Although peas have proven an excellent feed source for all classes of livestock, they are particularly suited to swine. Peas are used in marketing and breeding stock diets, as well as for sensitive weanling diets. They are accepted at high proportions in the diet and the resulting performance has "graduated" peas from an alternative feedstuff to a mainstream ingredient around the globe. Industry acceptance has been driven, in part by extensive scientific research establishing pea nutrient levels and the availability of peas. The objective of this review is to summarize the research in a format suitable for industry nutritionists using peas, government agencies funding research and researchers.

While protein is key in diet formulation, specific amino acid digestibility is a more useful concept. In general, peas are characterized by high ileal digestibility of lysine and methionine, and comparatively low digestibility of cysteine and tryptophan (Fan and Sauer 1994, Leterme et al.1990). Table 1 summarizes ileal, true and real digestibility values from recent research trials.

Implications of the high ileal digestibility of amino acids in peas are that the crude protein content of the diet can be lowered when using peas with the addition of supplemental amino acids. Thus resulting in lower nitrogen excretion and less metabolic cost with less pollution while still maintaining pig growth performance (Gatel and Grosjean 1992). Since a ratio of 50:50 essential:non essential AA allows for adequate N utilization in growing pigs, the addition of peas along with supplemental synthetic AA may improve diet performance (Lenis 1999 as cited by Shelton et al.2001). Peas provide additional N for non essential AA synthesis which enhances the growth of pigs beyond that attainable when diets are balanced solely with synthetic essential AA.

In peas, protein is laid down early in seed development (Daveby et al.1993) and the amino acid composition is dependent upon the proportion of the three major soluble protein fractions found in the seed - legumins, vicilins and albumins of the 11S, 2S and 7S fractions respectively.

¹Spring variety grown in Canada
²Round white flowered (tannin free) - Baccara 1&2, Victor 1&2, Brevent, Blizzard, Neve, Cheyenne, Froidure, Aravis, Rafale, Radley, Baroness
³Amino & Finale = spring varieties commonly grown in France, Australian variety and Frilene = winter variety commonly grown in France
⁴apparent ileal digestibility
⁵ standardized ileal digestibility (true) a,b For impala peas, AID in the same row followed by different letters are significantly different (P<0.05) cd For impala peas, SID in the same row followed by different letters are significantly different (P<0.05)

Legumins and vicilins together constitute the major storage protein, globulin (Fig.1) (Creveau 1999). Albumins are rich in sulfur AA and tryptophan. There is a linear relationship between CP and most AA, the exception being tryptophan which is uniformly low (Fan and Sauer 1999).

Amino acid utilization by the pig begins with protein solubilization. Pea protein solubility is very dependent upon pH, in comparison to wheat, soybean meal (SBM), meat and bone meal or sunflower meal. With peas, 80% of the protein is solubilized at pH 9 and less than 30% at pH 3. At pH 4.5, pea albumins resist precipitation (Le Guen et al.1993). Effective breakdown of vicilin occurs by trypsin and chymotrypsin in the intestine (Neilsen et al.1988. This indicates that when peas comprise the major dietary protein, they are not effectively solubilized by gastric juices containing pepsin, which is typically active at pH 2-3. As the protein moves towards the small intestine, pH rises and more proteins are degraded (Cone 1993) as evidenced by higher activity of pepsin hydrolases in the ileum (Salgado et al.2002b). A portion of vicilin and albumin polypeptides survive digestion in the piglet, as Salgado et al. (2002) found protein bands appearing at the terminal ileum belonged mainly to soluble proteins of the 7S family and other low MW components. The resistance of legumins to hydrolysis was not ascertained due to procedural discrepancies.

These findings are consistent with others where endogenous nitrogen losses (ENL) increased with the inclusion of peas in the diet. The general conclusion was that increased ENL depressed apparent AA digestibilities as a consequence of inherent fibre or antinutritional factors such as trypsin inhibitors or tannins (Le Guen et al.1995b, Buraczewska et al.1989, Jondreville et al.1992; Fan et al.1994).

More recently these conclusions have been questioned. Leterme et al. (1997) observed that pea fibre added to a N-free diet significantly increased ENL without disturbing the protein digestion process as suggested by N retention. Hess et al. (2000) (cultivars Solara and Amino) determined that significant differences in endogenous losses observed for some AA did not translate into differences in real digestibility. These results support an effect on the mucosal surface rather than a disturbance of digestion per se (Leterme et al.1998). Salgado et al. (2002) identified the endogenous proteins as serine proteases (trypsin 25kDa, chymotrypsin 30 kDa, and blood co-agulation factor IX 27kDa) not of dietary origin. A trial by Grosjean et al. (2000, 1998) compared different peas types - smooth, white flowered versus smooth (FP), coloured flowered (CoP) or wrinkled, white flowered (WP) peas. Large differences in fibre content did not affect protein or amino acid digestibility. Rather, they concluded the main factor affecting digestibility of protein was phenolic compounds, even at low levels.

It is known that tannins reduce nutrient digestibility through complexing dietary protein and CHO, as well as inhibiting digestive enzyme activity and/or increasing endogenous protein secretion. They may also have detrimental effects on the intestinal mucosa (Salgado et al.2002). Yet, in a later study of tannin free varieties, Grosjean et al. (1998), demonstrated that CP and AA dig were negatively and linearly reduced by trypsin inhibitor, expressed as absolute activity (TIA) or trypsin inhibitor activity per unit protein (TIAP) (r² >0.75). Pea TIA may decrease enzyme activity or alternatively, higher TIA may reflect higher resistance to digestion.

