Drinking Behavior of Lactating Dairy Cows and Prediction of Their Water Intake

Article Outline

• Abstract
• Introduction
• Materials and Methods
  • Animals, Housing, and Diet
  • Drinking Behavior
  • Statistical Analysis
    • Daily Drinking Behavior
    • Effect of Stocking Density
    • Water Intake Prediction
• Results and Discussion
  • Water Intake and Its Prediction
    • Daily FWI
    • Predicting FWI
  • Drinking Behavior
    • General Drinking Behavior
    • Daily Drinking Behavior
    • Effect of Stage of Lactation
    • Number of Cows Sharing the Water Bowls
• Conclusions
• Supplementary data
• References
• Copyright

Abstract 

The water intake of 41 lactating dairy cows managed according to current dairy farm practices was individually and continuously monitored to 1) investigate drinking behavior and 2) determine factors affecting water intake. The cows were housed in a free-stall barn and fed once daily with a corn silage and concentrate-based total mixed ration (48% dry matter content; 20.6±3.3 kg/d of dry matter intake). Cows were milked twice daily, with a yield of 26.5±5.9 kg/d. The daily free water intake (FWI) was 83.6±17.1L, achieved during 7.3±2.8 drinking bouts. The drinking bout water intake was 12.9±5.0L. Almost three-fourths of the FWI occurred during working hours (0600 to 1900h). Consumption peaks corresponded to feeding and milking times. More than one quarter of the daily FWI was met during the 2h after each milking. About 75% of the present cows visited the watering point at least once during the 2h after the evening milking. It is probable that drinking behavior evolved with lactation, but further studies are required to identify the relationship between lactation stage and drinking behavior. The most relevant factors affecting the daily FWI of lactating cows were best combined according to the following predictive equation: (R²=0.45; n=41 cows, n=1,837): FWI, L/d=1.53×dry matter intake (kg/d)+1.33×milk yield (kg/d)+0.89×dry matter content (%)+0.57×minimum temperature (°C)-0.30×rainfall (mm/d)-25.65. The results obtained using these equations were in agreement with the equations developed by other researchers.

Key words: drinking behavior, prediction of water intake, lactating cow

Back to Article Outline

Introduction 

Water supplies for both humans and livestock are becoming a subject of increasing importance. Indeed, climate change and drinking water deficits in certain areas have meant that supplies of clean water for livestock are becoming problematic, at least during certain periods of the year. Water is considered the most important nutrient for health and performance in dairy herds (NRC, 2001), and water deprivation can markedly affect the health, behavior, and performance of animals. Low water intake increases hematocrit and blood urea (Steiger Burgos et al., 2001), reduces the respiratory rate and rumen contractions (Little et al., 1980), reduces BW and milk yield (Little et al., 1980; Steiger Burgos et al., 2001), and provokes aggressive behavior around waterers (Little et al., 1980). Unfortunately, the water intake of dairy cows is rarely considered a potential limiting factor for milk production in modern dairy farms. Despite the attention paid to other nutrients, the quantity and quality of water are not sufficiently considered (Beede, 2005).

Several factors affect free water intake (FWI). Some of the most frequently cited parameters include DMI (Holter and Urban, 1992; Dado and Allen, 1994), milk yield (Dahlborn et al., 1998; Meyer et al., 2004), dry matter content (Dahlborn et al., 1998), and different expressions of climate conditions (Murphy et al., 1983; Meyer et al., 2004), and to a lesser extent, BW (Dado and Allen, 1994; Meyer et al., 2004) and sodium intake (Popovici et al., 1971; Meyer et al., 2004).

Although a considerable body of knowledge was developed on the prediction of water intake, very few studies have addressed the drinking behavior of dairy cattle. Indeed, dairy cows require large quantities of water, but there is little data to indicate how often and when dairy cows drink, or to describe the relationship between their behavior and the amount of water ingested. It is essential to answer these questions to ensure sufficient water supplies for all cows in a high-yielding dairy herd.

The aims of this study were to 1) verify information provided by the literature to validate factors affecting FWI and determine a prediction model, 2) investigate the drinking behavior of dairy cows managed in a modern dairy farm, and 3) clarify links between behavior and water intake.

