Effect of supplementing tropical tannin-free and tanniniferous legumes to grass-fed sheep on the utility of their manure as nitrogen fertiliser

T T Tiemann, B Hincapie**, E Frossard*, M Kreuzer and H D Hess*** 

ETH Zurich, Institute of Animal Science, Universitaetstrasse 2, CH-8092 Zurich, Switzerland

*ETH Zurich, Institute of Plant Science, Eschikon 33, CH-8315 Lindau, Switzerland

**Tropical Grass and Legume Project, CIAT, Cali, Colombia

***Agroscope Liebefeld-Posieux, Research Station ALP, Tioleyre 4, CH-1725 Posieux

dieter.hess@alp.admin.ch

Abstract 

A pot experiment was conducted to assess the fertiliser value of faeces from sheep fed with a diet supplemented with legumes free of or containing condensed tannins. Ten animals received the same five diets in different experimental runs of a double Latin Square design, all consisting of a low quality tropical grass ration, supplemented with 450g/kg legumes. Legume treatments were either solely Vigna unguiculata (a tannin free herbaceous legume) or mixtures of V. unguiculata with either Calliandra calothyrsus or Flemingia macrophylla (both shrub legumes rich in condensed tannins) in ratios of 2:1 and 1:2. Faeces from animals receiving the same diets were pooled and applied at two levels (20 and 80 mg N/kg soil) to pots with seedlings from the cultivar Mulato II, a Brachiaria grass hybrid. An acidic infertile tropical soil was used. Aerial biomass yield and plant N content were evaluated.

The results seem to indicate that the N fertiliser value of faeces from sheep receiving a diet supplemented with tanniniferous legumes is not substantially reduced compared to diets free of condensed tannins. However, the data also indicate that nitrogen was not the most limiting element in the soil used since mineral fertilisation without extra nitrogen was efficient, too.

Key words: condensed tannins, faeces, fertiliser value, minerals, shrub legumes

Introduction 

Manure is a cheap and valuable fertiliser for field cropping and grassland management, particularly in the poorer regions of the world, including many tropical countries, where small farmers often can not afford the more expensive mineral fertilisers. A characteristic of many tropical, not pasture-based, livestock systems is that only faeces are collected for use as fertiliser while urine is lost or discarded. As a result of different amounts and kinds of indigestible components in the diet, the quality of faeces as a fertiliser could depend on the type of feed the animal consumes. Therefore variations in the dietary N content may change the proportion of plant available N in faeces (Tiemann et al 2008). Condensed tannins (CT) are known to form complexes with proteins making them unavailable to ruminal digestion. The consequences of this attribute have been extensively discussed for ruminant nutrition where it was shown that in some cases high proportions of the feed protein leave the animal undigested (Carulla et al 2005; Tiemann et al 2008). Concerning the fertiliser value of the faeces, Carulla et al (2005) demonstrated that the addition of tannins to ruminant diets reduces the proportion of N excreted in urine and increases the proportion and total amount of N in faeces thus overall increasing the N return to the soils when only faeces are spread. However, the extent to which these CT-protein complexes, which are obviously transferred to the faeces, are available to the soil microbes is still unclear. Powell et al (1994) showed that tannins may decrease the proportion of water-soluble N in the faeces and increase the proportion of N bound to the fibre fraction, resulting in slower rates of faecal N mineralisation in the soil. Additionally, Cadisch and Giller (2001) demonstrated that leaves of Desmodium heterocarpon cv. ovalifolium, having a high CT-content, when added to the soil lead to a high carbon sequestration in the soil and to a slow release of N just after the application which is not compensated for later by a correspondingly higher N mineralisation. They ascribed the latter to the formation of stable CT-protein complexes in the soil. If CT lead to slower mineralisation this could reduce N losses from manure to the environment and favour C accumulation while, in case the complexes with protein in the soil are permanently, this would reduce the value of the faeces as fertiliser.

The hypothesis to be tested in the present study was that feeding high-CT tropical plants reduces the N fertiliser value of the faeces of such animals compared to those fed diets free of CT. This was investigated by conducting a pot experiment growing a common tropical grass on a typical low fertile tropical soil. The effects of faeces on plant growth and N content was assessed in comparison to zero fertilisation, mineral fertilisation and urea or both as controls.  A typical low fertile tropical soil and a common tropical grass were used to determine the effects of the faeces on the plants. For later explanation of the fertiliser value, the faeces of the different feeding origin were analytically characterised for their contents of macronutrients (including undigested fibre fractions from feed) and various minerals.

Material and methods 

An animal experiment was conducted with ten African Breed sheep in a double Latin square design at the research station of Santander de Quilichao of the International Centre for Tropical Agriculture (CIAT), Colombia. In each experimental run the sheep received one of five diets composed of 550 g/kg dry matter (DM) of Brachiaria humidicola and of 450g/kg DM of a legume supplement. These supplements consisted of Vigna unguiculata alone, or of 1:2 and 2:1 mixtures of V. unguiculata with either Calliandra calothyrsus or Flemingia macrophylla. Daily, faeces were collected separately from urine and stored at –20°C. At the end of the experiment, faeces of all animals and experimental runs with the same diets were pooled and lyophilised yielding a total of five composite faeces samples (one per experimental diet). Faeces were stored in a dry, dark and cool place until being either used for the subsequent pot experiment or subjected to analysis.

Each of the five experimental faeces types was applied at rates providing either 20 or 80 mg N/kg soil. Each faeces treatment was done with and without additional fertilisation with an N-free mineral mixture adding (in mg/kg soil), P, 139; Ca, 306; K, 107; Mg, 167; S, 13; Zn, 2.4; Cu, 3; B, 0.32 and Mo, 0.14. Additionally, there were four control treatments, two without any N fertilisation (also here with and without the extra N-free mineral fertilisation) and two with urea as N source being also provided at rates of 20 or 80 mg N/kg soil. This added up to a total of 24 different treatments. Each treatment was investigated in four replicates. As an experimental soil, an Oxisol (typical isohyperthermic Caolinitic Haplustox, pH <5 from the Eastern Planes (Matazul, Meta, Colombia)) with low natural fertility and high aluminum saturation had been chosen. The soil had been transferred to CIAT’s headquarter, homogenised and sterilised with steam for 24 h. In the experiment, 4 kg of soil were used per pot. The N-free mineral fertiliser was added in water soluble form when appropriate. With a ball mill pulverised faeces and, when foreseen in the treatment, urea were mixed and dissolved in 150ml of water and administered by watering the plants. Each pot was placed on a plate and excess water was collected and added again to the treatment.

The test plant was the Brachiaria hybrid Mulato II (Brachiaria ruziziensis × B. decumbens × B. brizantha; CIAT 36087), which was chosen due to its known high susceptibility to small changes in soil quality (Dr. Idupulapati Rao, CIAT, pers. communication). Mechanically scarified seeds were grown to seedlings in a greenhouse for 10 days. Afterwards four seedlings were transplanted into each pot. Half of the determined fertiliser amounts (faeces or urea) were added to the pots immediately, the other half was applied 21 days later. The pots were kept in the greenhouse under natural light regime at 25°C, and the plants were watered every 3 days. After 6 weeks, total aerial mass was harvested and weighed to determine dry matter yields. The material was frozen at –20°C and lyophilised.

Faeces were ground before analysis in a Wiley laboratory mill with a 0.1-mm screen. Faeces and harvests were analysed for DM (24 h at 105 °C) and N (San+ Autoanalyzer, Skalar Analytical, Breda, The Netherlands; analysed after Kjeldahl digestion, faeces additionally for total ash (3 h at 500 °C), P (San+ Autoanalyzer; also analysed after Kjeldahl digestion), as well as neutral detergent fibre (NDF; Van Soest et al 1991). Furthermore indigestible acid detergent fibre (IADF) was determined by preparing the culture media and the reducing agent as described by Theodorou et al (1994). Briefly, 0.5 g plant material was combined with 40 ml of culture media and 2 ml of reducing agent. After the inoculation with 10 ml of fresh rumen fluid obtained from two fistulated crossbred Zebu steers, tubes were flushed with CO₂ and immediately closed with a plug with gas release valve. Tubes were incubated at 39 °C for 144 h. Digestion was stopped through the addition of 0.6 ml of HgCl₂ solution (50 g/l). The content of each tube was subsequently subjected to the acid detergent fibre analysis procedure (Van Soest et al 1991). The carbon content of the faeces was determined by dry combustion using a CHN analyzer (LECO model CHN 600, Leco Corp. St Joseph, MI, USA). Contents of Ca, Mg, K, Na, Cu, Zn and Mn in faeces and (some of them) in the soil were determined after nitric-perchloric digestion (Benton and Jones 1989) with atom absorption/emission spectroscopy. Sulphur was analysed after extraction with calcium phosphate via turbidimetric measurement with barium chloride (Wall et al 1980). Additionally to that, in the soil pH was measured in water 1:1, OM was determined by colorimetric measurement after KCr₂/H₂SO₄ oxidation, mineralisable NH₄ by the San+ Autoanalyzer, Al by KCl 1M extraction and NaOH 0.1M titration and Al-saturation (Al/ECEC * 100, where ECEC=Al + Ca + Mg + K).

The N uptake ratio of the plants, i.e. the amount of N recovered in the harvest per unit of N delivered by fertilisation was calculated as (N_sample-N_control)/N_added. All data was subjected to analysis of variance using the GLM procedure of SAS (version 9.1.3; 2006) with N-free fertilisation, N level, origin of N-source and their respective interactions as sources of variation. The uptake ratio was analysed with two separate models for treatments with and without N-free fertilisation, with N level, origin of faeces and the respective interaction as sources of variation. All multiple comparisons among means were performed with the Duncan test. The standard errors of the means (SEM) are shown for all variables as are the P values for the effects.

Results and discussion 

The experimental faeces showed only small differences in organic matter (OM), fibre and N content (Table 1). The C:N ratio ranged between 27:1 and 31:1 which can be considered very high compared to other studies (Powell et al 2006). Adding the faeces alone provided (in mg/kg soil) P, 12/48; Ca, 20/80; K, 3/12; Mg, 5/20; S, 5/20; Zn, 0.4/1.4; and Cu, 0.1/0.4 at the low/high N level, with relatively small variation among faeces origins. The soil properties were: pH, 4.8; OM, 26 g/kg; mineralisable NH₄, 0.72 µg/g; P, 2.43 µg/g, K; 0.4 mmol/kg; Ca, 0.4 mmol/kg; Mg, 0.3 mmol/kg; Al, 16.3 mmol/kg; Al-saturation, 93.5%.

Without the addition of either urea or N-free fertiliser or faeces, plant growth was very poor (Table 2). A high increase (P < 0.001) in DM yield was obtained simply by the addition of the N-free fertiliser. The application of faeces, particularly at the high level, also improved plant growth compared to the treatments without application of the mineral mixture and those with low or without faeces application. The combination of faeces and N-free fertiliser numerically increased DM yield by up to 25% relative to supplementing the N-free fertiliser alone, but this was not statistically significant. This weak effect of extra faecal N suggests that other nutrients apart from N were at least co-limiting. From the soil analysis it seems likely that P supplementation accounted for most of the effects. Unexpectedly, DM yield was numerically higher by about 20% with the treatments including faeces plus mineral mixture as fertiliser compared to the treatments with urea as N source, but the total N amount harvested was still highest (P < 0.05) with the high urea treatment. This indicates that the minerals present in the faeces may have contributed to the differences in DM yield as discussed below in more detail.

The N content was clearly higher in plants from treatments without mineral mixture, but the totally incorporated N in the plants was markedly increased (P < 0.001) by mineral addition. It is well documented that limited growth due to limitation of essential plant nutrients leads to a higher concentration of N in the plant (Gastal and Lemaire 2002) while, without this limitation, N gets diluted by other plant matter. It is probable, that the differences in N content in the different mineral treatments were a result of this effect. Higher levels of faeces application also led to higher (P < 0.001) total N amounts harvested. In the presence of the N-free mineral fertiliser without urea or manure addition, plants took up 74 mg N/plant, representing the amount of N which could be mineralised from the soil. Considering this control treatment, N added in the form of urea was almost completely recovered in the plants, both in the 20 mg N/kg treatment with a total plant N uptake of 98.3 mg N/kg and in the 80 mg N/kg treatment with 165.5 mg N uptake/kg). The proportionate N uptake from the manure treatments was much lower. In the treatments with faeces and without mineral mixture, the uptake ratio, i.e., the N amount recovered per amount of N delivered by manure, was only influenced (P<0.05) by the amount of N added. The origin of the faeces, i.e., the diet of the animal, had no effect which means, that the action of CT in these faeces seems to be not relevant anymore and the differences in composition are to small to have a significant effect. However, while a higher N amount supplied with the faeces increased uptake ratios in treatments without extra N-free fertiliser, it had rather the opposite effect (although not significant) in treatments supplemented with extra minerals. At 20 mg N/kg soil and without N-free mineral fertiliser, plants recovered less N than those of the N-free fertiliser control, meaning that the net N mineralisation was lower in the presence of the low manure dose. At the high faeces level (80 mg N/kg soil) the N uptake was equivalent to that of the N-free fertiliser control. Calderón et al (2004) found that manures with C/N ratios of 19 or higher caused immobilisation of soil N, whereas manure with average C/N ratios of 16 or lower facilitated N mineralisation. Heal et al (1997) found that C/N ratios <20 lead to rapid decomposition of organic compounds in soils and that C forms (hemicellulosis, cellulosis, lignin) can highly influence decomposition and N mineralisation. These results support the assumption that N immobilisation occurred in the absence of the N-free fertiliser in the present experiment.

Another explanation for differences in both, DM and N yield of the plant, might be given by the large differences in minerals available in treatments with and without mineral mixture. The amount of minerals added by faeces application alone was 2-8 times lower than the amount added with the mineral mixture. Only for sulphur the quantity was in between the high and low faeces fertilisation treatments. In an ample study on dairy cattle requirements, excreta composition and its effect on fodder plant growth and composition, Powell et al (2006) found few significant relationships among the chemical properties of applied faeces and subsequent soil inorganic N, plant DM, and N uptake while the effects were depending on plant species and soil type. In contrast to the temperate soils used by Powell et al (2006) the soil in the present study was very limited in plant available minerals. Particularly the three to eleven times higher phosphorus content of treatments with mineral mixture compared to those with faeces only (high and low application, respectively) as well as the 8 to 34 times higher amounts of K and Mg might have been also responsible for the differences in DM yield.

Although these differences result in a 5 to 20 times higher concentration of basic compounds in the soil, the result that treatments with mineral mixture and urea had similar DM yield as treatments without mineral mixture but with high faeces application indicates that the positive effect on growth was related to the lack of crucial nutrients rather than to high pH and low base saturation. The same treatments showed on the other hand an about 20% lower DM yield than the treatments with high amount of faeces being applied together with the mineral mixture. Taken into account that the amount of minerals provided by the mixture was 2-8 times higher than in the high faeces treatment alone, the low gain in growth compared to the treatment without mineral mixture indicates that the amount of nutrients provided by faeces is getting close to the mineral requirements of the plant under the given soil conditions and this independently of the forage fed to the animals. Clear (P < 0.01) interactions for all variables occurred between the application of mineral mixture and the applied faeces level, reflecting the importance of a balanced nutrient supply.

The CT in the tested legumes must be considered as strong binding compounds, which do not liberate bound substances easily (Tiemann et al 2008). Although it is not possible to make a clear statement about their effects on the N availability of the faeces, their overall effect on nutrient remineralisation in the faeces seems to be negligible. Even on infertile soils as the one used in this experiment, where besides from N also minerals seemed to be limiting no differences between the different faeces origin, containing different proportions of tanniniferous plants, were detectable. This indicates that less aggressive CT of other plant species used as animal feeds should have no relevant effect under similar conditions on the fertility of animal manure.

