Livestock Science 162 (2014) 141153

Contents lists available at ScienceDirect

Livestock Science
journal homepage: www.elsevier.com/locate/livsci

An evaluation of the performance and efficiency of nitrogen

utilization in cattle fed tropical grass pastures
with supplementation
Edenio Detmann a,n, riton E.L. Valente a, Erick D. Batista a, Pekka Huhtanen b
a
Departamento de Zootecnia, Universidade Federal de Viosa, Viosa, Minas Gerais, Brazil
b
Department of Agricultural Research for Northern Sweden, Swedish University of Agricultural Sciences, Ume, Sweden

a r t i c l e i n f o abstract

Article history: The objective of this study was to evaluate the efficiency of nitrogen utilisation (ENU) and
Received 25 September 2013 animal performance in cattle fed tropical grass pastures and receiving nitrogen supplemen-
Received in revised form tation, using a meta-analytical approach. The dataset used to evaluate the ENU was compiled
29 January 2014
from 10 experiments carried out in Brazil, which were published between 2009 and 2012,
Accepted 30 January 2014
totalling 47 treatment means. To compose the dataset for animal performance evaluation,
mean treatment values were collected from 44 experiments involving grazing beef cattle,
Keywords: which were published in Brazil from 2001 to 2012, totalling 182 treatment means. The
Mixed models relationships between variables were evaluated using linear or non-linear mixed models,
Nitrogen balance
taking into account the random variations among experiments. There was no association
Nitrogen status
between ENU and the balance of nitrogen in the rumen and the efficiency of microbial
Nitrogen supplementation
Tropical forage synthesis in the rumen (P40.05). The ENU was linearly associated with the dietary content
of digestible organic matter in the diet (DOM) and relative production of microbial nitrogen
in the rumen (NMICR), which was positive with DOM (Po0.04) and negative with NMICR
(Po0.03). The relationships between ENU and dietary crude protein (CP) and the
concentration of rumen ammonia nitrogen (RAN) were described using hyperbolic models
(Po0.05). The estimated values that correspond to the apparent equilibrium point (i.e.,
ENU0) were 108 g CP/kg of dry matter and 6.30 mg RAN/dL of rumen fluid. There was a
positive response on weight gain as the amount of supplement increased (Po0.01).
However, this relationship was interactively affected by CP content in the supplement
(Po0.02) and in the forage (Po0.01). Through this way, the response of weight gain to
supplement is improved by increasing the CP content of the supplement, but decreased as
the CP in the forage increases. From this it can be conclude that there is a positive response
to nitrogen supplementation with regards to ENU and animal performance. The main benefit
of supplementation is the improvement of the nitrogen status in the animal's metabolism.
& 2014 Elsevier B.V. All rights reserved.

1. Introduction
Tropical grasses are rarely available as a balanced diet for
grazing cattle because they exhibit nutritional constraints
n
Corresponding author. Tel./fax: 55 31 3899 2252.
E-mail address: detmann@ufv.br (E. Detmann).
that limit pasture intake and digestibility. For this reason,
there is a demand to identify the nutritional limitations of
tropical pastures, in order to avoid constraints on animal
production via an appropriate supplementation programme.
During the dry season, there is a drastic decrease in the
nutritional quality of tropical grasses, as mainly indicated
by decreased crude protein (CP) content. The limited CP
availability has been recognised as the critical threshold

http://dx.doi.org/10.1016/j.livsci.2014.01.029
1871-1413 & 2014 Elsevier B.V. All rights reserved.
142 E. Detmann et al. / Livestock Science 162 (2014) 141153
for adequate microbial growth on the fibrous carbohy-
drates in basal forage (Hennessy et al., 1983; Leng, 1990),
which results in decreased intake and animal performance
(Egan and Doyle, 1985; Leng, 1990; Paulino et al., 2008).
Under these circumstances, the supplementation with
nitrogenous compounds is the primary nutritional tool to
improve the utilisation of low-quality forage by grazing
cattle (Hennessy et al., 1983; Leng, 1990; Sampaio et al.,
2010; Souza et al., 2010).
On the other hand, during the rainy season the avail-
able forage is of better quality than in the dry season.
However, despite improved animal performance, the uti-
lisation of basal forage during the rainy season is not
considered to be optimal. According to Detmann et al.
(2010), there is a nutritional unbalance in tropical pastures
during rainy season which is characterised by a relative
excess of energy in relation to available CP (Detmann et al.,
2010). Therefore, there is an unused potential growth of
approximately 200 g/animal/d, which can be achieved by
using supplements (Poppi and McLennan, 1995; Paulino
et al., 2008).
For that reason, it can be inferred that nitrogen can be
considered the main component of supplements for graz-
ing cattle in the tropics throughout the year. The efficiency
of nitrogen utilisation become an important parameter
either to define supplement composition or to understand
the efficiency of animal production.
Therefore, the objective of this study was to evaluate
the efficiency of nitrogen utilisation and animal perfor-
mance in cattle fed tropical grass pastures and supple-
mented with nitrogenous compounds by using a meta-
analytical approach.
2. Material and methods
2.1. Data acquisition
2.1.1. Nutritional characteristics and efficiency of nitrogen
utilisation
The dataset used to evaluate the nutritional character-
istics and the efficiency of nitrogen utilisation was com-
piled from 10 experiments carried out in Brazil, published
between 2009 and 2012, totalising 47 treatment means
(Appendix 1). The following variables were considered:
digestible organic matter (DOM) and CP concentrations in
the diet, average daily concentration of rumen ammonia
nitrogen (RAN), forage dry matter intake (FDMI), DOM
intake (DOMI), nitrogen intake, production of microbial
nitrogen in the rumen, total nitrogen flow to the aboma-
sum, and apparent balance of nitrogen in the animals'
body (nitrogen intake faecal nitrogen urinary nitro-
gen). All experiments were conducted with grazing cattle
(Brachiaria sp.) or cattle fed with tropical grass hays
(Brachiaria sp. or Cynodom sp.) that were cut from pasture
areas. The experiments were carried out according to Latin
square designs using Zebu or crossbred (Zebu  Holstein)
animals. Each study included a control treatment without
supplementation. In all experiments highly rumen degrad-
able protein sources were used in the supplements (casein,
albumin, urea, soybean meal or a mixed of more than one
source). The average daily RAN concentration was calcu-
lated as the mean of the collections that were made along
a 24 h period. The production of microbial nitrogen was
estimated by using either purine flow in the abomasum or
by the urinary excretion of purine derivatives. The flow of
nitrogenous compounds in the abomasum was only eval-
uated in six experiments, totalling 28 treatment means.
The nitrogen balance in the animal's body was not eval-
uated in one of the experiments. The forage intake in
grazing experiments was estimated using one external
(titanium dioxide) and one internal (indigestible neutral
detergent fibre) marker. The diet sampling in grazing
experiments was performed using hand plucked samples.
To minimise the interference due to differences in the
intake and size of the animals, the variables associated to
the efficiency of nitrogen utilisation were expressed as
ratios according to the following equations:
NMIC
NMICR 1
NI
NI  NFA
RNB 2
NI
NB
ENU 3
NI
NMIC
EMS 4
DOMI
where NMICR is the relative production of microbial
nitrogen in the rumen (g/g ingested N), NMIC is the
production of microbial nitrogen in the rumen (g/d), NI
is the nitrogen intake (g/d), RNB is the apparent relative
nitrogen balance in the rumen (g/g ingested N), NFA is the
nitrogen flow to the abomasum (g/d), ENU is the apparent
efficiency of nitrogen utilisation in the animals' body (g/g
ingested N), NB is the apparent nitrogen balance in the
animals' body (g/d), EMS is the efficiency of microbial
synthesis in the rumen (g microbial N/kg DOM), and DOMI
is the DOM intake expressed as kg/d.
In two experiments there were some treatments where
supplements were provided in the abomasum. In these
cases, the nitrogen infused in the abomasum was not used
to estimate RNB.
It must be emphasised that nutritional dataset is not
broad enough to allow inferring about class variables as
sex, genetic group and season. Despite this, the season
variation was indirectly taken into account by using
continuous variables associated with forage quality (i.e.,
CP and DOM content).
2.1.2. Animal performance
To compose the dataset for the animal performance
evaluation, mean treatment values were collected from
experiments with grazing beef cattle. These experiments
were carried out in Brazil between 2001 and 2012. The
minimum prerequisite for including a study in the dataset
was for it to provide information concerning the average
daily gain (ADG), the amount of supplement fed to the
animals (SUP), the crude protein content in the supple-
ments (SCP), the body weight (BW) of the animals at the
beginning of the experiment and the length of the
 E. Detmann et al. / Livestock Science 162 (2014) 141153
experiment. Furthermore, it was mandatory that a control
treatment without supplementation was included in each
experiment. Following this statement, data was collected
from 44 experiments published in 42 papers, totalling 182
treatment means (Appendix 2).
When an experiment was chosen for the dataset,
several additional characteristics were aggregated: animal
category (heifers, steer or bulls), genetic group of the
animals (European, Zebu or Crossbred), type of forage in
the pasture (Brachiaria sp., Cynodon sp., Panicum sp. or
sub-tropical pastures), forage availability in the pasture
(FAV), and the CP content in the available forage (FCP).
Actually, not all of that information could be found in
every experiment, which created a small heterogeneity in
variables regarding the total number of observations
(treatment means).
The feed compositions of the supplements were quite
variable as the dataset was composed by experiments
carried out in different regions of Brazil (which always
have different types of local feeds and by-products) and
with several types of feed comparisons. This pattern did
not allow using supplement feed composition as a class
variable in the dataset because of the high risk of con-
founding on the estimates.
During acquisition, the animal performance dataset
was also stratified according to the season in which the
experiments were carried out. This was defined by the
following classification: dry season, dry-to-rainy transition
season, rainy season, and rainy-to-dry transition season
(Detmann et al., 2010).
Due to the high heterogeneity among animals, seasons,
locals, years, etc., the values of ADG were expressed as the
differential weight gain (DWG) in relation to the control
treatment for each experiment. To decrease the influence
of the different animal weights, the values of the supple-
ment intake were used as a fraction of BW, as seen here
 
