J. Anim Sci.
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J. Anim Sci. 2007. 85:984-996. doi:10.2527/jas.2005-587
© 2007 American Society of Animal Science
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
jas.2005-587v1
85/4/984    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Krizsan, S. J.
Right arrow Articles by Randby, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Krizsan, S. J.
Right arrow Articles by Randby, A. T.

ANIMAL NUTRITION

The effect of fermentation quality on the voluntary intake of grass silage by growing cattle fed silage as the sole feed1

S. J. Krizsan2 and Å. T. Randby

Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences, NO-1432 Ås, Norway


   Abstract
Top
 Abstract
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 IMPLICATIONS
 LITERATURE CITED
 
This study was designed to separate the effect of fermentation quality on voluntary intake of grass silage from other feed factors affecting intake. Variations in DMI were quantified, and the impact on intake was modeled. The relationships between individual silage components and intake were examined. A partially balanced changeover experiment with 30 Norwegian Red steers (137 ± 16.4 kg of BW) was carried out to determine the intake of 24 silages and of hay harvested from the same parent crop within 60 h. Five forages were fed at a time in each of five 3-wk periods. Every 3-wk period was preceded by 2 wk of feeding a standard silage. Silage DMI ranged from 1.79 to 2.65, with a mean of 2.38 kg·100 kg of BW–1·d–1. Hay DMI averaged 2.43 kg·100 kg of BW–1·d–1. Ranges (mean) for the composition of silages were as follows: DM, 166 to 237 (213) g/kg; water-soluble carbohydrates, 16.3 to 70.9 (33.0) g/kg of DM; acetic acid, 11.5 to 64.7 (28.6) g/kg of DM; propionic acid, 0 to 5.2 (1.0) g/kg of DM; butyric acid, 0 to 25.1 (6.0) g/kg of DM; lactic acid, 2.2 to 102 (49.3) g/kg of DM; and NH3-N (not corrected for additive-derived N), 89.3 to 255 (153) g/kg of total N. Silage DMI was closely (P < 0.05) related to DM, ADL, VFA, lactic acid, total acids, the lactic acid:total acids ratio, ADIN, NH3-N (not corrected), histamine, tryptamine, cadaverine, and the total sum of amines (the explained variation in intake ranged from 14 to 53%). The 2 best models describing silage DMI included concentrations in the silage of propionic acid, butyric acid, and lactic acid, and these models explained 75 and 84% of the variation in DMI. The strong correlation (r = 0.84, P < 0.05) between total NH3-N and butyric acid concentrations in silages indicates that these variables described the same variation pattern. The inclusion of NH3-N in the equations describing the effect of fermentation quality on DMI of low-DM grass silage was less useful than that of butyric acid. This was due to the confounded relationship between the NH3-N concentration in silages and the use of ammonium-containing preservatives and to difficulties in correcting for the added ammonium.

Key Words: cattle • fermentation quality • grass silage • prediction equation • voluntary intake


   INTRODUCTION
Top
 Abstract
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 IMPLICATIONS
 LITERATURE CITED
 
The variation in fermentation quality of grass silages affects the voluntary intake of cattle (Huhtanen et al., 2002Go). Grassland management, crop, and weather conditions influence ensiling. Different ensiling techniques also contribute to variation in fermentation quality. The composition and concentration of fermentation end-products among silages displaying secondary and extensive lactate fermentation are highly variable (McDonald et al., 1991Go). There has been no agreement on which indices of fermentation quality should be included when assessing silage DMI (SDMI; Rook and Gill, 1990Go; Steen et al., 1998Go; Huhtanen et al., 2002Go).

A modified form of the SDMI index developed by Huhtanen et al. (2002)Go has been applied in Norway to estimate the relative intake potential of silages. Based on the digestible OM in the DM and on the fermentation characteristics, the silage is given an index relative to that of a well-fermented silage of high digestibility with a SDMI of 100%. The fermentation parameters in the model chosen by Huhtanen et al. (2002)Go were adapted for silage analyses available to Finnish farmers.

Electrometric titration gives an inclusive total for formic acid and lactic acid (LA). Further, propionic acid (PA) and butyric acid (BA) are included in the determination of acetic acid (AcA), such that the method does not provide values for the individual VFA (Moisio and Heikonen, 1989Go). When using the index of Huhtanen et al. (2002)Go, concentrations in silages of more than 80 g of total acids (TA)/kg of DM and of more than 50 g of NH3-N/kg of total N are predicted to lead to reductions in SDMI. With these criteria, it is not always possible to differentiate among silage fermentation qualities. In Norway, HPLC is routinely used to analyze silage quality. With HPLC, concentrations of individual organic acids are determined separately.

In the current study, the production of silages was designed to achieve a wide variation in silage fermentation quality. Organic acids, N-fractions, and biogenic amines were assayed as fermentation quality parameters. The objectives of this experiment were to quantify and model the variation in voluntary intake of low-DM grass silages varying in fermentation quality.


   MATERIALS AND METHODS
Top
 Abstract
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 IMPLICATIONS
 LITERATURE CITED
 
Animals and Experimental Design
The procedures in the study were accepted by the local responsible Laboratory Animal Science specialist who was under the supervision of the Norwegian Animal Research Authority and registered by the Authority. This experiment was therefore carried out in agreement with the laws and regulations controlling experiments on live animals in Norway.

Twenty-four silages plus hay were fed to 30 Norwegian Red steers (initial BW 137 ± 16.4 kg) in an experiment with a partially balanced changeover design. The 30 steers were offered 5 forages at a time during 5 periods throughout the experiment. Each experimental period consisted of 2 wk of feeding all steers the same standard silage, followed by 3 wk of feeding the treatment silages. This enabled the variation in intake attributable to carryover effects to be removed and provided a tool to correct intake for between and within animal variation (because the same steers were used in successive experimental periods). The experimental design was such that 6 steers offered the same silage in one period were distributed across the 5 silages in the next period. Thus, 2 steers from each previous treatment group were placed together in the same treatment in the next period. The 2 steers together in the same treatment were split into different treatment groups in the next period. Four silages and hay were fed in the last period. The steers were allocated at random to treatment groups, and the feeding order of the silages was randomly assigned to treatment group and period before the experiment started.