Conclusions 

• The lack of clear differences among faeces of different feeding origin appears to suggest that the basal hypothesis that CT-plants lead to a reduced N fertiliser value of the faeces would have to be rejected. Our data indicate that manure from sheep fed CT-plants did not limit the plant available N compared to manure from sheep fed plants free of CT. The proportion of tanniniferous legumes in the diet had no effect as well, and there was no evidence that manure from sheep fed CT-plants retards N mineralisation in the soil. However, since N was not clearly turning out to be the first limiting plant nutrient, effects of faeces type might have been masked. This restricts the statement that feeding legumes with CT does not limit the value of animal faeces as N source for plant fertilisation in tropical soils to situations when the soil is deficient in several plant nutrients. Faeces, at least when applied at the high level, apparently provided sufficient amounts of plant available minerals to maintain plant growth at the same level as found with the application of the mineral mixture indicating that plant mineral requirements were sufficiently covered by faeces application to allow normal plant growth.

Acknowledgements 

We wish to express our gratitude to Rafael Restrepo from the CIAT Forage Group for his great help with the greenhouse experiment and to Aristipo Betancourt for donating the seeds. We are also grateful to Idupulapati Rao and Jaumer Ricaurter from CIAT, Tropical Soil Biology Project, for their advice and help in the experimental design.

References 

Benton J B and Jones J 1989 Plant analysis techniques. Benton-Jones Laboratories. Georgia, USA.

Cadisch G and Giller K E 2001 Soil organic matter management: the roles of residue quality in C sequestration and N supply. pp 97–111. In: R M Rees, B C Ball, C D Campbell and C A Watson editors. Sustainable Management of Organic Matter. CAB International, Wallingford, Oxon UK.

Calderón F J, McCarty G W, Van Kessel J S and Reeves J B 2004 Carbon and Nitrogen dynamics during incubation of manured soil. Soil Science Society of America Journal 68: 1592–1599

Carulla J E, Kreuzer M, Machmüller A and Hess H D 2005 Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nitrogen in forage-fed sheep. Australian Journal of Agricultural Research 56: 951–969

Gastal F and Lemaire G 2002 N uptake and distribution in crops: an agronomical and ecophysiological perspective. Journal of Experimental Botany 53: 789–799

Heal O W, Anderson J M and Swift M J 1997 Plant litter quality and decomposition: An historic overview. pp. 3–32. In G Cadisch and K E Giller editors. Driven by Nature: Plant Litter Quality and Decomposition. CAB International. Wallingford, UK.

Powell J M, Fernández-Rivera S and Hofs S 1994 Effects of sheep diet on nutrient cycling in mixed farming systems of semi-arid west Africa. Agriculture, Ecosystems and the Environment 48: 263–271

Powell J M Wattiaux M A, Broderick G A, Moreira V R and Casler M D 2006 Dairy diet impacts on fecal chemical properties and nitrogen cycling in soils. Soil Science Society of America Journal 70: 786–794.

Theodorou M K, Williams B A, Dhanoa M S, McAllan A B and France J 1994 A simple gas production method using a pressure transducer to determine the fermentation kinetics of ruminant feeds. Animal Feed Science and Technology 48: 185–197

Tiemann T T, Lascano C E, Wettstein H-R, Mayer A C, Kreuzer M and Hess H D 2008 Effect of the tropical tannin-rich shrub legumes Calliandra calothyrsus and Flemingia macrophylla on methane emission and nitrogen and energy balance in growing lambs. Animal 2: 790–799

Van Soest P J, Robertson J B and Lewis B A 1991 Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. Journal of Dairy Science 74: 3583–3597 http://jds.fass.org/cgi/reprint/74/10/3583.pdf

Wall L L, Gehrke C W and Suzuki J 1980 An automated turbidimetric method for total sulfur in plant tissue and sulphate sulfur in soils. Communications in Soil Science and Plant Analysis 11: 1087–110

Received 18 October 2008; Accepted 18 December 2008; Published 10 March 2009

Go to top