tropical and subtropical areas of Ecuador, Colombia, Venezuela and Brazil. It is also produced in Australia, Kenya, Zimbabwe, Philippines, and several Pacific islands (Morton, 1987). Two main varieties are cultivated: yellow (Passiflora edulis var flavicarpa Degener) and red (var purpurea). Brazil is the largest producer in the world (Ferrari et al., 2004). Ecuador only began exporting passion fruit in 1986. However, by 1990, it was already the second largest exporter of concentrated juice from passion fruit in the world after Colombia. Passion fruit juice is exported to the European Union, United States, Chile, and Argentina, where it is valued for its flavor and unique aroma. It is rich in vitamin A, vitamin C, niacin, and flavonoids.Quantity
In 2000, Ecuador exported 27,000 tons of concentrated juice (CORPEI, 2001). In 2003, a production of 168,569 mT of fruit (approximately 50,000 tonnes of juice) represented $USD40 million in exports (Luna, 2005). Ferrari et al. (2004) estimated worldwide production at 364,000 tonnes of passion fruit.
Passion fruit is an alternative crop suited for small areas. Its cultivation is labor intensive because vines must be supported by wires, pollination is done by hand, and harvesting requires pulling and sorting of the fruit. In Ecuador, about 50 thousand families directly or indirectly benefit from this activity (CORPEI, 2001).
Although it is grown year-round, there are two production peaks: April through June, and the month of October. Market values of the fruit in the late 1990s plummeted in 2000, pushing farmers toward other crops. Subsequently, prices have increased followed with a better-planned supply (Luna, 2005).
With the stable demand for passion fruit juice, passion fruit seed meal (PFSM), a co-product of passion fruit processing for juice, has emerged as a potential feed ingredient worthy of attention. Quantities produced assure a reliable seed supply for meal manufacturing, and the economy of its production and international markets is promising enough to speculate on long-term availability. Thus, a promising feed ingredient has emerged.Processing and handling
Fruits after harvest need to be rapidly transported to the processing plant to avoid water losses and prevent spoilage from rapid ripening (Morton, 1987). At the extraction plant, after washing the fruit, the seedy pulp is separated from the peels. Afterwards juice is extracted and pulp residue with seeds is the resulting co-product.
Typical fruit composition is: peels 50 to 60%, juice 30 to 40% and seeds 10 to 15% (Morton, 1987; Texeira, 1994). Yield of seeds varies with the variety, but is between 6 and 12% in red passion fruit and 2 to 12% in yellow passion fruit.
Leftover pulp with seeds contains about 60% moisture. The process by which PFSM is made starts by drying the leftover pulp. First, a water separator is used to reduce moisture to 30 to 40%, then a cylindrical drier is used until a final moisture content of 12% is obtained.
Once dried, there are two alternatives: either extract the oil or process the whole seed for animal feed. In the first process, 20 to 23% of oil is mechanically extracted from the seeds and the remaining 77 to 80% is left with the meal. For the whole PFSM, dried seeds and leftover pulp are ground and bagged. Physical characteristics of PFSM follow:
Mean particle size 511 ± 80 μm
Density 550 ± 50 g/L
Analytical results and physical characteristics suggest that some limitations on handling of the feed ingredient should be considered. High content of unsaturated oil in a small particle size necessitates rapid rotation of the ingredient at the processing plant and then at the feed mill to maintain oxidative stability.STEP 2: IDENTIFY BASIC NUTRIENT PROFILELiterature review
Little information has been published about the use of PFSM for animal feeding. Otagaki and Matsumoto (1958) and Teixeira (1994) very briefly describe the use of rinds as a feedstuff for ruminants. Texeira quotes Akiki et al. (1977) who used up to 8% seed meal in broiler diets.
Morton (1987) briefly describes the use of chopped and dried rinds combined with molasses as feed for cattle or pigs, but mentions that seeds are not suitable for consumption, with no further explanation. On the contrary, Dos Reis et al. (2000) clearly showed that residues from passion fruit are alternative feed sources for ruminants. They tested silages made of various passion fruit residue-grass mixtures ( 100 vs 25, 50 or 75%) and found that only 100% passion fruit residues in silage reduced feed intake. Interestingly, passion fruit residues showed better nutrient digestibility than elephant grass. In addition, Ferrari et al. (2004) reported improvements in milk production when dairy cattle were fed passion fruit residues.