Silage Preparation
The crop was harvested on June 3 to 5, 2002, from a 12-ha, second-year ley at Hellerud Research Station in Norway (60° N, 11° E) that had been fertilized in the spring with 122, 23, and 58 kg/ha of N, P, and K, respectively. The sward consisted of timothy (Phleum pratense), meadow fescue (Festuca pratensis), and red clover (Trifolium pratense) in proportions of 0.81, 0.16, and 0.02, respectively. Other species and weeds made up 0.01.

The 24 experimental silages were prepared within 60 h. After mowing, the grass was wilted for 1.5 to 19 h to achieve a target DM content of 220 to 240 g/kg of fresh material. The longer wilting times were for grass cut in the afternoon one day and harvested the next day. The DM content in the cut herbage determined the harvest times for the silages. At least 2 different grass samples were collected and dried in a microwave oven to ensure that the DM content was within the target ranges before harvest. Weather conditions before and during harvest were sunny, with no rainfall recorded.

The methods used in preparing the 24 experimental silages are summarized in Table 1Go. Six of the silages were stored in duplicate 6-m3 tower silos, 2 were prepared in a common stack silo with plastic sheeting at the bottom and between the silages to separate them in the stack, and the remaining 16 were produced as round bale silages (5 bales per treatment). A flail-harvester (Serigstad FS134, Bryne, Norway), a precision chopper (Taarup 602B, Kerteminde, Denmark), and a self-loading wagon (Krone HSL 2503, Spelle, Germany) without knives were used to collect forage for the silages preserved in the tower silos and the stack silo. The round baler (Vicon RF 130, Ontario, Canada) had a fixed chamber and 14 knives. The knives were removed during the baling of 4 of the silages.


View this table:
[in this window]
[in a new window]

  		Table 1. Preparation of the experimental silages designed to represent wide variation in fermentation quality
 
Different additives and application rates were allocated to the experimental silages. No additive was used for 4 of the silages. As a fermentation stimulant, molasses was used alone or in combination with an inoculant. Biomax (Chr. Hansen A/S, Hørsholm, Denmark), containing Lactobacillus plantarum and Pediococcus pentosaceus, was applied to 3 experimental silages. Feedtech Silage II (Medipharm AB, Kågeröd, Sweden), containing L. plantarum, Lactococcus lactis, Enterococcus faecium, Pediococcus acidilactici, the enzyme cellulase, and sodium benzoate (7.5% of the total), was used for 1 additive treatment. Pure L. plantarum was used in combination with lactose permeate, sodium propionate, and propionic acid in fermented lactose permeate, a developmental product from Yara Formates AS (Lysaker, Norway). The bacterial strain in the fermented lactose permeate was added directly before application. The acid-based additives GrasAAT (645 g of formic acid/kg, 60 g of NH3/kg), GrasAAT Plus (641 g of formic acid/kg, 93 g of PA/kg, 19 g of benzoic acid/kg, and 54 g of NH3/kg), and GrasAAT Lacto (780 g of formic acid/kg, 20 g of lactose/kg, and 70 g of NH3/kg) were used (Yara Formates AS) in 8 silages. Another fermentation inhibitor used in the experiment was Kofasil Ultra (130 g of sodium benzoate/kg, 106 g of sodium nitrite/kg, 72 g of hexamethylenetetramine/kg, and 48 g of sodium propionate/kg; Yara Formates AS). The additives were applied manually in the tower silos during filling and compaction. For the round bales, the additives were applied at the pickup on the baler using an electronically regulated pump (Serigstad DP2000, Bryne, Norway).

To minimize differences caused by maturity and the ensiling conditions for all silages, both of the harvesting machines used for silo filling and the round baler were operated at the same time in the field. Bales were transported from the field and wrapped with plastic at the storage site. Before wrapping, the weight (717 to 812 kg) and diameter were recorded for each bale, and a core sample was taken. The calculated densities for the bales were from 80 to 107 kg of DM/m3. The grass in the stack silo was compacted with a New Holland TS 100 tractor (Essex, UK), with a total weight of 6,640 kg. The compaction time corresponded to 6 to 10 min per tonne, with the shorter time for the untreated long grass.

To produce more variation in the fermentation pattern, 11 of the experimental silages were exposed to aerobic stress by keeping the herbage in the wagon 10 to 16 h before ensiling, or by not wrapping some round bales until the following day (>24 h) or until a temperature rise above 40°C was recorded inside the unwrapped bales. Grass samples were collected from every load transported to the tower silos and stack silos.

Herbage cut for hay production from June 3 to 5 was left in the field until June 7, turned twice, and thereafter further barn dried to a DM content of 860 g/kg. There was no precipitation during the field wilting. The standard silage offered to the steers during the 2-wk periods was produced as GrasAAT-treated (4 L/ton) round bales on June 14, 2002, from a sward dominated by timothy and meadow fescue. Silages produced as round bales, the silage in the stack silo (which was harvested with the self-loading wagon), and the hay were processed through a tub grinder (Serigstad RBK 1200, Bryne, Norway) before feeding commenced in the experimental periods to reduce the effects of the chopping length on feed intake.

Animal Management
All steers were castrated at an age of 2.5 to 3 mo. During the experimental periods, the steers were fed only the experimental silages plus a supplement of a commercial mineral-vitamin mixture. The steers were fed silage ad libitum (10% refusals) and offered feed twice daily at 0730 and 1400 h. Four weeks before the experiment, the steers were group-housed in slatted-floor pens fitted with Calan electronic feeding doors (Holma van Melle, Brussels, Belgium). During this period, they were trained to operate the doors. The steers were treated prophylactically for traumatic gastritis by placing magnets in the rumens of all steers. All steers were weighed at 1000 to 1100 h for the same 3 consecutive days in the last week of the preexperimental periods and in the first and the last weeks of the experimental periods.

