Knowing your oats.

Abstract

211 Knowing your oats J.B. Rowe1, P.J. May2 and G. Crosbie3 1 2 3 Animal Science, University of New England, Armidale NSW 2351 Extensive Agriculture Branch, Tasmania Depar tment of Primary Industries, Water and Environment, PO Box 46, Kings Meadows Tas 7249 Agriculture Western Australia, Baron_Hay Cour t, South Per th WA 6151 jrowe@metz.une.edu.au Summary Although oats is widely used as a feed grain for ruminants and horses its value is often discounted because of the variability in its nutritional value. This paper investigates the major factors leading to variable digestibility and animal performance when oat grain is fed to sheep and cattle. Commercial samples of oat grain can contain significant quantities of free groats or free hulls and it is necessary to first to determine digestibility of the grain based on the `cleanness' of the grain and its general appearance. This is best done by measuring the content of insoluble non starch polysaccharide (NSP) or acid detergent fibre (ADF) and using this to predict digestible ener gy (DE). There are a number of calibrations for this prediction using wet chemistry or near infrared reflectance spectrometry (NIR). The second factor influencing DE is the lignin content of the hulls. The lignin content is genetically determined and it is possible to select high_ or low_lignin cultivars with a degree of confidence such that, all other factors being similar, low_lignin oat grain will have a DE value approximately 1.5 MJ/kg dry matter higher than the equivalent high_lignin oats. The third factor that influences DE of oat grain is the level of intake. When oat grain forms the major component of a production diet with intake approximately 2% of live weight, the DE value will be approximately 0.9 MJ/kg dry matter less than that measured at a maintenance level of feeding (approximately 1% of live weight per day). When oats constitute the sole component of the diet, feed intake may be significantly higher for animals fed high_lignin oats than low_lignin oats. On the other hand animals fed low_lignin oats compensate for reduced feed intake by having a better feed conversion efficiency and less gut fill (higher dressing percentage) than animals fed high_lignin oat grain. It is recommended that producers purchasing oat grain for livestock feeding should be prepared to pay a premium for low_lignin oats compared to high_lignin oats. There do not appear to be any detrimental effects on plant growth or disease resistance associated with hulls with low lignin and it is suggested that oat breeders should select cultivars with that attribute. Key words: oats, lignin, digestible energy, ruminants Introduction Oat grain has long been recognised as an excellent ingredient in the diet of humans and animals. Shaw (1907) declares that 'viewed from the standpoint of adaptation for feeding live stock no cereal grain grown in [the USA] compares with the oat'. Oats has been used extensively as a human food since cereals were first cultivated some 10000 years ago, and are considered to contain a very good balance of amino acids, oils and minerals as well as highly digestible starch (Sch�rch 1989). Even the high non_starch polysaccharide (NSP) content of oat grain is considered to be a benefit in human nutrition in assisting in the prevention of constipation and diverticular disease. In human nutrition only the groat is used and there appear to be no concerns about variability of nutritional value. On the other hand whole grain is fed to ruminants and horses and its variable nutritional value is the main reason it is discounted against most other grains considered to be of more consistent quality. Variability in the nutritional value of oat grain has been the subject of numerous studies (e.g. Pickering et al. 1982; Crosbie et al. 1985; Margan et al. 1987; Rowe and Crosbie 1987; Oddy et al. 1990) which show clearly that there are a number of factors to be considered if we are to accurately predict the performance of animals given this feed. This paper concentrates on the nutritional value of whole oat grain for ruminants. The purpose is to draw together results from a number of studies both published and unpublished to provide an overview of the most important factors influencing its nutritive value. Recent Advances in Animal Nutrition in Australia, Volume 13 (2001) 212 Rowe et al. Factors that determine the nutritive value of whole oat grain Digestible energy (MJ/kg DM) 18 17 Ratio of groats to