HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brosh, A. Articles by Aharoni, Y. Search for Related Content PubMed PubMed Citation Articles by Brosh, A. Articles by Aharoni, Y.

Energy cost of cows’ grazing activity: Use of the heart rate method and the Global Positioning System for direct field estimation¹

A. Brosh^*,2, Z. Henkin^*,, E. D. Ungar, A. Dolev, A. Orlov^*, Y. Yehuda and Y. Aharoni^*

^* Beef Cattle Section, ARO, Newe Yaar Research Center, P.O. Box 1021 Ramat Yishay 30095, Israel; and Department of Natural Resources, ARO—the Volcani Center, P.O. Box 6, Bet Dagan 50250, Israel; and MIGAL—Galilee Technological Center, Qiryat Shemona, P.O. Box 90000, Rosh Pinna 12100, Israel

	Abstract

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

 
This study with grazing beef cows on the range was designed to explore the possibility of determining incremental energy expenditure (EE) in standing, traveling, and grazing relative to that in lying down, by means of continuous monitoring of EE, location, and activity by the heart-rate method, with Global Positioning System (GPS) collars, and by motion sensors in the GPS collars, respectively. Cows were observed on Mediterranean foothill rangeland covered with herbaceous vegetation through 4 seasons of the year. There were 2 stocking rate treatments, and 14 statistical models were evaluated, including one that was a stepwise model. Total daily EE (TEE) was affected by many interdependent factors apart from activity, including season, stocking rate, herbage quality, standing biomass, and reproductive state of the cow. Each model included all activity variables, plus some of the other factors. Across seasons and treatments TEE, in kJ/(kg of BW^0.75 · d), ranged from 469 in densely stocked, nonlactating cows in June to 1,092 in sparsely stocked, lactating cows in April. The cows’ daily vertical and horizontal movements ranged from 75 to 174 m and from 1.5 to 4.2 km, respectively. Within a day, time spent traveling (without grazing) ranged from 0 to 32 min, and grazing time ranged from 4.4 to 12.1 h. Cows spent less time grazing (P < 0.001) in the summer, when herbage quality was low, than in winter and spring. Relative to the baseline EE while lying down, the daily increment incurred by grazing ranged from 13 to 48 kJ/(kg of BW^0.75 · d), and that incurred by grazing, standing, and traveling combined ranged from 38 to 74 kJ/(kg of BW^0.75 · d) or 5.8 to 11.4% of TEE. In conclusion, the estimates of activity costs yielded by 11 of the models were similar to one another, whereas those yielded by the stepwise model and the remaining 2 models were 20% smaller. The cost of grazing activity was estimated to be 6.14 J/(kg of BW^0.75 · m), and that of locomotion during grazing was 6.07 J/(kg of BW^0.75 · m), which agree with values obtained for animals and humans by means of a treadmill. The experimental and statistical approach tested here yielded fairly reliable estimations of energy costs of activities in grazing cows.

Key Words: cattle • energy cost • energy expenditure • global positioning system • grazing • heart rate

	INTRODUCTION

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

 
Grazing activity may increase the energy requirements of animals on pasture compared with those of confined animals of a similar status. Published estimates of such increases range from minor (DiMarco and Aello, 1998) to a 50% increase (Osuji, 1974; Havstad and Malechek, 1982).

Extrapolation of treadmill-derived energy expenditures (EE) to estimate grazing EE (Lachica et al., 1997b; DiMarco and Aello, 1998) is problematic because animals on a treadmill walk without grazing. Moreover, calculation of the entire-day energy cost of activities requires records of activities throughout the entire 24 h because cows are active at night. Modern electronic miniaturization enables attachment of a Global Positioning System (GPS) receiver and motion sensors to the cows, to record their activities, walking distances, and location in the grazed area (Ungar et al., 2005). Combining such a method with EE determinations derived from heart rate (HR) measurements (Brosh et al., 2004) might enable estimation of the energy cost of each activity and calculation of the entire day cost of activities of free-ranging cows.

The validity and accuracy of such estimations depend on the extent to which the analysis model accounts for the many factors that affect the total daily EE (TEE) in free-ranging conditions. Various approaches to this problem may introduce a variety of biases into their estimations. Our hypothesis is that energy costs of activities of grazing cows could be determined from individual cow EE and partitioned daily activities.

The objectives of the current study were 1) to characterize activities and EE throughout the day in grazing cows under various specified typical conditions; 2) to compare energy costs of specific activities, as estimated by several partial models and by the stepwise model; and 3) to quantify the contribution of the other factors to the TEE of free-ranging cows.

	MATERIALS AND METHODS

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

 
Experimental Site

The experiment was conducted at the Karei Deshe Experimental Range, situated in Lower Galilee, in Israel, near the junction of the Jordan River and the Sea of Galilee (latitude 32 ° 55' N, longitude 35 ° 35' E, altitude 80 to 150 m above sea level). The topography is hilly, with slopes generally < 10%. The site has a Mediterranean climate characterized by wet and mild winters with mean minimum and maximum temperatures of 7 and 14°C, respectively. The average annual rainfall is 570 mm, falling mostly from November to March. The summer is dry and hot, with mean minimum and maximum temperatures of 19 and 32°C, respectively. There are at least 6 mo (May to October) of dry weather, with little or no rain.

Soils in the experimental area are brown basaltic protogrumusols of variable depth but seldom deeper than 40 cm and with a rock cover of about 30% (Gutman and Seligman, 1979). The number of species in the sward was large: a total of 166 species were identified in the experimental plots during 1994 to 1997. Major species were Hordeum bulbosum L., Echinops spp., Psoralea bituminosa L., and many annual species, including such palatable pasture plants as Avena sterilis L., Bromus spp., Trifolium spp., and Medicago spp. (Seligman et al., 1989). These were also the major species grazed by the cows. Under the heavy stocking rate treatment, slight reductions were measured in the cover rate of tall grasses; this was correlated with an increase in the cover rate of prostrate annual legumes and less palatable plant groups, such as annuals and perennial thistles, crucifers, and forbs (Sternberg et al., 2000).

Stocking Rate Treatments and Measurement Months

In 2003, 2 plots of 27.5 and 28.2 ha each were selected from a long-term grazing experiment initiated in 1993, in which the first plot was managed at a high stocking rate (HSR) and the other at a low stocking rate (LSR) of 0.9 and 1.8 ha/cow, respectively, for the nonlactating cow, and for cow-calf pairs when the cows were lactating. Actual stocking rates in the current study were less than the programmed ones, and were 1.02 and 2.01 ha/cow in HSR and LSR plots, respectively, because 6 cows in the groups that were assigned to the experiment, 4 in the HSR group and 2 in the LSR group, were found not to be in late pregnancy just before the grazing season and therefore were excluded from the experiment. The maximal vertical differences in the 2 plots were 70 m.

Each grazing season began in the winter, about 2 mo after the first significant rainfall (which occurs on the average in mid November), when the green herbage biomass had reached a level of about 800 kg of DM per hectare. Grazing of the HSR treatment was generally terminated in August, when the cows showed a BW loss in 2 consecutive weighings, within a 21-d interval, which generally happened when the remaining herbage mass had diminished to 400 to 700 kg/ha. In the LSR treatment, grazing was terminated at the end of October, either when the herbage biomass was small or when the rain began. When not in the grazing plots, the cows were fed in confinement.