Back to Article Outline

Materials and Methods 

Animals, Housing, and Diet 

Data from 41 (13 primiparous and 28 multiparous) Holstein cows were collected over 3 consecutive winters (November to April) within 70-d periods at our experimental farm. No data were collected during the weekends; thus, a maximum of 50 cow-days were possible per animal (Table 1). Some cow-day data were deliberately excluded when abnormal circumstances existed (e.g., clinical lameness or mastitis). Cows were continuously housed in a free-stall barn. Data were not collected in the summer, because cows were grazing part of the day and the facilities, and spatial and temporal conditions were too different from data recorded during winter. The number of animals present in the pen ranged from 11 to 36, depending on calving dates and the numbers included in the different trials. The size of the barn was permanently adjusted to the animal numbers to ensure that 1 stall would be available for each animal. Cows were bedded in stalls with straw on rubber mats, and scrapers automatically cleaned the concrete floor 4 times daily. The layout and dimensions of the free-stall barn used (2-row tail-to-tail) and the placement of waterers are shown in Figure 1.

Table 1. Summary of the available data

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Layout and dimensions of the experimental barn (dashed line) used for free water intake studies.

At all times, the care of animals complied with the General Guidelines of the Council of the European Community (1986, no. 86/609/EEC).

Animals were milked twice daily (0600 and 1700h) in a double- 6 herringbone milking parlor (M2100, Bou-Matic, Saint-Nom-la- Bretêche, France). Cows were in mid-lactation (the average DIM at the start of the experiment was 74.2±17.8) and their milk yield was recorded automatically at each milking (Isalait 2045 system, BouMatic) and summed to produce a daily milk yield (MY).

Maximum temperature, minimum temperature (MINT), and rainfall (RF) were recorded at an official station located 2km from the farm. Mean temperatures were calculated as the average of maximum and minimum temperatures. Sixty-five days or 655 cow-days (35.7%) of the total 1,837 cow-days used in the analysis included a rainfall event.

The animals enrolled were simultaneously used for feeding trials evaluating feed additives. All cows received a similar diet based on corn silage, wheat straw, and concentrate (cracked wheat and soybean meal). Samples were composited every 4 wk and the average composition is presented in Table 2. Because the number of cows differed between years, weighted means and SD were calculated rather than simple means. Gross composition of the diet was identical for all of the cows evaluated during the experiment at the same time. Feed additives (yeast products or amino acids) resulted in very small changes in ration nutrient composition between feeding trials. Nevertheless, diet composition could differ slightly between years because of the quality of the silage. But no cow was subjected to different diets because no cow was used for more than one year. The CP content followed European standards, but was slightly lower than commonly recommended in American feeding strategies.

Table 2. Nutrient and ingredient composition of the diet fed to the cows (DM basis)

Item	Weighted mean	Weighted SD
Energy and nutrients1
DM, %	47.9	5.6
CP, %	14.5	0.7
NFC,2%	36.9	3.1
Starch, %	24.4	2.1
Fat, %	3.0	0.0
NDF, %	41.1	3.2
ADF, %	19.4	0.9
Ash, %	4.4	0.2
NEL, Mcal/kg of DM	1.6	0.0
Ingredients
Corn silage, %	64.4	2.3
Wheat straw, %	2.8	2.8
Protein concentrate,3%	12.1	3.3
Energy concentrate,4%	19.2	5.4
Minerals, %	1.5	0.1

The ration was mixed and offered for ad libitum intake as a TMR once daily at 0900h. Electronic feeding gates (Sefer Co., Neuville-le-Poitou, France) were used to measure individual feed consumption. Feed deliveries and refusals were weighed, dried (48h at 85°C), and recorded daily to determine the DMI and DM content (DMC).

Animals were allowed free access to 2 automatic drinking bowls (F60, La Buvette, Charleville-Mézières, France), located 4m from the feed bunk (Figure 1). The water bowls were fixed at a height of 65cm. The watering point was equipped with barriers at each side to prevent cows from disturbing those drinking. Because the number of water bowls remained the same throughout the experimental period, the number of cows per bowl ranged from 5 to 18. Even if most of the cows were familiar with other cows of the herd (except for primiparous cows) and were adapted to commingling in previous experiments, repeated mixing may have affected FWI and drinking behavior. It was not possible to separate the potential effects of commingling on FWI and drinking behavior.