non_groat material The dominant factor determining nutritional value of any particular sample of oat grain is the ratio of groats to non_groat material. The non_groat fraction consists mainly of the fibrous hull but can also include variable amounts of other material such as weeds and seeds from other plants as well as chaffed head, leaf and stem material not separated from the grain in the harvesting process. Samples of oat grain can also contain significant quantities of free groats if the hulls and groats are separated during harvesting and when this occurs the nutritive value of the sample can be improved considerably. The amount of non_groat material and the presence of free groats in an oat sample can be seen easily, but unfortunately it is not always accounted for in ascribing variability to oat grain quality. Increasing amounts of non_groat material tend to decrease the density of an oat sample and is almost certain to decrease its nutritional value (Pickering et al. 1982). For this reason oat grain is sometimes traded on the basis of hectolitre or bushel weight. Oddy et al. (1990) reported that bulk density measured as kg/hL or as weight per hundred grains was poorly correlated with acid detergent fibre (ADF) and these authors suggested bulk density is of little or no value in predicting oat grain quality. This conclusion regarding hectolitre weight, and therefore groat to hull ratio, is based on a range of diets for sheep constructed from oat fractions to match the extreme range of chemical compositions of oats grain found `in the field' (Oddy et al. 1990) and so overlooks the value of grain density as a predictor of groat to hull ratio in clean samples of oats. The study of Oddy et al . (1990) with the constructed diets shows the importance of accounting for the variable hull to groat ratio. Although the relationship with digestible energy is very good (Figure 1) it is important to realise that the most digestible `oat' diet in this study was in fact pure groats and that the least digestible `oat' contained approximately 75% oats and 25% additional oat hulls. Therefore while this is a very important predictive relationship when buying or selling oats with either a lot of free groats or an unusually high proportion of non_groat material, it does not necessarily provide an accurate description of differences among `normal' oats with only small contents of free groats or non_groat material. The problem of using ADF to predict the digestibility of `normal' oats is shown in Figure 2. The data have been taken from three experiments with sheep in which the digestibility of different oat samples were measured; two samples were the true oat grain samples described by Oddy et al. (1990), eight values come from the data of (Margan et al. 1987) and two come from Rowe and Crosbie (1988). The challenge is to understand the causes of the variability shown. 16 15 14 y = - 0.021x + 18.9 R2 = 0.94 13 12 0 50 100 150 200 250 300 Acid de te r ge nt fibr e (g/k g) Figure 1 Relationship between ADF and digestible energy of oat diets measured in sheep feeding experiments as repor ted by Oddy et al. (1990). 18 17 16 15 14 y = -0.012x + 16.7 13 12 50 100 150 Acid de te r g e nt fibr e (g/k g) 200 250 R2 = 0.18 Figure 2 Relationship between ADF and digestible energy of oat grain without added groats or hulls. Two values are the grain samples with 180 g ADF/kg and 25 g ADF/kg of Oddy et al. (1990), eight are taken from the data of Margan et al. (1987) and two from Rowe and Crosbie (1988). Hull lignin as a factor influencing digestibility Crosbie et al . (1985) reported that there was considerable variation in hull lignin content between different cultivars of oat grain and that these differences are predominantly genetically controlled. Most cultivars were found to be either `high' lignin with around 3% lignin in the whole grain and around 6_10 % in the hull fraction, or `low' lignin with around 1% lignin in the whole grain and 1_3% in the hull. This discovery provided a possible explanation for the experience of many livestock managers that certain cultivars of oat grain were better for livestock production than others. Rowe and Crosbie (1988) then showed that the differences in hull lignin content in high_ and low_lignin oat cultivars had a very significant effect not only on the digestibility of the hulls but of the whole grain. With Digestible energy (MJ/kg DM) Knowing your oats 213 ADF, the ratio of hulls to groats, and the concentrations of protein