Measurements on cows were made during 4 periods in 2003: February 9 to March 6 (February), April 7 to April 24 (April), June 3 to June 26 (June), and August 3 to August 21 (August). The herbage biomass and composition in the plots were determined for the same months. A protein supplement of poultry litter (CP, 20.3%; NDF, 33.7%; and calculated ME, 6.52 MJ/kg of DM; Brosh et al., 2004) was offered ad libitum in the plots during June and August.

Animals

All procedures involving animals were approved by the Israeli committee for animal care and experimentation.

Mature medium-frame-size cows approximately 7 yr of age were used in the study; they were Simford cross-bred, a Simmental x Hereford cross, with some blood of local breeds from the eastern Mediterranean region. Twenty-seven and 14 cows were assigned to the HSR and LSR plots, respectively, and good data recordings were obtained from a total of 21 cows: 12 in the HSR plot and 9 in the LSR. Four to 6 cows were monitored under each combination of season and stocking rate treatment; in the HSR and LSR treatments, their mean BW across seasons were 432 ± 15 kg and 429 ± 17 kg, respectively.

All cows were in late pregnancy in February, all were lactating in April, and all were dry in June and August. The grazing season began on January 16, 2003, when all cows were in late pregnancy, and ended on September 1, 2003; all the calves were weaned before June 1, 2003. In the HSR and LSR treatments, ADG of the cows throughout the grazing season were 83 and 88 g/d, respectively; the average calving dates were February 25, 2003, and March 3, 2003, respectively; the lactation periods were 87 and 89 d, respectively; the calf weaning BW were 105 and 100 kg, respectively, and their BW gains were 868 and 812 g/d, respectively.

Three reproductive states were defined for the data analysis: 1) nonlactating, nonpregnant, or pregnant from 1 to 180 d, 2) pregnant from 180 d up to calving, and 3) lactating. Pregnancy of 0 to 180 d was included in the definition of state 1 because a previous study (Brosh et al., 2004) showed that in terms of HR, EE, or oxygen consumption (VO₂) per heartbeat (O₂ pulse), this state did not differ from the nonpregnant state.

Measurements on Cows

Energy Expenditure Measurements. Energy expenditure was measured by using the HR method as described by Brosh et al. (1998, 2002, 2004). The method is based on measuring HR throughout several days and then relating it to EE by multiplying HR by the ratio of oxygen consumption to HR, the O₂ pulse, as measured simultaneously over a short time (10 to 15 min), and by the energy dissipated (heat production) by the animal in consuming a unit of oxygen (kJ/L of O₂). The HR was recorded with a Polar instrument (Polar Electro Oy, Kempele, Finland), a model T51H HR transmitter, and a watch model S610 data logger. The devices were attached to the thorax behind the forelegs by means of a specifically designed elastic belt (Pegasus, Eliad, Israel). Heart rate was measured continuously for 4 d in each measurement month; the data logger was programmed to record HR at 1-min intervals.

Oxygen consumption was measured with a facemask open-circuit respiratory system (Fedak et al., 1981). The accuracy of the system was checked gravimetrically by injecting nitrogen (N₂ recovery) into the facemask, as described by McLean and Tobin (1990). The measurements on each cow were carried out in a cattle squeeze chute located at the boundary of the 2 plots, between 0900 and 1500 in each of the 4 mo immediately after the 4-d HR measurement. Cows from the 2 plots were measured alternately.

During the O₂ pulse measurement, the cow was standing, and the exhaled air was sucked from its face and passed through an air filter and a model 50D-15 mass flow meter (McMillan, Georgetown, TX). Changes in oxygen concentration in relation to the environmental concentration were measured with a Servomex model 1400 paramagnetic oxygen analyzer (Servomex, Crowborough, East Sussex, UK), and the temperature and relative humidity of the expired air were measured with HygroClip S electronic sensors (Rotronic AG, Basserdorf, Switzerland) to enable calculation of VO₂ under standard conditions of temperature, pressure, and zero humidity. These data were recorded and saved simultaneously by the dataTaker model DT 50 (dataTaker, Rowville Melbourne, Victoria, Australia) data acquisition and data logging system and by laptop, at 5-s intervals; HR was recorded in parallel at the same 5-s intervals.

Each cow’s O₂ pulse was calculated, and then the EE throughout the day was calculated by multiplying the HR (beats/min) as measured throughout the day by O₂ pulse (L/beat), as measured over a short interval, and by the constant value of 20.47 kJ/L of O₂ consumed (Nicol and Young, 1990). The cow’s EE for each 5-min interval during the day was calculated for the metabolic BW (MBW, kg of BW^0.75) as kJ of · MBW^{– 1} · d^{– 1}. The use of a constant daily value of O₂ pulse, measured individually in each month, is supported by our previous findings (Brosh et al., 1998, 2002, 2003; Arieli et al., 2002; Aharoni et al., 2003).

Anderson and Kothmann (1977) reported a slow walking speed (1 to 3 km/h) and a short daily walking distance (3.5 to 5 km) for free-ranging cows. In the current study, we measured the speed of locomotion during grazing as 138 to 238 m/h, which suggests that periods of intense physical exertion that could modify the O₂ pulse (Brosh et al., 1998) are rare and, therefore, may be ignored in the calculation of the EE of these cows. The assumption that slow walking during grazing does not affect the O₂ pulse is now supported by Berhana et al. (2005).

Determination of Location and Activities. Cows undergoing EE measurements were equipped with Lotek GPS collars of the 2200 Series (Lotek Engineering Inc., Newmarket, Ontario, Canada). These collars included 2 captive-ball tilt switches (CW 1600 Series, Comus International, Clifton, NJ) to sense left-right and forward-backward movements. The collars were programmed to store GPS locations and motion sensor counts at 5-min intervals. Four physical activities of the cows were defined: lying, standing; traveling (i.e., walking without grazing), and grazing. The activities were determined from the GPS location and the motion sensor with the aid of calibration equations developed by Ungar et al. (2005). Horizontal and vertical locomotion distances during each 5-min interval were computed with the ArcView 3.2 software (Geographic Information Systems), from layers containing the GPS and topographic data. Vertical locomotion during each 5-min interval could not be estimated accurately because of the limited accuracy of the vertical component of the GPS location; therefore, we assumed that vertical distances corresponding to slopes of more than 20% were not correct, and we set those distances at one-fifth of the horizontal distance traveled during that interval. Small horizontal and vertical movements that were recorded occasionally in cows that were known to have been lying or standing were ignored.

Habitats

A Geographic Information Systems layer was also prepared in which the experimental plots were allocated to 5 habitat classes according to slope and rock cover (Henkin et al., 2003). The classes were 1) not stony, flat land; 2) stony, gentle slopes; 3) stony, steep slopes; 4) rocky, steep slopes; and 5) boulders on steep slopes. This classification was used to identify the habitat occupied by the cow during each 5-min interval.

Herbage Analyses

During each measurement month, herbage was sampled by clipping it to ground level in sixteen 0.25 x 0.25-m quadrates along each of 5 transects (i.e., 80 samples per plot). The location of each sampling point was determined by GPS during the clipping, and the samples were chosen to represent the various habitat classes. Samples were dried at 60°C for 48 h. All clipped quadrates were used for the statistical analysis conducted to characterize the plot (treatment) biomass. For the laboratory chemical analyses, all herbage samples clipped from each habitat class in each plot in each season were averaged by weight to represent one habitat sample. This was done because only one independent value could be used to represent a state of habitat, treatment, and month in the statistical analysis. The total number of combinations used for the analysis was 40: 5 habitat classes x 2 plots x 4 seasons.

Analyses were carried out for DM (105°C), ash (550°C), CP (Kjeldahl, AOAC, 1980), NDF, IVDMD, and in vitro OM disappearance (IVOMD). The IVDMD and IVOMD were analyzed with the Ankom Daizy system (Ankom, Fairport, NY). The NDF analyses were performed with the Ankom Fiber Analyzer according to Van Soest et al. (1991); in Ankom’s modification the sample is enclosed in a filter bag and remains in it throughout the process. The biomass was calculated on the basis of 100% DM in the sample weight (Table 1). The average standing biomass, its average composition, and calculated ME concentrations in each of the plots in each month (Table 1) were calculated by 3 procedures. The first procedure aimed at representing the average biomass weight of the whole plot and was calculated from all of the transect samples of that plot. The second procedure aimed at representing the herbage according to the time spent by the cows in each habitat within the plot; therefore, the standing biomass and its composition were averaged separately for each of the 5 habitats within the plot; then these values were assigned to each 5-min interval record of each cow, according to the identified habitat occupied by this cow during that interval. The third procedure was similar to the second procedure, except that only time the cows spent grazing in each habitat was used instead of the whole time during which the cow occupied the habitat. Because the 3 calculations were based on the same data source, it was not possible to statistically examine the differences among their results.

Database for Model Analyses

A total of 24,666 records were collected, each of which included the following categorical factors: cow identification (21 levels), treatment (2 levels), habitat (5 levels), month (4 levels), time of day (24 hourly levels), reproductive state (3 levels), and activity (4 levels). Each record included the following continuous variables: standing biomass DM (kg/ha); ash, NDF, and CP concentrations in the herbage DM (g/kg); ME content in DM (MJ/kg); horizontal and vertical distances traveled (meters per 5-min interval); and EE (kJ · MBW^{– 1} · d^{– 1}). Plot herbage data were calculated for each habitat, treatment (plot), and season combination.

In addition, 1 factor and 1 variable were added to the database. The 2-level reproductive state (RS2), indicating lactating or nonlactating, replaced the 3-level reproductive state in some analyses because we (Brosh et al., 2004) have not previously found indications that the EE/MBW in nonlactating, late-pregnancy cows is significantly greater than that of nonpregnant cows. This finding apparently contradicts the findings of Moe and Tyrrel (1972), who observed increasing ME requirements in cows as pregnancy progressed: up to 75% above maintenance at term and about 40% above maintenance 1 mo before term. However, when the period of late pregnancy occurs simultaneously with the period of grazing on lush, high-quality herbage (late winter and spring), increased energy requirements of the cows do not necessarily demand increased herbage intake, and fetal requirements might be met at the expense of body energy retention, with no or little change in ME intake and EE. Therefore, we decided to leave the 2 options open, to compare the influence of inclusion of either 3 or 2 reproductive states on the estimation of activity effects. The variable positive vertical distance was added because it had been reported for goats (Lachica et al., 1997a) and pigs (Lachica and Aguilera, 2000) that walking downhill did not use less energy than walking on level ground; for this variable all the measured negative vertical distances were replaced with zero values. This variable replaced the vertical distance in some of the models tested.

Analysis Models

Total daily EE is strongly affected by level of energy intake (Brosh et al., 2004), and this, in turn, depends on both herbage quality and availability, and on reproductive state of the cows (Aharoni et al., 2004). In addition, the pattern of the cows’ EE changes during the day is affected by factors such as weather, day length, and pasture conditions, many of which are interrelated. The usual approach to estimate the effects of possibly interrelated variables is using the step-down multiple regression (stepwise model) that firstly includes all the assumed effects together. In this procedure, every variable that is not found to be significant in a full model is excluded from the final best model. However, such a model that takes into consideration the contributions of all these factors together to a baseline EE, and attributing additional EE to activity is still expected to yield estimates that are biased because of confounding effects among the remaining factors and variables. On the other hand, if some of these factors are not taken into account by the model, then estimations of energy costs of the cows’ activity may also be biased, but differently so. Therefore we constructed several partial models in addition to the stepwise model to get a range of estimates of activity cost across these models.

All analyses were performed with the Genstat software, seventh edition (Lawes Agricultural Trust, 2003). Effects on HR and EE were analyzed by the REML procedure, with fixed effects of treatment, month, time of day, and all their interactions. This analysis did not address effects of activity on HR or EE. Models that analyzed effects of activity on EE together with other factors and variables are presented in Table 2, which summarizes the variables and factors addressed by each of the models; the formal equations of these 14 models are listed in the Appendix. The REML procedure was used in all of these models. Because the herbage availability and quality depended strongly on the treatment, month, and habitat, the so-called independent variables and factors were divided into 4 groups. The first group, M1, was composed of treatment, month, habitat, time of day, and all the interactions among treatment, month, and time of day. Because strong correlations were evident among some of the chemical composition variables and ME concentrations (Table 3), we divided models that ignored the treatment and season effects, and referred to herbage availability and quality instead, into 3 further groups: M2, which referred to biomass and chemical composition; M3, which referred to biomass and ME concentration; and M4, which referred to chemical composition and ME concentration, together with biomass. Activity and distance variables were included in all the models as additional independent variables; the dependent variable in all the models was the EE. The random effect of the individual cow was also included in all the models.

Treatment, habitat, and month effects were not included in groups M2, M3, or M4, but the effect of time of day was included in all of them to ensure that these models did not account for the interactions of time of day with treatment and month. Because the diurnal pattern of EE changed with season some loss of the explained variation was expected.

Model M2 tested the effects on EE of time of day, standing biomass, chemical composition, (ash, NDF, and CP concentrations), reproductive state, nature of activity, and horizontal and vertical distances. In Model M2a, positive vertical distance replaced vertical distance; Models M2b and M2c also used positive vertical distance, but the effect of reproductive state was omitted from M2b, and RS2 replaced reproductive state in Model M2c.

In M3, ME replaced the chemical composition variables that were included in the M2 group. Reproductive state and vertical distance were included in M3, but positive vertical distance replaced vertical distance in M3a, M3b, and M3c; reproductive state was omitted from M3b, and RS2 replaced reproductive state in M3c.

Chemical composition and ME content were included in M4, and all the reproductive state factors and vertical distance variables in M4a, M4b, and M4c corresponded to those in M2a, M2b, and M2c.

The last model listed in Tables 2, 5, and 6 was designated as the stepwise model; it was achieved by stepwise multiple regression. In the first stage we included all the "independent" factors and variables as fixed effects in the REML analysis, with a random effect of cow, and then we excluded from the model only those factors and variables that were not found to have significant effects on EE (i.e., reproductive state, and CP and ME concentrations). The analysis was then performed again without these factors and variables. It should be noted that no assumption was made in this approach as to possible confounding effects among the remaining independent components of the model.