Drinking Behavior 

Drinking activity was recorded continuously and simultaneously for randomly enrolled cows using a data acquisition system based on an individual radio frequency identification collar (Nedap, Groenlo, the Netherlands) and a flow meter (Bürkert, Ingelfingen, Germany) on each water bowl. During each of the 3 periods of data collection, other cows were in the experimental barn (running a feed trial), but their drinking behavior was not monitored. Different variables were recorded: a “visit” corresponded to the identification of a cow at the watering point (whether to drink or not). A “drink” or “drinking action” corresponded to the delivery of water to an identified cow present at the watering point. Finally, 2 separate “drinking bouts” (DRB) were defined when at least 4min was spent without any drinking, as suggested by Dado and Allen (1994). One drinking bout could be made up of one or several drinking events.

The FWI, and the number and volume of DRB (i.e., volume recorded by the flow meter) were recorded daily and linked to the identified collar number. Variables for each DRB included the starting time and volume of each drinking action.

Statistical Analysis 

Distributions of FWI and DRB within a day, and distribution of waterer use (i.e., the number of cows drinking at the watering point per hour) were analyzed using the MIXED procedure of SAS (version 9.1, 2001, SAS Institute Inc., Cary, NC). Percentage of daily water intake per hour, number of DRB per hour, and number of cows visiting the watering point per hour were the dependent variables, and time of the day (on an hourly basis) was the independent variable. In this approach, hours were included as repeated measures option of the MIXED procedure of SAS and cow was used as the subject effect. Differences between hours were determined with Tukey's test.

Furthermore, 3 periods were defined before the experiment to compare drinking patterns during the day: 1) a period of supposedly “high” drinking activity (2h after each milking and feeding); 2) nighttime drinking activity (from 2000 to 0559h); and 3) a period of “normal” daylight drinking activity (hours of daylight except for the 2h after each milking and feeding). Water intake, volume of DRB, and number of DRB were the dependent variables of interest, and period of the day was the independent variable. In this approach, the 3 periods were included as repeated measures option of the MIXED procedure of SAS, and cow was used as subject effect. Differences between each period were analyzed with contrasts.

Concerning the effects of stocking density, 3 classes were compared: 11 to 20 cows in the pen, 21 to 30 cows in the pen, and 31 to 40 cows in the pen. Dependent variables of interest were daily FWI, volume of DRB, and number of DRB. The effect of the number of animals in the barn sharing the 2 drinking bowls was determined by ANOVA (GLM procedure of SAS), in which density and cow were independent variables. Differences between each density were determined with Tukey's test.

Means, coefficients of variation, and correlation coefficients were calculated using the DESCO procedure under SPAD (SPAD Soft 6.0, Decisia, Levallois- Perret, France). A prediction model was constructed using multiple linear regression in a stepwise manner with the VAREG and FUWIL procedures of SPAD. The best model was chosen using the lowest value of the Akaike information criterion (Akaike, 1969), and the variance estimation of the parameters of the model was evaluated by jackknife variance estimation (Shao and Wu, 1989) using Tukey's conjecture. According to this predictive model, daily FWI was the dependent variable and DMI, MY, DMC, temperature, and RF were independent variables.

Specific effects of year or tested feed additive were not integrated in the statistical analyses. These effects were not studied but were integrated indirectly via the aforementioned factors. Furthermore, feed additives did not modify in any way the nutrient and ingredient composition of the diet and, to our knowledge no publication has reported any metabolic pathway inducing a specific effect of feed additives on water intake or drinking behavior. It should be noted that no cows participated in the study for more than 1 yr.

Table 3 gives an overview of the experimental conditions for each trial. Data from the different trials were pooled into a single data set. The reasonableness of this approach was assessed using the overall FWI prediction model and plotting residuals against predicted FWI for each trial. Unless otherwise specified, treatment effects were declared significant at P<0.05.

