The nutrient profile of peas indicates that they are suitable for use in high-density diets such as those used in modern poultry production. However, incomplete poultry nutrient data for feed peas, combined with limited industry experience using this feedstuff for poultry, has prevented peas from replacing more traditional feed ingredients such as soybean meal. The following review is intended to outline what is currently known about the properties of western Canadian feed peas and to provide insight into ways in which peas can be most successfully incorporated into poultry diets.

Peas contain moderate levels of crude protein (CP) and metabolizable energy (ME), making them suitable for use in the diets of many types of poultry. Peas are deficient in methionine but contain relatively high levels of lysine, which compliments the amino acid profile of locally produced cereals and canola meal. Studies of Western Canadian pea proximate composition are summarized in Tables 1 and 2. This data agrees with that from analyses performed previously (Sosulski and Holt 1980).

Moisture analysis of 1100 pea samples by Norwest Labs for the preceding four years indicated that the average moisture content of feed peas is 12.2% with a standard deviation (SD) of 2.75 (Ken Mrazek, 2004 Norwest Labs).

Western Canadian studies have measured average feed pea crude protein contents between 23 and 24% (dry matter (DM)-basis, Table 1). Analysis of 1100 samples compounded over four years from throughout western Canada by Norwest Labs confirmed a feed pea crude protein level of 23.9% (DM-basis, SD= 3.17). Samples of a single pea variety ranged from 14.5 - 28.5% CP (dry, dehulled basis), indicating that environment has a large impact on pea protein variability (Reichert and MacKenzie 1982).

Several observations emerged from a study of yellow- (n=8), green- (n=2) and brown- (n=2) seeded pea cultivars grown in Manitoba (Igbasan et al. 1997). The peas had a broad range of protein levels (20.8 - 26.4% CP DM-basis), but this was not correlated to seed size or colour. Seed coat colour was also not related to amino acid concentrations. The concentration of most essential amino acids, including lysine, methionine and threonine, did not increase proportionately with CP level (Igbasan et al. 1997); supporting the concept that attempts to genetically select for increased protein content in peas would likely result in an erosion of protein quality (Gatel 1994).

Pea CP levels are determined on the basis of nitrogen content. Non-protein nitrogen was present in the seeds, which reduced the true nitrogen-to-protein conversion factor (5.25) from that normally used (6.25). However, it is recommended that all CP calculations continue to use the 6.25 factor, lest peas appear at a disadvantage to the similarly inaccurate protein calculations of other feed ingredients (Sosulski and Holt 1980; Mosse 1990).

The constituent proteins of feed peas (ie: globulins, albumin, insolubles) and their sub-fractions are highly variable in terms of their digestiblities. Protein structure including hydrophobicity, glycosylation, beta-sheets, compact tertiary structure and disulphide bonds can have a negative impact on protein hydrolysis (Crevieu-Gabriel 1999). Although vicilin is one such protein in peas, unheated vicilin was far more digestible than equivalent proteins in dry beans or soybean (Deshpande and Damodaran 1989; Nielsen et al. 1988). However, the pea protein fractions that were less susceptible to digestion at the terminal ileum of 3-week-old broilers represented only a small proportion of total dietary protein (Crevieu et al. 1997).

Apparent protein digestibility (APD) differed significantly (P≤0.05) for a yellow-, a green- and a brown-seeded cultivar (76.7, 71.5 and 60.1%, resp.; Igbasan and Guenter 1996) when measured using 10-day-old male broilers. The APD of Trapper peas was 81% in 17-day-old broilers (Brenes et al. 1993). These APD values are reasonably close to the apparent amino acid digestibility values shown for 4-wk-old broilers (Table 2). Apparent protein digestibility for colostomized laying hens was 78.5% (Gruhn and Zander 1990). Chicken and porcine pepsinogen activities differ significantly, depending upon pH, incubation time and type of protein analysed; therefore, in vitro assessment of pea protein digestibility should approximate poultry gastric conditions (Crevieu-Gabriel et al. 1999).

High fibre levels in feedstuffs are not desirable because this component is not well digested or utilized by poultry. Fibre levels for western Canadian feed peas measured by Norwest Labs in 1100 samples spanning a four-year period were: crude fibre - 58 g kg^-1 DM; acid detergent fibre - 103 g kg^-1 DM; and neutral detergent fibre - 180 g kg^-1 DM. A European study (37 samples; Grosjean et al. 1999) determined that crude fibre was the only measured fibre component that had a significantly negative correlation (P<0.01) with AME in mash diets. Fibre components measured in pelleted diets did not have a significantly negative correlation with AME. Conversely, a high crude fibre level (99 g kg^-1) was implicated in the reduced AMEn content of peas for birds in-lay (2525 kcal kg^-1, as fed; Askbrant and Hakansson 1984).

Starch is the main storage polysaccharide in feed peas. It exists in two primary forms: amylose and the more readily digestible amylopectin (Grosjean et al. 1999). The amylose content of smooth-seeded peas such as those grown in Canada is roughly 40% of total starch (Lloyd et al. 1996). Wrinkled peas, harvested green for the human food trade, have a mutation reducing the content of amylopectin and total starch, making them less suitable as an energy source for monogastric feeds (Grosjean et al. 1999).