Otagaki and Matsumoto (1958) reported that rats fed a mixture of 80% skim milk and 20% passion seed oil for 9 days had weight gain not significantly different from rats fed 80% skim milk and 20% cottonseed oil.Proximate analysis
Table 3 shows the proximate composition of PFSM from our lab and from Pruthi (1963) compared with that of corn and soybean meal. Proximate composition is rather stable and any variation is mainly attributable to differences in moisture. Therefore, drying processes must be improved and controlled. The protein content is of marginal value, but is high enough to allow the ingredient to be considered as more than just a fiber and energy source. Mineral content is low as shown by low ash, calcium and phosphorus content. Ferrari et al.
(2004) reported that extracted PFSM contains 12.4% carbohydrates (~9.5% in full fat PFSM). The obvious ingredient strength is the high oil content that makes it an energy source. These nutrient combinations would be of interest in diets for layers, breeders, sows, and ruminants.Table 3. Chemical composition of passion fruit seed meal compared to corn and soybean meal (% as is).To enlarge the image, click here
1Analyzed following AOAC methods (AOAC, 1995)
2PRONACA Laboratory Database 2003 – 2005
4Nutrient Requirements of Poultry, 9th Revised Ed. 1994
5Assayed at the Department of Poultry Science, University of Georgia
6Cellulose contentCost effectiveness
Once a basic nutritional composition has been obtained for the ingredient, formulation exercises will reveal shadow prices and tentative inclusion levels in diets. To determine whether the cost per unit of nutrients is economically attractive, the ingredient cost can be divided by the main nutrient content. More accurate estimates can be obtained by using formulation software to determine the value of the nutrient combination.
STEP 3. PERFORM PRELIMINARY ANIMAL ASSAY
Once the ingredient is determined to be cost effective, the next step is to perform a preliminary, small-scale feeding test to assess major concerns such as obvious toxic effects, palatability problems, anti-nutritional factors, and major nutrient deficiencies not evidenced in the proximate analysis. The easiest and least expensive way to conduct such an evaluation is in a 21-day trial in batteries with day-old broilers. Baby chickens are susceptible animals that would easily show the effects listed above. At this point, energy values are estimated using regression equations. Diets are usually formulated to cover requirements or even have a marginal deficiency in the nutrients that are of major concern so as to add an extra nutritional challenge that will further magnify any deficiency or problem that the ingredient may produce.
A control with no test ingredient together with two or three inclusion levels of the ingredient are typically used.
The range of levels tested must be greater than that expected in eventual formulations. These levels may pose a dilemma to the researcher on whether to balance the diet or test high inclusions that might have the unintended effect of unbalancing the ration. Minor differences in protein and energy between treatments should not be a concern since only major differences are of interest at this stage.
PFSM was initially tested in 100 birds housed in a metallic five-floor battery with each floor divided into two cages. One-day-old broilers were randomly distributed with 10 birds per cage. Treatments consisted of 0, 10, and 17% PFSM inclusion in a typical broiler starter feed. The latter level was the highest inclusion that could be obtained without incurring major nutrient differences between treatments. Broilers were fed the experimental diets for 21 days. Metabolizable energy of PFSM was estimated at 1950 kcal/kg based on the energy value of the oil and protein and somewhat penalizing the high fiber content. Experimental feeds and nutrient composition are shown in Table 4 and results in Table 5.Table 4. Ingredients and nutrient composition of diets used in preliminary trial.10.3% sodium bentonite, 0.11% mineral premix, 0.10% vitamin premix, 0.10% mold inhibitor, 0.05% choline chloride and 0.015% antioxidant.
Table 5. Performance of broilers to 21 days fed diets with passion fruit seed meal.
There were no significant differences between treatments for any of the performance variables (Table 5). Results showed that there are no apparent deleterious effects of PFSM at the levels used in this experiment. A high mortality in the 10% PFSM treatment was not related to dietary treatments. Although higher levels of PFSM increased dietary fiber content, they did not affect feed consumption. This preliminary trial indicated that passion fruit can be included in broiler diets.
Metabolizable energy was underestimated and once it is corrected and other nutrients included in the formulation, practical PFSM inclusion levels will be lower. This first trial showed that PFSM did not have any major deleterious effects on chickens and that subsequent trials should focus on ingredient inclusions of not more than 10% of the diet.STEP 4. PERFORM DETAILED NUTRIENT CHARACTERIZATION
More detailed nutritional characterization is required to further study potential ingredients and run more detailed formulation simulations. Typically, amino acid composition, fatty acid composition, and metabolizable energy are determined. Efforts in chemical analysis should be focused on the nutrients of major importance.