and ash all held constant between two cultivars, Rowe and Crosbie (1988) found that a difference of approximately 20 g lignin per kg grain resulted in an improved digestibility of the order of 15%. An independent study by Margan et al. (1987) reported significant differences in digestibility between the grains from two cultivars, Coolabah and Cooba. They also differed in lignin content, though that was not the reason they were chosen for the study, and so their results add considerably to appreciation of the importance of hull lignin content. Table 1 summarises the effect of lignin content on digestibility of oat grain in sheep fed principally oats at a rate of approximately 1.6% of body weight per day. While the ADF content of grains of the high_lignin cultivars is slightly higher, this difference would only explain approximately one third of the change in DE with ADF indicated by the predictive equation shown in Figure 1. Level of feeding With all feeds it is accepted that as the level of feeding increases there is a slight decrease in the efficiency of digestion, and DE/kg of feed consumed is reduced. For most feeds the decrease in digestibility with increasing level of feed intake is relatively minor. However, in the case of whole oats, level of intake appears to be an important factor as shown in Figure 3 which uses data from Margan et al. (1987). Combining hull lignin and level of intake with ADF to predict DE When we account for both level of feed intake and lignin content of oat hulls in the prediction of digestible energy of oats grain we are able to explain much of the variability shown in Figure 2. Figure 4 illustrates the same data as those used in Figure 2, but level of feed intake has replaced ADF as the independent variable Table 1 Analysis of the major components (g/kg dr y matter) of four samples of oat grain and the digestible energy (DE, MJ/kg dr y matter) obtained with sheep fed these diets at rates equivalent to approximately 1.6% of body weight per day. Data for `Cooba' and `Coolabah' were der ived from the study of Margan et al. (1987) and for `Murray' and `Mor tlock' from Rowe and Crosbie (1988). Cultivar Low_lignin Murray Cooba High_lignin Mor tlock Coolabah Average difference between high_ and low_lignin cultivars 17.5 17 y = - 0.85x + 16.9 R2 = 0.95 Lignin ADF DE 8 10 133 110 15.6 15.4 23 30 144 150 14.0 13.5 25.5 1.8 Digestible energy (MJ/kg dry matter) 16 Digestible ener gy (k g/MJ DM) 16 15 y = -0.85x + 15.6 R2 = 0.81 15 14 14 13 13 y = - 0.87x + 15.6 R2 = 0.67 12 0 0.5 1 1.5 2 2.5 3 3.5 Level of intake (g/100 g body w eight) 12 0 0.5 1 1.5 2 2.5 3 3.5 In tak e (g o at s /100 g bo dyw e ig h t) Figure 3 Decrease in digestible energy value of oats grain with increasing intakes by sheep. Data are for Coolabah oats as repor ted by Margan et al. (1987). Figure 4 Prediction of digestible energy of oat grain taking into account level of feeding and grains as high_ lignin (F) or low_lignin (Q). As for Figure 2, eight values (plain triangles and circles) are taken from the data of Margan et al. (1987), two (closed triangles with a single horizontal line) are from Oddy et al (1990), and two (triangles and circles with an X) are from Rowe and Crosbie (1988). 214 Rowe et al. and grains have been separated on the basis of their lignin content. The two data points representing 18% ADF and 25% ADF oats are shown as the filled triangles each with a single line through them; that they fall either side of the line of best fit suggests that the prediction of DE based on feed intake and lignin could be further improved by considering ADF as a measure of groat to non_groat material. A robust method of predicting DE would be to use the predictive equation based on the results of Oddy et al. (1990), add 1.5 MJ/kg if the grain is low lignin, and then adjust DE for level of intake by 0.86 MJ/kg for each 1% of body weight consumed above or below the 1% base level used by Oddy et al. (1990). Oat grain in feeding for production The conclusions of the preceding section suggest it would be logical to expect that an oat cultivar with a low content of lignin in the hull would promote better liveweight gain and feed conversion efficiency than a grain with higher lignin. Not only would there be benefits from improved digestibility and nutrient availability, but it is generally accepted that feed intake by ruminants increases with increasing feed digestibility. We conducted two experiments to test the hypothesis that low hull lignin would have