Measurements and Chemical Analyses
Silage samples were collected on d 2, 5, 7, 9, 12, and 14 during the 2-wk feeding of standard silage, and also on d 16, 19, and 21 when the experimental feed was fed. Samples from d 2, 5, and 7 within each period were composited and analyzed for DM content. Additional composite samples collected 3 d per week were held separate. A portion of each sample was freeze-dried and used for determination of NPN and true soluble protein (TSP) with tungstic acid (Licitra et al., 1996Go). Total N in the freeze-dried samples was determined by the Kjeldahl method on a Kjeltec Auto 1030 (Tecator AB, Höganäs, Sweden) using a Cu catalyst. The OM degradability (OMD) was determined by analyzing NDF residues of the samples for ash content before and after in vitro gas production during 72-h incubations, as described by Hetta et al. (2004)Go. Another portion of each sample was oven-dried for 24 h at 67°C, equilibrated to room humidity, and milled with a Retsch SM 1 impeller-type cutting mill (Retsch GmbH & Co. KG, Haan, Germany) to a 0.75-mm particle size before analyses of NDF, ADF, ADL, ADIN, DM, and ash.

The concentration of NDF was determined according to Mertens et al. (2002)Go, ADF and ADIN were determined according to AOAC Method 973.18 (AOAC, 1995Go), and ADL was determined according to Van Soest et al. (1991)Go. Ether extract was analyzed according to AOAC Method 7.056 (AOAC, 1980Go). Contents of DM were determined by further oven-drying at 104°C for 4 h (Malkomesius and Nehring, 1951Go). Ash contents were determined by ignition of the dried sample at 600°C for 4 h (AOAC, 1975Go; Method 7.010).

The content of water-soluble carbohydrates (WSC) was analyzed according to Smith and Grotelueschen (1966)Go. Fresh silage material was extracted with water and filtered for analysis of WSC. The extracts were hydrolyzed with sulfuric acid in a water bath, neutralized with caustic soda, and then deproteinized with zinc sulfate/barium hydroxide. A sample of the resulting solution was oxidized with potassium ferricyanide in excess, and then back-titrated iodometrically. The content of WSC was calculated as glucose. Further, extracts of fresh silage were used for determination of pH, total N, NH3-N, LA, individual VFA, ethanol, and amines. Total N (Cu used as a catalyst) and NH3-N were determined by the Kjeldahl method on a Kjeltec Auto 1030 (Tecator AB). The volatile N fraction was distilled by heating the solution at pH >7 with MgO. Silage pH was determined with a glass electrode after homogenization of 10 g of fresh silage with 40 mL of distilled water. Organic acids and ethanol were analyzed by HPLC using a BioRad HPX-87H organic acid column (mobile phase, 0.00375 M H2SO4 at 0.8 mL/min) at 30 or 50°C with a UV spectrophotometric detector [formic acid (30°C), PA (30°C), BA (50°C)] or with a refractive index detector [ethanol (30°C), AcA (50°C), LA (50°C)].

The amines 2-phenyl-ethylamine, histamine, tryptamine, tyramine, putrescine, and cadaverine in silage samples were quantified by capillary zone electrophoresis using a P/ACE MDQ system (Beckman Coulter, Inc., Fullerton, CA) with silica capillaries and a UV spectrophotometer for detection. Putrescine and cadaverine were analyzed by the indirect UV method with a Waters UV-Kat-3 (Waters Corp., Milford, MA) as the electrolyte.

Herbage samples collected during ensiling were analyzed for DM, ash, total N, NDF, ADF, ADL, and WSC using the procedures described for silages. Buffering capacity was determined in the herbage samples according to Playne and McDonald (1966)Go.

Oven DM contents of grass silage were corrected for volatile losses by addition of 80% of the measured concentrations of formic acid, AcA, PA, and BA and 100% of the ethanol measured in wet samples to the DM contents determined by oven-drying (Nørgaard Pedersen, 1967Go; Mo and Tjørnholm, 1978Go). Ammonia-N values and CP values were corrected for silages to which NH3-containing additives were applied. The corrections were done on a wet weight basis, subtracting 80% of the applied N (assuming 20% field losses) from the NH3-N and total N determinations.

Silage DMI were recorded daily throughout each experimental period. Data from the final 14 d of each experimental period and from the final 7 d of each 2-wk period with standard silage were used in the statistical analysis.

Scaling of Intake with BW
No general agreement exists on whether animals of different size eat according to their BW or their metabolic BW. Mertens (1994)Go suggested that physical or physiological mechanisms of intake regulation determine the BW base that should be used to decrease the variation among steers. To best account for the variation in intake associated with differences in animal size, the best BW base for this trial was determined using a logarithmic relationship between intake and BW during the 2-wk periods of feeding standard silage using the MIXED procedure of SAS (SAS Inst. Inc., Cary, NC). Fixed effects of period, as single and quadratic, were included in the model. An autoregressive covariance structure was applied with regard to the dependent residuals owing to repeated measurements on the same steers during consecutive periods. The estimated BW base was 1.09, with a 95% confidence interval of 0.87 to 1.32. Based on this result, we decided to express feed intake on the basis of BW, and not on the basis of BW raised to a power of less than 1 (i.e., metabolic BW).

Statistical Analyses
Statistical analysis was carried out using the MIXED procedure of SAS according to the model Yij – ß (Xij – X) = µ + ti + {varepsilon}ij, where Yij is the dependent variable for treatment i on animal j; µ is the overall mean; ti is the effect of treatment i; ß is the regression of Y on X; Xij is the dependent variable for treatment i for animal j for the covariate period; and {varepsilon}ij is the residual. The model was fitted to include an effect of silage with the intake data from the 2-wk periods feeding standard silage as a covariate according to Abrams et al. (1987)Go. Linear and quadratic relationships between individual silage parameters, or groups of parameters, and SDMI were calculated using SAS. Correlations between all variables were also calculated. Multiple regression relationships between silage components and intake were evaluated using the REG procedure of SAS with forward and stepwise selection. The significance level of entry into the model was set at 0.05 for the forward and the stepwise selection, and the significance level for staying in the model was set at 0.05 in the stepwise selection method. Formic acid was not included in these regression analyses.