Metabolizable energy determination in vivo is only recommended if the ingredient is primarily considered to be an energy source (ME >2500 kcal/kg) and inclusion in diets would be >5%. Otherwise, energy values estimated with equations are probably exact enough for practical use.
In the case of PFSM, one of the most valuable nutrients is fat. The fatty acid profile of the oil is shown in Table 6. About 65 to 70% of the oil present in the seed is linoleic acid, an essential fatty acid for animals.
The fatty acid profile of the oil is somewhat similar to that of soybean oil, except for a higher linoleic acid content. Due to its high unsaturation, it is expected to have a high digestibility, as has been confirmed by Otagaki and Matsumoto (1958) who reported a 98.4% digestibility coefficient in rats.
Favorable characteristics of the oil are drawing the attention of the industry, which is currently shifting from the production of passion fruit juice and full-fat PFSM to the production of juice, oil, and defatted seed meal.Amino acids
Table 7 shows the PFSM amino acid profile of four different samples. Samples were taken every four months during 2003 and 2004. Notice that the protein content varies from sample to sample with no apparent explanation. When expressed as percentage of protein content, essential amino acid composition of PFSM is slightly lower than that of corn except for methionine and arginine. Arginine, phenylalanine, and glutamic acid contents are higher than those of corn and soybean meal. The variation in the data is largely due to both the small number of samples analyzed and an unstable drying process of the seeds.Table 6. Quality parameters, fatty acid and ester composition of passion fruit seed oil (% as is).To enlarge the image, click here
1Measured at 25oCTable 7. Amino acid content of passion fruit seed meal, corn and soybean meal.
To enlarge the image, click here
1Pronaca database 2004. Samples of passion fruit seed meal were analyzed by Degussa Analytical Lab, Hanau, Germany.
2Amino acid values from Aminodat. Degussa, Hanau, Germany.Restrictions on use?
Many co-products have limitations either due to a high content of an undesirable nutrient (Table 2) and/or because they have toxic compounds and anti-nutritional factors.
Passion flower has a complex chemical composition, which is responsible for its unique organoleptic and pharmacological characteristics (Taylor, 2005). The plant is rich in three main groups of active chemicals: alkaloids, glycosides and flavonoids. These compounds are present in different concentrations in flowers, leaves, stems, and fruit. We hypothesized that although most of them would be present in the seed, concentrations may be low enough to avoid the characteristic effects, such as sedation, relaxation, diuretic, antispasmodic and analgesic responses that have been described by Taylor (2005) for passion flower.
Chassagne et al. (1996) described five cyanogenic glycosides in passion fruit pulp and peels. Glycoside levels are higher in unripened and rather low in ripened fruits to be considered of toxicological significance (Texeira, 1994). Regardless, reports do not indicate whether glycosides are present in seeds and whether continuous feeding of animals during extended periods of time (i.e. fattening of swine and beef cattle, layers during a production cycle) poses a risk of accumulation that may affect performance.
Volatile compounds that produce the aroma of the juice were recognized by organoleptic analysis in the seed meal. They are likely to be mainly hexyl caproate since ethyl butyrate, ethyl caproate, and hexyl butyrate disappear during the first three days of storage (Texeira, 1994). Whether animals would be attracted or repelled by the smell is an issue to be considered during later evaluation.
None of the above compounds have been reported in high enough concentrations to be considered toxic.
However, it is not known whether low concentrations may affect animal performance.
Finally, effects of passion fruit seed fiber fractions have been described regarding in vitro absorption of glucose and retardation of amylase activity, because of their water- and oil-holding capacities that could be of some significance in vivo (Chau and Huang, 2003).STEP 5: CONDUCT A CONTROLLED ANIMAL TRIAL
The quick battery trial indicated that the ingredient PFSM is of interest. With the more detailed nutritional information, the next step is to run a well controlled and carefully designed trial at an experimental facility.
Proper replication, treatment and experimental design are required to obtain reliable results. Replication depends on variability of the characteristics to be measured and the power to detect significant differences.