Starch values from twelve Canadian pea cultivars averaged 415 g kg^-1 ± 16.8 (mean ± SD; DM basis; Igbasan et al. 1997). This is lower than the average starch values (DM-basis) of 459 g kg^-1 (Conan and Carrè 1989), 453 g kg^-1 (Cerioli et al. 1998) and 501 g kg^-1 (Grosjean et al. 1999) measured in extensive European surveys.

Starch digestibility in peas is affected by factors that limit its accessibility. The structure of pea starch granules may be partially responsible, as indicated by the significant differences between the digestibilities of purified starch granules from high-amylose peas, regular peas and maize (75.2, 94.4 and 98.8%; P<0.05, respectively). In addition, the digestibility of starch granules was lowest in the largest particles of ground peas, suggesting that the matrix surrounding the starch granules restricted their accessibility. Total tract pea starch digestibility in 3-wk-old broilers was 95.7% in pea particles below 100 µm in diameter, and 84.4% for pea particles greater than 100 µm in diameter (Carrè et al. 1998).

Human nutritionists have examined the relatively slow rate of starch breakdown and glucose absorption from peas because it moderated the peak in post-meal insulin production. This reduced glycemic effect was emphasized with the high amylose/amylopectin ratio pea genotypes (Skrabaja et al. 1999). The corollary was also examined in poultry nutrition. A broiler trial compared diets in which starch was degraded rapidly and absorbed from the jejunum (ie: corn, tapioca, rice) versus diets in which a proportion of the starch was also degraded in the ileum (ie: peas). Recent research indicated that broilers consuming diets which contained a minimum quantity of slowly degraded starch, specifically that available from peas (Weurding et al. 2001), had improved growth rates and feed conversion relative to isocaloric control diets containing rapidly-degraded starch (Weurding et al. 2003). These birds had an improved response to elevated lysine levels (Weurding et al. 2003b). It was suggested that this was due to the relatively continuous glucose supply: 1) sparing the catabolism of gluconeogenic amino acids, 2) providing a direct source of energy for posterior sections of the gut, 3) stimulating prolonged insulin production and therefore protein deposition, and 4) reducing the energetically inefficient conversions between glucose and its storage forms when glucose absorption spikes following a meal of rapidly-degraded starch. An in vitro method to predict starch digestion rate has been developed (Weurding et al. 2001b). This is an exciting area of research and may advance the goal of formulating diets to meet animal requirements at the metabolic level.

A comprehensive study of 12 pea cultivars was performed in Alberta (Jaikaran et al. 1995), the mineral values from which are shown in Table 1. A similar range in values was seen in four samples of a single cultivar of peas (cultivar Trapper) ranging from 14.5 - 28.5% CP (dehulled, DM-basis; Reichert and MacKenzie 1982 in Table 1), indicating that both cultivar and environment affect the mineral content of feed peas. Mineral analyses performed by Norwest Labs on 1100 samples over 4 years concurred with values from the previous authors (Table 1).

The energy value of peas for poultry changes drastically in response to processing, therefore the energy value used for feed formulation should be matched to the processing conditions intended for the diet. The average AME values for eleven feed peas ground (3000 rpm. 3-mm screen) and included in either mash or cold-pelleted diets of Isa Brown cockerels were 2851 and 3150 kcal kg^-1 (DM-basis, SD= 123 and 72, respectively; Grosjean et al. 1999). In addition, Carrè et al. (1991) determined that steam-pelleting (4x30mm die, flow rate 2.85 kg/min, inlet 75°C / outlet 76°C) pea-based diets had a strong positive effect on apparent metabolizable energy (AMEn; P<0.001) in both adult (3069 vs. 2813 kcal kg^-1 DM) and 19-day-old (3016 vs. 2762 kcal kg^-1) broilers, when compared to mash (2-mm hammer mill sieve) diets. Brenes et al. (1993) measured significantly lower pea AMEn values in mash diets with 2-week-old broilers (2189 kcal kg^-1, assume as-fed basis).

The true metabolizable energy (TMEn) values of yellow-, green- and brown-seeded peas (n=12) administered as the sole ingredient to adult leghorn cockerels averaged 2916 kcal kg^-1 (Igbasan et al. 1997) which is in agreement with other values published for peas (Farhoomand and Pirmohammad 2002).

The AME of feed peas (500 g kg^-1) for non-laying adult hens was 3060 kcal kg^-1 DM (Perez-Maldonado et al. 1999). The AMEn from a yellow-, a green- and a brown-seeded pea cultivar (2420, 2460 and 1980 kcal kg^-1, respectivley) was determined at 50% dietary inclusion using 10-day-old male broiler chicks (Igbasan and Guenter 1996). The TMEn values of the same pea samples administered as the sole ingredient to adult leghorn cockerels were 2796, 3011 and 2629 kcal kg^-1, respectively (Igbasan et al. 1997). On the basis of this data it appears that high dietary levels of peas are better utilized by older birds; however details regarding sample preparation (ie: ground vs. whole) for older birds, and regarding the format in which energy levels were reported (ie: DM basis vs. as-fed) were not reported. Conversely, other researchers reported that bird age did not appear to affect the energy value of peas. Carrè et al. (1991) reported that a single sample of peas had similar mean AMEn and standard deviation (SD) values for adult and 19-day-old birds in either mash (2762 vs. 2813 kcal kg^-1 DM) or steam-pelleted (3069 vs. 3016 kcal kg^-1 DM) diets.