DIE
ABW iBW  ADG 5 was evaluated using a canonical discrimination analysis.
2
SUPM
SUP  1000 6 should be merged. The DWG, FAV, FCP and SUP were used
ABW
SUPM  SCP
SCPI 7
ABW
where ABW is the average BW of the animal (kg), iBW is
the initial BW of the animal (kg), DIE is the number of days
in the experiment, SUP is the daily amount of the supple-
ment consumed by the animal (g/kg BW), SUPM is the
absolute mass of the supplement consumed by the animal
(kg/d), SCPI is the supplemental crude protein intake (g/kg
BW), and SCP is the crude protein content in the supple-
ment (g/kg).
2.2. Statistical methods and analyses
The data was analysed using meta-analysis techniques
(St-Pierre, 2001; Van Houwelingen et al., 2002), using the
mixed procedures of SAS (version 9.3) for linear (MIXED)
and non-linear (NLMIXED) models (Littell et al., 2006).
The random effect of the different subjects (experi-
ments) was considered in the regression parameters. The
143
adequacy of the models and the best (co)variance struc-
tures were evaluated using the corrected Akaike's infor-
mation criterion. All variance components in the linear
models were estimated using the restricted maximum
likelihood method.
When a relationship was found to be non-linear, a
random variance component was associated to each para-
meter of the model and the variation among experiments
(subjects) was taken into account during the adjustment.
When the random variance associated to a specific para-
meter was found to be non-significant (P40.05), a new
model was adjusted, excluding that specific random effect.
The dual quasi-Newton method was used as optimisation
technique. Additionally, the integration was performed
using the adaptive Gaussian quadrature (Littell et al., 2006).
The coefficients of determination (r2/R2) of the adjusted
models were calculated as the square of the correlation
between predicted and observed values.
In the dataset for evaluating the nutritional characteristics
one treatment mean for RAN concentration behaved as an
outlier and was not taken into account in the analyses.
The linear regression models for describing DWG were
adjusted in accordance with the influence of the class
variables on intercept and slope(s). However, it must be
pointed out that the forage type effect was not considered
in the analysis. This decision was influenced by the fact
that most information was obtained using Brachiaria sp.
(n 147). Some of the forage types had a low number of
observations and could not cover an adequate range of
DWG. This could confound the results. Furthermore, some
sexes were not evaluated in all seasons and the class
variable sex was not considered in the analyses to avoid
biases in the estimates. There were a low number of
treatment means associated to European cattle (n 3).
Therefore, the genetic group effect was evaluated only
for Zebu and crossbred cattle.
Prior to meta-analysis, the animal performance dataset
This analysis was performed to verify if there was a
difference between the four seasons, or if some seasons
as discriminating variables and the analysis was conducted
using the CANDISC procedure of SAS (version 9.3). Gen-
eralised Mahalanobis' distances were then calculated
between seasons and the significance was evaluated using
a F-test approach.
All statistical evaluations were performed using 0.05 as
the critical (asymptotical) level for the occurrence of type
I error.
3. Results
3.1. Nutritional characteristics and efficiency of nitrogen
utilisation
The ten experiments used to compose this dataset allowed
coverage of a broad range of diets. These included low-quality
forages presenting CP content below 60 g/kg DM, ruminal
conditions characterised by a severe nitrogen deficiency
(i.e., with highly negative RNB and NMICR greater than unity)
144 E. Detmann et al. / Livestock Science 162 (2014) 141153

or an excess of nitrogen (i.e., highly positive RNB and high The FDMI increased linearly as both dietary DOM
concentrations of RAN) and very low and very high ENU (Po0.01) and RAN concentration (Po0.03) increased.
(Table 1). This kind of pattern, in turn, represents a necessary On the other hand, the FDMI presented a quadratic pattern
need for interpreting nutritional characteristics. according to variations in dietary CP content (Po0.02) and
CP:DOM (Po0.01). The maximal responses were esti-
mated at 145 g CP/kg DM and 288 g CP/kg DOM, respec-
Table 1
tively (Table 2).
Overall description of the dataset used for the analysis of efficiency of
nitrogen utilisation. The RAN concentration was positively and linearly
associated with CP:DOM (Po0.01). There was a tendency
Variablea Statistics (Po0.10) for a linear positive association between RAN
and dietary DOM (Table 2). The RAN concentration
Mean Minimum Maximum s n
increased exponentially (Po0.05) as the CP of the diet
Whole dataset increased (Table 3).
DOM 469 268 606 88.1 47 The DOMI increased linearly as the dietary contents of
CP 116 45 205 36.8 47 DOM (Po0.01), CP (Po0.01) and the CP:DOM (P o0.02)
CP:DOM 251 103 428 74.9 47
increased. Furthermore, there was a quadratic response
RAN 9.47 2.53 17.93 4.381 46
FDMI 19.5 12.8 25.7 3.01 47 (Po0.01) of DOMI according to RAN concentration with
DOMI 9.7 3.5 15.0 3.15 47 the maximal response being estimated at 13.1 mgRAN/dL
NMICR 0.672 0.310 1.630 0.2744 47 of rumen fluid (Table 2).
RNB  0.044  0.987 0.603 0.3374 28
The NMICR decreased linearly as dietary DOM decreased
EMS 23.8 14.4 43.0 6.67 47
ENU 0.116  1.068 0.680 0.4332 42
(Po0.02; Table 4). The relationships between NMICR and
the dietary CP and CP:DOM were described by hyperbolic
Only forages
DOM 433 270 572 97.2 10
models (Po0.05) containing only one parameter (Table 3).
CP 82 52 128 24.8 10 This represents the value of CP (69 g/kg DM) or CP:DOM
CP:DOM 190 136 261 37.2 10 (157 g/kg), corresponding to a unitary NMICR that is the
RAN 5.30 2.74 9.77 2.016 10 apparent equilibrium between nitrogen intake and micro-
FDMI 18.9 12.8 25.7 4.07 10
bial production. The pattern of NMICR as a function of NAR
DOMI 8.4 3.5 13.6 3.62 10
NMICR 0.868 0.500 1.400 0.3197 10 was described by a negative non-linear power function
RNB  0.235  0.331  0.153 0.0748 6 model (Po0.05; Table 3). There was a quadratic and
EMS 23.3 16.2 37.0 6.66 10 negative relationship between NMICR and RNB (Po0.01).
ENU  0.552  1.068 0.147 0.4507 9 Evaluating the intercept of the adjusted model, when a
a theoretical state of nitrogen equilibrium in the rumen would
DOM, digestible organic matter content in diet (g/kg DM); CP, crude
protein content in diet (g/kg DM); CP:DOM, ratio between CP and DOM be achieved (i.e., nitrogen intake and nitrogen outflow from
contents (gCP/kgDOM); RAN, rumen ammonia nitrogen concentration rumen are equal, or RNB0), 62% of the nitrogenous
(mg/dL of ruminal fluid); FDMI, forage dry matter intake (g/kg BW); compounds flowing out to the abomasum would be of
DOMI, DOM intake (g/kg BW); NMICR, relative production of microbial microbial origin (Table 4).
nitrogen in the rumen (g microbial N/g ingested N); RNB, apparent
relative nitrogen balance in the rumen [(g ingested N- g N flowing out to
There was no influence (P 40.05) on RNB from dietary
abomasums)/g ingested N]; EMS, efficiency of microbial synthesis DOM and CP:DOM (Table 4). The RNB increased linearly
(g microbial N/kg DOM); ENU, apparent efficiency of nitrogen utilisation (Po0.02) as the dietary CP increased, and the apparent
in the animal body (gN apparently retained in the body/g ingested N). equilibrium point (i.e., RNB 0) was estimated at

Table 2
Summary of the linear models for describing the pattern of forage dry matter intake, concentration of rumen ammonia nitrogen and intake of digestible
organic matter.