In the authors’ experience, rarely are the number of replications sufficient to satisfy statistical requirements (typically >15 replicates per treatment are needed). Nevertheless, we recommend at least eight replicates per treatment in experiments with broilers. The objective of these trials is to determine narrow ranges of ingredient use, performance issues in animals (interactions, toxic compounds, anti-nutritional factors), and the economics of using an ingredient based on production variables.
At this time PFSM supply is restricted. A formulation study was conducted to determine in which formulas the co-product will be best used for its nutrient content and price. The study showed that PFSM was more cost efficient in layer diets. A controlled trial was designed and conducted in Pronaca’s experimental house near the city of Quevedo, Ecuador. A complete randomized block design was used whereby 616 Hyline brown layers of 41 wks of age were randomly distributed in 14 replicates, each replicate consisting of 11 cages with 4 hens each. Treatments consisted of a control diet without PFSM and a diet that included increasing amounts of PFSM starting at 2% inclusion, increasing to 4% at 3 wks, and 6% at 6 wks. Rations shown in Table 8 were formulated to be isonitrogenous and isoenergetic and to comply with nutritional requirements suggested by Hyline (2000). No significant differences were found in production variables during each of the three periods corresponding to 2, 4, and 6% dietary inclusions.
Therefore averages across the 12 wks of the experiment are presented (Table 9). During the last 6 weeks of the trial, weekly samples of feces were obtained from each replicate and moisture content determined. Egg quality parameters presented in Table 9 were measured every 3 weeks.Table 8. Ingredients and nutrient composition of diets in layer trial.
10.25% sodium bicarbonate, 0.1% salt , 0.1% mineral premix, 0.12% vitamin premix, 0.05% mold inhibitor, 0.15% choline chloride, 0.015% antioxidant and wheat middlings.
Table 9. Performance of layers fed diets with passion fruit seed meal.
There were no differences (P>0.10) in any of the variables measured in the experiment (Table 9). Animals did not reject feed; thus any flavor and/or scent provided by PFSM were not deleterious. Nutritional composition of diets (Table 8) varied in the amount of total fat and crude fiber, both being higher as inclusion of PFSM increased. These differences were within a range that did not affect body weight and quality of feces. No differences in egg shell quality or Haugh units were observed. In conclusion, results showed that PFSM is an alternative ingredient that can be included in rations for layers up to 6% without any negative effect.STEP 6. CONDUCT FIELD TRIAL
The last step before the ingredient can be regularly used in formulation is to run a large-scale field trial. Ideally, the design should include replication and two treatments: diet without the ingredient and diet with the ingredient at the inclusion level determined by previous trials.
However, due to house–to-house variations, delivery of animals, and other practical issues, exact replication is difficult to obtain. The objective of the field trial is to study more subtle effects that would be difficult to evaluate at the experimental house. Research facilities have generally higher sanitation, better control over environmental fluctuations and other challenges, and the small number of animals involved could mask negative effects of low incidence. Typical variables to study, besides performance measures, are uniformity, quality of feces, quality of carcasses, feathering, diseases and behavior. None of these issues were found to be affected by the use of PFSM at the levels studied.
STEP 7. ADDRESS MANUFACTURING ISSUES
The flow of the feed ingredient in the manufacturing process at the feed mill must be considered before an ingredient is adopted. So far, it has been shown that PFSM is of interest and can be used successfully as a feed ingredient in diets for broilers and layers.Storage
In the case of PFSM, it was found that due to the high oil content, bags could not be piled over 10 bags deep on pallets, otherwise the excessive weight would squeeze the oil from the bags located at the bottom. We also recommend that pallets be placed over a finished floor that would allow cleaning with detergents to remove eventual oil spills.Meal stability
When initial experiments were conducted it was noted that whole seed meal had a rapid oxidation within the first 3 days after production. Low pH of the juice as well as release of several enzymes such as stearases, polyphenolases, catalases and peroxidases from the peel were attributed the phenomena (Texeira, 1994). Figure 1 shows the rapid increase that occurs in peroxides during the first 30 days of storage; afterwards they decrease.
Free fatty acids increased with time showing that oil is very unstable. In order to improve quality and extend ingredient shelf life a series of changes in the process were worked out with the supplier that included the removal of the dried pulp from the seeds, as in the rice polishing process. Meal produced in this way typically lasts up to eight days at 80% moisture and 30oC with peroxides under 5 meq/L. It was also found that meal would gradually lose its characteristic odor during storage yet still provide the final feed a slight scent. Use of antioxidants in the meal at the time of manufacturing is advised.