Table 3. Summary of the experimental conditions in the consecutive trials

	Feed trial 1	Feed trial 2		Feed trial 3
Dates	02/12/2003 to 04/23/2003	11/19/2003 to 01/30/2004	01/07/2004 to 04/02/2004	11/22/2004 to 01/14/2005
Primiparous cows,1 n	3	3	5	2
Multiparous cows,1 n	3	8	14	3
Total cow-days, n	275	492	876	194
Mean cow-days/cow	45.8	44.6	46.2	38.8
Range in cow-days/cow	43 to 50	44 to 46	43 to 48	37 to 40
Mean stocking density2	21.0	26.6	25.0	17.8
Range in stocking density	11 to 34	15 to 36	14 to 36	17 to 18
Mean milk yield, kg/d	27.5	23.7	26.4	33.2
Range in milk yield, kg/d	9.8 to 38.1	7.3 to 33.8	10.6 to 42.6	13.2 to 46.3
Mean DMI, kg/d	20.0	19.7	21.0	21.9
Range in DMI	9.8 to 28.4	7.6 to 26.1	7.6 to 33.9	5.4 to 31.0
Mean DM content, %	41.3	51.7	50.1	39.3
Range in DM content, %	36.3 to 45.7	48.1 to 57.0	39.6 to 55.5	37.6 to 41.6
Mean ambient temperature,°C	5.8	3.2	3.9	2.4
Range in ambient temperature,°C	−5 to 16.2	−2.7 to 11.2	−1.8 to 13.6	−6.8 to 9.2
Mean daily rainfall, mm/d	2.6	5.5	3.3	2.1
Range in daily rainfall, mm/d	0.2 to 6.2	0.2 to 33.2	0.2 to 33.2	0.2 to 12.2
Cow-days with rainfall, n	43	253	253	105
Tested feed additives in the diet	Amino acids	Yeasts	Yeasts	Yeasts

Back to Article Outline

Results and Discussion 

Water Intake and Its Prediction 

Minimum, mean, and maximum values, SD, and correlation coefficients to FWI for each studied variable are presented in Table 4.

Table 4. Means, ranges and correlation coefficients to free water intake (FWI)

Variable	Mean	SD	Minimum	Maximum	r to FWI
FWI, L/d	83.6	17.1	23.5	143.3	1
DMI, kg/d	20.6	3.3	5.4	33.9	0.540**
Milk yield, kg/d	26.5	5.9	7.3	46.3	0.531**
Dietary DM content, %	48.1	5.0	36.3	57.0	0.088**
Minimum temperature,°C	0.1	4.0	−11.1	9.1	0.097**
Rainfall, mm/d	1.3	3.9	0.0	33.2	−0.064**
Ambient temperature,°C	3.8	4.4	−6.8	16.2	0.081**
Drinking bouts/d	7.3	2.8	1	27	0.044*
Drinking bouts FWI, L	12.9	5.0	2.5	50.6	0.412**
Cows/pen, n	24.1	6.8	11	36	−0.049*
Lactation number	2.4	1.4	1	7	0.252**
DIM	108.4	26.2	35	173	−0.141**

Taking account of the differences between experiments, the daily FWI (83.6±17.1 L/d) generally supported results of other studies, and FWI was similar to 82 L/d of Meyer et al. (2004) and 84 L/ d of Melin et al. (2005). Melin et al. (2005) had a similar DMC (45%), a slightly increased DMI (22.9 kg/d), and a greater MY (34.3 kg/d). Meyer et al. (2004) reported that cows ingested (20.5 kg/d) amounts that were similar to those we recorded, although those cows received a drier diet (54.5±9.5%) and had more MY (31.1 kg/d).

The slightly greater FWI (89.2 kg/d) reported by Murphy et al. (1983) could have been due to high salt intake (74 g/d), because increased salt content increased FWI (Popovici et al., 1971). Moreover, the MY (33.1 kg/d) was greater than in our study. The lower FWI of 13 L/d observed by Holter and Urban (1992) seems surprising. The DMC in their ration was slightly greater than in ours (50 vs. 48%), and the cows ate less (18.7 kg/d) but produced more milk (34.6 kg/d).