Evidence against TI interference is provided by Grala et al. (1999b). In this study, true ileal digestibility of CP in peas (94.0%) was similar to SBM (96.5%) (P>0.05) however, apparent ileal digestibility was 6% lower. This indicates that TI did not interfere with protein digestion, still excessive endogenous nitrogen losses (ENL) were incurred. It is possible that inherent pea fibre may increase passage rate, hereby reducing digestion time, decreasing apparent digestibilities and increasing ileal flows (Grala et al. 1999). Unexpectedly the higher ENL did not result in higher urinary excretion of N as compared to SBM. Total N retention was slightly but significantly lower. The researchers speculated that the high level of the fermentable pea CHO shifted N excretion from urine to faeces or alternatively, the CHO could influence gut protein turnover resulting in direct uptake of AA by the intestinal gut wall which would limit the inefficiency of endogenous N synthesis, recycling and urinary N losses.

Fan and Sauer (1999) reported no correlation between TIA and AA digestibility (P<0.05), except for tryptophan. They did report that apparent ileal digestibility values of most AA were positively correlated (P<0.05) with their respective levels in the diet. Buraczewska et al. (1989) also found a tendency to increased true ileal digestibility with increasing protein content in peas. This suggests that the discrepancies between TIA, fibre and phenolic effect on N digestibility may be a result of varying AA and CP levels in the assay diets, that could be overcome by reporting digestibilities on a true basis. Huisman et al. (1992) found that even, in young pigs (7.5 - 8.5kg), true ileal digestibility was very high (0.951 and 0.929) indicating that the protein of raw peas was almost completely digested in the small intestine.

Beyond the implications of TIA, phenolics and fibre, the low AID of pea protein may also reflect the compact structure of legume proteins which impairs hydrolysis due to differences in the stability or flexibility of globulin structure (Salgado et al.2002, Helander et al.1996). As well, differences among pea varieties may affect endogenous loss and digestibility, where in some cases digestibility differs but losses do not (Hess et al.1998).

Another theory for lowered AA digestibilities in peas is antigenicity of pea protein (LeGuen et al.1991b). As seen with various soybean products, an immunological response may have a detrimental effect on the small intestine and affect pig performance in the first few weeks following weaning. Salgado et al. (2002c) revealed an immunological reaction to vicilin proteins of peas, fababeans, chickpeas and lupins in weaned piglets (8.54kg kg, 28 days of age). Response was lowest for peas, with no response emitted to any legumin proteins despite their close structural relationship to vicilin. Le Guen et al. (1993) however did report a response to both legumin and vicilin (winter pea Frijaune, piglets 5.7 -7.5kg), independent of whether or not the sow's diet contained peas. Absorbance values for piglets in prior contact with pea proteins either indirectly from maternal immunity or direct consumption of the mother's diet were higher than for those piglets with no contact (+77% for legumin and +172% for vicilin). For both studies, the observed reaction indicates an immune response rather than a food allergy. With a true food allergy, gastrointestinal distress would have included diarrhea, weight loss and poor growth, none of which was apparent.

Despite uncertainty in the explanation for reduced AA digestibility it is noteworthy that when comparing pea starch to maize starch, both portal and arterial amino acid concentrations tended to be higher for pea starch (van der Meulen et al.1997). Also the availability of essential amino acids was significantly higher for the pea starch diet (P<0.05) (Bakker et al. 1998).

Peas can be fractionated into constituents of protein, fibre and starch. Pea protein concentrate (PPC) is derived from the air classification of ground peas; pea protein isolate is produced either by ultrafiltration or acid precipitation of solubilized pea protein (Le Guen et al.1993). Both products contain higher crude protein than raw peas. Pea protein isolate is devoid of carbohydrates and antinutritive factors whereas the simpler process of air classification has no apparent effect on either. Pea protein isolate may contain as much as 90% globulins with albumins reduced to 10%. This results in a lower sulfur amino acid content but more digestible product as globulins have a higher digestibility than albumins (Fan and Sauer 1994). For example, nitrogen ileal digestibility increased for pea protein isolates produced from spring pea Finale (69.1 to 83.7%) and winter pea Frijaune (69.5 to 85.4%) (Le Guen et al.1993).

When fed to 21 day old piglets (5kg) PPC fed animals grew more slowly due to a significantly lower feed intake (262g d^-1 vs. 448g d^-1 on the soybean meal diet). Although PPC generally contributes to pellet hardness, in this case the PPC pellets were more friable making the feed powdery and unappealing. Poor digestion of the PPC may have also resulted from the concentration of oligosaccharides or other ANF. The recommendation by the authors was that PPC should be restricted to 10% or less of the total diet for weanling pigs (Christison and Solano 1982).

Pea protein isolate (PPI) contains a similar nutrient profile to spray dried porcine plasma (SDPP), at a fraction of the cost. Using 16 day old (5.8kg) pigs, Baidoo et al.(1999) demonstrated that pigs fed PPI outgained (+4.7%) pigs fed SDPP at identical inclusion rates (75 and 45g kg^-1) over a 41 day period. (It is prudent to note that during the first two phases SDPP outperformed PPI but substantial compensatory growth occurred during Phase III with a commercial starter. - Editor's note) In another trial, with younger pigs (10 day old, 3.8kg, microbially challenged with enterotoxic Escherichia coli) equivalent performance in ADG, FI and gain:feed (P>0.05) was seen when 100g kg^-1 PPI replaced an equal amount of SDPP. Distressingly, mortality increased from 6.6% to 40% with the PPI diet. The addition of either egg yolk antibodies, zinc oxide, fumaric acid or carbadox to the PPI diet returned mortality rates to acceptable nonsignificant levels without affecting performance. Intestinal morphology was compromised in the PPI diet without an antimicrobial. The conclusions from this study still allow for the use of PPI in early weaned pigs however antimicrobial supplementation must be considered in a microbially challenged environment (Owusu-Aseido et al.2003).