Particle size and pellet quality The most efficient particle size for suppliers is around 1000 μm, which can be adequately used by layers in mash diets. However, when used in diets for broilers, swine or ruminants that will be pelleted, PFSM needs to be ground to around 500 μm to obtain good pellet durability.STEP 8. ENGAGE IN CONTINUOUS QUALITY IMPROVEMENT
By successfully completing Steps 1 through 7, we have shown that PFSM can be successfully used in the formulation of diets for broilers and layers. Results from other trials, not reported in this paper, have shown that it can also be used in diets for ruminants and swine.
Once an ingredient is introduced, it is recommended that periodic sampling and chemical analyses be performed. Frequency of analysis will depend on variability of the ingredient and reliability of the supplier.
PFSM has been shown to have a fairly stable nutrient composition. Thus, moisture and perhaps crude protein and ether extract are analyses that should be completed on a regular basis. Other analyses should be performed every four to six months. Results should be entered in a database so continual monitoring of trends can be performed to determine when changes in nutritional matrices are required.Figure 1. Peroxide and free fatty acid changes during two months of storage of PFSM.
Even if we think that we will be making better use of a material with relatively low value, the final decision to do so rests on economics. Co-products must be inexpensive because they usually provide marginal levels of nutrients. Thus, any effort to maintain lower prices and improve nutrient quality and digestibility will help ensure a continued market for the co-product.
For example, during the processing of passion fruit seeds, time and energy are spent in drying the pulp.
More economical processing alternatives have been tested to separate seeds from pulp, including wet screening and centrifuging. These methods have had little success due to the viscous nature of pulp that adheres to the seed. This gelling effect has been attributed to galacturonic acid from pectin (Texeira, 1994) and to starch (Morton, 1987). An alternative enzymatic process that disrupts the sacs surrounding each seed was suggested by Morton (1987).
In addition, leftover pulp itself may turn out to be another co-product that is currently not used due to the high cost of drying. Air-dried pulp has the following nutrient composition: 12.7% moisture, 14.6% crude protein, 10.4% crude fat, 29.5% crude fiber and 49.6% neutral detergent fiber (Dale, personal communication).
Several alternatives have been proposed for upgrading low nutritional quality ingredients or co-products with mechanical, thermal, or chemical processes. The latter generally relies on the use of enzymes. The approach in using enzymes applied to the diet with the objective of improving nutrient digestibility and/or ameliorating detrimental effects may not be appropriate when dealing with co-products of low inclusion. Lack of sufficient substrate for enzyme catalysis may explain poor responses to enzyme addition in trials conducted at our facilities and others. Perhaps a better approach, one that the authors are currently evaluating, is the application of enzymes directly onto the ingredient at some later steps of processing as is done with enzymes for feather meal (Hanley et al., 1998). Bedford (2002) advises that trials evaluating enzyme application should be carefully planned particularly considering interactions between dietary ingredients, dietary nutrient density, and substrate.
We have proposed a holistic approach to evaluating new feed ingredients. Steps recommended in the evaluation of a new feed ingredient include a study of its availability in sufficient quantities to secure a steady supply to the feedmill. A quick nutritional characterization by proximate analysis and use of equations to estimate energy will suffice to evaluate whether the ingredient is of interest. Then a quick animal assay in day-old chickens using a wide range at two or three concentrations of the ingredient will provide information about potential problems and probable levels of use in diets. Once a potential use is determined, a detailed nutrient analysis, including amino acids, fatty acids, digestibility and metabolizable energy evaluation, should be performed. These data can be used to plan a well designed and replicated trial at an experimental facility. This trial will provide the information on the adequate levels at which the ingredient should be used, that will be later verified in a field trial. Having determined that the ingredient can be successfully used in animal diets is not enough. Logistic, storage, and processing issues during feed manufacturing should be studied and evaluated. Finally, once the ingredient is used on a regular basis, monitoring of nutritional quality and economic advantages must be continued. New improvements in processing and the increased availability of enzymes may further justify use of the feed ingredient.
By following this protocol, it has been show that PFSM is a valid alternative feed ingredient in diets for layers and broilers. The use of enzymes could potentially upgrade this ingredient and further justify its use in formulation of animal diets.Acknowledgements
The authors express their appreciation to Dr. Nick Dale, Department of Poultry Science, University of Georgia, for his valuable contribution to this paper.