The variables best correlated to daily FWI (Table 4) were DMI, MY, lactation number, MINT, DMC, and RF. The daily FWI of lactating cows was best predicted by combining these variables in the following equation (R²=0.45; P<0.01; n=41 cows, 1,837 data; Table 5):

Table 5. Estimated value, SE and increase in R² for each variable used in the water intake prediction model (n = 1,837 observations)

Minimum temperature and RF explained only a very small part of the variability of FWI in the model (1.6 and 0.2%, respectively; P<0.01; Table 5). Our predictive equation for FWI confirmed the main variables cited in the literature as affecting FWI: DMI, MY, DMC, and weather (Dado and Allen, 1994; Dahlborn et al., 1998; Meyer et al., 2004).

The ratios of FWI/kg of DMI (4.1 L/kg) or per kg of MY (3.1 L/kg) were slightly greater, but supported values in other recent studies: 3.4 (Dado and Allen, 1994), 3.6 (Melin et al., 2005), and 3.9 (Meyer et al., 2004) for FWI:DMI and 2.3 (Melin et al., 2005), 2.4 (Meyer et al., 2004), and 2.6 (Dado and Allen, 1994) for FWI/MY.

Rainfall was cited for the first time as an independent variable affecting FWI when feeding TMR. But RF and high humidity reduced FWI by cattle on pasture (Castle, 1972). Other studies feeding TMR observed negative correlations between relative humidity and FWI (Meyer et al., 2004). Rainfall and ambient temperature are directly linked to relative humidity. Nevertheless, no consistent relationship was found between RF and ration DMC at the time of TMR preparation. Ration DM was evaluated just after delivery to the feed bunk and it is not known whether DMC changed during the course of a day with RF. Thus, the effect of RF on FWI may be due to 1) RF decreasing the ration DM and thereby, reducing demand for FWI; 2) RF decreasing the ambient temperature and FWI decreasing due to the positive correlation between FWI and temperature; or 3) RF increasing relative humidity, which in turn decreased thirst during the winter months. Nevertheless, it should be noted that the correlation between RF and FWI was small (Table 4).

The FWI prediction model was satisfactory regarding the explained part of variability (R²=0.45). However, this could probably be improved in 2 ways: 1) by measuring other factors such as BW or relative humidity in the barn; and 2) by running the experiment over a longer period to obtain a broader range of values. Although the range of our prediction for FWI is similar to others (Table 6), it appears higher than other predicted values. Some differences related to the experimental conditions and animals involved might explain this.

Table 6. Comparison of predicted free water intake (FWI; L/d) with models developed by different authors

Drinking Behavior 

The frequency of 7.3 DRB/d supported the range of values reported in the literature from 5.2 (Jago et al., 2005) to 9.4 DRB/d (Huzzey et al., 2005). The average intake per DRB (12.9L) was very similar to (Jago et al., 2005)Jago et al. (2005; 14.9L) during a study conducted in free stalls. The water intake during DRB was correlated (P<0.01) with lactation number (r=0.41), MY (r=0.37), and DMI (r=0.35).

In support of Dado and Allen (1994; r=−0.8), the number of the DRB was negatively correlated to their size (r=−0.77). With an average FWI near our observations (77.6 L/d), they noted an increased drinking frequency (14 DRB/d) under tie-stall conditions, with a reduced FWI per DRB (6.4L). This observation could have been due to the watering system used. Indeed, during their experiment, the drinking rate was low (4.3 L/min) compared with other reports (Pinheiro Machado Filho et al., 2004; 18.1 L/min). Andersson et al. (1984) observed that increasing the flow rate of waterers from 2 to 12 L/min led to a 25% reduction in drinking frequency (40 vs. 30 DRB/d) and increased FWI.

Considering the relationship between daily FWI and drinking behavior, it appeared that variations in FWI were mainly due to variations in the quantity of water drunk during a DRB (r=0.41; Table 4). Indeed, the number of DRB was lowly correlated to daily FWI (r=0.04). These results agree with those obtained by Dado and Allen (1994), where the FWI was correlated with the volume of DRB (r=0.47), but lowly correlated with the number of DRB (r=0.04).