The daily energy cost of each kind of activity equals the activity cost (i.e., the EE in excess of that in the lying down state) multiplied by the time the cow devotes to this activity throughout the day. In calculating the daily locomotion (horizontal and vertical) cost, horizontal and vertical distance measurements replaced time measurements, and daily horizontal and vertical distances replaced the daily time duration of activity in the calculation. Data calculated by models M1a, M2a, M3a, and M4a (Table 5) for the activity energy costs above that of the lying down state, and the data of daily time duration of the activity, and of daily horizontal and vertical distances (Table 7), were used to calculate the daily energy cost of each kind of activity (Table 8). The EE to be attributed to activity and locomotion, expressed as a fraction of the TEE, was calculated for each month within each treatment (Table 8). Analysis of variance was used to test effects of treatment and model on the estimated daily energy cost of all the activities (Sum-Eac) and their proportion within the entire daily EE (Sum-Eac per TEE). The design of this ANOVA test was factorial, 2 treatments on 4 models, with months as blocks.

	RESULTS

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

The average daily HR and EE in the HSR and LSR treatments during 4 mo are listed in Table 4. Average daily EE was high in February, even greater in April (P < 0.001), and low in June and August (P < 0.001), when the quality of the herbage was low, compared with February and April, when the herbage was lush. Cows’ average daily EE was greater (P < 0.01) in the LSR than in the HSR treatment in all months. Similar findings were received for HR except in August, when the cows’ HR in the HSR treatment was smaller (P < 0.001) than that of the cows in the LSR treatment.

Typical 1-d records of the activities, the locomotion distances during 5-min intervals, and EE for an individual cow are depicted in Figure 1 for April (1a) and June (1b). The statistics on the dependency of EE on activity are presented below in Table 5. It can be observed that in both months increased EE was associated with grazing (square symbols) and locomotion distance, and decreased EE was associated with time of rest, when the cows lay down (diamond symbols). A unique, previously unreported occurrence of grazing activity in the middle of the night was recorded from 0145 to 0240 in April, and from 2240 to 2310 in June. This hitherto unknown behavior was recorded for all cows in all seasons and treatments; it was also expressed in a small increase of EE in the middle of the night (Figure 2).

View larger version (42K): [in this window] [in a new window]   Figure 1. Activity records of cow 19 (A) and cow 394 (B) during one day: April 21 and June 10 for figures (A) and (B), respectively; both cows were in the high-stocking-rate plot. Energy expenditure (thick line) and locomotion distance (thin line) are depicted on the 24-h scale. Body activity symbols: lying; • standing; traveling (walking without grazing); grazing. EE = energy expenditure; MBW = metabolic BW.

View larger version (33K): [in this window] [in a new window]   Figure 2. Diurnal patterns of energy expenditure (EE) by cows grazing at 2 stocking rates (HSR = high stocking rate; LSR = low stocking rate) during 4 mo of the year. Symbols: HSR February ; LSR February •; HSR April ; LSR April ; HSR June ; LSR June ; HSR August ; LSR August . Vertical bars on the HSR February curve represent SEM ( ± 17.5) for the interaction treatment x season x time of day (P < 0.001).

Estimations of Energy Cost of Activity

Tables 5 and 6 present EE attributed to cows’ activities, herbage biomass, and chemical composition, as estimated by the 14 models. Because there were many variables and factors affecting EE according to these models, we have presented the results of the same analyses in 2 consecutive tables. Table 5 presents effects associated with the animal: activity and reproductive state, Table 6 presents effects associated with the plot: herbage standing biomass, chemical composition, and ME content.

Energy cost of activities, locomotion, and reproductive state, as estimated by 14 models, are listed in Table 5. The statistics across models, presented in the lower part of Table 5, were performed on 13 models; the stepwise model, which used a different statistical approach from that of the other models, was excluded. However, activity effects estimated by the stepwise model were almost identical to those estimated by M1 and M1a, which suggests that inclusion of herbage biomass and composition together with components of plot, month, and habitat in the same model had negligible influence on estimation of activity effects. Mean values of estimated coefficients of the EE cost of standing, traveling, and grazing in addition to the EE of lying, respectively (CV in parentheses), were 46.1 (6.0%), 60.9 (14.9%), and 89.6 (9.5%) kJ/(MBW · d). The CV of the horizontal locomotion coefficient, averaged across these 13 models, was 8.8%. The CV of the estimated effects of vertical locomotion was small, but it was based on only 3 models, whereas that of positive vertical locomotion (n = 10) was 43%. The CV of reproductive state effects, expressed at 3 levels (n = 5), were also high, which sheds doubt on the reliability of these estimates. The CV of the estimate of the EE added by lactation to that of a nonlactating cow was small, but the reliability is also questionable because only 3 estimates of this factor were available.

Differences among the estimated effects of activity according to the 14 models were tested by grouping the models into 5 groups: the M1 group (n = 2), M2 group (n = 4), M3 group (n = 4), M4 group (n = 3), and the stepwise model (n = 1). Analysis of variance in a completely randomized design was used to test the differences among these groups in estimates of activity and in the fraction of variance accounted for by the entire model (R²).

Coefficient of determination (R²) differed among models (average of 0.731 ± 0.042) and was greater (P < 0.05) for the M1 group and for the stepwise model than for the M2, M3, and M4 groups. The estimated energy cost of grazing, standing, and horizontal locomotion yielded by the stepwise model and by the M1 group were smaller than those yielded by the M2 group (P < 0.001), M3 group (P < 0.001), and M4 group (P = 0.002), by approximately 20% on the average.

Estimations of the Effects of Herbage Quality and Standing Biomass on EE

Effects on EE of the standing biomass and composition (chemical and ME) of the herbage, as estimated by 11 models, are listed in Table 6. The CV of all these variables were high, above 45% across the models. However, within each model the SE of most of these estimates were small. There were obviously trade-offs among the interrelated effects of biomass, chemical composition, and ME content, and also between these variables and the reproductive state. Estimates of the effects of ME and chemical composition were considerably smaller when these variables were considered together (the M4 group) than those yielded by models that included only chemical composition (M2 group) or only ME (M3 group). For the models that included 3 levels of reproductive state, the coefficients of the effects of biomass were estimated to be not different from zero in models M2, M2a, and M4a, and significant (P < 0.001) in M3 and M3a. The estimated effect of biomass was high (P < 0.001) with models that did not include the reproductive state factor (M2b, M3b, and M4b) and was moderate but still highly significant (P < 0.001) with those that included the 2-level reproductive state factor (M2c, M3c, and M4c). These findings indicate considerable interaction between the biomass variable and the reproductive state factor.

Estimations of Effects of Treatment, Habitat, and Month on EE

The estimated effects on EE of treatment, habitat, month, and the interactions of treatment x month and treatment x month x time of day were essentially the same according to models M1 and M1a, with the same small SE. According to both models, the TEE of LSR cows was estimated to be greater (P < 0.001) than that of HSR cows by 123 ± 15 kJ/(MBW · d). The TEE levels in April, June, and August, in comparison with that in February, were estimated to be +283, – 509, and – 371 (SEM ± 10.7) kJ/(MBW · d), respectively (P < 0.001). In relation to that in habitat 1, the EE in habitats 2, 3, 4, and 5 were – 40, – 20, – 19, and – 14 (SEM ± 4.0) kJ/(MBW · d), P < 0.001, P < 0.01, P < 0.05, and P < 0.10, respectively. The SE for the interactions of treatment x month and treatment x month x time of day, respectively, were 21 and 17 kJ/(MBW · d), and both interactions were highly significant (P < 0.001).