The apparent metabolizable energy (AMEn) of pea flour "chips" from a yellow and a green-seeded pea cultivar (2748 and 2696 kcal kg^-1, respectively) was determined at 450 kcal kg^-1 dietary inclusion using 10-day-old male broiler chicks (Igbasan and Guenter 1996).

The amino acid content determined for feed peas by various researchers is shown in Table 2. Feed peas are an excellent source of lysine, but contain lower levels of the sulphur amino acids, methionine and cystine. The digestibilities of feed pea amino acids, as measured by several researchers, are also shown in Table 2. Igbasan et al. (1997) measured true amino acid availabilities for yellow- and green-seeded pea samples (n=10), and values ranged from 78.3% for total sulphur amino acids to 90.5% for glutamic acid. Brown-seeded cultivars had average lysine, threonine and total sulphur amino acid availability values of 71.9, 72.4 and 64.0%, respectively, indicating elevated tannin levels may have impaired digestibility. Amino acid content (Table 2) was determined for 12 cultivars of peas and amino acid prediction equations were developed (Table 3) (Jaikaran et al. 1995). Values from a European study of true and apparent ileal amino acid digestibilities (Table 2) are similar to the others listed, with the exception of valine and histidine. However, it was indicated that these ileal analyses were difficult to perform without overestimating true digestibility (Perez et al. 1993).

Mossé (1990 in Gatel 1990) indicated that a linear relationship existed between CP and amino acid levels, with methionine, cystine and tryptophan accumulating slowly in comparison to nitrogen-rich residues like asparagine and glutamine (Huet et al. 1987). Regression equations used for the prediction of some limiting essential amino acid values are shown below (Table 3). Alternatively, near-infrared reflectance spectroscopy (NIRS) calibrations developed by Degussa (Fontaine et al. 2001) were used to accurately measure the levels of most amino acids in peas

^a (AA, g kg-1 DM) = a * (CP, g kg-1 DM) + b
^b (AA, % of DM) = a * (CP, % of DM) + b

2.a Isolated pea components

Air classification and other methods of separating peas into their component fractions produces by-products that are suitable for use in the poultry feed industry. Coarse "pea chips" were examined in diets for male broilers up to 3 weeks of age. The AMEn was 2748 and 2696 kcal kg^-1 (DM basis) for yellow and green pea chips, respectively (Igbasan and Guenter 1996), which agrees with AMEn values of whole peas (Igbasan and Guenter 1996). Coarse milling did not affect the amino acid distribution within the different size classes of pea particles (Leterme et al. 1990). A technique designed to isolate the central part of pea cotyledons resulted in a product with elevated starch and reduced protein and fibre levels (Otto et al. 1997; Kosson et al. 1994).

Pea protein concentrate, a high-protein fraction derived from air classification of ground pea particles, has been used as a nutritive pellet binder (≥ 25g PPC kg^-1 diet). However, diets using pea protein concentrate as the primary protein source should never be fed unpelleted because, at high dietary levels, the particles will adhere inside the beak, causing difficulty eating and beak necrosis (Brown, 1991).

Peas are used extensively for poultry diets in Europe (Gatel 1994), and studies from these countries offer information on cultivars other than the smooth-seeded, white-flowered, spring varieties commonly grown in Canada. Wrinkled peas contained higher CP levels than smooth-seeded peas (264 vs. 239 g kg^-1 DM, resp. (Gueguen and Barbot 1988) but had lower yields so that protein production per hectare was similar (Monti 1983). Compared to spring seeded varieties, winter-sown peas had higher levels of trypsin inhibitor (Conan and Carrè 1989) but generally lower AMEn and APD (Carrè et al. 1991).

Producers who grow and utilize their own feedstuffs should be aware that fertilization practices can affect the nutrient profile of peas. Sulphur deficiency reduced the production of legumin, a storage protein that contains methionine and cystine (Evans et al. 1985), but sulphur fertilization increased the total sulphur amino acid content of peas by 10% (Eppendorfer and Eggum 1995). Nitrogen and phosphorus fertilization increased the seed content of these elements, and inconsistently improved seed yield (Browning and George 1981). These plants were grown in sterilized soil, and consequently a similar nitrogen-related improvement may not be seen in peas inoculated with Rhizobium bacteria. Normal fertilization practices are likely to support the growth of peas with maximized nutrient values.

Many studies have been performed to determine the optimum inclusion rate of feed peas in poultry diets. Unfortunately, some feeding trials identified only very specific conditions under which the feed ingredients could be used, and if these conditions were not outlined clearly in the research paper (ie: mash vs. pelleted diets, enzyme vs. no enzyme in cereal-based diets, energy values assigned to ingredients for feed formulation), even this limited information could not be put to practical use by industry. It could be concluded that the elementary nutrient availability information required to properly utilize peas had not yet been gathered at the time of some of this review, and that the use of peas in poultry diets could yet benefit from basic nutrient-defining research.