Fifteen percent of DRB were made up of several drinks. In that case, the first drink was the largest in volume (11.0±4.7L); subsequent drinks averaged 2.8±1L. Nevertheless, consumption during DRB including several drinks was greater than that seen during DRB involving only 1 drink (14.0 vs. 12.1L; P<0.01). Regarding their distribution within the day, no difference was observed between DRB made up of several drinks and DRB involving only 1 drink.

Most of the daily FWI (72.7±14%) was achieved during working hours on the farm (0600 to 1859h), supporting other studies (Nocek and Braun, 1985; Sowell et al., 1999; Osborne et al., 2002). Consequently, nocturnal drinking behavior (i.e., between 1900 and 0559h) accounted for only a small percentage (27%) of the FWI. Osborne et al. (2002) found similar results (25% of daily FWI met during night), although their animals were fed twice daily. In this situation, cows could have been attracted to the feed bunk by the second delivery of fresh feed in the afternoon and may consume water later in the night.

In our experiment, 2 main consumption peaks were observed between 0900 and 1059h and 1700 and 1859h (Figure 2). Three smaller peaks could be observed between 0700 and 0759h, 1400 and 1559h, and 1900 and 1959h (Figure 2). As described by others, drinking bouts occurred at all times of the day and night, but consumption peaks were associated with feed delivery and milking times (Nocek and Braun, 1985; NRC, 2001; Osborne et al., 2002). In contrast, low drinking activity was observed during the period 0300 to 0559h, whereby 1, 0.8, and 0.7% of the daily FWI was met, respectively, during each of these 3h. Furthermore, 7, 22, and 32% of the cows never visited the watering point during each of these 3h, respectively (data not shown). Our results confirm that 1) feeding and drinking are mostly synchronized activities, and 2) with respect to feeding (Shabi et al., 2005; DeVries et al., 2003), drinking is a diurnal activity.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  Distribution of free water intake (FWI) during a day in hourly intervals. Bars represent the SEM.

During our experiment, 75% of the cows came to the waterer within 2h of the evening milking (Figure 3). The strength of the relationship between milking and drinking time was reported differently by Wieclaw et al. (1973): they reported that 90% of cows drank within 2h postmilking. In the context of field measurements, Beede (2005) found that 50 to 60% of the daily water needs were met immediately after milking. In our study, only 27% (22.3L) of the daily FWI was consumed during the 2h after each milking, but considering the 2h after milking and feeding, our observation of 42.3% of daily FWI corresponded to the results of Osborne et al. (2002), who found that 40% of the FWI was met within this period.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  Proportion of cows drinking during a day in hourly intervals. Bars represent the SEM.

On average, half (54%) of the herd frequented the waterer within 1h following the evening milking (Figure 3). This means that up to 10 cows used the same water bowl within 1h. No water trough was available at exit from the milking parlor, which could explain why water consumption was distributed differently than in Beede (2005). Furthermore, the large number of cows visiting the watering point within 1h of leaving the milking parlor seems to indicate a rush at the watering point.

Based on information in the literature, 3 periods were defined before the experiment to compare drinking patterns during the day using a contrast test: 1) a period of supposedly high drinking activity (2h after each milking and feeding); 2) night-time drinking activity (from 2000 to 0559h); and 3) a period of supposedly normal daylight drinking activity (the rest of the day). The average FWI during DRB that occurred within 2h of feeding and milking was greater than the FWI during other daylight hours or at night (13.6, 11.9, and 11.6L, respectively; P<0.01). This may have been due to dehydration caused by milk output or eagerness caused by the rush to the watering station.

Furthermore, cows visited the watering point 7.6±3.4 times/d, which means that cows went to the watering point but did not drink 0.3±1.1 times/d. This behavior was observed at least once with each cow during the experiment, but only 4.9% of these cows (2 of 41) tended to adopt it once a day or more. During the evening high drinking activity period (1700 to 1859h; data not shown), the frequency of nondrinking visits to the watering point decreased, which indicated that cows visited the waterer at that time to actually drink.

Because DMI, MY, and hence, FWI evolve throughout lactation, we can expect that drinking behavior varied with lactation stage. Indeed, the water intake of DRB was correlated to MY (r=0.37; P<0.01) and DMI (r=0.35; P<0.01). Nevertheless, during our experiment, DIM was only lowly correlated with DRB (r=0.09; P<0.01 and r=−0.09; P<0.01, for the number of DRB and DRB water intake, respectively). This could be because the cows were in mid-lactation or because the duration of the experiment was relatively short (i.e., even MY and DIM were slightly correlated: r=−0.2 (P<0.01).

The effect of the number of animals in the barn sharing the 2 drinking bowls could not be investigated clearly. However, although our statistical power was insufficient to reach a threshold of significance, a slight but clear trend was observed regarding number of cows sharing the 2 drinking bowls and drinking behavior (Table 7). No other variable (DMI, DMC, MY, or lactation number) was affected significantly by stocking density. Social effects due to commingling cattle may play a role in varying DRB volume.

Table 7. Means and SD of the studied variables when considering stocking density (pooled data from all trials, except number 3)

	Stocking density
Item	11 to 20	21 to 30	31 to 40
Volume of drinking bouts, L	12.2±4.8	12.9±4.4	13.9±4.9
Number of drinking bouts/d	7.9±2.8	7.2±2.3	6.8±2.7
Free water intake, L/d	82.6±12.1	82.6±12.0	83.0±14.4
Milk yield, kg/d	25.2±5.1	25.6±4.8	27.0±5.4
DMI, kg/d	19.8±2.2	20.6±2.4	20.2±2.7
DM content, %	49.5±4.5	49.4±3.6	48.5±3.7
Lactation number	2.5±1.5	2.5±1.4	2.5±1.4
Cow-days within category, n	340	956	347

As for high drinking activity periods, increasing the number of cows in the barn tended to reduce the number of nondrinking visits to the waterer (0.51, 0.32, and 0.20/d; P=0.07). Because the observed effect of the number of animals was low when compared with individual variations, about 100 cows would have been necessary to demonstrate statistical differences. The NRC (2001) recommends 1 water bowl per 10 cows, whereas during our experiment, the maximum number of cows per drinking bowl was 18. Andersson et al. (1984) observed that when pairs of tied cows shared the same water bowl, submissive cows drank 7% less than dominant ones. These elements support the suggestion that overstocking of animals may exert an effect on drinking behavior.

Back to Article Outline

Conclusions 

Our results confirm that environmental factors such as diet composition and climatic conditions and individual factors such as MY or the amount of DMI can affect drinking behavior, water intake, or both. The predictive model can efficiently predict FWI for lactating cows using easily recordable factors. Our results supported those in the literature: most drinking activity occurred during daylight hours and was correlated with feeding and milking events. Yet, the description only concerns cows in mid-lactation. This study was conducted during the winter months and behavioral effects may be different during the heat of summer when cows need more water per day.

Back to Article Outline

Supplementary data 

Back to Article Outline

References 

1. Akaike H. Fitting autoregressive models for prediction. Ann. Inst. Stat. Math. 1969;21:243–247
   • View In Article
2. Andersson M, Schaar J, Wiktorsson H. Effects of drinking water flow rates and social rank on performance drinking behaviour of tied-up dairy cows. Livest. Prod. Sci. 1984;11:599–610
   • View In Article
3. Beede DK. The most essential nutrient: Water. In: Proc. 7th Western Dairy Management Conf. Reno, NV. Kansas State Univ., Manhattan. 2005;p. 13–32
   • View In Article
4. Castle ME. A study of the intake of drinking water by dairy cows at grass. Grass Forage Sci. 1972;27:207–210
   • View In Article
5. Dado RG, Allen MS. Variation in and relationships among feeding, chewing, and drinking variables for lactating dairy cows. J. Dairy Sci. 1994;77:132–144
   • View In Article
   • Abstract
   • Full-Text PDF (1055 KB)
   • CrossRef
6. Dahlborn K, Akerlind M, Gustafson G. Water intake by dairy cows selected for high or low milk-fat percentage when fed two forages to concentrate ratios with hay or silage. Swed. J. Agric. Res. 1998;28:167–176
   • View In Article
7. DeVries TJ, von Keyserlingk MAG, Beauchemin KA. Short communication: Diurnal feeding pattern of lactating dairy cows. J. Dairy Sci. 2003;86:4079–4082
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (345 KB)
   • CrossRef
8. Holter JB, Urban WE. Water partitioning and intake prediction in dry and lactating Holstein cows. J. Dairy Sci. 1992;75:1472–1479
   • View In Article
   • Abstract
   • Full-Text PDF (701 KB)
   • CrossRef
9. Huzzey JM, von Keyserlingk MAG, Weary DM. Changes in feeding, drinking, and standing behavior of dairy cows during the transition period. J. Dairy Sci. 2005;88:2454–2461
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (407 KB)
   • CrossRef
10. Jago JG, Roche JR, Kolver ES, Woolford MW. The drinking behaviour of dairy cows in late lactation. Proc. N. Z. Soc. Anim. Prod. 2005;65:209–214
    • View In Article
11. Little W, Collis KA, Gleed PT, Sansom BF, Allen WM, Quick AJ. Effect of reduced water intake by lactating dairy cows on behaviour, milk yield and blood composition. Vet. Rec. 1980;106:547–551
    • View In Article
    • MEDLINE
12. Melin M, Wiktorsson H, Norell L. Analysis of feeding and drinking patterns of dairy cows in two cow traffic situations in automatic milking systems. J. Dairy Sci. 2005;88:71–85
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (933 KB)
    • CrossRef
13. Meyer U, Everinghoff M, Gädeken D, Flachowsky G. Investigations on the water intake of lactating dairy cows. Livest. Prod. Sci. 2004;90:117–121
    • View In Article
14. Murphy MR, Davis CL, McCoy GC. Factors affecting water consumption by Holstein cows in early lactation. J. Dairy Sci. 1983;66:35–38
    • View In Article
    • Abstract
    • Full-Text PDF (292 KB)
    • CrossRef
15. National Research Council. 2001. In: Nutrient Requirements of Dairy Cattle. 7th rev. ed. Washington, DC: National Academy Press; 2001;p. 178–183
    • View In Article
16. Nocek JE, Braun DG. Effect of feeding frequency on diurnal dry matter and water consumption, liquid dilution rate, and milk yield in first lactation. J. Dairy Sci. 1985;68:2238–2247
    • View In Article
    • Abstract
    • Full-Text PDF (1154 KB)
    • CrossRef
17. Osborne VR, Hacker RR, Mc Bride BW. Effects of heated drinking water on the production responses of lactating Holstein and Jersey cows. Can. J. Anim. Sci. 2002;82:267–273
    • View In Article
18. Pinheiro Machado Filho LC, Teixeira DL, Weary DM, von Keyserlingk MAG, Hötzel MJ. Designing better water troughs: Dairy cows prefer and drink more from larger troughs. Appl. Anim. Behav. Sci. 2004;89:185–193
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (123 KB)
    • CrossRef
19. Popovici D, Raitaru M, Jurencova G. Efectul excesului de sodiu in apa potabila asupra secretiei si compositiei laptelui la rumegatoare. Stud. Cerc. Biol. Ser. Zool. 1971;23:573–580
    • View In Article
20. Shabi Z, Murphy MR, Moallem U. Within-day feeding behavior of lactating dairy Cows measured using a real- time control system. J. Dairy Sci. 2005;88:1848–1854
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (302 KB)
    • CrossRef
21. Shao J, Wu CFJ. A general theory for jackknife variance estimation. Ann. Stat. 1989;17:1176–1197
    • View In Article
22. Sowell BF, Branine ME, Bowman JG, Hubbert ME, Sherwood HE, Quimby W. Feeding and watering behavior of healthy and morbid steers in a commercial feedlot. J. Anim. Sci. 1999;77:1105–1112
    • View In Article
    • MEDLINE
23. Steiger Burgos M, Senn M, Sutter F, Kreuzer M, Langhans W. Effect of water restriction on feeding and metabolism in dairy cows. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2001;280:R418–R427
    • View In Article
    • MEDLINE
24. Wieclaw B, Czajkowski Z, Tupaj C, Voelkel J. An influence of drinking on milking capacity and certain blood properties of cows. Zeszyty Naukowe Akademii Kolniczej w Szczecinie. 1973;41:63–71
    • View In Article