Daily Activities of Cows on Pasture

The daily time (h/d) spent by cows lying, standing, traveling, grazing, and the horizontal and vertical distances of locomotion are listed in Table 7, according to treatment and season. Horizontal and vertical distances are listed separately according to their association with grazing or traveling, and their totals are listed as well. In both treatments and during all months, the cows spent very little time traveling (locomotion without grazing), but they covered considerable distances during this activity. Cows spent 4.5 to 8.3 h/d lying, with no difference among seasons or treatments; they spent 5 to 7 h/d standing during February and April, compared with 10 h/d or more during June and August (P < 0.05), and spent more daily hours grazing (P < 0.05) during February and April than during June and August. The treatments showed no differences in grazing time between LSR and HSR cows, despite the much smaller standing biomass available to the HSR cows. Average travel distance of the LSR cows tended (P = 0.13) to be larger than that of the HSR cows. It is interesting to note that the distances traveled did not differ between spring and summer, despite the marked decreases in grazing time. These observations suggest that other factors besides standing biomass governed the grazing behavior of the cows and their ability to collect their desired amounts of feed. The most obvious factor is herbage nutritional quality. The amount of total vertical positive locomotion, which was inaccurately estimated by the GPS for the short recording time, ranged from 75 to 174 m daily, and was greater (P < 0.05) in winter and spring than in summer.

Energy Costs of Daily Activities

Daily energy costs of specific activities throughout the months (Table 8) were calculated by multiplying the coefficients of the energy costs of specific activities and distances (Table 5), as estimated by 4 of the models, by the daily time spent on each activity and distances traveled (Table 7). For the calculations, one model from each group—M1, M2, M3, and M4 (Table 5)—was chosen to represent a range of putative baselines from which the activity costs were estimated. All the represented models included the same measures: type of activity, horizontal movement, and positive vertical movement. Thus, differences among models in estimated daily costs of activity (Table 8) arose from different coefficients estimations (Table 5), all multiplying the same times of activity and distances (Table 7).

The estimated Sum-Eac was affected by treatment (P < 0.001) and model (P < 0.001) with no interaction between them (P = 0.97); the respective SEM were 1.274, 1.802, and 2.548. Sum-Eac as estimated by M1a was smaller (P < 0.01) than that estimated by the other models with no difference among M2a, M3a, and M4a. The Sum-Eac per TEE was not affected by treatment (P = 0.785), but it was affected by the model (P < 0.001), with no interaction of treatment on model (P = 1.0); the respective SEM were 0.181, 0.256, and 0.362. The Sum-Eac per TEE that was estimated by M1 was smaller (P < 0.01) by 19% than that of the other models, with no differences among estimations by M2a, M3a, and M4a.

	DISCUSSION

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

 
To our knowledge, the current study is the first in which the behavior and EE of grazing cows were measured simultaneously throughout the day. Thus, some of our findings, such as the occurrence of late-night grazing, cannot be compared with those of previous studies. However, some of the present findings can be compared with reported indirect estimations of the energy cost of grazing and with previous laboratory-based direct measurements of the energy cost of traveling.

Herbage Quality, Plot Biomass, and Cows’ Behavioral Responses

Table 1 presents the general dietary conditions provided by each treatment in each month of measurements. The cows’ intake preferences and plot biomass variations could affect the intake quality and biomass quantity in each grazed area. For example, in August the standing biomass in one part of the HSR plot was 700 kg of DM/ha, whereas the whole-plot average and the habitat averages weighted according to the duration of cow presence or time spent grazing were 452, 353, and 375 kg of DM/ha, respectively, indicating that the biomass in more frequently grazed areas was smaller than that in areas less frequently grazed. Standing biomass and nutritional values that were derived from samples by means of calculations weighted according to the time the cows occupied the habitats or grazed on them were almost identical. This suggests that effects of herbage conditions on EE are expected to yield similar estimations whether they are based on calculations made according to the time the cows spent staying in each habitat or according to the time spent grazing each habitat.

Cows in both treatments spent more time grazing (Table 7) in the seasons of lush herbage (February and April) than in those of dry herbage (June and August), mostly at the expense of standing time. This finding contrasts with previous reports that heifers (Nestis, 1979) and sheep (Arnold, 1960) increased their daily grazing time when forage availability decreased. Di-Marco et al. (1996) also reported that the EE of grazing cattle was greater when herbage was scarce than when it was abundant. Our observations that grazing time and horizontal distances of the LSR cows were not different from those of the HSR cows suggest that the time required for feed collection did not limit the feed intake of the HSR cows. It is also possible that the greater standing biomass in the LSR plot was associated with increased grazing time because of the more varied feed selection in this plot, which decreased the feeding rate and increased the difference in diet quality between that calculated from clipped samples and that of the actually grazed diet. This interpretation is indirectly supported by Aharoni et al. (2004), who measured the mean retention time of digesta in grazed cows under conditions similar to those of the current study. The smaller EE of the HSR cows compared with that of the LSR cows during all months (Table 4) suggest that their feed intake was also smaller (Brosh et al., 2004). However, this hypothetically reduced feed intake would have resulted from a combination of low biomass and low-quality herbage, which may have discouraged the cows from grazing.

Coefficients of the Energetic Costs of Activity

The estimated energy costs of activity differed only to a small extent among the 14 models. The energy costs of standing, grazing, and horizontal locomotion had a CV of less than 10% across models, with the CV of the cost of traveling being 14.9% (Table 5). This observation suggests that there were only small interactions between these variables and other factors in the models, such as season, treatment, and herbage quality. However, the estimations of the energy spent on the main activities, grazing, standing, and horizontal locomotion (Table 5) yielded by models M1, M1a, and by the stepwise model were 19% smaller than those of the other models. The diurnal EE pattern that was used as the baseline for the calculations of EE cost of activities in these first 3 models differed from that calculated by the other models by accounting for hour of the day, treatment, month, and all their interactions, unlike the baseline in the other models that ignored effects of month and treatment on the diurnal EE pattern. Because these diurnal patterns changed among months and treatments (see Figure 2), the values of R² yielded by M1, M1a, and the stepwise model were greater than those obtained from the other models (Table 5). In other words, the fraction of the variance that could be attributed to herbage biomass and composition by model groups M2 to M4 was smaller than that explained by month, treatment, and their interactions with hour of day in group M1 and in the stepwise model. This structural difference between the M1 group and the stepwise model on one hand and the other models on the other hand may explain the differences in activity cost estimations, but nevertheless, we cannot say that the greater R² indicates more accurate estimations of the EE costs of activity. Therefore, we suggest that the estimations of the energy costs of activity according to each of these models should be considered reliable within a range of ± 10%. Vertical distances within the plots were small, and the estimations of the energy cost of vertical locomotion or positive vertical locomotion showed a high CV across models, in addition to their unacceptably low values. In studies with goats (Lachica et al., 1997a) and with pigs (Lachica and Aguilera, 2000), the energy costs per meter of vertical locomotion were estimated to be about 10 times greater than those of horizontal locomotion. Therefore, it appears that our present models greatly underestimated the costs of vertical locomotion. This underestimation, the large SE within each model and the high CV across models probably resulted from the very small extent of vertical locomotion, together with the poor estimation of the vertical locomotion itself (i.e., the inaccuracy of the evaluation of vertical locomotion for short time intervals by the Lotek GPS device under our experimental conditions). In addition it is possible that grazing cows prefer to walk on level or moderate slopes unless they are forced to change their behavior due to the location of water sites or other management reasons. Consequently, it is also possible that the daily energy costs of vertical locomotion are small and negligible compared with other activities in moderate hilly grazing areas.