   RESULTS AND DISCUSSION
Top
 Abstract
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
IMPLICATIONS
 LITERATURE CITED
 
Chemical Composition of the Herbage, Hay, and Silages
The ranges (mean) for composition of the fresh crop were as follows: DM, 179 to 254 (207) g/kg; OM, 895 to 922 (911) g/kg of DM; CP, 166 to 208 (184) g/kg of DM; NDF, 538 to 631 (594) g/kg of DM; ADF, 287 to 347 (310) g/kg of DM; ADL, 41.8 to 44.7 (43.2, n = 2) g/kg of DM; WSC, 50.3 to 144 (95.4) g/kg of DM; and buffering capacity, 305 to 453 (360) mEq/kg of DM. The chemical composition of the hay and standard silage is given in Table 2Go. The standard silage was well preserved. The contents of AcA, PA, BA, and LA were 9.5, 0.0, 0.0, and 39.0 g/kg of DM, respectively. Ammonia N, corrected for additive-derived N, was 88 g/kg of total N, and 115 g/kg of total N without correction.


View this table:
[in this window]
[in a new window]

  		Table 2. Chemical composition of the hay fed as experimental feed and of the standard silage provided before the periods with the experimental feed1
 
The chemical composition of the 24 silages is given in Table 3Go. The intended lower range in DM content was not achieved; the maximum value was 237 g/kg, but the minimum value was as low as 166 g/kg. One reason for this could be the difference in drying rate in the field during the daytime. The time used to produce these silages was intended to be as short as possible, keeping the growth stage at harvest comparable among the silages, and small deviations from the intended DM ranges were therefore accepted during harvest. Silages 4, 5, and 6 had the lowest DM contents (Table 3Go). The fermentation characteristics of these silages suggest that the low DM concentrations could have been caused by DM losses because of secondary fermentation (McDonald et al., 1991Go). The contents of NDF and ADL varied widely, ranging from 476 to 601 and 28.2 to 63.9 g/kg of DM, respectively. The wide range of NDF concentrations in the silages suggests that the process of ensiling influenced this feed fraction. Concentrations of NDF in the silages that are lower than the original herbage reflect the breakdown of hemicellulose during ensiling, which provides additional substrate for the fermentation. Hemicellulose also can be degraded through hydrolysis by organic acids produced during fermentation or can be applied with silage additives (McDonald et al., 1991Go). Nitrogen bound to NDF can be broken down during ensiling and can alter the concentration of NDF compared with the initial herbage (Rinne et al., 1997Go). An increase in the concentrations of NDF and ADL in silage compared with the herbage could be due to effluent losses of soluble nutrients or DM losses from secondary fermentation (McDonald et al., 1991Go). Increases in the ADL concentration could also be due to synthesis of Maillard polymers, which have physical and chemical properties similar to lignin (Van Soest, 1994Go). In this study, ADL showed the strongest positive correlation with ADIN among all silage parameters in the correlation matrix shown in Table 6Go. The changes in the fiber fractions attributable to fermentation in this experiment could influence digestibility, as implied by the negative correlations observed with OMD.


View this table:
[in this window]
[in a new window]

  		Table 3. Chemical composition of the experimental silages fed to steers1
 

View this table:
[in this window]
[in a new window]

  		Table 6. Correlation coefficients between silage parameters in the experimental silages fed to steers1
 
Intake Data and Fermentation Characteristics of the Experimental Silages
The daily DMI and the carbohydrate fermentation characteristics for each silage quality are presented in Table 4Go. Table 5Go gives the amine content, NH3-N, and N-fractions for all silages. The silages showed a large variation in fermentation characteristics, with the exception of ADIN. Additionally, few silages had pH below 4.2 and total N as NH3 lower than 80 g/kg. The average daily DMI of the 24 silages was 2.38 kg/100 kg of BW, ranging from 1.79 to 2.65. A comparison of intakes for all silages with the silage giving the highest intake (silage 24) indicated reductions (P < 0.05) between 6 and 32%, with an average reduction of 13%. Hay DMI was 2.43 kg/100 kg of BW daily. A comparison of all silages with the hay gave 4 significant (P < 0.05) reductions in the intake of silage, ranging from 7 to 26%, with an average reduction of 16%. When the intake of silage was better than that of hay, the improvements amounted to 7 and 9% for 2 silages. The reduced intake frequently observed with cattle offered grass silage compared with animals offered fresh grass or hay prepared from the same sward has been attributed to the end-products of fermentation (Gill et al., 1988Go). The reported intake reductions have varied widely. In this study, silages 1 and 6, which gave the largest reductions in intake compared with hay, showed signs of secondary fermentation. The poor preservation was most pronounced in silage 6. Silage 2 represented silage with typical extensive LA fermentation, but it also demonstrated reduced intake compared with hay. Dulphy and Van Os (1996)Go reviewed the DMI of dairy cows fed silage of good quality compared with hay prepared from the same parent crop; they reported a 6% increase for 8 comparisons in the literature and a 5.5% decrease for 3 comparisons. Fermentation quality is of little importance for intake, compared with classical characteristics of forages such as the rate of digestion, N, and cell wall contents, when the silage is well preserved (Dulphy and Van Os, 1996Go). The reductions achieved in this experiment, when all silages were compared with the silage giving the highest intake and when all silages were compared with hay, indicated that criteria related to fermentation quality should be included in assessments of the DMI potential of grass silages.


View this table:
[in this window]
[in a new window]

  		Table 4. Silage DMI and carbohydrate fermentation characteristics of the experimental silages fed to steers1
 

View this table:
[in this window]
[in a new window]

  		Table 5. Nitrogen fraction characteristics (g/kg of total N) and amine contents (g/kg of DM) of the experimental silages fed to steers
 
Simple Regressions on Intake and Correlations Between Silage Parameters
Simple correlation coefficients for all pairs of the composition data are given in Table 6Go. In general, silage quality parameters were highly correlated. Dry matter concentration showed the strongest negative correlations with VFA and amines, and DM was positively correlated with the LA:TA ratio. The fiber components (NDF, ADF, and ADL) were negatively correlated with NPN, WSC, and LA, and positively correlated with ADIN, pH, and OM. Acid detergent fiber and ADL were more positively correlated with the amines, NH3-N (for both uncorrected and corrected values), AcA, and BA, and were more negatively correlated with OMD and the LA:TA ratio than with NDF. Acid detergent fiber and ADL also showed negative correlations with DM. Organic matter degradability was positively correlated with WSC and LA. There were strong relationships between the N-fractions with the exception of TSP, which did not show significant correlations with any other silage parameters. Apart from NPN, the N-fractions were generally negatively correlated with the LA:TA ratio and positively correlated with individual and total VFA (TVFA). There were also strong correlations between the organic acids. Individual VFA and TVFA showed a strong negative correlation with the LA:TA ratio, and AcA showed a strong positive correlation with PA.