The level of nutrients provided by peas is well suited to the requirements of laying hens and studies quoted elsewhere in this review suggest that the digestive physiology of the adult bird will allow optimum utilization of this feedstuff. Recent North American trials indicated moderate quantities of peas were suitable for use in the diets of young birds, but that hens could achieve excellent performance at higher dietary levels (Table 4). A well-defined study in which diets were formulated on predetermined nutrient values indicated laying hen performance did not differ from a corn-soy control diet at pea inclusion levels of up to 400 g kg^-1 (Igbasan and Guenter 1997). A European study indicated that performance was not affected by 350 g kg^-1 wrinkled (assumed) peas, but that daily egg production declined significantly from week 3 (98%) to week 8 (90%) at a 500 g kg^-1 inclusion rate. As seen by other authors (Ivusic et al. 1994, Ivusic 1989), average egg weight increased with dietary pea concentration (Castanon and Perez-Lanzac 1990). These diets did not contain added methionine, and this may have affected the results. Hen-day egg production averaged 84.1 eggs/100 hens for weeks 25 to 65 with least-cost formulated diets containing 250 g kg^-1 peas (Perez-Maldonada et al. 1999). Another trial indicated that 375g kg^-1 peas sustained production equivalent to a fishmeal control, but egg size was reduced; however, the oat-based diets contained only 2400 kcal kg^-1, making the data inapplicable to modern laying strains (Davidson 1980). Further work is needed to corroborate the studies shown below.

A high crude fibre level in peas (9.9% DM) was implicated in reduced pea ME levels for birds in-lay (2806 kcal kg^-1 DM; Askbrant and Hakansson 1984). However, antinutritive effects of viscous polysaccharides in the barley, oat (320g kg^-1 each) and wheat (100g kg^-1) components of the diet may have been attributed to low pea ME. Another study found pea AMEn values were 4.6% higher in test diets containing maize versus wheat, but attributed this effect to the higher dietary lipid content of the maize diet (Carrè et al. 1987). It is possible that viscous polysaccharides in the small-seeded cereals exacerbated the alleged inhibitory effect of pea cotyledon structure on digestibility.

A few recent studies have documented the performance of broilers consuming pea-based diets (Table 5). While results are somewhat inconsistent, it is obvious that broilers respond well to moderate levels of peas in both starter and grower rations. Recent research indicated that broilers consuming pea-supplemented (341 g kg-1 diet) corn diets had improved body weight (1823 vs. 1729 g @ 38d.) and feed conversion (1.73 vs. 1.77) relative to isocaloric tapioca-soy control diets. This improvement in performance was attributed to the beneficial effects of slowly degraded starch found in peas (Weurding et al. 2003 see section 1.f.). Brenes et al. (1993,) indicated that the addition of peas did not affect the performance of young broilers consuming corn-SBM diets. Other workers (Igbasan and Guenter 1996a, b) observed reductions in gain and feed conversion (FC) with dietary pea levels of 400g kg^-1 or greater, but only FC was affected at 230g kg^-1 peas (Guenter et al. 1992 in Castell et al. 1996). Further investigation showed equivalent performance for mash diets containing 100, 200 and 400 g kg^-1 peas if CP and EAA were supplemented to 115% of NRC requirements (Igbasan and Guenter 1996). Whether birds are better able to utilize ingredients with reduced digestibility at low levels in the diet, or whether the erosion in performance is so small as to not be detected statistically, is unclear. A pea/ whole canola blend caused slight reductions in gain and FC, but the ingredient responsible for this effect was not determined (Fasina and Campbell 1997).

High dietary levels of peas did not support broiler performance equivalent to controls (Table 5) but this may not accurately reflect the nutrient value of peas. Young birds digested only 79% of pea amino acids on average, and values for lysine and methionine were only 80 and 60%, respectively (Table 2).

Therefore, diet formulation on the basis of available amino acids may have enhanced the performance of birds consuming the higher inclusion levels of peas. A European study formulated diets on an available amino acid basis (1.25, 0.82, 0.7 and 0.2% available lysine, methionine, threonine and tryptophan). Broilers consuming the pea-based diet (300 g kg^-1) had higher weight gains (43.4 vs. 41.7 g day^-1) but poorer FC (1.60 vs. 1.58 g gain g^-1 feed) than the casein-based control (Huisman et al. 1990).

European and Australian studies supported the use of peas at 300g kg^-1 diet in broiler diets (Wurzner et al. 1988; Farrell et al. 1999). Brenes et al. (1989) even indicated that levels of 600 and 800g kg^-1 peas improved performance over corn - soy isolate diets; however, the latter trial added oil only to the pea diets, and this may have adversely affected consumption of the control if it was unpelleted.

The 'Miranda' variety of pea replaced SBM as the sole protein source in starter, developer and breeder rations fed to ISA Vedette dwarf broiler breeder hens. Performance was measured to 45 weeks of age and indicated weight gain, hen-day egg production, egg weight, fertility and hatchability of pea-based diets was equivalent to the corn-SBM control (Kill and Savage 1992).

The semen quality and reproductive performance of broiler breeder cockerels fed corn-SBM or corn-yellow pea diets were similar for both caged and floor pen environments. High (16% CP, 2850 kcal kg^-1) and low (7% CP, 2400 kcal kg^-1) density pea diets supported fertility and hatchability equivalent to diets containing SBM (Rakphongphairoj and Savage 1988; Bootwalla et al. 1988).