a beneficial effect on feed intake, liveweight gain and feed utilisation in sheep and cattle fed diets with oat grain as the major component of their diet. Two other factors considered in the design of the experiment were protein content of the grain and acid insoluble ash content of the hulls. An earlier study in Western Australia (J.L. Suiter, pers. comm.) had shown a positive relationship between protein content of oat grain and liveweight gain of sheep fed solely on the grain. Acid insoluble ash was included because the study of Crosbie et al. (1985) had found a negative correlation between acid insoluble ash and pepsin_cellulase digestibility of oat hulls. Sheep production experiment Samples of grain from the cultivars Mortlock (high lignin) and Murray (low lignin) were collected from commercial growers throughout the south western land division of Western Australia. There were approximately 30 samples, each weighing around one tonne, and measurements were made of free groats and hulls and total nitrogen in each of these. Hulls were separated manually for measurement of lignin and acid insoluble ash. Based on these measurements composite diets were formulated to create five levels of protein and five levels of acid insoluble ash for both the high lignin and low lignin cultivar, giving a total of 10 diets. The groat to hull ratios were almost identical for each diet. In order to evaluate the inherent importance of grain protein without N being a limiting nutrient for microbial fermentation we added non_protein N as urea and ammonium sulphate (9:1) to each diet to achieve a constant level of 14% crude protein. All diets also contained 1% limestone to ensure that Ca availability did not limit performance. Each diet was fed ad libitum for a period of 10 weeks to six individually housed Merino sheep. The weight of the sheep at the start of the experiment was 31 kg. Measurements were made of daily feed intake and animals were weighed each week. Mid_side patches (10 cm x 10 cm) were clipped after three weeks on the diets and again at the end of the study. The amount of wool grown between weeks 3 and 10 was used to estimate daily growth of clean wool. The feed intakes and liveweight changes of sheep on the 10 diets are summarized in Figure 5. It is clear that there was no effect of either protein content of the grain or acid insoluble ash in the hull on feed intake or liveweight gain. There were however significant effects due to the lignin content of the hull on both feed intake and liveweight gain that were the reverse of what was expected at the outset of the experiment. The intake of high_lignin, Mortlock grain was higher than that of the low_lignin Murray cultivar. Similarly liveweight gain of sheep consuming the high_lignin grain was better than those on the low_lignin grain. The results for feed intake, weight gain and wool growth with respect to lignin content are summarized in Table 2. It is clear that the higher intake of high_lignin grain was the major factor affecting liveweight gain and wool growth. The higher digestibility of the low_lignin Murray grain did not compensate in terms of liveweight gain. The reduced intake of low_lignin grain compared to the high_lignin cultivar was unexpected. It is, however, consistent with the results of Margan et al. (1988) who reported higher intakes by sheep offered the high_lignin Coolabah cultivar than those offered the low_lignin Cooba cultivar. Since the same phenomenon has been observed for two different pairs of high_ and low_lignin cultivars in two independent studies it appears that this is more than a ` random ' factor associated with the Murray low_lignin cultivar but may be a general factor. One possible reason for this phenomenon could be that low lignin oats contain higher levels of non_lignin phenolics than the high_lignin oats. This is speculative, but it is consistent with the observation that most low_lignin oats have a slightly darker hull colour than the high_lignin cultivars and it is known that a number of phenolic compounds are pigmented. Another possibility is that the additional digestible energy available to the animals fed low_lignin oats created a demand for extra nutrients such as amino acids at the tissue level and that an imbalance of nutrients caused the reduction in feed intake. This hypothesis was tested in a preliminary experiment in which additional protein in the form of fishmeal was provided to sheep on the low_ and high_lignin diets (J.B. Rowe and