Availability to ruminants of nitrogen in senesced C₄ tropical grasses

Keywords: acid detergent insoluble N, C₄ grasses, forage digestibility, forage N, low quality forages, N-deficiency, ruminants, tropical forages.

There has been extensive investigation of ruminant digestion of the total nitrogen (TN) fractions in forages to understand their breakdown and use as substrates by rumen microbes and, following passage of some forage N to the post-ruminal tract, whether various fractions are digested or are excreted in faeces (Webster 1987; AFRC Technical Committee on Responses to Nutrients 1992). This has often been investigated through use of surgically prepared animals to partition the sites of digestion, and synthetic fibre bags to measure digestion in the rumen and post-ruminally. A laboratory approach to estimating the amounts and proportions of the TN fractions in forages has been used to determine the solubility of TN under various specified conditions (Sniffen et al. 1992; Licitra et al. 1996). There is consensus that the forage N that is insoluble in acid detergent fibre (ADF) solution (i.e. ADIN) represents the TN that is not digested in the rumen or post-ruminally, and hence provides a reliable measure of the TN that is indigestible and not available to the ruminant (Van Soest 1982; Broderick 1994). It is generally agreed that ∼90% of the TN consumed in forages is digested and thus available in ruminants (Webster et al. 1982; AFRC Technical Committee on Responses to Nutrients 1992; Sniffen et al. 1992). These approaches have, with some variations, been incorporated into evaluations of the nutritional value of feeds and feeding standards (AFRC Technical Committee on Responses to Nutrients 1992; CSIRO 2007; National Academies of Sciences, Engineering and Medicine 2016).

Investigation of the digestion of forage TN fractions has focused on improved and sown species of grasses and legumes during vegetative growth, or soon after physiological maturity, when forages are usually grazed or harvested in higher input ruminant production systems (Wilson and Strachan 1981; Waters et al. 1992; Wales et al. 1999; National Academies of Sciences, Engineering and Medicine 2016). Relatively little information is available on the digestion of TN fractions in mature, senesced and weathered grasses, particularly C₄ grasses, that usually have low TN concentrations and are highly fibrous. Nevertheless, this is important in some low-input rangeland systems, and crop–livestock systems in low-income countries, where ruminants are highly dependent on senesced C₄ grasses and stovers of low nutritional value as feeds during some seasons of the annual cycle. Also, there is evidence that the N availability in at least some C₄ grass hays is lower than in C₃ grass hays of comparable low nutritional value (Hogan et al. 1989; Hogan 1996). Because senesced forages, particularly C₄ grasses, are usually marginal or deficient in TN for ruminants, it is important to understand the availability of this TN to the ruminant.

The present study examined the concentrations of TN, N insoluble in neutral detergent fibre (NDF) solution (i.e. NDIN), ADIN, and fibre components in C₄ grasses, as well as in some C₃ grasses and legumes, grown in subtropical or tropical environments of northern Australia. Forages sampled from a wide range of seasons and sites were used to examine the hypothesis that a substantial proportion of the TN in low-quality tropical C₄ grasses is present as indigestible ADIN. In addition, the study collated published measurements of TN and ADIN in C₄ grasses grown in the subtropics or tropics.