The single reference available indicated that peas were an excellent protein source for all stages of turkey rations. Nicholas Large White turkeys were fed 250 (0-4 wk), 300 (4-8wk), 350 (8-12 wk), 500 (12-16 wk) and 550 g kg-1 (16-20 wk: toms only) peas in methionine-supplemented diets with (CSP+F) and without (CSP-F) added fat. In two separate trials, CSP+F supported growth and FC equal or superior to the corn-SBM control. The CSP-F rations contained 50 (starter) to approximately 140 kcal kg^-1 (finisher) less energy, but only FC was inferior to CSP+F and corn-SBM rations (Savage et al. 1986).

Geese are noted for their ability to utilize dietary fibre. An experiment in which White Italian geese (3-8 wk of age) were fed rations containing 200 g kg^-1 peas indicated digestibility of neutral detergent fibre, acid detergent fibre, cellulose and hemicellulose was 41.7, 17.6, 20.9 and 53.8%, respectively (Jamroz et al. 1992). The growth performance (63.4 g day^-1) of geese consuming pea diets was among the highest measured in the study (Wiliczkiewicz et al. 1992).

The AMEn of whole peas for mature pigeons was 3348 ± 98 SD (moisture content not stated). The total-tract digestibilities of crude protein and ether extract were 85.70 ± 1.41 SD and 82.59 ± 5.97 SD, respectively (Hullar et al. 1999).

Many authors have reviewed the nutritional value of feed peas, and the various chemical and physical treatments used to improve these values (Gatel 1994; Castell et al. 1996; van Barneveld et al. 2000). The following sections review the state of this knowledge as it pertains specifically to poultry.

Analysis of digesta from the proximal ileum of 3-wk-old broilers indicated that semi-purified starch from peas (90.3%) was less digestible than that from wheat (97.3%). However, pea starch digestion was comparable to wheat (94.4 vs. 97.6%, resp.) at the distal ileum. Other pea cell components had been removed from surrounding the starch during purification; therefore, the digestibilities are thought to reflect varying susceptibilities to enzyme hydrolysis in pea and wheat starches (Yutse et al. 1991).

Significant improvements in organic matter, CP, crude fat and starch ileal digestibilities in broiler chicks were observed with fine grinding (0.3 mm sieve) versus coarse rolling of feed peas (Daveby et al. 1998); therefore, nutrient accessibility appears inversely related to particle size. Other workers had noted a correlation (r=0.97, P<0.05) between starch and protein digestibility (Conan and Carrè 1989), and it was observed that the majority of undigested starch was located in the large particles (>0.5mm) in the excreta (3 wk. old). It was hypothesized that this was due to the physical inaccessibility of protein and starch in the large particles, a situation improved by steam pelleting (Carrè et al. 1991).

Although fine grinding of peas would be cost-prohibitive for the commercial feed industry; steam pelleting appears to be a practical means of improving nutrient accessibility. Diets were based on ground peas (2 mm sieve), or on ground peas that were subsequently steam pelleted (4x30 mm die) and reground (2.5 mm sieve). Pelleting produced a strong positive effect (P< 0.001) on AMEn and starch digestibility across age groups (adult and 3-week-old broilers) and a significant improvement in protein digestibility (P<0.05) within the young birds (Table 6). Digestibility was not affected by regrinding of the steam-pelleted peas, because the mean diameter of particles ground once or twice was similar (0.434 vs. 0.507 mm, respectively; Carrè et al. 1991).

Legumes are known to contain antinutritional factors (ANFs) which interfere with digestive processes, thereby reducing the nutritional value of pulse crops for monogastric animals. Potential pea ANFs include amylase, trypsin/chymotrypsin inhibitors, tannins (proanthocyanidins), phytic acid, saponins (hypocholestolemic factors), hemagglutinins (lectins) and oligosaccharides. Fortunately, the white-flowered, spring-seeded peas grown in Canada have relatively few ANF issues when compared to other pea varieties (Valdebouze et al. 1980) and pulse crops (Bond and Smith 1989; Lalles and Jansman 1998). An extensive study of feed legume ANFs in Europe supports this observation (Cerioli et al. 1998). Poultry are therefore able to enjoy a relatively large proportion of peas in their diets.

Trypsin and chymotrypsin inhibitors have been presumed responsible for reduced protein digestibility in pea diets (Pisulewski et al. 1983) because each trypsin inhibitor molecule is capable of forming a stable complex with both a trypsin and a chymotrypsin molecule (Le de la Sierra 1999). The production of cystine-rich trypsin and chymotrypsin then increases (Lhoste et al. 1998), and this may place further stress on birds consuming pea-based diets deficient in sulphur amino acids (Huisman and Tolman 1992). However, studies indicated that (chymo)trypsin inhibitors in peas had little practical effect on poultry performance. Even when data was pooled between spring peas and high-TIA winter peas, improved protein digestibility was non-significantly correlated to a reduction in TIA (r=0.71, Conan and Carrè 1989; r=0.79, Carrè and Conan 1989). Pancreatic enlargement (0.21 vs. 0.18% live body weight) was noted in broiler chicks (0-28d of age) consuming pea-supplemented diets (200 mg kg^-1), in addition to increased weight gain, intake and FC (P<0.05) (Huisman et al. 1990). However, reviewers Lalles and Jansman (1998) suggested undigested fragments from many protein sources, including feed peas, may result in pancreatic stimulation.