G.B. Crosbie, unpublished). While there was a slight increase in intake in response to the additional protein the differences between Mortlock and Murray were not reversed. Knowing your oats 215 Cattle production and carcass yield in cattle fed oat grain A feeding experiment was conducted in cattle (May et al. 1989) using the same low_ and high_lignin oat cultivars, Murray and Mortlock respectively, as used in the sheep feeding study described above This study (Table 3) demonstrated a significant effect of high_lignin oats on gut fill and dressing percentage, thus explaining the apparent paradox of a feed with lower digestibility producing better live weight gain and feed conversion efficiency than a similar feed of higher digestibility. As in the sheep experiment, feed intake was higher (P = 0.07) in cattle fed Mortlock than for those given the higher digestibility Murray grain. Liveweight gain and feed conversion based on liveweight were also significantly better. However, due to the accumulation of low digestibility roughage in the rumen and a much larger rumen in cattle fed high_lignin Mortlock grain, carcass gains for the two grains were similar. Feed conversion based on carcass weight change was marginally (P = 0.09) better for the low_lignin cultivar and this is consistent with its improved digestibility. Oat lignin content, gut fill and carcass yield in sheep Following the above study in cattle we designed an experiment in sheep to determine the effect of hull lignin content on gut fill and carcass weight change to determine if the reduced intake of a low_lignin oat cultivar was compensated for by increased digestibility. In this experiment (Rowe and Coss 1992) sheep of approximately 33 kg liveweight were offered either 1 kg/d of chaffed wheat hay or 800 g/d of a high_ or low_lignin oat grain (Mortlock or Murray, respectively). Both the chaffed hay and oat grain contained 10 g/t of urea/ammonium sulphate (respectively 9:1 w/w). Twenty sheep had received each diet for three weeks when 10 sheep from each group were slaughtered and carcass and reticulo_rumen weights were measured. The remaining 10 animals were slaughtered and similar Table 2 Intakes and liveweight gains of sheep fed either Mor tlock (high_lignin) or Murray (low_lignin) oat grain balanced for protein, insoluble ash and free groats. Mortlock Intake (g/d) Weight gain (g/d) Wool growth (mg/cm2/d) 759 123 0.81 Murray 599 87 0.66 SE P<0.05 20 8 0.03 * * * 1200 Fe e d intak e (g/d) Fe e d intak e (g/d) 1000 800 600 400 200 0 8 200 Live w e ight gain (g/d) 10 12 1200 1000 800 600 400 200 0 3 200 Live w e ight gain (g/d) 150 100 50 0 8 10 12 3 4 5 6 Ins oluble as h conte nt (%) 4 5 6 150 100 50 0 Pr ote in conte nt (%) Figure 5 Feed intake and liveweight gain of Mer ino sheep fed low_lignin oat grain (cultivar Murray, Q) or high_lignin (cultivar Mortlock, b) with var ying levels of protein in the whole grain and var ying concentrations of acid insoluble ash in the hulls. 216 Rowe et al. measurements were made 7 weeks later. The results are summarized in Table 4. The results of this experiment confirm the higher feed intake by the animals fed the lower digestibility high_lignin oat grain that was observed in previous experiments. Measurements of carcass weight and the reticulo_rumen also confirm the results of the cattle experiment showing that high_lignin oat grain produces a significant increase in gut fill that masks the reduced rate of carcass gain compared to low_lignin oats. In this study the higher digestibility of the low_lignin grain produced a higher rate of carcass gain than the high lignin grain even although approximately 20% less feed was consumed. The study supports the conclusion that low_lignin oat grain provides a higher digestible energy per kg dry matter than the high_lignin, and that animals consume more of the high_lignin grain. Oat grain as a supplementary feed The effects of high_ and low_lignin oat cultivars summarized above apply to situations where oat grain comprises the major part of the diet with only small amounts of minerals added to balance the diet. This is rarely the case in practical feeding applications where oats is more commonly used as a supplement for grazing animals with access to poor quality paddock roughage. Alternatively oats are fed at restricted levels to provide a maintenance diet for animals during an extended dry period when there is insufficient paddock roughage available. Under conditions of supplementary feeding it is considered unlikely that there would be any reduction in feed intake associated with the use of low_ lignin oat cultivars and we conducted a further experiment to test this hypothesis. Table 3 Diet composition (g/kg dr y matter) and perfor mance of cattle fed oat grain with high (Mor tlock cultivar) or low (Murray cultivar) lignin contents in the hulls (from May et al. 1989a). Mor tlock Murray Significance of difference (P) Crude protein Acid detergent fibre Lignin 124 124 24 110 104 11 In vivo dry matter digestibility (%) Measured in sheep (maintenance) Measured in cattle (intake approx. 2.2% of body weight) Days on feed Feed intake (kg/d) Relative dry matter intake (% of live weight) Average live weight gain (kg/d) Dressing % Rumen contents (% of live weight) Average carcass gain (kg/d) Feed conversion live weight (kg feed/kg live weight gain) Feed conversion carcass (kg feed/kg carcass weight gain) 70.2 64.1 107 6.34 2.29 1.06 49.8 1.79 0.58 6.05 11.15 82.4 71.6 104 5.86 2.20 0.90 51.8 0.57 0.58 6.60 10.19 0.03 0.09 0.07 NS 0.003 0.003 0.001 Table 4 Feed intake, liveweight and carcass weight changes in sheep fed chaff or oat grain with high (Mor tlock cultivar) or low (Murray cultivar) lignin contents in the hulls (from Rowe and Coss 1992). Chaff Mean Initial liveweight (kg) Average feed intake (kg/d) Live weight change (kg) Carcass weight (kg) Carcass weight change# (kg) Reticulo_rumen (kg) 33.0 0.913 1.56 13.6 1.3 4.83 SE 0.58 0.015 0.31 0.34 0.16 0.21 Mortlock Mean 34.1 0.664 2.63 14.1 1.44 4.70 Murray Mean 32.7 0.547 3.00 14.6 2.74 3.71 SEM 0.8 0.044 0.57 0.28 0.29 0.31 P NS 0.001 NS NS 0.01 0.01 # Carcass weight change measured between weeks 3 and 10. NB: feed intake was used as a covariate in the analysis P Indicates significance of differences between the two oat grains Knowing your oats 217 In this experiment, described in more detail by Rowe and Coss (1994), high and low_lignin oats were fed to Merino sheep at the three rates of 150, 300 and 450 g/day in addition to chaffed hay that was available ad libitum. There were l0 sheep per treatment, and a further 20 sheep were fed chaff only; average liveweight at the start of the experiment was 38 kg. Feed intake was measured daily and all sheep were weighed weekly. At the end of 7 weeks on the experimental diets all sheep were slaughtered and measurements were made of dressing percent and gut contents. The results are shown in Figure 6. Only the level of grain feeding had a significant (P<0.001) negative effect on the intake of chaff; there was no significant effect due to oat cultivar. While there was a significant (P<0.05) increase in dressing percent with increasing intake of low_lignin oats there was no increase with increasing intakes of the high_lignin oats. grinding and mixing equipment are available. We are aware of three studies examining this question and the results are summarized in the Table 5. The consistent finding in all three studies was that there is little or no improvement in digestibility or animal performance in response to processing oat grain by rolling or hammer milling. Toland (1976) reported improvements of between 60 and 100% in the digestibility of wheat and barley starch as a result of rolling compared to feeding whole grain, but found no significant benefits in the case of oats. It is therefore safe to conclude that oat grain can be fed whole to cattle without any risk that it will be digested or utilized inefficiently for production. The data of May et al. (1989b) even suggest possible benefits in whole grain feeding to achieve slightly higher consumption. As is the case with all cereal grains, there is no benefit in rolling or grinding oats before feeding it to sheep. Processing oat grain for animal feeding The question of whether or not there is any advantage to be gained from processing oat grain before feeding it to cattle is a very important one in determining its suitability for on_farm use in situations where no Using oat hulls in animal feeds The preparation of groats for the human and poultry markets produces quantities of hulls as a by_product that are often included as a source of roughage in mixed feeds for ruminants. Round (1988) investigated the nutritional characteristics of oat hulls on their own and in pelleted diets for sheep. In measurements of hull digestibility he reported that one source of hulls was very much more digestible (47.7% digestibility of dry matter) than hulls from three other commercial sources (average 37% digestible). While