Effects on EE of Seasons, Reproductive State, and Herbage Quality and Availability

In contrast to the estimates of the energy costs of activities, those of the effects of standing biomass, herbage composition, and ME concentration exhibited very high CV across models. The effect of biomass was estimated as nonsignificant or very limited in models that included 3 reproductive states, greater in models that included 2 reproductive states, and most prominent in models that did not include a reproductive state factor. The effect of advanced pregnancy (180 d to calving) on EE was estimated to be negative (compared with that of a nonpregnant, nonlactating cow) in models that included herbage composition but not ME, and also in the model that included both chemical composition and ME, but its effect was positive in models that included ME but not chemical composition. The estimated effects of ME were smaller when chemical composition was included in the model than when it was not. Similarly, the estimated effects of chemical composition were greater in models that did not include ME than in those that did.

All these observations indicate considerable interactions among the effects of the variables of biomass, chemical composition, and ME, and between them and the effect of reproductive state. It is suggested, therefore, that the effect of ME concentration, which presumably determines the level of ME intake by the cows (Brosh et al., 2004), cannot be separated from those of chemical composition, which also determines the level of ME intake. On the basis of our present observations, in which both herbage quality and reproductive states were seasonal, no distinction could be made between the effects of herbage quality and those of reproductive state. Nevertheless, any combination of factors that was used in constructing the models that were implemented in this study yielded a baseline EE from which reliable estimates of the additional energy costs of activities could be obtained.

Energy Costs of Locomotion and Grazing

In this study the energy cost of locomotion was treated as an addition to that of each activity (i.e., standing, traveling, or grazing). Thus, there was no distinction between the cost of locomotion and that of the concurrent activity. The cost of horizontal movement, averaged over 5 min, which is 1/288 of a whole day, was estimated at 0.836 kJ · m^{– 1} · MBW^{– 1} per 5 min (Table 5); transformation of this value to energy cost per meter gives a value of 2.90 J/(MBW · m). However, our models attributed some additional energy cost to the activity. Therefore, to estimate the total energy cost of grazing activity, in addition to that of standing, in terms of cost per meter of distance covered, the specific additional cost of grazing, per hour of activity, over and above that of standing, should be multiplied by the grazing time and divided by the distance that was walked while grazing. The result should be added to the cost of locomotion per meter to obtain the total costs of the grazing activity. Such a calculation yielded an estimated cost of grazing of 12.21 ± 1.39 J/(MBW · m) over and above the cost of standing, averaged over all the combinations of treatment x season [range 10.5 to 16.0 J/(MBW · m)]. This value includes the energy cost of walking and of the additional activities associated with grazing (i.e., head movement and chewing). The costs of the activities of locomotion and of grazing-during-locomotion could be estimated as the difference between the specific costs of traveling and of grazing [averages over all models: 60.9 and 89.6 J/(MBW · m), respectively; Table 5]. The EE attributed to the grazing activity per se, to be discounted from the total activity EE during grazing, was calculated for all the treatment x season combinations; this was done by multiplying this difference by the fraction of the day that was occupied by grazing, and the net EE in locomotion activity was estimated by subtracting this discount from the total activity EE and dividing the result by the distance. The estimated net cost of locomotion according to this calculation was 6.07 ± 0.65 J/(MBW · m), averaged over all the combinations of treatment x season (range 5.49 to 7.36). The estimated average additional cost of grazing over the net cost of locomotion was 6.14 ± 1.26 J/(MBW · m) (ranged from 5.03 to 8.64).

Several studies have measured the locomotion energy costs of animals and humans on a treadmill. Most of these studies reported the energy cost per kilogram of BW; therefore, the values quoted here are transformations of the originally reported values to MBW. Lachica et al. (1997a) found a cost of 8.15 J/(MBW · m) for goats (35 kg of BW); Lachica and Aguilera (2000) reported costs of 7.9 and 8.2 J/(MBW · m), respectively, for light (41 kg) and heavy (84 kg) pigs; Pearson et al. (1998) measured costs of 4.3 and 4.7 J/(MBW · m), respectively, for donkeys and ponies (both having approximately 200 kg of BW); DeJaeger et al. (2001) reported a cost of 4.5 J/(MBW · m) for children of various ages and adults (BW ranging from 18 to 65 kg) walking at their most efficient speed. This most efficient speed increased with age, but its energetic cost did not differ among ages. Among these quoted values, the smallest related to the heaviest animals, but a comparison of costs for goats, pigs, and humans did not reveal systematic relations between BW and the energy cost of locomotion. This finding is in accordance with those of Taylor et al. (1982), who found almost equal costs of locomotion across a very wide range of BW of avian and mammalian species. Thus, our present estimation of the cost of walking on pasture of 6.07 J/(MBW · m) falls well within the range of reported costs obtained from treadmill measurements. To the best of our knowledge, the current study is the first to have estimated the additional cost of the grazing activity over traveling [6.14 J/(MBW · m)] by using data obtained from animals grazing in their natural environment. Although Osuji (1974) estimated the energy cost of eating by ruminants on pasture, his estimations were based on data that were not collected directly in the field.

Daily Energy Costs of Activity

There are considerable differences among the reported estimations of the daily energy cost of animals’ activities in the pasture. Blaxter (1967) suggested that energy requirements increased by 11 and 15% for grazing sheep and cattle, respectively, at maintenance level. Osuji (1974) suggested that activities in the pasture might increase energy requirements of ruminants by 25 to 50% above those of confined animals and that a considerable part of the increase should be attributed to the cost of herbage collection in addition to the locomotion activity. Havstad and Malechek (1982) reported that the EE of grazing heifers was 46% above that of confined heifers consuming similar forage. Lachica et al. (1997b) reported increases of only 14.2 and 8.7%, respectively, in the energy requirements of goats that grazed and traveled distances of 5.8 km/d in summer and 3.5 km/d in autumn. More recently, Lachica et al. (1999) found that the grazing activity of goats maintained on a mountain range caused their energy requirements to increase by 46.6 and 31.6%, respectively, in the summer and autumn, when they traveled distances of 13 and 8 km/d, respectively. DiMarco and Aello (1998) concluded that the energy cost of walking in the pasture could have only a minor effect on the energy requirements of grazing cattle.

Some of these differences among the findings of various studies may be attributed to differences in activity among species, to differing conditions of the pastures, to differences in the hours of activity, and in the horizontal and vertical distances walked in the various studies. However, we suggest that some of these differences arise from the application of activity coefficients obtained on a treadmill to activities that were observed in the field. Our present estimations of the energy costs of activities were obtained under pasture conditions, and the activity coefficients and the activity records were calculated from data obtained simultaneously in the field. Our estimated values of the costs of activity of 38 to 74 kJ/(MBW · d) under the various combinations of months and stocking rates (Table 8) correspond to total energy costs of 8.5 to 16.5% above the maintenance requirement, which was assumed to be 450 kJ/(MBW · d) in confinement. These estimations, obtained from cows that traveled 1.5 to 4.2 km/d, are comparable with those determined by Lachica et al. (1997b) in goats that traveled 3.5 to 5.8 km/d. We suggest that some of the above studies overestimated the energy costs of grazing activities, probably because of differences among the estimation methods.

Relations of Energy Cost of Activity to Total Energy Expenditure

There was a positive association between the daily EE for activity and the TEE of cows (Tables 4 and 8). Thus, the total daily EE for activity (Sum EAC, kJ/(MBW · d)) in HSR (51.6) was smaller (P < 0.001, SEM = 1.274) than in LSR (58.7) cows in most of the months (Table 8); this is in accordance with the smaller TEE of cows in the HSR treatment. However, the fraction that accounted for the daily activity per TEE increased from a range of 5.8 to 7.6% in the lush herbage season, when ME intake and EE, and therefore TEE, were high, to values of 7.1 to 11.4% in the dry herbage season, when ME intake and EE, and therefore TEE, were low (Table 8).

	IMPLICATIONS

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

 
Fourteen models were tested for their ability to isolate the energy costs attributable to specific activities of cows maintained on plots that offered a wide range of conditions. Energy expenditure was affected by many factors other than activity; many of these factors were interdependent. All the models yielded similar estimates of the activity costs; therefore, it is concluded that energy cost of various activities has little interaction with other tested variables, and consequently all the models yielded fairly reliable estimations of those costs. The locomotion energy cost obtained in the current study was consistent with estimates obtained in treadmill studies on various animal species and humans. All the activity costs, taken together, accounted for 5.8 to 11.4% of the daily energy expenditure across seasons and treatments; this proportion increased as herbage quality decreased. Cows responded to decreased herbage quality by decreasing their grazing time and energy expenditure.

	APPENDIX

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

 
Formal descriptions of models analyzed by the REML procedure of Genstat 7:

Model M1

where

K

constant;

A_i

random effect of cow i, i = 1 to 21;

T_j

fixed effect of treatment j, j = 1 to 2;

M_k

fixed effect of month k, k = 1 to 4;

H_l

fixed effect of hour l, l = 0 to 23;

HB_m

fixed effect of habitat m, m = 1 to 5;

AC_n

fixed effect of activity n, n = 1 to 4;

DH

horizontal distance traveled (m per 5 min);

c1

regression coefficient for horizontal locomotion;

DV

vertical distance traveled (m/5 min.);

c2

regression coefficient for vertical locomotion; and

e

random residual effect.

Model M1a

Identical to Model M1, except that the variable DV was replaced with the variable DVP = positive vertical distance traveled.

Model M2

where

K

constant;

A_i

random effect of cow i, i = 1 to 21;

H_j

fixed effect of hour j, j = 0 to 23;

RS_k

fixed effect of reproductive state k, k = 1 to 3;

AC_l

fixed effect of activity l, l = 1 to 4;

BM

standing biomass (kg of DM per ha);

Ash

ash concentration (g/kg of DM);

NDF

NDF concentration (g/kg of DM);

CP

crude protein concentration (g/kg of DM);

DH

horizontal distance traveled (m per 5 min);

DV

vertical distance traveled (m per 5 min);

c1 to c6

regression coefficients for BM, ash, NDF, CP, DH, and DV, respectively; and

e

random residual effect.

Model M2a

Identical to model M2, except that the variable DV was replaced with the variable DVP = positive vertical distance traveled.

Model M2b

Identical to M2a, except that the RS factor was omitted.

Model M2c

Identical to M2a, except that the factor RS_k (k = 1 to 3) was replaced with RS2_k (k = 1 to 2).

Model M3

where

K

constant;

A_i

random effect of cow i, i = 1 to 21;

H_j

fixed effect of hour j, j = 0 to 23;

RS_k

fixed effect of reproductive state k, k = 1 to 3;

AC_l

fixed effect of activity l, l = 1 to 4;

BM

standing biomass (kg of DM per ha);

ME

ME concentration (MJ/kg of DM);

DH

horizontal distance traveled (m per 5 min);

DV

vertical distance traveled (m per 5 min);

c1 to c4

regression coefficients for BM, ME, DH, and DV, respectively; and

e

random residual effect.

Model M3a

Identical to Model M3, except that the variable DV was replaced with the variable DVP = positive vertical distance traveled.

Model M3b

Identical to M3a, except that the RS factor was omitted.

Model M3c

Identical to M3a, except that the factor RS_k (k = 1 to 3) was replaced with RS2_k (k = 1 to 2).

Model M4a

where

K

constant;

A_i

random effect of cow i, i = 1 to 21;

H_j

fixed effect of hour j, j = 0 to 23;

RS_k

fixed effect of reproductive state k, k = 1 to 3;

AC_l

fixed effect of activity l, l = 1 to 4;

BM

standing biomass (kg of DM/ha);

Ash

ash concentration (g/kg of DM);

NDF

NDF concentration (g/kg of DM);

CP

crude protein concentration (g/kg of DM);

ME

ME concentration (MJ/kg of DM);

DH

horizontal distance traveled (m per 5 min);

DVP

positive vertical distance traveled (m per 5 min);

c1 to c7

regression coefficients for BM, ash, NDF, CP, ME, DH, and DVP, respectively; and

e

random residual effect.

Model M4b

Identical to M4a, except that the RS factor was omitted.

Model M4c

Identical to M4a, except that the factor RS_k (k = 1 to 3) was replaced with RS2_k (k = 1 to 2).

Stepwise Model

where

K

constant;

A_i

random effect of cow i, i = 1 to 21;

T_j

fixed effect of treatment j, j = 1 to 2;

M_k

fixed effect of month k, k = 1 to 4;

H_l

fixed effect of hour l, l = 0 to 23;

HB_m

fixed effect of habitat m, m = 1 to 5;

AC_n

fixed effect of activity n, n = 1 to 4;

DH

horizontal distance traveled (m/5 min);

DVP

positive vertical distance traveled (m/5 min);

BM

standing biomass (kg of DM/ha);

Ash

ash concentration (g/kg of DM);

NDF

NDF concentration (g/kg of DM);

c1 to c5

regression coefficients for DH, DVP, BM, Ash and NDF, respectively; and

e

random residual effect.

	Footnotes

 
¹ Contribution from the Agricultural Research Organization, Bet Dagan, Israel, No. 474/05.

² Corresponding author: brosha{at}volcani.agri.gov.il

Received for publication June 14, 2005. Accepted for publication February 2, 2006.

	LITERATURE CITED

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

 


Aharoni, Y., A. Brosh, P. Kourilov, and A. Arieli. 2003. The variability of the ratio of oxygen consumption to heart rate in cattle and sheep at different hours of the day and under different heat load conditions. Livest. Prod. Sci. 79:107–117.[CrossRef]

Aharoni, Y., A. Brosh, A. Orlov, E. Shargal, and M. Gutman. 2004. Measurements of energy balance of grazing beef cows on Mediterranean pasture, the effects of stocking rate and season: 1. Digesta kinetics, faecal output and digestible dry matter intake. Livest. Prod. Sci. 90:89–100.[CrossRef]

Anderson, D. M., and M. M. Kothmann. 1977. Monitoring animals travel with digital pedometers. J. Range Manage. 30:316–317.

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

Arieli, A., A. Kalouti, Y. Aharoni, and A. Brosh. 2002. Assessment of energy expenditure by daily heart rate measurement—validation with energy accretion in sheep. Livest. Prod. Sci. 78:99–105.[CrossRef]

Arnold, G. E. 1960. The effect of the quantity and the quality of pasture available to sheep on their grazing behaviour. Aust. J. Agric. Res. 11:1034–1043.

Berhana, T., R. Puchalaa, A. L. Goetsch, and R. C. Merkel. 2005. Effects of walking speed and forage consumption on energy expenditure and heart rate by Alpine does. Small Rumin. Res.