Ya Xa Intercept Linear term Quadratic term RSDc r2/R2

Estimate s.e.b P value Estimate s.e. b

P value Estimate s.e. b
P value

FDMI DOM 8.0 1.44 o 0.001 0.024 0.0034 o 0.001 1.82 0.950
FDMI CP 11.4 2.36 0.001 0.125 0.0399 0.012  0.00043 0.000167 0.016 1.63 0.772
FDMI CP:DOM 9.1 2.87 0.011 0.075 0.0197 0.004  0.00013 0.000035 0.001 1.57 0.759
FDMI RAN 17.7 1.24 o 0.001 0.203 0.0725 0.021 1.62 0.774
RAN DOM 1.28 4.381 0.777 0.018 0.0096 0.092 4.34 0.322
RAN CP:DOM  2.39 1.797 0.217 0.049 0.0085 o 0.001 2.74 0.693
DOMI DOM  7.2 2.34 0.013 0.036 0.0044 o 0.001 0.66 0.966
DOMI CP 6.1 1.03 o 0.001 0.030 0.0054 o 0.001 1.15 0.896
DOMI CP:DOM 7.2 1.27 o 0.001 0.010 0.0032 0.015 1.36 0.854
DOMI RAN 5.0 1.31 0.004 0.875 0.2301 0.004  0.03343 0.010840 0.005 1.26 0.880

c
RSD, residual standard deviation of the relationship.
a
DOM, digestible organic matter content in diet (g/kg DM); CP, crude protein content in diet (g/kg DM); CP:DOM, ratio between CP and DOM contents
(gCP/kgDOM); RAN, rumen ammonia nitrogen concentration (mg/dL of ruminal fluid); FDMI, forage dry matter intake (g/kg BW); DOMI, DOM intake
(g/kg BW).
b
s.e., standard error of the estimate.
 E. Detmann et al. / Livestock Science 162 (2014) 141153 145

Table 3
Summary of the different relationships described by non-linear models.

Ya Xa Model structure ASDc R2

Estimate Effect typeb Estimate Effect typeb

RAN CP Y^  e(  X) 3.80 Fixed 0.0075 Random 3.15 0.667

NMICR CP Y^ /X 68.67 Random 0.064 0.965
NMICR CP:DOM Y^ /X 157.29 Random 0.070 0.960
NMICR RAN Y^  X 1.82 Random  0.494 Random 0.100 0.892
RNB RAN Y^  (1  /X) 0.538 Random 9.19 Random 0.148 0.878
ENU CP Y^  (1  /X) 0.236 Fixed 107.99 Random 0.157 0.922
ENU RAN Y^  (1  /X) 0.324 Random 6.30 Random 0.165 0.894
ENU DOMI Y^  (1  /X) 0.123 Fixed 11.23 Random 0.153 0.926
ENU DOMI Y^  (1  /X) 0.482 Fixed 6.53 Fixed 0.399 0.893

a
DOM, digestible organic matter content in diet (g/kg DM); CP, crude protein content in diet (g/kg DM); CP:DOM, ratio between CP and DOM contents
(g CP/kg DOM); RAN, rumen ammonia nitrogen concentration (mg/dL of ruminal fluid); DOMI, DOM intake (g/kg BW); NMICR, relative production of
microbial nitrogen in the rumen (g microbial N/g ingested N); RNB, apparent relative nitrogen balance in the rumen [(g ingested N gN flowing out to
abomasums)/g ingested N]; ENU, apparent efficiency of nitrogen utilisation in the animal body (gN apparently retained in the body/g ingested N).
b
A random effect type means that a random variance component was associated to the parameter (Po 0.05). A fixed effect type indicates that the
random variance (among experiments) was not significant (P4 0.05).
c
ASD, asymptotic standard deviation of the relationship.

Table 4
Summary of the linear models for describing the pattern of the relative production of microbial nitrogen in the rumen, the apparent relative nitrogen
balance in the rumen and the efficiency of microbial synthesis.

Ya Xa Intercept Linear term Quadratic term RSDc r2/R2

Estimate s.e.b P value Estimate s.e. b

P value Estimate s.e. b
P value

NMICR DOM 1.561 0.3388 0.001  0.0019 0.00061 0.012 0.200 0.472
NMICR RNB 0.622 0.0794 o 0.001  0.344 0.0778 0.007  0.3672 0.12480 0.010 0.102 0.777
RNB DOM  0.059 0.0823 0.507 0.312
RNB CP  0.951 0.2265 0.009 0.0077 0.00204 0.013 0.226 0.641
RNB CP:DOM  0.138 0.0771 0.134 0.232
EMS DOM 36.3 6.12 o 0.001  0.027 0.0126 0.060 3.82 0.738
EMS CP 23.8 1.89 o 0.001 3.76
EMS CP:DOM 23.7 1.58 o 0.001 3.71
EMS RAN 23.8 1.88 o 0.001 3.64
EMS DOMI 30.3 3.67 o 0.001  0.686 0.3391 0.073 3.79 0.743
EMS NMICR 18.5 2.38 o 0.001 7.411 3.0543 0.038 3.46 0.794
EMS RNB 22.1 1.64 o 0.001 3.80

a
DOM, digestible organic matter content in diet (g/kg DM); CP, crude protein content in diet (g/kg DM); CP:DOM, ratio between CP and DOM contents
(g CP/kg DOM); RAN, rumen ammonia nitrogen concentration (mg/dL of ruminal fluid); DOMI, DOM intake (g/kg BW); NMICR, relative production of
microbial nitrogen in the rumen (g microbial N/g ingested N); RNB, apparent relative nitrogen balance in the rumen [(g ingested N  gN flowing out to
abomasum)/g ingested N]; EMS, efficiency of microbial synthesis (g microbial N/kg DOM).
b
s.e., standard error of the estimate.
c
RSD, residual standard deviation of the relationship.

approximately 124 g CP/kg DM (Table 4). The relationship
between RNB and RAN was described by a hyperbolic
model (P o0.05). In this case, the asymptotic value
obtained when X- 1 (parameter in this model)
represents the maximum theoretical value for RNB. On
the other hand, the parameter represents the value of the
independent variable that is equivalent to the equilibrium
point, i.e., when RNB is zero. Considering this, a minimal
RAN concentration of 9.2 mg/dL was needed to provide a
positive estimate of RNB (Table 4).
There EMS was not associated (P40.05) with dietary DOM
and CP, CP:DOM, RAN concentration and BNR (Table 4).
The EMS was linearly and positively associated with NMICR
(Po0.04). There was also a tendency (Po0.08) for a linear
and negative relationship between EMS and DOMI (Table 4).
There was no association between ENU and RNB and EMS
(P40.05). The ENU was linearly associated with DOM and
NMICR, being positive with DOM (Po0.04) and negative with
NMICR (Po0.03). The CP:DOM affected ENU according to a
quadratic pattern (Po0.02) with a maximum estimate at
312 g CP/kg DOM (Table 5).
The relationships between ENU and dietary CP, RAN
concentration and DOMI were all described using hyper-
bolic models (P o0.05), such as those used for evaluating
the relationship between RNB and RAN. The estimated
values that correspond to the apparent equilibrium point
146 E. Detmann et al. / Livestock Science 162 (2014) 141153

Table 5
Summary of the linear models for describing the pattern of the apparent efficiency of nitrogen utilisation in the animal body.

Ya Xa Intercept Linear term Quadratic term RSD c

r2/R2

Estimate s.e.b P value Estimate s.e.b P value Estimate s.e.b P value

ENU DOM  0.832 0.3763 0.054 0.0020 0.00078 0.034 0.396 0.229
ENU CP:DOM  2.390 0.0866 0.022 0.0169 0.00605 0.024  0.000027 0.000011 0.018 0.315 0.612
ENU NMICR 0.668 0.2134 0.012  0.9035 0.33970 0.029 0.352 0.468
ENU RNB 0.131 0.0914 0.212
ENU EMS 0.109 0.0739 0.176

a
DOM, digestible organic matter content in diet (g/kg DM); CP, crude protein content in diet (g/kg DM); CP:DOM, ratio between CP and DOM contents
(g CP/kg DOM); NMICR, relative production of microbial nitrogen in the rumen (g microbial N/g ingested N); RNB, apparent relative nitrogen balance in the
rumen [(g ingested N g N flowing out to abomasums)/g ingested N]; EMS, efficiency of microbial synthesis (g microbial N/kg DOM); ENU, apparent
efficiency of nitrogen utilisation in the animal body (gN apparently retained in the body/g ingested N).
b
s.e., standard error of the estimate.
c
RSD, residual standard deviation of the relationship.