there is no record of cultivar or lignin content for these sources of oat hulls it is possible that the hulls of higher digestibility came from cultivars with low hull lignin. An alternative explanation could be that the higher digestibility hulls contained more grain fragments than the hulls of lower digestibility. The data of Rowe and Crosbie (1988) summarized in Figure 7 suggest that high_lignin hulls are almost indigestible; less than 10% of their dry matter disappeared from nylon bags incubated in the rumen for 96 h compared with low lignin hulls of which almost 30% disappeared in 96 h. Although studies by Crosbie et al. (1987), Rowe and Crosbie (1988) and Crosbie and Rowe (1988) have suggested the digestibility of oat hulls is determined predominantly by lignin content, with a minor effect of silica (insoluble ash), the studies reported by Welch et al. (1983) and of A.G. Kaiser (pers. comm.) indicate that factors other than lignin may have an important influence on digestibility. We have examined the data of Welch et al. (1983) to try to explain why these authors did not find lignin to be as important as in the studies of Crosbie and his colleagues. Welch et al. (1983) examined the effects of various components of the oat hull on digestibility measured using the pepsin_cellulase method, and although these authors reported a correlation coefficient of lignin and pepsin_cellulase digestibility of _0.73 for 11 cultivars, the correlation for all samples studied 900 850 800 Ch aff in t ak e (g/d ) 750 700 650 600 550 500 450 400 0 100 200 300 400 500 Gr ain s upple m e nt (g/d ) 43 Dressed carcase (% of live w e ight) 42 41 40 39 38 0 100 200 300 400 500 Grain s upple me nt (g/d) Figure 6 Chaff intake and dressing percent of young Merino sheep fed chaffed wheat hay ad libitum and var ying amounts of oat grain of low lignin content (open circles) or high lignin content (closed circles). 218 Rowe et al. was less than _0.5 (R2 = 0.23, see Figure 8a). However a closer examination of their data shows a good relationship (R2 = 0.75) between hull protein and hull starch suggesting that the groats and hulls were not cleanly separated (see Figure 8b). This point is further illustrated in Figure 8c which shows a good relationship between starch content of the hulls and pepsin_cellulase digestibility (R2 = 0.60). The relationship between starch content and digestibility has one data point off the line of best fit and it is interesting that this point represents the only sample of hulls in that study with a low lignin content (less than one_third the mean value of all samples). A possible conclusion is that, apart from the one sample of low lignin oat hulls, most of the variation in digestibility observed by Welch et al. (1983) could be explained by the incomplete separation of hulls and groats. In commercially prepared samples of oat hulls, variable and incomplete separation from groats has an important effect on pepsin_cellulase digestibility. Table 6 summarises data for samples of oat hulls separated by hand and hulls from the same grain separated in commercial milling operations. The amount of starch in the hull fraction is an indicator of fragments of groat remaining in the hulls and of complete grains passing through in the hull fraction. The data of Round (1988) suggest that most modern milling operations produce a clean hull fraction with little starch and that under these conditions the characteristics determining hull digestibility (principally lignin) become even more important. Measuring lignin content of oat hulls Measurement of lignin content by wet chemistry is time consuming and expensive, and if the analysis is not made by an experienced technician it can yield uncertain results. Probably due to the painstaking task of measuring large numbers of lignin concentrations, combined with the difficult process of hand separation of hulls from groats, there are as yet no NIR prediction equations for hull lignin. Although not quantitative, a test developed by Crosbie and his co_workers provides a quick and accurate indication of whether samples of oat grain fall into the category of `high_lignin' or of `low_lignin'. The Crosbie oat hull assay is based on the pink colour that develops when a solution of phloroglucinol stains the lignin component of the hulls, and is most easily made by adding approximately 2 ml of a phloroglucinol solution to about 10 whole oat grains or to separated hulls from 3 to 5 grains. The solution is prepared by diss