The high energy value derived by pigs from peas may be attributed to its complex carbohydrate composition. Pea carbohydrates are mainly starch, concentrated in the endosperm and non starch polysaccharides (NSP) distributed throughout the cotyledon and hull (Table 2). Absolute values of NSP vary between arvense and hortense varieties and composition of NSP varies between cotyledon and hull (McCartney and Knox 2002). During early development of the pea seed, sucrose and glucose plus fructose are dominant while oligosaccharide concentration remains low. At maturity, oligosaccharides increase, while the former stabilize at a low level. In mature pea seeds, starch levels vary among feed (429-527g kg^-1 DM), garden (357-545g kg^-1 DM), coloured (327-506g kg^-1 DM) and wrinkled peas (186-346g kg^-1 DM) (Bastianelli et al.1998). The three primary NSP constituents, glucose, uronic acid and xylose, increase considerably during development (Reichert and MacKenzie 1982; Daveby et al.1993) of cotyledon cell walls, which are rich in arabinose-containing pectic substances with xyloglucans and cellulose occurring as minor components. Pea and bean cotyledons yielded only 6.9 and 7.0% cell wall, which is considerably lower than soybeans at 9.3% (Brillouet and Carre 1983). The cell wall material of pea hulls has high insoluble residues, comprising cellulose, acidic xylans, pectic polysaccharides (uronic acids) and lignin. Soluble NSP is low (15%) in pea hulls compared with cotyledons (47%) (Canibe et al.1997, 2001). The oligosaccharides of peas are mainly of α-galactosides, which have a sucrose moiety base and one to four additional residues of α-galactose yielding raffinose, stachyose, verbascose and ajucose respectively. The average content of α-galactosides on a dry matter basis are 32-46g kg^-1 for peas (Bach Knudsen 1997). The proportion of hull is negatively correlated with seed size, in contrast to internal fibre which is not affected by seed size. For example, feed peas and coloured peas have the same content of cotyledon cell walls despite a very different seed size. The negative correlation between starch and protein is not related to seed size but the negative relationship between starch and fibre is in part linked to total seed weight (Bastianelli et al.1998).

A major determinant of dietary energy is the digestibility of starch and NSP. Energy values along with corresponding ileal digestibilities for peas can be found in Table 3. Digestion of pea starch, like maize starch, is nearly complete by the time it reaches the terminal ileum. Bakker et al. (1998) observed 98.3% digestion of maize starch and 97.9% for pea starch (P>0.05). This is comparable to in vitro results for resistant starch reported by Martin et al. (1998). Others (Abrahamsson et al. 1993) found a significant reduction from 92.5% starch digestibility for a barley-SBM control diet to 90.7% for a diet containing 33% light colored peas. Similarly, Canibe et al. (2001) reported starch digestibility was significantly lower for a diet containing pea cotyledons and one containing a pea hull and pea starch mixture.

Even with high starch digestibility, Bakker et al.(1998) observed differences in glucose absorption. Pea starch was absorbed more distally in the gastrointestinal tract and the appearance of glucose was delayed but sustained for a longer period with pea starch as compared to maize starch. Van der Meulen et al.(1997) reported the highest portal glucose concentration one hour after feeding peas compared to 0.5 hours after feeding maize. The net portal glucose flux was significantly lower during the first six hours for peas and significantly higher during the last 4 hours. Over 12 hours, only 72% of ileal digested glucose was accounted for with pea starch as compared to 97% for maize starch.

Overall this decreased the amount of retained energy available to the pig, 212 versus 250 kJ per W ^0.75 (P<0.01) (Bakker et al.1998, van der Meulen et al.1997). The reduced glucose supply has yet to be explained satisfactorily. Rainbird et al. (1984 as cited by Dierick et al.1989) demonstrated that the addition of guar gum to a glucose solution halved the rate of glucose absorption, due to reduced diffusion from the intestinal lumen to the epithelial cells. Possibly fibre in peas similarly affects glucose uptake. It has been shown that VFA production provokes the release of enteroglucan, which inhibits gastric emptying, leading to gastric retention and slower glucose appearance (Van Der Meulen et al 1997).

Starch digestion and ultimately glucose absorption may be affected by the amylose:amylopectin ratio. Pea starch typically contains more amylose Table 4 (Martin et al.1998). Feed peas contain 32% amylose while much higher levels are found in wrinkled peas (63%) (Bastianelli et al.1998). Le Guen et al.(1995b) indicated that pea starch being one-third amylose, was not as well digested by monogastrics as amylopectin. Several other explanations invoked to explain the lower digestibility of pea starch compared with cereal starch include the entrapment of starch in fibrous, thick walled cells and the crystalline C pattern structure of pea starch granules which is more resistant to pancreatic amylase than the A pattern typical of cereal starch granules (Wursh et al.1986; Ring et al.1988; Gallant et al.1992 as cited by Canibe and Bach Knudsen (1997). Others have implicated α-amylase inhibitors to explain the relatively slow hydrolysis of pea starch (Gatel 1994 as cited by Salgado et al.2002). Nonetheless, however the delayed release of glucose may have a positive impact on protein and amino acid availability (Bakker et al.1998).

Pea NSP is well utilized by the pig compared to other fibre sources. Goodlad and Mathers (1991) suggested that the digestibility of galactose and uronic acids of wheat are enhanced by the inclusion of peas in the diet. In an experiment comparing soy concentrate, SBM or a mixture of toasted and untoasted SBM, peas had the highest digestibility of NDF (P<0.05) (Grala et al.1999). The pectic polysaccharides of the pea cotyledon are readily available to the microflora in the proximal segments of the large intestine, while the acidic xylans and cellulose from pea hulls are more resistant to digestion (Canibe et al.1997). Canibe et al. (2001) found that pea cotyledon fibre had the highest tissue swelling and water retention capacity (10-12g of water retained/g of matter Leterme et al.1998) compared to pea hulls, barley hulls or dehulled barley which decreased from ileum to feaces. This was accompanied by a considerable reduction in fibre particle size reflecting digestion of the water soluble NSP with transit down the GI tract. Leterme et al. (1998) offered the explanation that pea inner fibres were insoluble, with water retained only in the empty cells and claimed pectic substances found in the soluble fraction interfered with reabsorption of some endogenous secretions (Leterme et al.1996).