Conversely, feeding trials with isogenic pea lines differing only in trypsin inhibitor content have shown reduced rat body weight and protein utilization with levels as low as 2.7 - 3.9 trypsin inhibitor units gram^-1 DM peas (Hedemann et al. 1999). When other workers (Wiseman et al. 2003; Al-Marzooqi and Wiseman 2002) examined the same pea lines in 3-week-old poultry diets, the apparent ileal digestible content of methionine, lysine, threonine, leucine and histidine was significantly reduced in the high-trypsin pea diets (8.07 ± 0.42 vs 1.62 ± 0.17 TIU mg^-1 DM). Although significant (P<0.001), the differences between apparent ileal digestible lysine values were relatively small (6.64 versus 6.86 g kg^-1 in diets containing 500 g kg^-1 low- and high-TI peas), and this may explain why trypsin-inhibitor effects have not yet been measured in poultry performance trials.

A survey of Canadian feed pea co-op tests indicated trypsin inhibitor levels (2.3 to 10.3 TUI (range); n=35); mean=4.5; CV=20.5%; Slinkard and Tyler 1993 below the range shown to have no effect on nutrient digestibility for poultry (Conan and Carrè 1989; Carrè and Conan 1989). Similar TIA values were seen for spring peas grown at different locations (mean=4.4; n=21; Slinkard and Tyler 1993. These values were in agreement with Wang et al. (1998), whose measurements of western Canadian peas indicated that cultivar was more important than environment in explaining variance in TIA levels (55.3 vs. 17.7%, resp.).

Research attempting to predict trypsin inhibitor activity based on physical characteristics has met with mixed results. Analysis of European peas indicated that the trypsin inhibitor activity (TIA) of winter cultivars was twice that of those sown in spring, and that smooth peas had higher TIA than wrinkled varieties (Valdebouze et al. 1980). Other studies found that the level of TIA was not linked to seed character (round vs. wrinkled) or CP level (Griffiths 1984) but was significantly affected by cultivar and environment (Bacon et al. 1991). European pea cultivars had considerable variation in trypsin and chymotrypsin inhibitor levels, with average values similar to and double those of faba beans, respectively (Griffiths 1984).

Tannins, or proanthocyanidins, are polyphenolic compounds that inhibit the activity of digestive enzymes including trypsin, α-amylase and lipase (Longstaff and McNab 1991). They are found in the hull (Griffiths 1981) of colored-flowered peas, but are difficult to extract and quantify using current methodologies (Marquardt and Blackburn 1991). Studies involving dehulling confirmed that tannins did not affect the nutritional value of a white-flowered, spring-seeded pea cultivar (Brenes et al. 1993). Recent studies indicated that the yellow and green-seeded peas commonly grown in Canada are devoid of tannins, but that the brown-seeded cultivars may contain appreciable amounts of this ANF (<0.1, <0.1, and 11.5 to 41.0g condensed tannins kg^-1, resp; Igbasan et al. 1997; Brenes et al. 1993). A brown-seeded and a green-seeded cultivar had reduced TMEn, but tannins were not thought to be the only agent reducing ME since they were absent from the green-seeded sample (Igbasan et al. 1997). Antinutritive effects should not be seen with any practical inclusion level of yellow or green peas because faba bean-based diets (5g condensed tannins kg^-1) supported excellent growth rates in young broilers (5 - 26d of age) (Jansman et al. 1993).

Oligosaccharides consist of a sucrose moiety β-1,4-linked to one or more galactose subunits. Dehulled, Canadian-grown peas contained 44.2 to 56.1g kg^-1 oligosaccharides (DM basis; Reichert and MacKenzie, 1982). Monogastrics do not produce the α-galactosidase enzyme required to digest oligosaccharides, and these carbohydrates may cause digestive disturbances when microbially fermented in the hindgut of poultry (Saini 1989). However, studies found high digestibilities of oligosaccharides in cockerels (>90%) and chicks (>70%), and indicated the birds were capable of absorbing the organic acids that may have resulted (Carrè et al. 1995). Addition of pea oligosaccharide extract (56 and 28g kg^-1) to the diets of young broilers (7-28d of age) did not affect performance or digestibility of dietary nutrients (Trevino et al. 1990). Studies with pea protein concentrate indicated significant levels of endogenous α-galactosidase (Brown, 1991; Weins, 1992), which may explain the apparently negligible antinutritive effects of oligosaccharides in pea-based rations, and the lack of response by broilers to supplementation of pea-based diets with a commercial α-galactosidase.

Lectins, saponins and phytic acid are ANFs that occur in peas, but either have mild effects or receive less attention. The cotyledons of peas contain lectins (hemagglutinins), polysaccharide-binding agents that cause hyperregenerative villus atrophy in the small intestine. A general survey of legumes indicated that pea lectins had low reactivity and were non-toxic (Grant et al. 1983). Laying hens had enhanced performance when fed heated versus raw peas, which was attributed to the inactivation of lectins. However, a high-trypsin-inhibitor variety of peas (Maro) was used in these diets, and the improvement in performance may have been due to the inactivation of trypsin inhibitors (Davidson 1980).

Saponins are ANFs composed of sugar and steroid or triterpenoid moieties. Studies indicated that pea saponins were less hemolytic than those from field or soya beans, but intermediate in toxicity to guppy fish (Khalil and El-Adawy 1994). It is not known whether pea saponins affect poultry.