Simple linear and quadratic regressions for all silage parameters or groups of parameters are shown in Table 7Go and 8Go. There was a positive relationship between DM content and intake (P = 0.001, r2 = 0.37; Table 7Go). The increase in SDMI with increasing DM concentration is in agreement with findings in several studies conducted on cattle (Gill et al., 1988Go; Rook and Gill, 1990Go; Rook et al., 1991Go; Steen et al., 1998Go). Most of the fermentation end-products were strongly and negatively correlated with DM content in this study. Huhtanen (1993)Go reported a trend for a smaller mean increase in SDMI with increasing DM content in more recent years (from 23% in 1978 to 7% in 1980 or later) and related this to a more widespread use of effective additives in low-DM silages. The effect of DM content on intake, modeling with materials of widely diverse fermentation quality and with a comparable small range as in this experiment, would be mostly due to the end-products of fermentation. No significant relationship was found between SDMI and NDF, ADF, CP, OM, or in vitro OMD. Silages made from the same parent crop within a limited time period would not be expected to relate to intake for these parameters. However, intake was related to ADL (R2 = 0.29) in this experiment. Steen et al. (1998)Go reported a linear negative relationship between intake and NDF, ADF, and ADL. Their study comprised silages harvested at different stages of maturity. In this study, ADL was related to poor silage fermentation, as shown by the negative correlations with WSC, LA, and the LA:TA ratio, and positive correlations with NH3-N, individual VFA, TVFA, pH, and the amines. The negative relationship between intake and ADL for values above 42 g/kg of DM was most likely due to increases in ADL caused by DM loss in secondary fermentation and ADIN formation in Maillard reactions.


View this table:
[in this window]
[in a new window]

  		Table 7. Linear (y = a + bx) relationships between silage parameters (x) and silage DMI (y; kg of DM·100 kg of BW–1·d–1) by steers fed silage as sole feed
 

View this table:
[in this window]
[in a new window]

  		Table 8. Quadratic (y = a + b1x + b2x2) relationships between silage parameters (x) and silage DMI (y; kg of DM·100 kg of BW–1·d–1) by steers fed silage as the sole feed
 
The observed net negative linear or curvilinear effect (P = 0.04 or less) of individual VFA, TVFA, and TA on SDMI is consistent with other results (Gill et al., 1988Go; Rook and Gill, 1990Go; Steen et al., 1998Go; Huhtanen et al., 2002Go). The use of the quadratic model for BA accounted for more variation (adjusted R2 = 0.32) than the relatively weak, but significantly negative, linear relationship (r2 = 0.15). Propionic acid was predicted to decrease SDMI much more than all other organic acids, TVFA, and TA and explained the most variation in SDMI among the simple regressions (r2 = 0.53). Rook and Gill (1990)Go reported a direct effect, rather than an indirect effect, of VFA on intake. Attempts to quantify the direct effect of the individual VFA by intraruminal infusion have yielded divergent results. Mbanya et al. (1993)Go could not confirm a depression of voluntary intake when AcA and PA were infused individually. Anil et al. (1993)Go obtained a dose-related reduction in intake of hay and silage when AcA and PA were infused individually, but significant effects were obtained only when physiological ranges were exceeded. Gill et al. (1988)Go infused large doses of AcA intraruminally in cattle and obtained a dose-related reduction in the subsequent short-term intake of silage. Faverdin et al. (1992)Go and Peyraud et al. (1993)Go obtained significant reductions in long-term intake with infusions of mixtures of VFA. An effect of VFA on intake seems to be more likely when the VFA are provided in combination rather than individually. Further, Huhtanen et al. (2002)Go considered it improbable that the small concentrations of PA found in silage would directly influence intake. The strongest correlations observed for PA, except the correlation of PA and TVFA, were with the individual biogenic amines (Table 6Go). However, these correlations were not stronger than those present between the individual amines and AcA. It seems more likely that the correlations describe the process of secondary fermentation rather than indicating a causal effect of PA on intake.

The quadratic model relating SDMI to LA (P = 0.001) demonstrated maximum SDMI at 53 g of LA/kg of DM, and thereafter SDMI declined with increasing concentrations of LA. Gill et al. (1988)Go, Rook and Gill (1990)Go, Huhtanen (1993)Go, Steen et al. (1998)Go, and Huhtanen et al. (2002)Go have reported varied relationships between LA and SDMI (linear and curvilinear). Huhtanen (1993)Go and Steen et al. (1998)Go observed maximum intake at LA concentrations of 41 and 80 g/kg of DM, respectively. The variable relationship between LA and SDMI has been explained by various fermentation patterns dominating in the population of silages used in the statistical analysis (Huhtanen, 1993Go). Thomas et al. (1980)Go and Choung and Chamberlain (1993)Go showed a depressive effect of LA on intake by adding concentrations to silage that could be representative of levels found in extensively fermented silages. The quadratic model for LA:TA (adjusted R2 = 0.33) improved the fit compared with the linear model (r2 = 0.21). The effect of LA:TA on intake was positive, but it declined at higher proportions.

The negative linear relationship of NH3-N corrected for additive-derived N and SDMI tended to be significant (P = 0.07). Quadratic modeling improved the fit compared with the linear model, but none of the regression coefficients could be regarded as different from zero in this model (Table 8Go). Without the correction for additive-derived N, NH3-N was not linearly related to intake (P = 0.15), but the quadratic model for NH3-N (not corrected) was significant (P < 0.01) and the fit was relatively good (adjusted R2 = 0.33). The linear models for histamine, tryptamine, cadaverine, and total amines showed consistently negative relationships with intake (P < 0.04). Tryptamine decreased SDMI much more than did histamine or cadaverine per unit increase in concentration. Of the amines, tryptamine also showed the strongest negative relationship with SDMI (r2 = 0.43). The linear negative relationship of ADIN and intake was significant (P = 0.04). The linear models for silage NPN, TSP, pH, ethanol, and WSC on intake were not significant (P ≥ 0.17).