Authors: GERMAN ROMO1,2 and GUSTAVO NAVA1
Akiki, J., P.R. Toledo, C. Ruggiero and J.C. Oliveira. 1977. Aprovetamento de cascas desidratadas e sementes de maracuja amarelo (Passiflora edulis F. Flavicarpa Derg) na alimentacao de frangos de corte. Cientifica 3(3):340-343.
Association of Official Analytical Chemists. 1995. Official methods of Analysis. 16th ed. AOAC, Arlington, VA.
Bedford, M.R. 2002. The foundation of conducting feed enzyme reserarch and the challenge of explaining the results. J. Appl. Poult. Res. 11:464-470
Chassagne, D., J.C. Crouzet, C.L. Bayonove, and R.L. Baumes. 1996. Identification and quantification of passion fruit cyanogenic glycosides. J. Agric. Food Chem. 44 (12):3817-3820.
Chau, C.F., and Y.L. Huang. 2003. Characterization of passion fruit seed fibres – a potential fibre source. Food Chemistry 85(2):189-194.
CORPEI. 2001. Product profile: sour passion fruit concentrate. CORPEI – CBI Project “Expansion of Ecuador’s export commodities”. http:// www.sica.gov.ec/agronegocios/productos% 20para%20invertir/CORPEI/maracuyá.pdf
Dos Reis, J., P.C. De Aguiar Paiva, I. Von Tiesenhausen, and C.A. De Rezende. 2000. Chemical composition, voluntary intake and digestibilityof the silages of passion fruit residues with elephant grass cv. Cameroon and their combinations. Cienc. Agrotec., Lavras. 24 (1):2134-224.
FAO (Food and Agriculture Organization of The United Nations). 2005. Animal Feed Resources Information System. http://www.fao.org/ag/aga/agap/frg/afris/ default.htm
Ferrari, R.A., F. Colussi and R.A. Ayub. 2004. Characterization of by-products of passion fruit industrialization - utilization of seeds. Rev. Bras. Frutic. 26(1):1-4.
Hanley, F.A. Wildish, W. Johnson, E. Sonega, J. Harvey, J. Everington and S.L. Woodgate. 1998. The use of enzyme processing to improve the nutritive value of keratin-containing raw materials: case studies from around the world. In: Biotechnology in the Feed Industry, Proceedings of Alltech’s 14th Annual Symposium (K.A. Jacques and T.P. Lyons, eds). Nottingham University Press, UK, pp 575-586.
Hyline. 2000. Hyline variety Brown, commercial management guide, rev 01/00. Hyline International, West Des Moines, Iowa, USA.
Luna, O. 2005. Un ejemplo del ejercicio: el maracuya ante el TLC. Ministerio de Agricultura y Ganaderia, Ecuador. http://www.sica.gov.ec/agronegocios/ acceso_a_mercados/tlc_usa/tlc_maracuya.pdf Morton, J. 1987. Fruits of warm climates. Creative Resource Systems Inc, NC.
National Research Council. 1994. Nutrient Requirements of Poultry. 9 th rev. ed. Natl. Acad. Sci., Washington D.C.
Prasad, P.C. 1980. Pectin and oil from passion fruit waste. Fiji Agric. J. 42:45-48.
Prasad, P. C. and R. Chandra. 1980. Review of passionfruit research in Fiji. Fiji Agric. J. 42(2):19- 22.
Pruthi, S. 1963. Physiology, chemistry and technology of passion fruit. In: Advances in food research. (C.O. Chichester, E. Mrak and G. Stewart, eds.) New York Academic Press. Vol. 12, p. 203-282.
Otagaki, K.K. and H. Matsumoto. 1958. Nutritive values and utility of passion fruit by-products. Agricultural and Food Chemistry 6(1):54-57.
Taylor, L. 2005. The healing power of rainforest herbs. Square One Publishers, NY. http://www.rain-tree.com/ maracuja.htm
Teixeira, C. 1994. Maracuja: cultura, matera prima, processamento e aspectos economicos. 2da ed. Serie Frutas Tropicales 9. Instituto de Tecnologia de Alimentos. Campinhas, Brasil.
1 Centro de Investigaciones, Procesadora Nacional de Alimentos C.A., Quito, Ecuador
2 Programa de Medicina Veterinaria, Universidad San Francisco de Quito, Ecuador