The forages analysed (n = 194) were obtained from previous experiments. Most (n = 184) had been fed to cattle in individual pens during the development of faecal NIRS calibration equations (Coates and Dixon 2011) or in other, later pen experiments (Kennedy and Charmley 2012; McLennan 2014; RM Dixon and DB Coates, unpublished data) where both the forage and faeces were sampled. The forages were all mechanically harvested from pastures growing in the subtropics or tropics. The forages were considered as classes comprising native or improved species of C₄ grasses, improved species of C₃ grasses, tropical legumes and temperate legumes (Table 1). Subclasses were forage species represented by ≥10 samples, or combinations of several species when there were fewer samples. The native species of C₄ grass forages were generally cut with a forage harvester from swards of native pastures and classified into three subclasses: (1) dominated by Heteropogon contortus (black speargrass), (2) dominated by Astrebla spp. (Mitchell grasses), or (3) ‘mixed native C₄ grasses’. The mixtures of species included H. contortus, Bothriochloa pertusa (Indian couch) and Chrysopogon fallax (golden beard grass), and sometimes Urochloa mosambicensis that had invaded native grass paddocks. Hays of the introduced, sown C₄ grass species Cenchrus ciliaris and Chloris gayana were obtained from commercial farms, and the U. mosambicensis was harvested from swards dominated by this species. The subclass ‘Other improved C₄ grasses’ combined such grasses where there were <10 samples of a species; these included Pennisetum spp. (millets), Setaria spp., Brachiaria mutica (para grass), Cynodon dactylon (couch) and Stenotaphrum secundatum (buffalo couch). The C₃ grass forages were the cereal crop species Avena sativa (oats, n = 6), Triticum spp. (wheat, n = 3) and Hordeum vulgare (barley; n = 1), all in the vegetative growth stage. The tropical legume Stylosanthes included both Stylosanthes scabra and S. hamata, and ‘Other herbaceous legumes’ included Lablab purpureus (dolichos), Neonotonia wightii (glycine), Centrosema pascuorum (cavalcade), Clitoria ternatea (butterfly pea), Macroptilium bracteatum (burgundy bean) and Arachis pintoi (peanut). The temperate legume was Medicago sativa (lucerne). The C₃ grasses and the legumes had been cut as hays on commercial farms.

Table 1.  Dataset 1: composition (mean, s.d. in parentheses) of the forages (n = 194) in the sample set (subclass) of each species, or group when there were <10 samples of a species.

The forage samples were analysed for TN by Kjeldahl analyses. NDF and ADF contents, and NDIN and ADIN, were analysed as described by Pichard and van Soest (1977) and Licitra et al. (1996). The dry matter digestibility (DMD) of the forages was measured by near-infrared reflectance spectroscopy of faeces (F.NIRS) as described by Coates and Dixon (2011).

Data were collated from published papers reporting measurements of the TN and ADIN fractions in C₄ grasses, or mixtures containing predominantly C₄ grasses (n = 187), grown in the subtropics or tropics. Where a study also reported measurements in C₃ grasses (n = 45) or tropical herbaceous legumes expected not to contain tannins (n = 51) and grown in the same environment, those results were included. Measurements of wheat or rice straws as senesced C₃ grasses in the subtropics were also included for comparison. One additional study with C₄ grasses (Buckner et al. 2013), in which ‘total tract indigestible protein’ was measured using disappearance from synthetic fibre bags incubated in the rumen and during passage through the post-ruminal gastrointestinal tract, was also considered. Some of the papers also reported ADF, NDF and NDIN concentrations (105, 120 and 77 of the 187 C₄ grass forage samples, respectively). Samples from treatments within experiments where the pasture had been fertilised with N were excluded. The papers had been published by research groups in North America (n = 11), South America (n = 4), India (n = 7), Africa (n = 4), and elsewhere (n = 5), and often the C₄ grass forages had been measured at various maturity stages during the summer growing season. Genera most often reported and the number of samples were Cynodon (31), Cenchrus (21), Panicum (20), Andropogon (16), Digitaria (12), Pennisetum (11) Brachiaria, (9), Zea mays (8), Eragrostis (7), Paspalum (8), Setaria (5) and Sorghum (5). The dataset included one experiment where C. dactylon pasture had been ‘stockpiled’ after physiological maturity and then sampled several times over 3 months (Scarbrough et al. 2002), and another experiment with hay made from mature Andropogon spp. (‘bluestem range mixed forage’) (Hannah et al. 1991). One paper reported the changes in two C₄ and two C₃ grass species in the same environment through the summer growth period (Mitchell et al. 1997). Another study reported measurements in a range of grass species during the dry season in a southern African semi-arid environment (Gemeda and Hassen 2014). Nine papers included herbaceous tropical legumes (23 species and 51 samples). The forages were considered as groups comprising C₄ grasses, C₃ grasses, tropical legumes and temperate legumes.

Each class of forage within each dataset was divided into six categories of ADIN concentration (<1.00, 1.00–1.49, 1.50–1.99, 2.00–2.99, 3.00–3.99 and >4.00 g/kg DM) and six categories of ADIN/TN (<100, 100–149, 150–199, 200–299, 300–399 and >400 g/kg). In Dataset 1, the attributes of the various subclasses of forages were compared by one way analysis of variance. The relationships between the variables were examined by regression. The statistical package used was Genstat Release 16.1 (VSN International, Hemel Hempstead, UK).

The measured forage attributes generally did not differ (P > 0.05) among subclasses within the C₄ native and C₄ introduced grass species, or between these classes, and there was no consistent pattern (Table 1). Overall, the C₄ grasses (n = 143) averaged 9.7 g TN/kg DM, 1.45 g ADIN/kg DM, 170 g ADIN/kg TN, and (n = 139) 526 g digestible DM/kg, but each of these attributes had large variation and range. The C₃ grass and both legume classes were generally much higher in TN and ADIN concentrations, but they had lower ADIN/TN ratios, than the C₄ grasses. However, due to the small number of samples and few species, the C₃ grasses and temperate legumes did not broadly represent these classes of forages. The tropical legumes had high concentrations of TN and ADIN. The distributions of the categories of ADIN and ADIN/TN ratio in the forage classes in Dataset 1 are shown in Fig. 1. Among all C₄ grasses, 44% contained ≥1.50 g ADIN/kg DM and 62% had a ratio of ≥150 g ADIN/kg TN; among C₄ grasses with >400 g ADF/kg DM, 46% contained ≥1.50 g ADIN/kg DM and 74% had a ratio of ≥150 g ADIN/kg TN.

Fig. 1.  Dataset 1: distribution of samples of forage classes (left to right): C₄ grasses (n = 143), C₃ grasses (n = 15), herbaceous tropical legumes (n = 25), and temperate legumes (n = 11) across categories of ADIN concentration, and categories of ADIN/TN ratio.

Many of the measured forage attributes were correlated (P < 0.01), but the correlation coefficients (r) were generally <0.7 and did not allow useful prediction of one variable from another (Table 2). TN was negatively correlated with NDF and ADF (r = −0.81 and −0.77), and was positively correlated with DMD (r = 0.85). ADIN concentration was not correlated with either NDF or ADF content (both P > 0.05), and was only poorly correlated with TN and NDIN (r = 0.42 and 0.65, respectively).

Table 2.  Dataset 1: correlation coefficients between the measured variables and ratios from all four classes of forages (n = 194), as well as dry matter digestibility (DMD) and faecal N concentration (n = 141).

Among the C₄ grass forages, the ADIN/TN ratio increased exponentially with increasing ADF content (Fig. 2a). Forages containing <400, 400–450, 450–500 and >500 g ADF/kg DM had, on average, ratios of 86, 132, 175 and 230 g ADIN/kg TN, respectively (Table 3). Conversely, increasing TN concentration or DMD was associated with decreasing ADIN/TN ratio (Fig. 2b, c). In forages containing <5 g TN/kg DM, the ADIN/TN ratio averaged 257 (s.d. 109) g/kg, and in forages containing 5–10 g TN/kg DM, the ADIN/TN ratio averaged 173 (s.d. 82) g/kg (Table 3). As the TN concentration in forages increased to 20 g TN/kg DM, the ADIN/TN ratio decreased to 154 and then 76 g/kg. There was very large variation in the ADIN/TN ratio within each of these categories, especially at low forage TN concentrations. Similar large variation and changes between categories were also observed for four categories of DMD (<500, 500–530, 530–560 and >560 g/kg) (Table 3). In summary, these relationships indicated that the ADIN/TN ratio was increasing, and thus availability of the TN to the ruminant was decreasing, with decreasing nutritional value of the C₄ grass forages. ADIN (g/kg DM) and ADIN/TN ratio (g/kg) could be estimated using multiple regression models:

and:

Fig. 2.  Dataset 1: relationships in C₄ grasses between ADIN/TN ratio and (a) ADF concentration, (b) total N concentration, and (c) diet DM digestibility (n = 143). Equations are: (a) Y = 31.5 + 3.23exp(1.00786X) (RMSEP = 67.9, R² = 0.43, P < 0.001); (b) Y = 132.5 + 952exp(−0.4874X) (RMSEP = 74.0, R² = 0.33, P < 0.001); (c) Y = −3.8 + 8836exp(−0.00755X) (RMSEP = 74.3, R² = 0.33, P < 0.001).

Table 3.  Datasets 1 and 2: the percentage distribution of categories of acid detergent fibre (ADF) content and total N (TN) concentration in C₄ grasses, and the category mean measured concentration of N insoluble in ADF solution (ADIN) and ADIN/TN ratio.

The distributions of the categories of ADIN concentration and ADIN/TN ratio in the various forage classes of Dataset 2 are shown in Fig. 3, and the ADIN concentrations and ADIN/TN ratios in Table 3. On average, the C₄ grasses (n = 187) contained 1.93 g ADIN/kg DM and had a ratio of 147 g ADIN/kg TN. The more fibrous grasses with >400 g ADF/kg DM contained, on average, 1.78 g ADIN/kg DM and had a ratio of 189 g ADIN/kg TN, with 55% recording a ratio of ≥150 g ADIN/kg TN. The C₄ grasses containing <400 g ADF/kg DM averaged 1.53 g ADIN/kg DM and 123 g ADIN/kg TN. By comparison, 43% of C₄ grasses with <10 g N/kg DM had a ratio of ≥150 g ADIN/kg TN. In the entire sample set, the concentrations of ADIN were comparable in C₄ grasses, C₃ grasses and tropical legumes (1.93, 1.78 and 2.63 g ADIN/kg DM, respectively), but because the TN concentration was usually lower in the C₄ grasses, the proportion of forages with ratios ≥150 g ADIN/kg TN was higher among C₄ grasses (41%) than C₃ grasses (19%) or tropical legumes (17%). In summary, the profiles of forage constituents in this dataset derived from published measurements by many research groups showed concentrations and trends similar to those observed in Dataset 1.

Fig. 3.  Dataset 2: distribution of samples of forage classes (left to right): C₄ grasses (n = 187), C₃ grasses (n = 45), herbaceous tropical legumes (n = 51), and temperate legumes (n = 22) across categories of ADIN concentration, and categories of ADIN/TN ratio. Sources of data were: Blasi et al. (1991); Bowen (2004); Brown and Pitman (1991), Buckner et al. (2013), Chaurasia et al. (2006), Coblentz et al. (2004), da Silva et al. (2013), Das et al. (2015), Erasmus et al. (1990), Fondevila et al. (2002), Foster et al. (2012), Gemeda and Hassen (2014), Gupta et al. (2011), Hannah et al. (1991), Johnson et al. (2001), Juarez Lagunes et al. (1999), Kabuga and Darko (1993), McLennan (1997), Mitchell et al. (1997), Mupangwa et al. (2003), Negi et al. (1988a, 1988b), Nogueira Filho et al. (2000), Ogden et al. (2006), Peripoli et al. (2016), Prakash et al. (2003), Salazar-Cubillas and Dickhoefer (2021), Scarbrough et al. (2002), Singh et al. (2002, 2012), Wang et al. (2015), Zhao and Cao (2004).

There is consensus that, in grass and herbaceous legume forages, ADIN represents the N that is bound to plant cell walls, is indigestible and not available to the ruminant, and that it typically comprises ∼10% of the TN (Van Soest 1982; Webster et al. 1988; AFRC Technical Committee on Responses to Nutrients 1992; Broderick 1994). Dataset 1 in the present study clearly showed that in low-quality, senesced C₄ grasses the ADIN/TN ratio was substantial, averaging 190 g ADIN/kg TN in those with >400 g ADF/kg, and similar ratios in those with ≤10 g TN/kg or ≤560 g digestible DM/kg. Importantly, the ADIN/TN ratio in these groups was highly variable, ranging from negligible up to 495 g ADIN/kg TN, and was more closely correlated with ADF than with TN or DMD (Table 2). This is in agreement with associations between ADIN concentration and ADF content reported by Webster et al. (1982), AFRC Technical Committee on Responses to Nutrients (1992) and Waters et al. (1992). The efficacy of low TN concentrations or low DMD for identifying samples with high dietary ADIN/TN ratio was likely due primarily to the associations among these attributes.

The samples collated from published results as Dataset 2 (Table 3) generally included higher quality C₄ grass forages than Dataset 1 (Table 3). This was most obviously because the forages in Dataset 2 were largely harvested during their vegetative growth rather than after senescence. However, among the C₄ grasses in Dataset 2 that contained >400 g ADF/kg or <10 g TN/kg, the ADIN/TN ratio was comparable to the same classes of C₄ grasses in Dataset 1.

The results are in accord with other evidence of low rumen availability of the TN in some low-quality C₄ grasses, from observations of expectedly low rumen ammonia concentrations in sheep fed such grasses (Hogan et al. 1989; Hogan 1996). In summary, there is considerable evidence to support the hypothesis that a substantial proportion of the TN in low-quality C₄ grasses is present as ADIN and therefore indigestible in ruminants.

From the present study and other published studies, it is not possible to separate the effects of factors such as genotype, growth environment and stage of growth of the grasses on ADIN concentration or ADIN/TN ratio. Both Blasi et al. (1991) and Redfearn et al. (1995) reported that during vegetative growth both ADIN and ADIN/TN were consistently higher in two C₄ grasses (Andropogon gerardii and Panicum virgatum) than in a C₃ grass (Bromus inermis), and the former study also found that these attributes could vary widely between years. However, contrasting results were reported by Mitchell et al. (1997), who studied two C₃ grasses and two C₄ grasses grown in the same environment during one summer. First, values of ADIN/kg DM and ADIN/TN were generally higher than reported in the two studies cited above, and also increased with plant maturity; at the greatest maturity examined, there was a range from 226 to 408 g ADIN/kg TN among these C₃ and C₄ grasses. Second, there were large differences among the grass species such that values of ADIN/kg DM and ADIN/TN ranked in the order: A. gerardii (C₄) > P. virgatum (C₄) ≈ B. inermis (C₃) > Thinopyrum intermedium (C₃). In this experiment the differences between the two grass species within each of the C₃ and C₄ classes could be as great as differences between C₄ and C₃ classes. There is evidence that higher ADIN concentration and ADIN/TN ratio occur with low soil fertility and/or unfavourable environmental conditions for grass growth; Johnson et al. (2001) reported that the ADIN/TN ratio was generally low (<120 g N/kg TN) in C₄ grasses (C. dactylon, Cynodon nlemfuensis and Paspalum notatum) grown in a favourable environment. Differences between C₃ and C₄ grasses and changes with increasing plant maturity may be associated with the different anatomical structures and greater proportions of parenchymal bundle sheaths and epidermis in both the leaf and stem of the C₄ grasses (Wilson 1994), and with the generally higher content of ADF in C₄ grasses (Laetsch 1974; Akin and Burdick 1975; Akin 1989). It is clearly important to understand the circumstances under which the ADIN/TN ratio is high and TN is low in C₄ grass forages, but further investigation is needed to understand the reasons for the variation in the ADIN/TN ratio.

Limitations of the present study for examining the relationships between ADIN/TN ratio and other forage attributes were that neither sample set was designed for that purpose, and that the reported measurements were limited. The forages in Dataset 1 were harvested from pasture swards that usually contained a mix of C₄ perennial grass species in various stages of regrowth, or were hays obtained from commercial farms. The growth and physiological maturity of the grasses sampled varied with the grass species and with interactions such as with soils and rainfall events to provide a broad array of forages as required for the development of the F.NIRS calibrations. Similarly, Dataset 2 collating results for 305 forages from 30 published papers did not provide a rigorous or balanced sample set. Nevertheless the present study showed that, on average, lower quality C₄ grass forages containing ∼>400 g ADF/kg, ∼<10 g TN/kg or ∼<530 g digestible DM/kg had much higher ADIN/TN ratios than generally found in higher quality C₄ grass forages, C₃ grass forages or legumes. It was also clear that among the low-quality C₄ grasses there was great variability in the ADIN/TN ratio, ranging from 11 to 495 g ADIN/kg TN. Wilson (1982, 1994) has provided excellent qualitative descriptions of the structure and anatomy of C₄ grasses, C₃ grasses and dicotyledonous forages and how these are likely to be influenced by plant part, age of individual leaves and stem, water stress and temperature. However, it was not possible to relate such factors to the species, growing conditions or anatomical structure of the grasses in Datasets 1 and 2. High soil N availability during grass growth would be expected to increase the TN concentration and proportion of soluble N; however, as stated in the Materials and methods, the forages in Dataset 1 were generally not from fertilised soils. Also Dataset 2 did not include samples from treatments where the grass had been fertilised with N. It seems likely that better understanding of effects of forage species and environment will explain at least some of the variation in ADIN/TN ratio among C₄ grasses and provide the reasons why this ratio is high in highly fibrous C₄ grasses. This requires further investigation. Until further information is available it seems reasonable that for senesced C₄ grasses, defined, for example, as containing >400 g ADF/kg, <10 g TN/kg or <530 g digestible DM/kg, nutritional recommendations should assume that 20–30% of the TN is not available to the ruminant.

The finding that a substantial proportion of the TN in mature C₄ grasses is indigestible will be most important in circumstances where diets are based on such forages that are marginal or deficient in N. One such situation is in low-input, low-output tropical production systems where ruminants are highly dependent on physiologically mature or senesced C₄ grass forages for many months of the annual cycle. This often occurs in rangeland production systems in the seasonally dry tropics of the northern Australian, African and South American rangelands where the only pastures available during the dry season are physiologically mature or senesced and weathered (Winks 1984; Hogan 1996). Many of these C₄ grass forages, such as those in the present Dataset 1 and in the database of Norton (1982), contain <10 g N/kg DM and have low DMD, so that the ADIN/TN ratio is expected to be high. In addition, Aumont et al. (1995) reported that in a large, normally distributed database of forages from the humid tropics in the Caribbean and La Reunion (n = 1313), the means and standard deviations indicated that ∼30% of the forages contained >400 g ADF/kg DM or <11 g N/kg DM. High ADIN/TN ratios are also likely to occur in C₄ crop residues (e.g. maize, sorghum and millet stovers), which are important feedstuffs in many African crop–livestock systems (Duncan et al. 2013, 2016; Dejene et al. 2022). Senesced tropical pastures or crop residues are usually deficient in TN for ruminants (Winks et al. 1979; Hennessy and Williamson 1990; Kennedy et al. 1992; Coates and Dixon 2008; Dixon and Coates 2010). Lower than expected availability of the TN in such forages will require reconsideration of the need for, and responses of ruminants to, supplementary rumen-degradable N. In low-input systems such as those described above, the provision of N supplements usually incurs substantial increases in input costs and necessary management skills, and especially where urea non-protein N is used to provide supplementary N. The development and availability of F.NIRS in ruminants to measure diet attributes (Stuth et al. 1999; Coates 2004; Dixon and Coates 2009, 2015) provides opportunity for commercial farms to routinely monitor the ADF, TN, DMD and metabolisable energy of the diet of grazing cattle, and to use these measurements to estimate the proportion of indigestible N.

The study established that in highly fibrous (>400 g ADF/kg DM) and low TN (<10 g TN/kg DM) C₄ grass forages, the ADIN/TN ratio was highly variable, ranging from negligible to ∼50%, but averaging ∼20%. This greater than expected proportion of indigestible forage TN bound to plant cell walls requires consideration when evaluating the adequacy of TN in the diets of ruminants depending on such forages. Further investigation is needed to identify the grass species, environmental conditions and circumstances where the proportion of indigestible ADIN in TN in such forages is of practical nutritional importance.

The data that support this study may be shared upon reasonable request to the corresponding author if appropriate.

The authors declare no conflict of interest. Dr R M Dixon is an Associate Editor of Animal Production Science but had no role in the review or evaluation of the manuscript.

The research did not receive any specific funding.