In several studies, NH3-N has appeared to be the best variable for predicting an effect of silage fermentation quality on intake, with reported values of the coefficient of determination ranging from 0.10 to 0.44 (Gill et al., 1988Go; Rook et al., 1991Go; Steen et al., 1998Go). However, the reported relationship of NH3-N on intake has lacked agreement between experiments (Gill et al., 1988Go; Rook and Gill, 1990Go; Rook et al., 1991Go; Huhtanen et al., 2002Go). The inconsistency has been explained by different fermentation qualities represented in the data sets (Huhtanen et al., 2002Go). The use of additives in which the preservative contains ammonium will influence the concentration of NH3-N in the silages. Increased levels will be observed and the relationship with intake confounded. Clostridia are the main contributor to NH3-N (and BA) present in silage, but in some of the high-ammonia silages, clostridia is not detected, and entero-bacteria are known to be capable of producing large quantities of NH3-N during ensiling. Enterobacteria have also been shown to persist when formic acid is applied (McDonald et al., 1991Go). For these reasons, NH3-N concentrations above 50 g/kg of total N are not necessarily representative of poor fermentation quality, with a subsequent reduction in the intake potential of silages.

Ammonia N per se has not been implicated as the direct causal agent for reduced intake of silage (Rook and Gill, 1990Go). Rook and Gill (1990)Go proposed that VFA were the causative factor, whereas Steen et al. (1998)Go suggested a possible relationship between NH3-N and soluble-N in forms other than ammonia because the latter was closely related to intake in their study. Huhtanen et al. (2002)Go pointed out that the limiting effect on intake of NH3-N could be either direct or indirect because of a correlation with end-products of silage proteolysis. Van Os et al. (1996)Go and Stedilová and Kalac (2004)Go reported strong positive, significant correlations between NH3-N and amine content.

The negative effects of total amines and ADIN on intake were comparable in magnitude. Neumark et al. (1964)Go reported a negative correlation between tryptamine and the intake of silage by sheep and goats. A number of experiments have been carried out, either by infusion or as additives, to assess the effect of amines in the diet (Buchanan-Smith, 1990Go; Van Os et al., 1995aGo,bGo, 1996Go; Dawson and Mayne, 1995Go, 1996Go, 1997Go). Biogenic amines are naturally present in silage, and their presence in high concentrations most likely lowers feed intake in ruminants by reducing palatability or by influencing intermediary metabolism (Van Os et al., 1997Go; Phuntsok et al., 1998Go). A lack of impact on the intake of amines has been explained as a dietary adaptation (Van Os et al., 1997Go). Formation of ADIN during ensiling will increase the fraction of N resistant to degradation. An increased supply of protein to the small intestine has been suggested to stimulate SDMI (Huhtanen et al., 2002Go). In view of the extensive collinearity among silage components, leading to over- or underestimations of the strength of the relationships between individual parameters and intake, one should be careful about interpreting biological or causal relationships from simple regression equations.

Multiple Regression of Silage Parameters on Intake
All the silage quality parameters analyzed, and their quadratic term if the response was curvilinear rather than linear, were included in the pool of candidate regressors. The actual subset of regressors to be included in the models was chosen by forward and stepwise selection. Both selection methods terminated after adding the same 4 variables to the model. However, the bounds on condition number (15.5, 134) indicated a multicollin-earity problem; therefore, a model with 3 variables included was also evaluated. This model was identical for both selection methods, and the bounds on condition number (1.75, 13.6) did not indicate multicollinearity. The criteria for evaluating the subset regression models were the adjusted R2, root mean square error (RMSE), the predictive capability based on the PRESS statistic (Rp2), the root mean square error of prediction (RMSEP), and the variance inflation factors (VIF). The 2 best models were (SDMI in kg·100 kg of BW–1·d–1 and silage components as g/kg of DM): SDMI = 2.58 – 0.0815 PA + 0.0272 BA – 0.00168 BA2 – 0.0000378 LA2 [R2 = 0.838, RMSE = 0.0787, Formula = 0.729, RMSEP = 0.0996, VIF (PA) = 1.33, VIF (BA) = 15.5, VIF (BA2) = 15.2, VIF (LA2) = 1.47], and SDMI = 2.64 – 0.0789 PA –0.000531 BA2 – 0.0000418 LA2 [R2 = 0.747, RMSE = 0.0982, Formula = 0.513, RMSEP = 0.133, VIF (PA) = 1.33, VIF (BA2) = 1.75, VIF (LA2) = 1.44]. The high VIF values, in the equation including all 4 variables, for the BA and BA2 terms indicated a multicollinearity problem. Excluding the BA term in the second equation removed this problem. Near-linear dependence among regression variables is thought to impair the predictive ability of the regression model. Regardless, the equation including both BA and BA2 (Formula = 0.73) explained 20% more of the variability in predicting new observations than did the equation from which BA was removed (Formula = 0.51). The values of RMSE and RMSEP were also lower for the equation including both BA and BA2, indicating a better model (smaller error term) with improved precision to predict SDMI.

Attempts to isolate the specific fermentation products involved in controlling SDMI have failed to pinpoint conclusively any particular product as responsible for reducing SDMI. Huhtanen et al. (2002)Go aimed to include the most important end-products of fermentation in the SDMI index. The choice of model parameters describing fermentation quality was limited to those that could be measured by electrometric titration. The parameters of this model were a quadratic term for TA and the natural logarithm of NH3-N, which explained 47% of the variation in SDMI within the experiment. The best model in the work of Huhtanen et al. (2002)Go, with regression coefficients that could be regarded as different from zero (P < 0.05), included a quadratic effect of LA, an effect of PA, and the natural logarithm of NH3-N, all with a negative influence on intake. This model explained 51% of the total variation in SDMI within the experiment. There was a notable resemblance between this model and the equations obtained in our study. The main difference was the inclusion of BA instead of NH3-N. Ammonia N, which was not important in explaining intake variation in the multiple linear regression equations in this study, and which may only indirectly limit SDMI, was strongly and positively correlated with BA. Among the parameters positively correlated with NH3-N, BA with r = 0.80 (corrected) and 0.84 (not corrected) was much more important than any of the individual amines or the total sum of the amines (r = 0.46 to 0.67). This suggests that BA could mediate the indirect effect of NH3-N on intake. This is in line with the results by Rook and Gill (1990)Go, which suggested VFA as the causative factor based on a high, positive correlation with NH3-N. Ammonia N was always more important than VFA in the multiple regression models of Huhtanen et al. (2002)Go. The silages studied by Huhtanen et al. (2002)Go included mainly well-fermented silages, as reflected by a median value of BA of 0.5 g/kg of DM.