Blaxter, K. L. 1967. Pages 110–112 in The Energy Metabolism of Ruminants. Hutchinson, London, UK.

Brosh, A., Y. Aharoni, A. Degen, D. Wright, and B. A. Young. 1998. Estimation of energy expenditure from heart rate measurements in cattle maintained under different conditions. J. Anim. Sci. 76:3054–3064.[Abstract/Free Full Text]

Brosh, A., Y. Aharoni, and Z. Holzer. 2002. Energy expenditure estimation from heart rate: validation by long-term energy balance measurement in cows. Livest. Prod. Sci. 77:287–299.[CrossRef]

Brosh, A., Y. Aharoni, E. Shargal, I. Choshniak, B. Sharir, and M. Gutman. 2004. Measurements of energy balance of grazing beef cows in Mediterranean pasture, the effects of stocking rate and season: 2. Energy expenditure estimation from heart rate and oxygen consumption, and the energy balance. Livest. Prod. Sci. 90:101–115.

Brosh, A., D. Robertshaw, Y. Aharoni, R E. Rawson, M. Gutman, A. Arieli, E. Shargal, and I. Choshniak. 2003. Estimation of energy expenditure in free-living adult and growing ruminants by heart rate measurement: From cardiovascular system to whole animal. Pages 473–476 in EAAP Publication No. 109, 2003. Rostock-Warnemünde, Germany.

DeJaeger, D., P. A. Willems, and N. C. Heglund. 2001. The energy cost of walking in children. Eur. J. Physiol. 441:538–543.[CrossRef][Medline]

DiMarco, O. N., and M. S. Aello. 1998. Energy cost of cattle walking on the level and on gradient. J. Range Manage. 51:9–13.[CrossRef]

DiMarco, O. N., M. S. Aello, and D. G. Mendez. 1996. Energy expenditure of cattle grazing pasture of low and high availability. Anim. Sci. (UK) 63:45–50.

Fedak, M. A., L. Rome, and H. J. Sheeherman. 1981. One-step N2-dilution technique for calibrating open-circuit VO2 measuring systems. J. Appl. Physiol. 51:772–776.[Abstract/Free Full Text]

Gutman, M., and N. G. Seligman. 1979. Grazing management of herbaceous Mediterranean foothill range in the upper Jordan Valley and its effect on cattle and vegetation. J. Range Manage. 32:86–92.

Havstad, K. M., and J. C. Malechek. 1982. Energy expenditure by heifers grazing crested wheatgrass of diminishing availability. J. Range Manage. 35:447–450.

Henkin, Z., E. D. Ungar, M. Gutman, A. Dolev, and A. Brosh. 2003. Tracking beef cattle with GPS collars to study the impact of landscape and management on grazing behavior. Pages 872–875 in Proc. VIIth Int. Rangeland Congr., Durban, South Africa. N. Allsopp, A. R. Palmer, S. J. Milton, K. P. Kirkman, G. I. H. Kerley, C. R. Hurt, and C. J. Brown, ed.

Lachica, M., and J. F. Aguilera. 2000. Estimation of energy costs of locomotion in the Iberian pig (Sus mediterraneus). Br. J. Nutr. 83:35–41.[Medline]

Lachica, M., F. G. Barroso, and C. Prieto. 1997a. Seasonal variation of locomotion and energy expenditure in goats under range grazing conditions. J. Range Manage. 50:234–238.

Lachica, M., C. Prieto, and J. F. Aguilera. 1997b. The energy cost of walking on the level and on negative and positive slopes in the Granadina goat (Capra hircus). Br. J. Nutr. 77:73–81.[Medline]

Lachica, M., R. Somlo, F. G. Barroso, J. Boza, and C. Prieto. 1999. Goats locomotion energy expenditure under range grazing conditions: Seasonal variation. J. Range Manage. 52:431–435.

Lawes Agricultural Trust. (Rothamsted Experimental Station). 2003. Genstat^® for Windows^TM. 7th ed. VSN Int., Oxford, UK.

McLean, J. A., and G. Tobin. 1990. Page 276 in Animal and Human Calorimetry. Cambridge Univ. Press, Cambridge, UK.

Moe, P. W., and H. F. Tyrrel. 1972. Metabolizable energy requirements of pregnant dairy cows. J. Dairy Sci. 55:480–483.[Abstract/Free Full Text]

Nestis, A. S. 1979. Effects of forage availability on voluntary intake and feeding behavior of grazing heifers. PhD Diss. Utah State Univ., Logan, UT.

Nicol, A. M., and B. A. Young. 1990. Short-term thermal and metabolic responses of sheep to ruminal cooling: Effects of level of cooling and physiological state. Can. J. Anim. Sci. 70:833–843.

NRC. 1989. Nutrient Requirements of Dairy Cattle. 6th rev. ed. Natl. Acad. Press, Washington, DC.

Osuji, P. O. 1974. The physiology of eating and the energy expenditure of the ruminant at pasture. J. Range Manage. 27:437–443.[CrossRef]

Pearson, R. A., J. T. Dijkman, R. C. Krecek, and P. Wright. 1998. Effect of density and weight of load on the energy cost of carrying loads by donkeys and ponies. Trop. Anim. Health Prod. 30:67–78.[CrossRef][Medline]

Seligman, N. G., M. Gutman, Z. Holzer, I. Noy-Meir, and H. Baram. 1989. Stocking density of cattle and herbage production on Mediterranean grassland. J. Agric. Sci. 113:51–58.

Sternberg, M., M. Gutman, A. Perevolotsky, E. D. Ungar, and J. Kigel. 2000. Vegetation response to grazing management in a Mediterranean herbaceous community: A functional group approach. J. Appl. Ecol. 37:1–15.

Taylor, C. R., N. C. Heglund, and G. M. Maloiy. 1982. Energetics and mechanics of terrestrial locomotion. I. Metabolic energy consumption as a function of speed and body size in birds and mammals. J. Exp. Biol. 97:1–21.[Abstract/Free Full Text]

Ungar, E. G., Z. Henkin, M. Gutman, A. Dolev, A. Genizi, and D. Ganskopp. 2005. Inference of animal activity from GPS collar data on free-ranging cattle. Rangeland Ecol. Manage. 58:256–266.[CrossRef]

Van Soest, P. J., J. B. Robertson, and B. A. Lewis. 1991. Methods of 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:

				Y. Aharoni, Z. Henkin, A. Ezra, A. Dolev, A. Shabtay, A. Orlov, Y. Yehuda, and A. Brosh Grazing behavior and energy costs of activity: A comparison between two types of cattle J Anim Sci, August 1, 2009; 87(8): 2719 - 2731. [Abstract] [Full Text] [PDF]

				T. Sahlu, L. J. Dawson, T. A. Gipson, S. P. Hart, R. C. Merkel, R. Puchala, Z. Wang, S. Zeng, and A. L. Goetsch ASAS Centennial Paper: Impact of animal science research on United States goat production and predictions for the future J Anim Sci, January 1, 2009; 87(1): 400 - 418. [Abstract] [Full Text] [PDF]

				A. Brosh Heart rate measurements as an index of energy expenditure and energy balance in ruminants: A review J Anim Sci, May 1, 2007; 85(5): 1213 - 1227. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brosh, A. Articles by Aharoni, Y. Search for Related Content PubMed PubMed Citation Articles by Brosh, A. Articles by Aharoni, Y.