Since pigs lack the appropriate digestive enzymes to hydrolyze pea α-galactosides, they escape undigested to the hindgut where they are fermented into VFA (Salgado et al.2002). Resistant pea starch too may be fermented in either the stomach or distal small intestine (Van Der Meulen et al.1997, Martin et al.1998). Contributions of VFA to the total energy of growing finishing pigs amounts substantially to 15% of the NE of maintenance and may be higher for gestating sows and breeding boars (Dierick et al.1989).

Varietal comparisons with round white flowered (FP), coloured flower (CoP) and wrinkled white flower peas (WP) revealed energy digestibility for FP (0.89) was higher than WP (0.82) due to less starch and higher fibre, and was lower still for CoP (0.81) due to tannins (Grosjean et al.1998). The difference in digestibility of energy between FP and WP was only 6% for pigs but rose to 10% in poultry. This spread would be even larger in terms of NE because VFA have a lower energy efficiency in metabolism. Taverner and Curcic (1983) observed that fecal energy digestibilities of peas and SBM were equivalent, however 6% of the total digested energy occurred in the hindgut with peas, in contrast to SBM where there was 25% disappearance indicating a far greater energy contribution was conferred by the peas.

Peas have the same energy value for post weaning and growing pigs (Van Cauwenberg et al.1997). However, energy values derived with growing pigs underestimate the value for adult pigs, particularly sows (Table 5) (Noblet and Bourbon 1997). The origin of dietary fibre is important, as pectin and water-soluble NSP are more conducive to digestion than lignin and water insoluble fibre. Sows are able to ferment NSP as efficiently as starch and the inclusion of NSP may even enhance sow reproductive performance (Van der Peet-Schwering et al.2002).

The addition of enzymes to pea diets to reduce the flow of fermentable substrates to the large intestine, has had limited success. Results of recent trials are summarized in Table 6. In all instances, there was some improvement in digestibility of nutrients, sometimes even significant however, these improvements were not translated into overall pig performance. Considering the high oligosaccharide component in peas, potential improvements with enzyme addition seem likely. However, the naturally high digestibility of the NSP component of peas due to high water solubility may mitigate the benefits of adding enzymes. Endo glucanase and xylanase would not affect the pea component of the diet since peas do not contain endo 1,3(4)-B-D-glucan and very little arbinoxylan. The cellulose present in peas is sufficiently digested by the cellulase already present in the digestive tract (Nonn et al.1999). Effects seen in the first period post weaning may be important since adequate substrate specific microbial populations are not yet developed (Nonn et al.1999).

* commercial enzyme cocktail that provided protease, cellulase, xylanase, α-galactosidase, amylase, B- glucanase, and pentosanase

** α-galactosidase plus cellulase, endo 1,3(4)-B-D-glucanase, and endo 1, 4-B -xylanase

*** α-galactosidase plus enzyme complex Roxazyme = cellulase, endo 1,3(4)-B-D-glucanase, and endo 1, 4-B -xylanase

Peas provide considerable phosphorus (P) with higher availability (47%) than many plant sources. Inherent phytase activity of pea is 100 vs 200U kg ^-1 in barley (Helander et al.1996). Bone bending moment of metacarpal bones (20kg pigs) responded linearly to 3.0, 3.5 and 4.0g kg^-1 P supplied either from monosodium phosphate (p<0.01) or peas (p<0.05) but not to diets containing SBM (Ketaren et al.1980). Pallauf et al.(1994) reported that, in contrast to maize-SBM diets where all phytate P was indigestible, a diet based on peas, fababeans, wheat, and barley showed 37% of phytate P to be digested. Supplementation of microbial phytase to these diets (350 and 700U kg ^-1) increased P, Ca, Mg and Zn utilization in pigs (12 -16kg). Helander et al.1996 also found the addition of microbial phytase improved P digestibility in peas (0.468 to 0.689) in older pigs (48-75kg).

High fibre levels have been implicated in interfering with mineral absorption by pigs. Increasing pea hulls had no effect on apparent absorption of calcium, phosphorus or magnesium nor were these minerals available for absorption from the pea hull. However apparent absorption of sodium and potassium was significantly reduced (P<0.001) by pea hull addition. It was suggested that some fibre components (other than cellulose) may have a greater propensity to bind monovalent rather than polyvalent cations since there was higher excretion of sodium and potassium but not of calcium, phosphorus or magnesium (Stanogias et al.1994).

As is well documented, pulse crops are known to contain trypsin inhibitors, tannins, lectins (haemagglutinins) and saponins. The effects of these antinutritional factors (ANF) vary depending on the ANF in question. Pigs in general do not suffer serious consequences to their presence in the diet, however, age and size of pig are important considerations (Arentoft et al.1991). Antinutritional factors vary among cultivar. In peas, tannins are only found in arvense (dark colored flower) varieties (Grosjean and Gatel 1989) where they are concentrated in the testa (Griffiths 1981). Trypsin inhibitor activity (TIA) is greater in winter than spring and smooth than wrinkled pea varieties (Valdebouze et al.1980). Canadian peas, the majority of which are spring seeded, white flowered varieties are characterized by very low levels of ANF of little nutritional significance. In a Canadian pea variety trial only 8 out of 56 samples exhibited trypsin inhibitor units (TIA x 1.9) greater than six. The two highest cultivars contained 10.3 and 9.6 trypsin inhibitor units (100% DM). In comparison, half of the 16 Austrian winter pea (Maple pea) varieties tested were above 6.0, with the highest being 10.7 trypsin inhibitor units (100% DM) (Slinkard and Tyler 1993).

Despite the low levels, trypsin inhibitors remain the most troublesome ANF in peas. Both trypsin inhibitor activity (TIA) and chymotrypsin inhibitor activity are independent of protein content. In peas 10% of TIA is found in the hulls and 90% in the cotyledons, which is proportionate to their respective weight distributions in the whole seed. Chymotrypsin inhibitor activity is higher in peas than in beans whereas the reverse is true for TIA (Griffiths 1984). In peas TIA and chymotrypsin inhibitor activity are highly correlated (r=0.986) suggesting a double headed trypsin/chymotrypsin inhibitor analogous to the Bowman Birk inhibitor found in soybeans (Grosjean and Gatel 1989). Arentoft et al.(1991) reported that pea inhibitors are different from Bowman-Birk and Kunitz soybean trypsin inhibitors. They reported that young piglets (several days to 8 weeks of age) showed 2-10 times higher inhibition than full-grown animals. In addition they found the degree of inhibition to vary among species and stressed the importance of using animal species of interest at the correct age and determining 50% inhibition as the benchmark measurement. Boisen (1989) proposed that an extraction procedure at pH 2 with added pepsin followed by an incubation of the extract with porcine trypsin gave the best evaluation of trypsin inhibitors in all types of pig feeds. Using this method they found that porcine trypsin was inhibited less than bovine trypsin by either legume or cereal inhibitors.

Both pea and bean proteolytic enzyme inhibitors remain stable at temperatures at or below 80°C and their inhibitory properties are unaffected by pelleting. Both are slightly reduced at 100°C and completely denatured by extrusion (Grosjean and Gatel 1989) or autoclaving (Griffiths 1984; Fan and Sauer 1994). TIA does not decline with time, as levels remained the same after storage for one year (Grosjean and Gatel 1989).

An early explanation of the mode of action of protease inhibitors was that as a consequence of inhibition of proteolytic enzymes in the digestive tract, the pancreas was stimulated to synthesize and secrete more enzyme protein. Ultimately this resulted in not only a hyperactive pancreas but also an increased demand for amino acids by this organ. In addition the higher loss of inactivated proteolytic enzymes depresses the apparent absorption of sulfur amino acids, since both methionine and cysteine are relatively consistent amino acids of the proteolytic enzymes themselves (Griffiths 1984).

Lectins (haemagglutinins) are proteins with the capability of binding sugars present as glycoproteins. Lectins are typically measured according to their ability to agglutinate erythrocytes (red blood cells) through cross linking surface glycoproteins (Huisman et al.1990). Although this binding is not a physiological concern lectins similarly bind to the mucosa of the intestinal wall damaging the epithelial cells, depressing nutrient absorption, reducing activity of brush border enzymes, and causing hypersecretion of endogenous protein with the shedding of damaged cells. Increased production of mucins and a loss of plasma proteins to the intestinal lumen, also contribute to decreased nitrogen digestibility and retention. Occasionally scouring may further reduce weight gain and feed conversion (Huisman and Van der poel 1994). Despite this dire description, in a comparison of 8 pea cultivars, lectin in peas were only one tenth of that found in soybeans (Valdebouze et al.1980).

Saponins are complex material based on a sugar linked to a steroid or triterpenoid moiety. Saponins have detergent like effects as illustrated by their haemolytic capabilities when incubated with erythrocytes attributed to their interaction with cholesterol in the erythrocyte membrane. Legume extracts vary in their ability to lyse red blood cells. For peas, beans and soybean extracts, sheep red blood cells were more sensitive to saponin extracts than rabbit red blood cells. Survival of guppy fish was lowest for those exposed to crude saponin extracts from beans (3.8 minutes) as compared to peas (6.6 minutes) or soybean extracts (11.4 minutes). The different sensitivity according to saponin source is indicative of structural differences among saponins. Saponin extracts from peas were separated into seven fractions whereas the saponin extracts from beans or soybeans were separated into only six fractions (Khalil and El-Adawy 1994).

Overall, the effect of pea ANF on pigs appears relatively slight. The repeated lack of performance response to pea processing and the inconclusive results seen during protein digestion bring into question the importance of ANF consideration in pig diets. Using two varieties of Canadian peas with varying TIA (1.12 and 4.60) in semi synthetic diets, Gabert et al.(1996) found only minor effects on the quality of exocrine protein secretions relative to fababean (which have a higher content of tannins than peas) or SBM diets. In a second study using the same diets no differences in concentration, flow or composition of total, protein bound or free AA in the pancreatic juice of 18kg pigs were noted (Gabert et al.(1996). Following a series of experiments examining the effects of pea ANF in young (10-15kg) piglets, Le Guen et al. (1995; 1995b; 1993; 1991b; 1991c) found isolated ANF alone did not account for the entire decrease in digestibility observed when raw peas were fed in comparison to an extracted pea protein isolate devoid of ANF. Both diets were balanced for NE and amino acids. Alternatively, they suggested that ANF activity may depend on the protein source associated with them. Grosjean et al. (2000) suggested a maximal threshold of TIA= 4 TIU mg DM ^-1 or CP=16 TIU mg DM ^-1 (of which most varieties are below but can range from 2-15 TIU). Regression equations for amino acid digestibility based on TIA and TIAP were established (eg. dLYS=-0.1617TIAP+87.84 (r² = 0.84 RSD = 0.98 p<0.001)). The negative relationship between AA digestibility and TIA was recognized earlier, however the correlation was low (0.53-0.69) (Grosjean et al. 1998).

For older pigs (33-74kg) Buraczewska et al. (1991) observed no increase in pancreatic secretion with a diet containing 260g kg^-1 peas (white flowered Opal variety) compared to a 95% barley diet. Tryptic, chymotryptic and amylolytic activities were not appreciably altered for the barley or pea barley diets when measured as trypsin inhibitor units ml^-1 pancreatic juice. These results indicate that the trypsin inhibitor and tannin levels occurring in peas do not negatively affect pigs even at a body weight of <10kg.

In conclusion, is unlikely that ANF occurring in Canadian peas are of sufficient level and type to affect piglet performance. Additionally in practical diets their low concentration deems their impact negligible.

The pigment level of peas varies with protein content. Low protein peas appear very bleached (faded yellow) whereas high protein content peas were a bright yellow. Xanthophylls were the predominant carotenoids in these peas (Reichert and MacKenzie 1982).

Numerous processing procedures have been applied to peas including dehulling, toasting, extrusion, pelleting and expansion. In general, pigs elicit only a modest response to most processed peas and often no response at all. Removal of the outer hull, or dehulling, offered limited improvement for swine owing to the low proportion of hull matter (72-103g kg^-1 Bastianelli et al.1998) and the inherently high digestibility of this fraction. Thacker and Racz (2001) reported improvements in DM, CP GE digestibility (P=0.004, 0.003 and 0.004 respectively) over SBM for pigs (9.9-103kg) with dehulling, however this did not result in improved animal performance.

Results from extruding pea seeds appear more favourable. Extrusion improves starch digestion through gelatinization and increases protein digestion due to protein denaturization and ANF destruction (O'Doherty and Keady 2000). Bengala Freire et al.(1991) observed significantly increased starch digestibility in both spring (97.1 to 98.9%) and winter peas (94.4 to 99.1%) following extrusion. In a winter pea variety (Frilene) amino acid (AA) digestibilities increased by 10 percent. A marked increase in tryptophan and cysteine digestibility was found.

Extruding starter diets (150°C, 15 seconds) containing either winter or spring peas (450g kg^-1) significantly improved apparent ileal digestibility of dry matter and nitrogen for winter peas, with no effect on spring varieties. The same was true for nitrogen retention. It was suggested that the destruction of ANF was responsible for the improvements to winter pea varieties (Freire et al.1991). Extrusion of a spring seeded pea variety showed an improvement (P>0.05) in F:G, ADFI and ADG over a SBM control (pigs 33-100kg). More importantly, extrusion allowed for higher levels of peas to be included in the diet. Whereas pig performance and digestibility of DM, N and NDF and DE declined (P<0.05) when raw peas increased from 200 to 400g kg^-1, no differences were seen with extruded peas at the high level. Carcass yield was unaffected by extrusion. Extrusion reduced tannin (10.4 to 6.2 mg g^-1) and TIA content (2 to 0.4 mg g^-1) and increased the proportion of gelatinized starch (0.215 - 0.746), although the starch content decreased (359 to 297g kg^-1 DM). It is possible that extrusion degraded the starch into structural units not measured (O'Doherty and Keady (2000). In a study for very young SEW pigs, Landblom and Poland (2000) concluded that extruded peas should not exceed 200 g kg^-1 in the diet for piglets weighing 7.5kg, despite improvements over raw peas included at the same level.

Owusu-Asiedo et al.(2002) compared extrusion and micronization of peas in Phase I and II diets (pigs 4.5 - 20kg). Digestibility of heat treated amino acids improved significantly (Table1) on both an apparent and standardized ileal basis with little difference between extrusion and micronization (P>0.05). Although all treatments contained amylase (0.2%) and xylanase (0.1%), improvements were attributed to a reduction of ANF, conformational changes in storage proteins and possible cell wall solubilization allowing for easy access of endogenous and exogenous enzymes. No performance advantages were seen from heat treatment. However as compared to the SBM control plasma urea nitrogen (PUN) levels were reduced for all pea diets in both Phase I and II (P<0.05), the single exception being for raw peas in Phase I. Lower PUN indicates reduced urea synthesis and more efficient use of AA for body growth.

Toasting (130°C, 3-4 minutes) did not affect NSP composition, solubility, fermentability or digestibility (Canibe et al.1997). It did however affect the nature of starch digestion, increasing the rapidly digested starch fraction at the expense of the slower digested and resistant fractions. Toasting decreased the digestibility of α-galactosides as the heat destroyed the endogenous a-galactosidase of the pea (Canibe and Bach Knudsen 1997). Despite a loss of the crystalline structure of starch after toasting, the cell walls surrounding the starch granules remained intact.

Excessive heating has deleterious effects on protein availability. Native (unheated) vicilin was extremely susceptible to trypsin digestion yet in vitro digestibility of vicilin slowed to half previously observed after heating (99°C, 15 minutes). A high degree of structural integrity was maintained indicating strong hydrophobic interactions in stabilizing the protein structures, which reduced susceptibility to digestion (Deshpande and Damodaran 1989). A series of four papers examined the effects of heat treatment of peas on lysine digestibility, availability and utilization. As heating increased from 110°C to 165°C (15 minutes) there was little change in lysine apparent ileal digestibility over raw peas (0.79 - 0.56) (Barneveld and Batterham 1994a). However lysine availability declined from 0.96 to 0.47, resulting in poorer growth performance. The implication was that although the amino acid was absorbed it was poorly utilized after heat treatment (Barneveld and Batterham 1994b; Barneveld and Batterham 1994c). A large proportion of non-utilizable amino acids in heated field peas may be excreted from the pig via the urine in the form of a protein. Differences in pea cultivars were noted suggesting that some cultivars were more susceptible to heat damage than others (Barneveld and Batterham 1995). (Heating for fifteen minutes is far longer than most practical feed applications - Editor's note.)

The most difficult test for the acceptability of an ingredient is its use in diets for newly weaned pigs. At this critical stage, ingredients need to be not only highly digestible, but also highly palatable. In a Canadian trial, peas were included up to 200g kg^-1 in a diet for weaned piglets (5.8-6.9kg). No differences in feed intake (FI), average daily gain (ADG) or feed/gain (F/G) were found in isoenergetic and isolysine diets comparing peas and SBM (Kehoe et al.1995). In contrast, Landblom and Poland (2000b) recommended peas be limited to <200g kg^-1 if raw and not to exceed 200g kg^-1 if extruded for piglets. Yet higher levels (300g kg^-1) of peas were successfully incorporated into diets of 10.9kg liveweight pigs. Results were equal to a SBM control (325g kg^-1), however supplementation of the pea diet with 1.2g kg^-1 DL-methionine significantly increased all parameters (Gatel et al.1989). Similarly a German experiment concluded that peas could be included up to 300g kg^-1 in piglet diets (10-25kg) (Bohme 1988). Additionally, Nonn et al.(1999) found highly satisfactory daily gains of 412-469g of barrows (7.8kg) with 400g kg^-1 peas in the diet. With the same level of green or yellow peas, Grosjean et al.(1997) reported ADG ranged 497-612g in pigs 8 to 25kg (Table 7). These results indicated that performance of weanling pigs was not adversely affected even at a relatively high dietary inclusion level of peas (Pisum sativum hortense).

The performance of grower finisher pigs is, in most cases, unaffected by the inclusion of peas in isocaloric and isonitrogenous diets. Often due to the uncertainty of nutritionists peas have been restricted to levels of 200g kg^-1 or less in grower-finisher feeds. Newer plant varieties have alleviated some of the traditional problems of high trypsin inhibitor and tannin levels and higher inclusion levels have been successfully reported in recent research trials (Table 8). Stein (2001) reported equivalent performance to a corn-SBM control from feeding graded levels of peas in grower and finisher diets (180g kg^-1 and 360g kg^-1 respectively). Overall performance of pigs (33-114kg) fed peas (670g kg^-1) was equivalent to or significantly higher than diets composed of SBM, extruded soybeans, canola meal, peanut meal, sunflower meal, meat-meal, poultry by product meal (Shelton et al.2001). Landblom and Poland (1998) compared various energy grains, with peas (Profi) (350-400g kg^-1) as the sole protein source in each of 4 phase diets. The barley - pea fed pigs grew faster, required less days on feed (P<0.05) and were more efficient (P<0.01), although acceptable performance was found with naked oats, corn and combinations of all grains.

In three experiments comparing a total of 12 different diets representative of typical on-farm mixes and more complex commercial diets, the inclusion of peas (200 and 450g kg^-1) resulted in no significant differences in growth performance when amino acids and digestible energy were balanced. It was noted that in some grower diets, ADG lagged behind that of the control however, this was overcome by significantly (P<0.05) improved growth in the finisher (Gatel et al.1991). This response, possibly indicating a restriction of secondary amino acids (methionine, cysteine) had been observed before and prompted Jaikaran et al.(1995b) to establish a pea experiment with or without supplemental methionine and threonine. Even with pea at 500g kg^-1 for growing pigs (21-53kg) significant differences in ADG, FI and F/G were not found between the experimental diets and the control. However, significant improvements over the control were seen in ADG and FC with the addition of synthetic methionine but not threonine. Supplementation of the finisher (53-102kg) diet containing 400g kg^-1 peas resulted in no significant performance differences over the control. Over the entire grower-finisher period no differences were seen among any treatment suggesting that methionine supplementation of pea based diets is only beneficial to approximately 50kg liveweight.

The addition of canola either as whole seed or in meal form is a complementary source of sulfur containing amino acids often used to supplement pea diets. This combination offers the benefit of a complementary amino acid profile (Alert J II et al.1999) at a lower dietary crude protein concentration. A Finnish study using market pigs (25.5 -105kg) found daily gain and feed conversion ratio unaffected when protein content was lowered and the diet balanced with synthetic lysine and threonine in a 1/3 pea 2/3 rapeseed meal diet (Valaja et al.1993). Matre et al. (1989) found better growth rates when pea diets were supplemented with RSM or herring meal as opposed to synthetic amino acids. Kornewietz et al. (1997) found pigs fed a combination of 200g kg^-1 peas and 50g kg^-1 rapeseed meal performed equally with similar carcass characteristics to a SBM control. A Canadian study, using pea screenings and CM (Castell and Cliplef 1993) found overall growth and feed conversion were not different for market pigs (25-97kg) fed either a blend of pea screenings and CM (124 and 141g kg^-1 or 61 and 283g kg^-1 respectively) or pea screenings alone (424g kg^-1) or SBM (150g kg^-1) or CM (187g kg^-1).

Poorer performance was observed with the pea and CM diets in the grower phase (<63kg), however in the finishing period, all CM and pea diets were superior to the SBM diet. Using regression analysis the calculated maximum growth rate (in accordance with the experimental conditions) would have been achieved with a diet containing 180g kg^-1 peas and 100g kg^-1 CM.

The importance of ingredient analysis for proper diet formulation was demonstrated in a trial by Edwards et al. (1987). Relying only on published values for pea nutrients, the authors were unable to obtain performance equivalent to a SBM control, especially when peas made up 300 or 450g kg^-1 of the diet for grower pigs. Following chemical analysis, diets were reformulated and pigs fed diets containing 300g kg^-1 peas performed equally to their SBM counterpart. Since peas rarely undergo processing prior to feeding, their nutrient composition is subject to the natural variability of environment and genetics. Digestibility of peas was higher (P>0.05) when the crop received adequate rainfall, no nitrogen fertilizer or cattle slurry or when the seed had been protected against pea weevil (Sitona lineatus) (Grosjean et al.1998). Even so, a comparison of growing conditions alone did not give clear evidence to the causes of variability in the feeding value of peas (Grosjean et al.1998b). Therefore to ensure maximum pig performance it is imperative to analyse the nutrient composition of a particular lot of peas before feeding.

Regardless of whether peas are added to corn, wheat, barley, tapioca or naked oats pigs perform equally well with isocaloric and isonitrogenous diets. In grower finisher diets, peas provide growth performance equivalent to or greater than other protein sources such as SBM, canola meal, sunflower meal, meat and bone meal and poultry by-product meal.

Pea screenings are a derivative of the cleaning process when peas are prepared for human consumption. Composed of undersized, off-color or broken seeds, pea screenings are an excellent source of nutrients for pigs and do not vary in composition from peas unless the sample is diluted with cereal, weed or other seeds. Using pea screenings composed of only peas (0, 110, 220 and 330g kg^-1) from two Canadian pea varieties (B.C. Blues [green] and Century [yellow]) no significant differences in the live performance means or carcass yield between any of the treatments were found. Daily gains were (878, 887, 828, 857g) B.C. Blues and (899, 863, 859, 832g) for Century for pigs (24-85kg) (Castell et al.1988). These results lend credence to those interested in replacing imported SBM with locally grown ingredients.

Although current management systems in intensive swine production do not readily allow for the feeding of whole crop (entire plant including seed) ingredients, studies have evaluated the use of whole crop peas in swine diets. Lund et al. (1981) determined the ME value of whole crop peas to be 10.7 MJ kg^-1 DM with a protein content of 125g kg^-1 DM. Compositional changes, as the crop matures, were established by Graham and Aman (1987). They found nutrients in late harvest originate mainly from the pea cotyledon and are readily absorbed in the small intestine. However the remaining residue entering the large intestine contained mainly senesced and lignified leaves and stems that are very resistant to microbial attack. In earlier harvest, protein and starch (mainly enclosed within the cell walls of the stems and leaves) passed the ileum and were susceptible to fermentation in the large intestine conferring an undetermined advantage to the pig. The authors advised that maximum benefit of whole crop peas to pigs was achieved when seeds are near maturity. Carlson et al.1999 compared whole crop peα-barley silage and clover grass silage for growing pigs (30kg). Pea silage, harvested 4 weeks post barley heading, contained 250g kg^-1 whole crop peas. Daily dry matter intake of both roughages was equivalent (20%). Higher levels of xylose and cellulose for the barley pea silage, resulted in lower NSP digestibility (P<0.05), compared to the clover - grass silage with higher levels of uronic acids. Whether the difference in digestibility is due to absolute content or physicochemical properties is yet to be ascertained.

Various studies have looked at the implications of peas in sow diets (Table 9). Overall, reproductive performance such as piglet mortality, average number of weaned pigs, weaning to rebreeding interval and cull rates, along with sow condition (body weight, backfat thickness) remain unchanged with the addition of peas. Positive results were reported with peas included at up to 300g kg^-1 in both lactating and gestation diets (Landblom et al.2001, von Leitgeb et al.1994).

It is interesting to note that peas may be able to provide greater energy during gestation since sows are able to ferment NSP as efficiently as starch and there is the possibility that NSP even enhances sow reproductive performance (Reece, 1997 as cited by Van der Peet-Schwering et al.2002).

With the increase in availability of feed peas, information on carcass composition of pea fed pigs has become more readily available in the last 3 years (Table 10). In an extensive trial, using 96 Dutch Landrace pigs, Szabo et al. (2001) examined numerous meat quality parameters between pigs fed pea, sunflower meal, fish or soybean meal as a protein source. They found no significant differences in intramuscular fat, pH value, lightness, hue, and water loss or color stability. Carcass composition did not differ as lean meat percentage, empty body weight, organ weight and muscle and bone proportions were similar (P>0.05). Bone development was also unaffected by protein source. These results were confirmed in a second trial with 32 Dutch Landrace pigs where again no differences were found in ADG, gain:feed, intramuscular fat content, colour, hue, water loss or bone development (Szabo et al.2001).

In contrast, Shelton et al.2001) reported pigs fed 670g kg^-1 peas had slightly lower levels of acceptable quality lean (P<0.10) according to NPPC standards (52.51 vs. 53.93%) compared to a SBM corn based diet. Also fat free lean (%) was lower (P<0.10). Comparing corn, naked oats and barley as energy sources with peas (350-400g kg^-1) as the sole protein source, hot carcass weight, percent yield, fat depth, loin depth and fat free lean index did not vary due to dietary treatment. Mixtures of barley and naked oats or barley and corn had higher percent lean values (P=0.02) (Landblom and Poland 1998).

Despite the negative perceptions of fat, intramuscular fat content influences the tenderness and juiciness of meat (Szabo et al.1999). This is desirable in terms of consumer appeal. Using peas (250 or 140g kg^-1) in an organic pig diet resulted in a positive dramatic doubling of intramuscular fat without increasing fat area or backfat thickness. In accordance with organic certification this diet was not supplemented with synthetic amino acids, yet was isocaloric to the control. A reduction in feed intake and accompanying performance was seen, although the lower supply of amino acids resulted in more gradual growth of pigs with the accompanying desirable effect on meat quality (Sundrum et al.2000).

Variations in carcass quality seen from feeding peas, may be the result of genotypic differences within pigs. Bahelka and Fl'ak (2000) fed pigs of three different genotypes, diets containing either SBM (120g kg^-1) or a pea-rapeseed blend (100g kg^-1 and 25g kg^-1) and observed differing performance and carcass quality without one diet consistently being superior to another.

The versatility of peas as a feed ingredient is evidenced by the range of diets into which they can be successfully incorporated. From weanling pigs to sows, pea inclusion in the diet results in performance equal to or better than a traditional SBM diet. Even at elevated levels of inclusion (670g kg^-1) feed intakes, weight gain and feed conversion remain unaffected. Carcass composition too is unaffected. These positive performance results, coupled with increasing pea availability make peas a desirable and economical alternative that producers can utilize with confidence.