It is tempting to conclude that the equation including both BA and BA2, with the high predictive ability, was the best model for this data set. However, the indicated collinearity, and the knowledge of the restrictions of multiple linear regression as an appropriate statistical method for a data set consisting of several intercorrelated parameters, make the decision less straightforward (Montgomery et al., 2001Go). In situations like this, in which 2 alternative regression models have been developed, it would be useful to compare the prediction performance on new data.


   IMPLICATIONS
Top
 Abstract
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 IMPLICATIONS
LITERATURE CITED
 
In this study, in which an early harvested grass crop was ensiled, neither the nitrogenous compounds nor water-soluble carbohydrates were of importance in predicting silage dry matter intake. The individual organic acids propionic, butyric, and lactic acids were the best descriptors of feed intake in this study. However, because of the confounded relationship among ammonia values not corrected for, which added in the ammonium-containing preservatives and feed intake, and the difficulties in practice of correcting ammonia concentrations for added ammonium, butyric acid was considered more appropriate when describing silage intake. However, further work is recommended, either by modeling feed intake with more advanced statistical methods and validation techniques, or by collecting independent data to evaluate the predictive ability by a more conservative method.


   Footnotes
 
1 The Research Council of Norway provided funding for this study. The authors thank the personnel at Hellerud Research Station for assisting in experimental work, and L. Norell for conducting the part of the statistical analysis regarding the BW base related to intake. Back

2 Corresponding author: sophie.krizsan{at}umb.no

Received for publication October 11, 2005. Accepted for publication December 4, 2006.


   LITERATURE CITED
Top
 Abstract
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 IMPLICATIONS
 LITERATURE CITED
 


Abrams, S. M., H. W. Harpster, P. J. Wangsness, J. S. Shenk, E. Keck, and J. L. Rosenberger. 1987. Use of a standard forage to reduce effects of animal variation on estimates of mean voluntary intake. J. Dairy Sci. 70:1235–1240.[Abstract/Free Full Text]

Anil, M. H., J. N. Mbanya, H. W. Symonds, and J. M. Forbes. 1993. Responses in the voluntary intake of hay or silage by lactating cows to intraruminal infusions of sodium acetate or sodium propionate, the tonicity of rumen fluid or rumen distension. Br. J. Nutr. 69:699–712.[CrossRef][Medline]

AOAC. 1975. Official Methods of Analysis. 12th ed. Assoc. Off. Anal. Chem., Washington, DC.

AOAC. 1980. Official Methods of Analysis. 13th ed. Assoc. Off. Anal. Chem., Washington, DC.

AOAC. 1995. Official Methods of Analysis. 16th ed. Assoc. Off. Anal. Chem., Arlington, VA.

Buchanan-Smith, J. P. 1990. An investigation into palatability as a factor responsible for reduced intake of silage by sheep. Anim. Prod. 50:253–260.

Choung, J. J., and D. G. Chamberlain. 1993. Effects of addition of lactic acid and post-ruminal supplementation with casein on the nutritional value of grass silage for milk production in dairy-cows. Grass Forage Sci. 48:380–386.

Dawson, L., and C. S. Mayne. 1995. The effects of either dietary additions or intraruminal infusion of amines and juice extracted from grass silage on the voluntary intake of steers offered grass silage. Anim. Feed Sci. Technol. 56:119–131.

Dawson, L. E. R., and C. S. Mayne. 1997. The effect of infusion of putrescine and gamma amino butyric acid on the intake of steers offered grass silage containing three levels of lactic acid. Anim. Feed Sci. Technol. 66:15–29.

Dawson, L., and C. S. Mayne. 1996. The effect of intraruminal infusions of amines and gamma amino butyric acid on rumen fermentation parameters and feed intake of steers offered grass silage. Anim. Feed Sci. Technol. 63:35–49.

Dulphy, J. P., and M. Van Os. 1996. Control of voluntary intake of precision-chopped silages by ruminants: A review. Reprod. Nutr. Dev. 36:113–135.[Medline]

Faverdin, P., B. Richou, and J. L. Peyraud. 1992. Effects of digestive infusions of volatile fatty acids or glucose on food intake in lactating or dry cows. Ann. Zootech. 41:93.

Gill, M., A. J. Rook, and L. R. S. Thiago. 1988. Factors affecting the voluntary intake of roughages by the dairy cow. Pages 262–279 in Nutrition and Lactation in the Dairy Cow. P. C. Garnsworthy, ed. Butterworths, London, UK.

Hetta, M., A. M. Gustavsson, J. W. Cone, and K. Martinsson. 2004. In vitro degradation characteristics of timothy and red clover at different harvest times. Acta Agric. Scand. A 54:20–29.

Huhtanen, P. 1993. Factors influencing forage intake. Pages 103–127 in Proc. Nova Scotia Forage Counc. Symp. A. H. Freeden, ed. Dep. Agric. Marketing, Nova Scotia, Canada.

Huhtanen, P., H. Khalili, J. I. Nousiainen, M. Rinne, S. Jaakkola, T. Heikkila, and J. Nousiainen. 2002. Prediction of the relative intake potential of grass silage by dairy cows. Livest. Prod. Sci. 73:111–130.

Licitra, G., T. M. Hernandez, and P. J. Van Soest. 1996. Standardization of procedures for nitrogen fractionation of ruminant feeds. Anim. Feed Sci. Technol. 57:347–358.

Malkomesius, P. E., and K. Nehring. 1951. Chemische untersuchung von futtermitteln. Page 15 in Handbuch der landwirtschaftliegen Versuchs- und Untersuchungsmethodikt. R. Herrmann, ed. Neumann Verlag, Berlin, Germany.