(i.e., ENU 0) are 108 g CP/kg DM, 6.30 mgRAN/dL of Table 6

rumen fluid, and 11.2 g DOM/kg BW (Table 3). Overall description of the dataset used for animal performance analysis.

Variablea Statistics

3.2. Animal performance Mean Minimum Maximum s n

The dataset broadly covered the characteristics of the Whole dataset

DWG 0.15  0.64 0.83 0.286 136
production systems that are used to produce grazing ADG 0.59  0.17 1.46 0.330 182
cattle. The broad ranges of ADG, SUP, SCPI, and FCP SUP 3.0 0 15.1 3.23 182
indicate that the dataset comprises of low- to high- SCP 318 70 1100 169.3 136
quality pastures and low- to high-level production, which, SCPI 0.77 0 2.91 0.719 182
FAV 5.1 1.2 14.3 2.63 167
in turn, represents a necessary feature for interpreting the
FCP 86 35 210 36.8 182
cattle performance under grazing (Table 6). Dry season
The evaluation of the generalised Malahanobis' dis- DWG 0.23  0.19 0.83 0.244 50
tances between seasons gave evidence that both transition ADG 0.29  0.17 0.93 0.228 66
seasons (dry-to-rainy and rainy-to-dry) were similar to SUP 3.1 0 10.0 2.98 66
SCP 408 130 1100 195.4 50
each other (P 40.12), but differed from the dry and rainy SCPI 0.99 0 2.91 0.847 66
seasons. There was also a significant difference between FAV 5.1 2.2 8.1 1.93 56
the dry and rainy seasons (P o0.01). The variables involved FCP 56 35 104 15.0 66
in this multivariate discrimination analysis were chosen Rainy season
because it was assumed that they could represent the DWG 0.07  0.64 0.60 0.338 54
productive conditions (DWG, FAV, and FCP) and the ADG 0.78 0.37 1.39 0.242 73
SUP 2.8 0 15.1 3.58 73
management (SUP). From this result, the pattern of DWG
SCP 266 70 560 134.1 54
was evaluated by merging the information of both transi- SCPI 0.57 0 2.47 0.538 73
tion seasons (Table 6). FAV 4.8 1.2 14.3 3.47 68
Two different approaches were used to adjust the FCP 111 71 210 35.3 73
models for describing the DWG pattern. The first one Transition seasons
was based on the utilisation of season as a class variable, DWG 0.13  0.37 0.62 0.218 32
the second one utilised the FCP as an indicator of forage ADG 0.71 0.28 1.46 0.265 43
SUP 3.3 0 10.1 3.01 43
quality variation across seasons. However, the use of
SCP 266 80 700 114.3 32
season as a class variable did not give fair results as there SCPI 0.79 0 2.67 0.688 43
were no differences regarding DWG (P40.05). Such a FAV 5.5 2.0 8.7 1.76 43
pattern seems unlikely considering the marked differences FCP 89 40 143 28.0 43
amongst seasons in regards to forage quality and animal a
DWG, differential body weight gain in relation to non-supplemen-
performance. Accordingly, the models were then adjusted
ted animals (kg/d); ADG, average daily gain (kg/d); SUP, daily amount of
using the variable FCP as the main indicator of variation in supplement per animal (g/kg of body weight); SCP, crude protein content
forage quality. The models were parameterized to pass in the supplement (g/kg dry matter); SCPI, supplemental crude protein
through the origin because the response variable is a intake (g/kg of body weight): FAV, forage availability in the pasture (ton
differential value (i.e., DWG is zero when SUP and SCPI dry matter/ha); FCP, crude protein content in forage (g/kg dry matter).

are zero).
There was a positive DWG response to increased SUP
(Po0.01). However, there were interactions between SUP
and both SCP (Po0.02) and FCP (P o0.01). The response of
DWG to SUP improved by increasing the CP content of the
supplement, but this decreased as the CP in the forage
increased (Table 7; Fig. 1).
A second model was adjusted considering only the
amount of supplemental CP (SCPI). In a similar way to
 E. Detmann et al. / Livestock Science 162 (2014) 141153 147

Table 7 It must be emphasised that the FAV was not associated

Summary of the linear models for describing the pattern of differential (P 40.05) with the animals' performance.
weight gain according to the intakes of supplement or supplemental
crude protein.
4. Discussion
Itema Estimate Standard error P value
4.1. Nutritional characteristics and efficiency of nitrogen
DWG (SUP) utilisation
SUP 0.035 0.0106 0.002
SUP  SCP 0.000074 0.000029 0.015
SUP  FCP  0.00023 0.000073 0.004 The main factor influencing intake is the digestibility of
RSD 0.064 the pasture eaten, as digestibility increases so does intake
R2 0.968 (Poppi et al., 1987; Weston, 1996). In fact, the improve-
DWG (SCPI) ment of diet digestibility, expressed by dietary DOM
SCPI 0.198 0.0286 o0.001 content, caused an increase in forage intake (Table 2)
SCPI  FCP  0.00088 0.000354 0.016
and indicated that the low digestibility of the diets was a
RSDb 0.066
R2 0.965 constraint for pasture intake.
In contrast, the relationship between forage intake and
a
DWG, differential body weight gain in relation to non-supplemen- dietary protein reinforced the idea that increases in dietary
ted animals (kg/d); SUP, daily amount of supplement per animal (g/kg of CP content positively affect the intake when the diet is
body weight); SCP, crude protein content in the supplement (g/kg dry
deficient in nitrogen. At the same time, it can cause
matter); SCPI, supplemental crude protein intake (g/kg of body weight);
FCP, crude protein content in forage (g/kg dry matter).
harmful responses to intake if an excessive CP is provided
b
RSD, residual standard deviation of the relationship. by supplementation (Table 2). Quadratic responses to
supplemental protein have been have been reported in
regards to forage intake (DelCurto et al., 1990; Sampaio
et al., 2010) as well as cattle performance (Detmann et al.,
0.6 2004).
Dry season The positive effect of supplemental CP on intake of low-
0.5 Transition seasons
Rainy season quality forages is thought to be associated with two
0.4 different aspects: the supplying of nitrogen compounds
0.3 to rumen microbes and the adequacy of the available
DWG (kg/d)

absorbed nutrients (Detmann et al., 2010; Egan and Moir,

0.2
1965; Lee et al., 1987; Leng, 1990). The primary way of
0.1 improving the utilisation of low-digestibility forages by
0 ruminants is to optimise the availability of nutrients from
0 2 4 6 8 10 12 14 16 rumen fermentation. This can be achieved by ensuring that
-0.1
there are no nutrient deficiencies to the rumen micro-
-0.2 SUP (g/kg BW) organisms. This will allow them to grow efficiently and,
through fermentative activity, extract the maximum
possible amount of energy from forage carbohydrates
0.6
Dry season (Detmann et al., 2009; Leng, 1990). A ruminal nitrogen
0.5 Transition seasons deficiency is the main constraint for intake and digest-
Rainy season
0.4 ibility of low CP forages (Egan and Doyle, 1985; Lee et al.,
1987; Souza et al., 2010). Therefore, providing an adequate
0.3
DWG (kg/d)

rumen degradable protein supply and consequently an

0.2 adequate RAN concentration is the first priority for opti-
0.1
mising fermentative digestion of forage (Detmann et al.,
2009; Lee et al., 1987; Leng, 1990; MacRae et al., 1979;
0
0 0.5 1 1.5 2 2.5 3
Satter and Slyter, 1974). In fact, the dietary CP content was
-0.1 positively associated to RAN concentration (Table 3).
-0.2 SCPI (g/kg BW) In diets based on tropical grasses, the positive effect of
nitrogen supplementation on fibre degradation is only
Fig. 1. Relationship daily amount of supplement (SUP, a) and supple- supposed to be observed with RAN concentrations of up
mental crude protein intake (SCPI, b) and the differential weight body
weight gain in relation to non-supplemented animals (DWG) (The data
to 8 mg/dL (Detmann et al., 2009). Considering the equa-
points were adjusted for random study effects. See more details of tion used for describing the relationship between RAN and
relationship in Table 7). dietary CP (Table 3), this RAN concentration would be
obtained with a dietary CP content of 99 g/kg DM (X
the previous model, the DWG was positively affected by [ln(8/3.80)]/0.0075). Despite this, the stimulus of protein
the SCPI (P o0.01), however this response is decreased on intake was verified up to 145 g CP/kg DM (Table 2),
(P o0.02) as the FCP increases (Table 7; Fig. 1). According which indicates that metabolic or post-digestive effects of
to the adjusted model, positive responses to supplemental protein are positively affecting the intake, apart from their
protein would be theoretically observed with FCP up to effects on ruminal degradation. Such a pattern agrees with
225 g/kg DM (Table 7). the statement of Egan (1977), who found that voluntary
148 E. Detmann et al. / Livestock Science 162 (2014) 141153

intake was more strongly related to total OM digestibility a

than to ruminal OM digestibility. This means that intake is 25
an integrated mechanism (Weston, 1996) and different
nitrogen effects, other than decreasing rumen fill, are
determining the control of the intake of forages, even

FDMI (g/kg BW)

20
those of low-quality (Costa et al., 2011a; Detmann et al.,
2009; Egan 1965a).
The metabolic effects of protein on intake regulation
cannot be separately evaluated because an animal's meta- 15
bolism is based on integration of different mechanisms, on
availability of several substrates and metabolites and on a
complex hormonal and biochemical signalling and regula-
10
tion. Improvements in voluntary intake have been asso- 0 50 100 150 200 250 300 350 400 450
ciated with a better adequate protein status (Egan, 1965a; CP:DOM (g/kg)
Egan and Moir, 1965; Kempton et al., 1976). Theoretically,
the term protein or nitrogen status is defined as the b
availability of different nitrogenous compounds in both 20
quantity and quality for all the required physiological 18
functions in animal metabolism. 16
Using a protein to energy ratio (P:E) seems to be more 14

plausible for understanding the metabolic effects of pro-
tein on intake, because it is a more reliable indicator of the
metabolic adequacy of the animal. Furthermore, this ratio
is a recognised parameter that regulates the voluntary
intake in ruminants (Illius and Jessop, 1996). Several
authors have tried to associate the dietary P:E with
variations in the voluntary intake of pasture (Costa et al.,
2011a; Egan, 1977; Panjaitan et al., 2010). Considering this,
the CP:DOM ratio was used as an indicator of protein to
energy status. The response of the intake was similar to
that obtained with CP alone, with the maximum intake
observed at 288 g CP/kg DOM (Table 2; Fig. 2a).
During the growth season, the tropical pastures were
not thought to be deficient in CP (Poppi and McLennan,
1995). Despite this, a more recent approach has drawn
different conclusion. Detmann et al. (2010) evaluated data
from 20 experiments that had been carried out in Brazil
during the growth season of the grasses. They concluded
that there is an unbalanced P:E in tropical pastures with a
relative excess of energy. This type of unbalance decreases
the efficiency of the utilisation of metabolisable energy
and limits intake due to heat production from energy
excess (Illius and Jessop, 1996; Poppi and McLennan, 1995).
Therefore, the increase in dietary CP will be appropriate
for the CP:DOM and will optimise intake (Table 2; Fig. 2a).
In contrast, the surplus of nitrogen in relation to the
energy availability for animal metabolism has several
negative effects on voluntary intake. These include ATP
deficiency in liver metabolism due to excessive utilisation
of the urea cycle (Visek, 1984), increased body heat
production (Poppi and McLennan, 1995) and animal indis-
position due to excess ammonia in the blood (Detmann
et al., 2007). Those negative effects are derived from diet
unbalance and can decrease the voluntary intake of forage
(Table 2; Fig. 2a).
Egan (1977) evaluated the relationship between the P:E
in the diet and the forage intake in sheep and concluded
that positive responses on intake would be observed with
ratios up to 7.5 g of digestible protein/MJ of digestible
energy. Assuming that the apparent digestibility of pasture
CP would be 0.55 and the energy content of digestible
RAN (mg/dL)
12
10
8
6
4
2
0
30 60 90 120 150 180 210
CP (g/kg DM)
Fig. 2. Relationships between the ratio of CP on digestible organic matter
content in diet (CP:DOM) and the forage dry matter intake (FDMI) and
between crude protein (CP) content in the diet and concentration of
rumen ammonia nitrogen (RAN) (see details of the equations in Tables 2
and 3, respectively).
energy in DOM would be 18.4 MJ/kg, the Egan's estimate
would correspond to approximately 250 g CP/kg DOM,
which is very close to the estimate here obtained. On the
other hand, Poppi and McLennan (1995) stated that losses
of protein will occur when the CP content of the diet
exceeds approximately 210 g/kg DOM. This value is too
low compared to the estimate here obtained (288 g CP/kg
DOM; Table 2). Despite CP content in forage, responses to
supplemental protein are also dependent on forage type as
well as the cell wall structure of the forage (Bohnert et al.,
2011). This can explain the difference between both
estimates.
The RAN concentration presented an exponential rela-
tionship with the dietary CP content (Table 3; Fig. 2b).
Some authors have used broken-line models to describe
ammonia concentration according to dietary CP. In those
approaches, there is a threshold point from which ammo-
nia accumulation becomes more intense, due to the
saturation of microbial ammonia uptake. Such a threshold
point usually lies between 100 and 140 g CP/kg DM
(Detmann et al., 2009; Satter and Slyter, 1974). In spite of
using an exponential model, which provides a smoother
description of ammonia accumulation, it seems reasonable
that the adjusted function becomes more sloped from CP
contents near to 120 g/kg DM. This supports the previous
reports (Detmann et al., 2009; Satter and Slyter, 1974).
 E. Detmann et al. / Livestock Science 162 (2014) 141153
It should be noted that RAN concentration was linearly
increased as DOM content increased (Table 2). Such a
pattern could be at least partially caused by a positive
relationship between dietary DOM and CP. Despite this, it
also agrees with the reports presented by Kennedy and
Milligan (1978) and Kennedy et al. (1981), who claimed
that urea transference from blood to rumen is positively
regulated by the amount of OM digested in the rumen,
such an increase in ruminal digestion implies a greater
microbial growth and, as a consequence, a greater ruminal
nitrogen demand.
Despite the linear relationship between RAN and forage
intake, a quadratic effect of RAN concentration on DOM
intake was verified (Table 2). Such a pattern corroborates
the previous discussion about the effects of dietary CP
content on voluntary intake. The maximum DOM intake
would be obtained with a RAN concentration close do
13 mg/dL (Table 2). This estimate is similar to the concen-
tration suggested by Detmann et al. (2009) to maximise
fibre intake (15 mg/dL), but still lower than the value
described by Leng (1990) to maximise voluntary intake
(20 mg/dL).
Taking into account the aforementioned theoretical
concept of protein status, it could be suggested that the
nitrogenous compounds would be used in different meta-
bolic functions following the order of priority to the
animal. This order would be survival, maintenance, and
production (growth, reproduction, etc.).
One of the possible high-priority metabolic functions is
the recycling of nitrogen to the gastrointestinal tract. Such
a statement seems plausible because a continuous nitro-
gen supply for microbial growth in the rumen is a strategy
for animal survival (Egan, 1965b; Van Soest, 1994). When
there is a nitrogen deficiency, the animal is able to
decrease urinary nitrogen excretion and increase the
fraction of dietary nitrogen that is recycled to the rumen
(Hennessy and Nolan, 1988). When nitrogen deficiency
becomes more severe, the animal can mobilise tissue to
sustain the mass of recycled nitrogen (National Research
Council NRC, 1985; Rufino, 2011).
In a normal feeding situation, the amount of nitrogen
that is recycled through the rumen wall seems to be
relatively constant (Marini and Van Amburgh, 2003;
Maltby et al., 2005; Reynolds and Kristensen, 2008).
Therefore, there will be a greater percentage of the
ingested nitrogen will be directed to recycling and a lower
percentage of nitrogen will be available for production.
The pattern observed for RNB can support the assump-
tions about the metabolic priorities of nitrogen. Several
estimates of negative RNB have been obtained in experi-
ments carried out in the tropics, such as those used for this
meta-analysis (Table 1). This behaviour shows that nitro-
gen flow to the abomasum can be greater than the
nitrogen intake on several occasions. In these cases, there
is a more significant dependency on recycling events to
provide an adequate nitrogen supply to the rumen. Then,
the animal will decrease the efficiency of the utilisation of
metabolisable protein for gain and may also increase the
breakdown rate of muscle protein to supply the nitrogen
demands of a higher priority (Costa et al., 2011b; National
Research Council NRC, 1985; Rufino, 2011).
149
The RNB was positively associated with nitrogen avail-
ability (CP content and RAN concentration; Tables 3 and 4)
and was not influenced by diet digestibility (Table 4). This
pattern supports the idea that a dietary nitrogen defi-
ciency is the only cause for the lack of nitrogen equilibrium
in the rumen and a negative RNB can be overcome by
using a protein supplement (Costa et al., 2011b; Figueiras
et al., 2010; Lazzarini, 2011). According to the adjusted
models (Tables 3 and 4), a null estimate of RNB would be
obtained by providing approximately 124 g CP/kg DM or
9.2 mgRAN/dL. This seems to be the threshold level of
nitrogen availability, anything less would not assure a
good availability of nitrogen for the other purposes in
animal metabolism.
The evaluation of RNB must be connected to the
evaluation of NMICR, as both are thought to give informa-
tion about the nitrogen status of the rumen environment.
Estimates of NMICR greater than one indicate a severe
dietary deficiency of nitrogen and a much greater depen-
dency on recycled nitrogen to sustain microbial growth in
the rumen (Detmann et al., 2010). Under a deficient diet-
ary nitrogen supply, there would be a net gain of nitrogen
in the rumen due to the significant assimilation of recycled
nitrogen into the microbial protein (Batista, 2012; Egan,
1974; MacRae et al., 1979; National Research Council
NRC, 2001). This is one of the main causes of the negative
RNB, as well as the high NMICR. Once again, it should be
emphasised that a supplemental nitrogen supply should
be used to avoid this kind of ruminal deficiency.
Indeed, the NMICR was negatively associated with
dietary CP content, RAN concentration, and CP:DOM.
According to the adjusted models, it should be necessary
to provide at least 69 g CP/kg DM, 3.4 mgRAN/dL, or 157 g
CP/kg DOM to obtain a NMICR lower than one (Table 3;
Fig. 3a). Despite this, it must be taken into account that in
tropical pastures, a NMICR lower than one is not an
assurance for a good nitrogen status in the rumen itself.
This is because an excessive nitrogen gain via recycling
could still be happening and a non-optimised condition
concerning the nitrogen availability in metabolism would
not be avoided. An apparent equilibrium in the rumen
regarding nitrogen inflow and outflow seems to be a
logical aim to be associated with the pattern of NMICR.
A model was then adjusted to associate NMICR and RNB
(Table 4), which indicates that under an apparent zero
RNB, the microbial nitrogen would account for 62% of the
total nitrogen flow to the abomasum. Beever and Siddons
(1986) reviewed some experiments where ruminants were
fed with different forages and stated that the microbial
nitrogen accounted, on average, to 7172% of non-
ammonia nitrogen in the abomasum. Unfortunately, the
present dataset did not comprise the abomasum flow of
ammonia nitrogen. Despite this, the estimate of Beever
and Siddons (1986) would be similar to the estimate
obtained by assuming an ammonia flow equivalent to
1020% of the total nitrogen in the abomasum, which
seems likely for animals that are fed in tropical pastures.
The NMICR was decreased as dietary DOM content
increased (Table 4). Such a relationship may be supported
by the positive correlation between DOM and dietary CP
(r 0.465; Po0.01). Therefore, the low DOM diets could
150
1.8
1.6
1.4
1.2
NMICR (g/g)
1 DOM (Table 1). The Commonwealth Scientific and Industrial
0.8
0.6
0.4
0.2
0 EMS indicates how much energy is directed toward nitro-
0 50 100 150
35
EMS (g microbial N/kg DOM)
30
25
20
15
0.2 0.4 0.6
Fig. 3. Relationships between the ratio of CP on digestible organic matter
content in diet (CP:DOM) and the relative production of microbial
nitrogen in the rumen (NMICR) and between NMICR and efficiency of
microbial synthesis (EMS) (see details of the equations in Tables 3 and 4,
respectively).
also be low CP diets, in which the net gain of nitrogen in
the rumen should be more prominent. The net gain of
nitrogen in the rumen can also increase the EMS (National
Research Council NRC, 2001), as an increased portion of
microbial protein will be synthesised from non-dietary
nitrogen. This statement is supported by the positive
relationship between EMS and NMICR and by the negative
relationship between EMS and DOM content and intake
(Table 4). On the other hand, some decreases in EMS can
be observed when DOM content is increased without a
concomitant increase in nitrogen supply. In these cases,
the EMS will decrease (Bach et al., 2005), possibly due to
an uncoupling between energy and protein in the micro-
bial metabolism. Providing a relative excess of energy
compounds can increase the occurrence of energy spilling
reactions (Strobel and Russell, 1986). Under these circum-
stances, bacteria could sustain fermentation close to the
normal levels, but still show a significant decrease in cell
yield (Clark et al., 1992; Hespell and Bryant, 1979).
According to Dewhurst et al. (2000), sometimes the
EMS is supposed to be greater with higher quality forages.
However, these authors reviewed several research studies
and confirmed that the estimation of EMS presents some
confounding between season and physiological state of the
animals. The Commonwealth Scientific and Industrial
Research Organisation CSIRO, 2007 stated that in non-
tropical pastures, the differences between seasons, which
E. Detmann et al. / Livestock Science 162 (2014) 141153
also represent differences concerning quality, could have a
great influence on EMS. Despite this, such influence would
not be significant in tropical forages. The average EMS
obtained with forages (without supplements) was approxi-
mately 146 g of microbial crude protein (MCP) per kg of
Research Organisation CSIRO (2007) summarised estimates
obtained from 22 experiments with grazing cattle and found
that the EMS varied from 99 to 191 g MCP/kg DOM.
The lack of association between the RAN concentration
and the EMS (Table 4) can be supported by the fact that
200 250 300 350 400 450
gen assimilation by the microorganisms. Despite this, EMS
CP:DOM (g/kg)
is unable to predict how much available nitrogen is
actually being used by microbes (Bach et al., 2005). This
former association could be better represented by NMICR,
which was significantly associated with RAN concentration
(Table 3).
In general, the ENU was more strongly associated to the
nitrogen supply rather than to the energy content of the
diet (Tables 3 and 5). The response on ENU according to
the variation of dietary DOM was small, with a very poor
adjustment (Table 5). The EUN was much more responsive
to DOM intake, however the adjusted model presented
some biologic inconsistencies. The predicted minimal
DOM intake associated with a EUN greater than zero was
0.8 1 1.2 1.4 1.6 1.8
incompatible with the expected energy intake to main-
NMICR (g/g)
tenance level, also the maximal predicted ENU was very
low. A second model was then adjusted without consider-
ing the random variation amongst experiments. The DOM
intake associated with maintenance (i.e., ENU 0) was
biologically coherent, however, the maximum predicted
ENU stayed out of any plausible biological limits (Table 5).
Therefore, this relationship does not seem to have enough
confidence to be discussed.
In the first instance, a quadratic model was used to
describe the relationship between ENU and CP:DOM
(Table 5). However, there is also a biological inconsistency
in such a model, because it predicts that EUN will be
negative with a CP:DOM greater than 400 g/kg. A new
model was then adjusted considering a linear-response-
plateau shape, which predicted a minimal dietary CP:DOM
of 191 g/kg in order to provide a positive ENU (Fig. 4).
0.8
0.6
0.4
0.2
0.0
EUN (g/g)
0 50 100 150 200 250 300 350 400 450 500
-0.2
-0.4
-0.6 = -1.395 + 0.0073 X, X 218.8
= 0.202, X> 218.8
-0.8
s = 0.171
-1.0
-1.2
CP:DOM (g/kg)
Fig. 4. Relationship between the ratio of crude protein on digestible
organic matter content in the diet (CP:DOM) and the efficiency of
nitrogen utilisation in the animal body (EUN).
 E. Detmann et al. / Livestock Science 162 (2014) 141153
The increase in ENU caused by improvement in nitro-
gen status supports the aforementioned arguments.
A higher nitrogen availability increases the portion of total
nitrogen that can be used for anabolic purposes and
enhances the overall efficiency of the utilisation of meta-
bolisable protein. The hypothesis regarding the improve-
ment of nitrogen status seems to be more plausible than
the direct effect of an increased metabolisable protein
supply as the increments in ENU have been obtained with
either protein or non-protein nitrogen (Costa et al., 2011b;
Egan and Moir, 1965). The benefit of a better equilibrium in
the ruminal environment was noticed through the nega-
tive relationship between NMICR and ENU (Table 5).
Furthermore, the small response in ENU according to
the variation of dietary DOM seems to support the afore-
mentioned statement of Detmann et al. (2010) about the
relative excess of energy in grazed tropical grasses. The
improvement on ENU because of nitrogen is only possible
because some energy is already available in the metabo-
lism. Therefore, the nitrogen supplementation could also
improve the efficiency of the utilisation of metabolisable
energy (Leng, 1990; Poppi and McLennan, 1995). The sum
of these arguments allows the inference that nitrogen/
protein must be the first priority for a supplementation
programme, even when tropical grasses are thought to be
of medium- to high-quality.
4.2. Animal performance
The results obtained from the analysis of the animal
performance dataset propitiated a practical approach for
the discussion on ENU. In other words, animals will
respond to protein supplementation over the year
(Table 7; Fig. 1). The response to protein supplementation
will be inversely proportional to the CP content in the
pasture (Table 7), but still positive regardless of pasture
quality. Some authors have reported that the response to
protein supplementation with regards to weight gain
should be non-linear and asymptotical (Lana et al.,
2005), which seems biologically consistent. Despite this,
none curvilinear responses were detected in the present
dataset.
The greater response to protein with low-quality
forages can be explained by nitrogen improvements in
both ruminal fermentation and metabolic nitrogen status
(Detmann et al., 2009; Egan and Moir, 1965). On the other
hand, the response to protein supplementation with med-
ium to high-quality forages (Table 7) agrees with both the
Australian (Poppi and McLennan, 1995) and Brazilian
(Paulino et al., 2008) approaches. These approaches have
confirmed that an additional weight gain of 200300 g/d
could be obtained by using supplements alongside high-
quality forages.
According to Detmann et al. (2010), although it is of
lower quality, the chemical composition of the forage
available during the dry season is rather constant because
there is an almost complete interruption of plant growth.
Despite this, during the other seasons of the year the
chemical composition of the available forage is not con-
stant as plant growth is intense and variable throughout
the season. Furthermore, some interactions of the animal
151
with the pasture can cause significant changes in the
quality of the selected forage. From a nutritional point of
view, the overall rainy season can be divided into three
different phases: the dry-to-rainy transition period, the mid-
dle rainy season, and the rainy-to-dry transition period.
The result obtained with canonical discrimination
showed that transition seasons are different from the main
seasons of the year (rainy and dry season). This provides
strong evidence to support the adoption of supplementa-
tion programs for both transition seasons, with character-
istics that should be different from the supplementation
adopted either in the dry season or in the rainy season.
Theoretically, the relationship between forage avail-
ability in the pasture and animal performance could be
described by an asymptotical model (Mannetje and
Ebersohn, 1980; Poppi et al., 1987). There is an ascendant
part of the curve where the pasture mass influences the
response curve, primarily by altering pasture height and/or
pasture density. These are both components of the pas-
tures' structure, which, in turn, will define the rate of
pasture intake (Poppi et al., 1987) and consequently, the
animal performance. In these cases, production is limited
by the non-nutritional aspects of the pasture. When the
asymptote is achieved, there is no longer a relationship
between intake and forage mass and the nutritional
aspects of forage utilisation will become predominant for
production.
The absence of a response on performance with regards
to forage availability in pasture could mean that all
experiments in the dataset were carried out on the
asymptotical phase of the theoretical model described by
Mannetje and Ebersohn (1980) and Poppi et al., (1987).
This type of statement could agree with the fact that the
issue in tropical environments is more about diet quality
than feed supply (Poppi and McLennan, 2010). Despite
this, a single estimate of DM availability in the pasture
cannot give an accurate idea about the real availability of
forage resources for production. There are a lot of addi-
tional factors that can define the production, these take
into account the structural composition of forage mass
(e.g., fraction of green leaves and dead material), the forage
allowance (the ratio of available forage on live weight of
animals) and the potential of the available mass to be used
in the gastrointestinal tract of the animals (Mannetje and
Ebersohn (1980); Barbosa et al., 2006; Paulino et al., 2008).
Unfortunately, there was not enough information in the
dataset to allow an evaluation of a more complex relationship
between pasture characteristics and animal performance.
5. Conclusions
There is a positive response to nitrogen supplementa-
tion on forage intake and efficiency of nitrogen utilisation.
This positive response, to some extent, is due to some
improvements on digestibility. Despite this, the main
benefit of supplementation is based on the improvement
of the nitrogen status in animal metabolism. There is a
concomitant positive response to nitrogen supplementa-
tion on animal performance even with forages of medium
to high-quality.
152 E. Detmann et al. / Livestock Science 162 (2014) 141153
Conict of interest statement
We declare that there is not conflict of interests in this
project.
To whom it can concern, I declare on behalf of others
co-authors of the manuscript An evaluation of the per-
formance and efficiency of nitrogen utilisation in cattle fed
tropical grass pastures with supplementation that there is
no conflict of interest with regard submission and pub-
lication of that paper. Additionally, there is no financial or
other relationship with other people or organisations that
may inappropriately influence the authors' work.
Acknowledgements
The authors wish to thank the Conselho Nacional
de Pesquisa e Desenvolvimento Cientfico e Tecnolgico
(CNPq; Programa Cincia Sem Fronteiras) and INCT Cincia
Animal for providing financial support. A special high-
lighting should be done to the Department of Agricultural
Research for Northern Sweden at Swedish University of
Agricultural Sciences for giving support to Edenio Det-
mann during his sabbatical period.
Appendix. Supporting information
Supplementary data associated with this article can be
found in the online version at http://dx.doi.org/10.1016/j.
livsci.2014.01.029.
References
Bach, A., Calsamiglia, S., Stern, M.D., 2005. Nitrogen metabolism in the
rumen. J. Dairy Sci. 88, E9E21. (eletronic supplement).
Barbosa, M.A.A.F., Nascimento Jr., D., Cecato, U., 2006. Dinmica da
pastagem e desempenho de novilhos em pastagem de capim-
tanznia sob diferentes ofertas de forragem. Rev. Bras. Zootec 35,
15941600.
Batista, E.D., 2012. Ruminal and/or Abomasal Nitrogen Supplementation
in Cattle Fed High-Quality Tropical Forage (M.S. thesis). Departa-
mento de Zootecnia/Universidade Federal de Viosa, Viosa.
Beever, D.E., Siddons, R.C., 1986. Digestion and metabolism in the grazing
ruminant. In: Milligan, L.P., Grovum, W.L., Dobson, A. (Eds.), Control of
digestion and metabolism in ruminants, Academic Press, New York,
pp. 479497.
Bohnert, D.W., DelCurto, T., Clark, A.A., Merrill, M.L., Falck, S.J., Harmon, D.L.,
2011. Protein supplementation of ruminants consuming low-quality
cool- or warm-season forage: differences in intake and digestibility. J.
Anim. Sci. 89, 37073717.
Clark, J.H., Klusmeyer, T.H., Cameron, M.R., 1992. Microbial protein
synthesis and flows of nitrogen fractions to the duodenum of dairy
cows. J. Dairy Sci. 75, 23042323.
Commonwealth Scientific and Industrial Research Organisation - CSIRO,
2007. Nutrient Requirements of Domesticated Ruminants. CSIRO
Publishing, Collingwood.
Costa, V.A.C., Detmann, E., Paulino, M.F., Valadares Filho, S.C., Carvalho, I.P.C.,
Monteiro, L.P., 2011a. Intake and digestibility in cattle under grazing
during rainy season and supplemented with different sources of
nitrogenous compounds and carbohydrates. Rev. Bras. Zootec. 40,
17881798.
Costa, V.A.C., Detmann, E., Paulino, M.F., Valadares Filho, S.C., Henriques,
L.T., Carvalho, I.P.C., 2011b. Total and partial digestibility and nitrogen
balance in grazing cattle supplemented with non-protein and, or true
protein nitrogen during the rainy season. Rev. Bras. Zootec. 40,
28152826.
Detmann, E., Paulino, M.F., Zervoudakis, J.T., Cecon, P.R., Valadares Filho,
S.C., Gonalves, L.C., Cabral, L.S., Melo, A.J.N., 2004. Crude protein
levels in multiple supplements for finishing crossbred beef cattle at
pasture during dry season: productive performance and carcass
characteristics. Rev. Bras. Zootec. 33, 169180.
Detmann, E., Paulino, M.F., Valadares Filho, S.C., Lana, R.P., 2007. Fatores
controladores de consumo em suplementos mltiplos fornecidos ad
libitum para bovinos manejados a pasto. Cad. Tc.Vet. Zootec. 55, 7393.
Detmann, E., Paulino, M.F., Mantovani, H.C., Valadares Filho, S.C., Sampaio,
C.B., Souza, M.A., Lazzarini, I., Detmann, K.S.C., 2009. Parameterization
of ruminal fibre degradation in low-quality tropical forage using
MichaelisMenten kinetics. Liv. Sci 126, 136146.
Detmann, E., Paulino, M.F., Valadares Filho, S.C., 2010. Otimizao do uso
de recursos forrageiros basais. In: Proceedings of the 3rd Interna-
tional Symposium on Beef Cattle Production, Viosa, Brazil. pp.191
240.
DelCurto, T., Cochran, R.C., Corah, L.R., Beharka, A.A., Vanzant, E.S.,
Johnson, D.E., 1990. Supplementation of dormant tallgrass-prairie
forage: II. Performance and forage utilisation characteristics in graz-
ing beef cattle receiving supplements of different protein concentra-
tions. J. Anim. Sci., 68; 532542.
Egan, A.R., 1965a. Nutritional status and intake regulation in sheep. III.
The relationship between improvement of nitrogen status and
increase in voluntary intake of low-protein roughages by sheep. Aust.
J. Agric. Res. 16, 463472.
Egan, A.R., 1965b. The fate and effects of duodenally infused casein and
urea nitrogen in sheep fed on a low-protein roughage. Aust. J. Agric.
Res. 16, 169177.
Egan, A.R., 1974. Protein-energy relationships in the digestion products of
sheep fed on herbage diets differing in digestibility and nitrogen
concentration. Aust. J. Agric. Res. 25, 613630.
Egan, A.R., 1977. Nutritional status and intake regulation in sheep. VIII.
Relationships between voluntary intake of herbage by sheep and the
protein/energy ratio in the digestion products. Aust. J. Agric. Res. 28,
907915.
Egan, A.R., Moir, R.J., 1965. Nutritional status and intake regulation in
sheep. I. Effects of duodenally infused single doses of casein, urea,
and propionate upon voluntary intake of a low-protein roughage by
sheep. Aust. J. Agric. Res. 16, 437449.
Egan, J.K., Doyle, P.T., 1985. Effect of intraruminal infusion of urea on the
response in voluntary food intake by sheep. Aust. J. Agric. Res. 36,
483495.
Dewhurst, R.J., Davies, D.R., Merry, R.J., 2000. Microbial protein supply
from the rumen. Anim. Feed Sci. Technol. 85, 121.
Figueiras, J.F., Detmann, E., Paulino, M.F., Valente, T.N.P., Valadares Filho,
S.C., Lazzarini, I., 2010. Intake and digestibility in cattle under grazing
during dry season supplemented with nitrogenous compounds. Rev.
Bras. Zootec. 39, 13031312.
Hennessy, D.W., Nolan, J.V., 1988. Nitrogen kinetics in cattle fed a mature
subtropical grass hay with and without protein meal supplementa-
tion. Aust. J. Agric. Res. 39, 11351150.
Hennessy, D.W., Williamson, P.J., Nolan, J.V., Kempton, T.J., Leng, R.A.,
1983. The roles of energy-or protein-rich supplements in the sub-
tropics for young cattle consuming basal diets that are low in
digestible energy and protein. J. Agric. Sci. 100, 657666.
Hespell, R.B., Bryant, M.P., 1979. Efficiency of rumen microbial growth:
influence of some theoretical and experimental factors on YATP.
J. Anim. Sci. 49, 16401659.
Illius, A.W., Jessop, N.S., 1996. Metabolic constraints on voluntary intake
in ruminants. J. Anim. Sci. 74, 30523062.
Kempton, T.J., Nolan, J.V., Leng, R.A., 1976. Principles for the use of non-
protein nitrogen and by-pass proteins in diets of ruminants. World
Anim. Rev. 22, 210.
Kennedy, M.P., Milligan, L.P., 1978. The effects of dietary sucrose and the
concentrations of plasma urea and rumen ammonia on the degrada-
tion of urea in the gastrointestinal tract of cattle. Br. J. Nutr. 43,
125140.
Kennedy, P.M., Clarke, R.T.J., Milligan, L.P., 1981. Influences of dietary
sucrose and urea on transfer of endogenous urea to the rumen of
sheep and numbers of epithelial bacteria. Br. J. Nutr. 46, 533541.
Lana, R.P., Goes, R.H.T.B., Moreira, L.M., Mancio, A.B., Fonseca, D.M., Tedeschi,
L.O., 2005. Application of LineweaverBurk data transformation to
explain animal and plant performance as a function of nutrient supply.
Liv. Prod. Sci. 98, 219224.
Lazzarini, I., 2011. Nutritional Performance of Cattle Under Grazing During
Dry and Rainy Seasons Supplemented with Nitrogenous Compounds
and/or Starch (Ph.D. thesis). Departamento de Zootecnia/Universi-
dade Federal de Viosa, Viosa.
Lee, G., Hennessy, D.W., Nolan, J., Leng, R.A., 1987. Responses to nitrogen
and maize supplements by young cattle offered a low-quality pasture
hay. Aust. J. Agric. Res 38, 195207.
 E. Detmann et al. / Livestock Science 162 (2014) 141153
Leng, R.A., 1990. Factors affecting the utilisation of poor-quality forages
by ruminants particularly under tropical conditions. Nutr. Res. Rev. 3,
277303.
Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., Schabenberger, O.,
2006. SASs for mixed models, 2nd ed. Cary, SAS Institute Inc..
MacRae, J.C., Milne, J.A., Wilson, S., Spence, A.M., 1979. Nitrogen digestion
in sheep given poor-quality indigenous hill herbages. Br. J. Nutr. 42,
525534.
Maltby, S.A., Reynolds, C.K., Lomax, M.A., Beever, D.E., 2005. Splanchnic
metabolism of nitrogenous compounds and urinary nitrogen excre-
tion in steers fed alfalfa under conditions of increased absorption of
ammonia and L-arginine supply across the portal-drained viscera.
J. Anim. Sci 83, 10751087.
Mannetje, L.T., Ebersohn, J.P., 1980. Relations between sward character-
istics and animal production. Trop. Grassl. 14, 273280.
Marini, J.C., Van Amburgh, M.E., 2003. Nitrogen metabolism and recycling
in Holstein heifers. J. Anim. Sci. 81, 545552.
National Research Council NRC, 1985. Ruminant Nitrogen Usage.
Academic Press, Washington, DC.
National Research Council NRC, 2001. Nutrient Requirements of Dairy
Cattle, 7th ed. Academic Press, Washington, DC.
Panjaitan, T., Quigley, S.P., McLennan, S.R., Swain, T., Poppi, D.P., 2010.
Intake, retention time in the rumen and microbial protein production
of Bos indicus steers consuming grasses varying in crude protein
content. Anim. Prod. Sci. 50, 444448.
Paulino, M.F., Detmann, E., Valente, E.E.L., Barros, L.V., 2008. Nutrio de
bovinos em pastejo. In: Proceedings of the 4th Symposium on
Strategic Management of Pasture, Viosa, Brazil. pp. 131169.
Poppi, D.P., McLennan, S.R., 1995. Protein and energy utilisation by
ruminants at pasture. J. Anim. Sci. 73, 278290.
Poppi, D.P., McLennan, S.R., 2010. Nutritional research to meet future
challenges. Anim. Prod. Sci. 50, 329338.
Poppi, D.P., Hughes, T.P., LHuillier, P.J., 1987. Intake of pasture by grazing
ruminants, Livestock Feeding on PastureSociety of Animal Production,
Hamilton, New Zealand5564 (occasional publication no.10).
153
Reynolds, C.K., Kristensen, N.B., 2008. Nitrogen recycling through the gut
and the nitrogen economy of ruminants: An asynchronous symbiosis.
J. Anim. Sci. 86, E293E305.
Rufino, L.M.A., 2011. Ruminal and/or Abomasal Nitrogenous Supplemen-
tation in Cattle Fed Tropical Forage (M.S. Thesis) Departamento de
Zootecnia/Universidade Federal de Viosa, Viosa.
Sampaio, C.B., Detmann, E., Paulino, M.F., Valadares Filho, S.C., Souza,
M.A., Lazzarini, I., Paulino, P.V.R., Queiroz, A.C., 2010. Intake and
digestibility in cattle fed low-quality tropical forage and supplemen-
ted with nitrogenous compounds. Trop. Anim. Health Prod. 42,
14711479.
Satter, L.D., Slyter, L.L., 1974. Effect of ammonia concentration on rumen
microbial protein production in vitro. Br. J. Nutr. 32, 199208.
Souza, M.A., Detmann, E., Paulino, M.F., Sampaio, C.B., Lazzarini, I.,
Valadares Filho, S.C., 2010. Intake, digestibility, and rumen dynamics
of neutral detergent fibre in cattle fed low-quality tropical forage and
supplemented with nitrogen and/or starch. Trop. Anim. Health Prod.
42, 12991310.
St-Pierre, N.R., 2001. Integrating quantitative findings from multiple
studies using mixed model methodology. J. Dairy. Sci. 84, 741755.
Strobel, H.J., Russell, J.B., 1986. Effect of pH and energy spilling on
bacterial protein synthesis by carbohydrate-limited cultures of mixed
rumen bacteria. J. Dairy Sci. 69, 29412947.
Van Houwelingen, H.C., Arends, L.R., Stijen, T., 2002. Advanced methods
in meta-analysis: multivariate approach and meta-regression. Stat.
Med 21, 589624.
Van Soest, P.J., 1994. In: Nutritional Ecology of the Ruminant2nd ed.
Cornell University Press, Ithaca.
Visek, W.L., 1984. Ammonia: its effects on biological systems, metabolic
hormones, and reproduction. J. Dairy Sci. 67, 481498.
Weston, R.H., 1996. Some aspects of constraint to forage consumption by
ruminants. Aust. J. Agric. Res. 47, 175197.