Peas have been reported to contain 22 g kg^-1 phytic acid, a cyclohexane compound with six phosphate groups (Blatny et al. 1995) commonly found in cereal grains. Antinutritional effects associated with this compound include mineral-complexing and inactivation of digestive enzymes, although it is not known what effects pea phytate has on poultry. Addition of commercial phytase to poultry diets will reduce the antinutritional effects of cereal phytates.

Results from various methods of fibre determination are shown in Table 1. Initial reports indicated a high level of pea hull fibre digestion by cockerels (Longstaff and McNab 1987) but further study showed pea hull digestion was very low (6.1%), and did not greatly exceed the available carbohydrate content (30.95 g kg^-1) (Longstaff and M^cNab 1989). Indigestibility of the hulls may reduce the ME (Jørgensen et al. 1996) and digestible nutrient content of whole peas; however, if available as a low-cost byproduct of the pea processing industry, the hulls of low-tannin peas have potential application as a diluent in broiler-breeder rations.

Peas are a high-quality feedstuff for use in poultry diets, capable of supporting excellent performance and efficiency in all types of birds. As with any feed ingredient, knowledge regarding the factors that affect the nutrient value of peas will allow its optimum utilization. The poultry studies that have examined the effects of various processing techniques on pea nutritive value are outlined in the following sections.

Several forms of heat treatment have been documented to effectively eliminate many pea ANFs. Cooking at 100°C destroyed antitrypsin activity (Savage and Deo 1991), and autoclaving at 130°C and 170 kpa for 3 minutes significantly reduced TIA levels (Conan and Carrè 1989) in peas. Tannins, lectins and trypsin inhibitors in round-seeded spring peas were virtually eliminated at minimum extrusion conditions of 105°C and 20.3% moisture (Poel et al. 1992). However, dry heat (≤100°C, 24 h) was ineffective as a means of reducing (chymo)trypsin inhibitor activity in peas (Griffiths 1984), or as a means of improving digestibility of peas in rat diets (Canibe et al. 1997).

The practical effects of heat treating peas for use in poultry diets were less encouraging. The TMEn of autoclaved peas (121°C, 15 psi, 30 min) was lower than that of heated peas (121°C, 30 min), and cooking (simmered, 20 min) caused a non-significant reduction in starch digestibility due to retrograde starch formation (Longstaff and McNab 1987). In addition, the vicilin storage protein of peas became more resistant to in vitro digestion upon heating, possibly due to changes in secondary and quaternary structure (Deshpande and Damodaran 1989). Autoclaving did not improve the protein digestibility or AMEn value of low-tannin Trapper peas, although a significant improvement was seen when the treatment was duplicated for high-tannin Maple peas (Brenes et al. 1993). Micronization (110-115°C for 55 s) improved AMEn, APD and starch digestibility relative to untreated peas fed to broilers (3-12 d of age; Igbasan and Guenter 1996); however, the peas were tempered prior to micronization and then immediately roll flaked, and this additional processing may have affected nutrient accessibility.

The process of grinding has a great effect on the nutrient value of peas. Relative to whole peas, fine grinding (<1mm diam.) significantly increased (P<0.001) the starch digestibility (75.6 vs. 88.1%) and TMEn (9.91 vs. 11.38 kJ g^-1) in adult white leghorn cockerels; however, further improvements were not seen when peas were heated, autoclaved or dehulled prior to grinding (Longstaff and M^cNab 1987). Reduced particle size in ground faba beans (0.5 vs 0.16mm mean diameter) enhanced starch digestibility (P<0.01) and AMEn in faba beans (P<0.05) consumed by 3-week-old broilers (Lacassagne et al. 1991). An improvement in ileal protein digestibility was also seen when chicks consumed fine versus coarse-ground peas (P<0.05; Crevieu et al. 1997). Starch was more digestible in pea seeds moistened before grinding (85.7% at 82.5% DM) than in seeds ground as received (65.9% at 92.5% DM). The three-fold increase in energy required to grind moistened seeds was thought to alter cell wall or starch granule structure, thereby improving starch digestibility (Carrè et al. 1998). Under similar grinding conditions (hammermill at 75% capacity; 3.2mm screen), peas ground 8.2% slower than corn (Behnke et al. 2002). It is fortunate that pelleting further enhances the nutritive value of ground peas because the time and expense associated with fine grinding of feedstuffs may be cost-prohibitive for the feed industry.

Steam pelleting has been shown to improve the utilization of peas by poultry, an effect some authors attribute to increased nutrient accessibility rather than ANF destruction. Although the AMEn values of cereals were not affected by pelleting, the AMEn value of peas for Rhode Island Red cockerels was enhanced by steam-pelleting pea-corn (3112 vs. 3056 kcal kg^-1) and pea-wheat (3028 vs. 2894 kcal kg^-1) rations. This effect was attributed to the improved digestibilities of pea protein (3.5%) and starch (5.4%). Pelleted pea-based diets did not show significant improvements in starch digestibility when they were repelleted (P>0.01; Carrè et al. 1987).

Pelleting improved starch digestibility for both young and mature birds and reduced AMEn variability so that the average value for each group was 3000 kcal kg^-1 DM. Steam pelleting improved the digestibility of spring pea protein without reducing the level of TIA (3.47 vs. 2.98mg TIA/g peas) (Carrè et al. 1991).

The gelatinization of starch that occurs during pelleting and expansion may improve nutrient accessibility by disrupting intra- and intercellular structure within peas. Steam pelleting (80°C) was comparable to expansion in terms of increasing starch gelatinization relative to that of untreated pea/whole canola-based diets (38.6, 40.9 and 18.5%, resp.). Pelleting and expansion produced similar improvements in weight gain and FC (P<0.05) and increased AMEn relative to that of the untreated pea diets (3087, 3028 and 2877 kcal kg^-1, respectively). However, the rations contained equal amounts of peas and whole canola seed (364.5 g kg^-1 each), and the treatments may have enhanced the nutritional value of both ingredients (Fasina et al. 1997). Even at lower steam pelleting temperatures (65°C, 1/8" grind), significant starch gelatinization occurred relative to the unpelleted control (30.1 vs 15.9%; Fasina and Campbell 1997).

Extrusion significantly reduced the measured levels of antinutritional factors and increased in vitro starch and protein digestibility. However, significant reduction in the availability of cystine and methionine suggested that the use of extruded peas required careful attention to amino acid supplementation (Alonso et al. 2000).

Addition of up to 300 g kg^-1 peas in corn-soy diets improved pellet durability by up to 10%; however, the data was not subjected to statistical evaluation (Behnke et al. 2002).

Dehulling improves the nutrient content of peas because it removes the indigestible, high-fibre hull and the ANFs it contains. Studies with 10- to 17-day-old broilers indicated that dehulling did not affect the AMEn or APD of low-tannin peas (Brenes et al. 1993), a result that was supported by performance trials (Igbasan and Guenter 1996). Conversely, dehulled cotyledons had improved starch digestibility and TMEn for adult cockerels (Longstaff and McNab 1987).

Peas have contained relatively low levels of ANFs, and have been considered to require little genetic selection against these traits (Monti 1983). Two European varieties, Maro and Progreta, were high in trypsin inhibitor activity, and it was suggested that feed peas used for monogastrics should not contain this parentage (Bond and Smith 1989). High tannin levels are pleiotropic with colored petals in peas (Bond and Smith 1989); however, most varieties grown in Canada are low-tannin, white-flowered varieties.

With the exception of phytate reduction through the use of phytase, attempts to improve the nutritive value of peas through enzyme supplementation have been largely unsuccessful (Broz and Beardsworth 2002; Campbell and van der Poel 1998; Wiryawan and Dingle 1999). Broiler and leghorn chicks tended to exhibit reduced weight gain and feed intake when crude enzyme preparations were added to low-tannin pea diets, but the same enzymes improved the FC of rations containing high-tannin peas (Brenes et al. 1993). Water extracts from high-tannin peas have been shown to inhibit trypsin, α-amylase and fungal cellulase (Griffiths 1981); therefore, selective binding of ANFs to the supplemental enzyme preparations may have contributed to improved performance.

A commercial cellulase preparation increased the digestibility of xylose from peas, but did not alter significantly pea TME value (Longstaff and M^cNab 1987). Neither performance, AMEn value (Igbasan et al. 2003) or ileal apparent digestibilities of protein, starch or fat (Daveby et al. 1998) was improved for broiler chicks consuming pea-based diets supplemented with crude pectinase or α-galactosidase. Some foreign studies (Jeroch et al. 1995; Keller and Jeroch 1997) and promotional materials (Charlton and Pugh 1995; Cowan et al. 1996) indicated an improvement in ME from enzyme supplementation of pea-based diets. The absence of statistical analysis made results from these studies difficult to interpret. Peas showed a significantly greater AMEn value (0.55 MJ kg^-1) when mixed with corn versus wheat (Carrè et al. 1987). Although adult cockerels were used in this study, the difference in ME may indicate that the viscous polysaccharides in wheat were reducing the nutrient availability of the peas. In conclusion, peas themselves do not require enzyme supplementation, but appear to benefit from the standard industry practice of adding viscosity-reducing enzymes to wheat and barley diets.

Many studies have been performed to identify factors that reduce the utilization of peas; however, peas are essentially well digested by poultry (Conan and Carrè 1989). Peas are ideal for commercial feed manufacturing because they contain low levels of ANFs plus the grinding, heating and pelleting processes used in commercial feed manufacture significantly improve pea nutrient digestibility. Steam pelleting also reduced the variability of starch digestion between cultivars (Carrè et al. 1991), and the benefits accrued from processing feed peas would be due to both improved digestibility and reduced variability.

Nutrient accessibility issues must be addressed in order for birds to access the full nutritional value of peas; however, the exciting work by Weurding et al. (2003) indicated that the moderate rate of pea starch digestion may provide a distinct advantage to feed peas as feed formulation moves toward assessing and meeting the metabolic requirements of poultry.

Brown, M.D. 1991. Oligosaccharides in pea protein concentrate. Undergraduate thesis. University of Saskatchewan. Saskatoon, SK. Canada.

Weins, I. 1992. A study on the effects of oligosaccharides in the diets of broiler chicks. Undergraduate thesis. University of Saskatchewan. Saskatoon, SK. Canada.