Mbanya, J. N., M. H. Anil, and J. M. Forbes. 1993. The voluntary intake of hay and silage by lactating cows in response to ruminal infusion of AcA or propionate, or both, with or without distension of the rumen by a balloon. Br. J. Nutr. 69:713–720.[CrossRef][Medline]

McDonald, P., A. R. Henderson, and S. J. E. Heron. 1991. The Biochemistry of Silage. Chalcombe Publications, Marlow, UK.

Mertens, D. R. 1994. Regulation of forage intake. Page 450 in Forage Quality, Evaluation and Utilization. G. C. Fahey, Jr., ed. ASA-CSSA-SSSA, Madison, WI.

Mertens, D. R., M. Allen, J. Carmany, J. Clegg, A. Davidowicz, M. Drouches, K. Frank, D. Gambin, M. Garkie, B. Gildemeister, D. Jeffress, C. S. Jeon, D. Jones, D. Kaplan, G. N. Kim, S. Kobata, D. Main, X. Moua, B. Paul, J. Robertson, D. Taysom, N. Thiex, J. Williams, and M. Wolf. 2002. Gravimetric determination of amylase-treated neutral detergent fiber in feeds with refluxing in beakers or crucibles: Collaborative study. J. AOAC Int. 85:1217–1240.[Medline]

Mo, M., and T. Tjørnholm. 1978. Losses of carbon-containing substances during dry matter determination by oven-drying. Acta Agric. Scand. 28:196–202.

Moisio, T., and M. Heikonen. 1989. A titration method for silage assessment. Anim. Feed Sci. Technol. 22:341–353.

Montgomery, D. C., E. A. Peck, and G. G. Vining. 2001. Introduction to Linear Regression Analysis. 3rd ed. John Wiley & Sons, New York, NY.

Neumark, H., A. Bondi, and R. Volcani. 1964. Amines, aldehydes and keto-acids in silages and their effect on food intake by ruminants. J. Sci. Food Agric. 15:487–492.

Nørgaard Pedersen, E. J. 1967. Alkohol i ensilage. Tidsskr. Planteavl. 71:355–358. (Abstr. in English)

Peyraud, J. L., B. P. Widyobroto, and P. Faverdin. 1993. Effects of addition of water-filled bladders or volatile fatty acids into the rumen on digesta passage in the dairy cow eating ad-libitum. Ann. Zootech. 42:187.

Phuntsok, T., M. A. Froetschel, H. E. Amos, M. Zheng, and Y. W. Huang. 1998. Biogenic amines in silage, apparent postruminal passage, and the relationship between biogenic amines and digestive function and intake by steers. J. Dairy Sci. 81:2193–2203.[Abstract]

Playne, M. J., and P. McDonald. 1966. Buffering constituents of herbage and of silage. J. Sci. Food Agric. 17:264–268.

Rinne, M., S. Jaakkola, and P. Huhtanen. 1997. Grass maturity effects on cattle fed silage-based diets. 1. Organic matter digestion, rumen fermentation and nitrogen utilization. Anim. Feed Sci. Technol. 67:1–17.

Rook, A. J., and M. Gill. 1990. Prediction of the voluntary intake of grass silages by beef cattle. 1. Linear regression analyses. Anim. Prod. 50:425–438.

Rook, A. J., M. Gill, R. D. Willink, and S. J. Lister. 1991. Prediction of voluntary intake of grass silages by lactating cows offered concentrates at a flat rate. Anim. Prod. 52:407–420.

Smith, D., and R. D. Grotelueschen. 1966. Carbohydrates in grasses. 1. Sugar and fructosan composition of the stem bases of several northern-adapted grasses at seed maturity. Crop Sci. 6:263–266.

Stedilová, S., and P. Kalac. 2004. The effect of lactic acid bacteria, inoculants, and formic acid on the formation of biogenic amines in grass silages. Arch. Anim. Nutr. 58:245–254.[CrossRef][Medline]

Steen, R. W. J., F. J. Gordon, L. E. R. Dawson, R. S. Park, C. S. Mayne, R. E. Agnew, D. J. Kilpatrick, and M. G. Porter. 1998. Factors affecting the intake of grass silage by cattle and prediction of silage intake. Anim. Sci. 66:115–127.

Thomas, C., M. Gill, and A. R. Austin. 1980. The effect of supplements of fishmeal and lactic acid on voluntary intake of silage by calves. Grass Forage Sci. 35:275–279.

Van Os, M., J. P. Dulphy, and R. Baumont. 1995a. The effect of protein degradation products in grass silages on feed intake and intake behavior in sheep. Br. J. Nutr. 73:51–64.[Medline]

Van Os, M., J. P. Dulphy, and R. Baumont. 1995b. The influence of ammonia and amines on grass silage intake and intake behaviour in dairy cows. Ann. Zootechnie 44:73–85.

Van Os, M., M. Jailler, and J. P. Dulphy. 1996. The influence of ammonia, biogenic amines and gamma amino butyric acid on grass silage intake in sheep. Br. J. Nutr. 76:347–358.[CrossRef][Medline]

Van Os, M., A. M. Van Vuuren, and S. F. Spoelstra. 1997. Mechanisms of adaption in sheep to overcome silage intake depression induced by biogenic amines. Br. J. Nutr. 77:399–415.

Van Soest, P. J. 1994. Nutritional Ecology of the Ruminant. Cornell University Press, Ithaca, NY.

Van Soest, P. J., J. B. Robertson, and B. A. Lewis. 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74:3583–3597.[Abstract]


This article has been cited by other articles:


Home page
 J DAIRY SCIHome page
P. Huhtanen, M. Rinne, and J. Nousiainen
Effects of Silage Soluble Nitrogen Components on Metabolizable Protein Concentration: A Meta-Analysis of Dairy Cow Production Experiments
J Dairy Sci, March 1, 2008; 91(3): 1150 - 1158.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
jas.2005-587v1
85/4/984    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Krizsan, S. J.
Right arrow Articles by Randby, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Krizsan, S. J.
Right arrow Articles by Randby, A. T.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS