- Email: [email protected]
Motivation for additional water use of growing-finishing pigs
Livestock Science 124 (2009) 112–118
Contents lists available at ScienceDirect
Livestock Science j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i v s c i
Motivation for additional water use of growing-finishing pigs Herman M. Vermeer ⁎, Nienke Kuijken, Hans A.M. Spoolder Animal Sciences Group of Wageningen UR, Lelystad, The Netherlands
a r t i c l e
i n f o
Article history: Received 5 November 2007 Received in revised form 23 December 2008 Accepted 14 January 2009 Keywords: Pigs Water Drinking Motivation Behaviour
a b s t r a c t Liquid fed growing-finishing pigs receive an amount of water in their ration that is more than their physiological requirement. For welfare reasons it can be argued that in addition to this diet, pigs may be motivated to obtain additional fresh water. The aim of the present experiment was to test the hypotheses that liquid fed pigs will work harder to obtain extra fresh water, compared to dry fed pigs which receive water in a conventional way. A consumer-demand technique was used, in which flow rate from an extra (test) drinker determined the ease with which pigs could obtain the water. The more persistent pigs were to obtain water from the test drinker (with declining flow rates), the harder they were assumed to work for it. Four treatments were divided over 48 pens of 12 finishing pigs in 2 batches (566 pigs). There was one Dry Feed treatment (D, with standard drinking nipple in a single space trough) and three liquid feeding systems: Long trough (LT); Sensor Feeding (S) and Variomix (V). Each pen had an additional drinker with a weekly randomly changing flow rate of 134, 356, 733 or 1041 ml/ min. From the extra drinker pigs used on average 3.39a (D), 0.76b (LT), 0.58bc (S) and 0.44c (V) litre per day (different superscripts indicate differences P b 0.05). The persistence to obtain water differed between the four treatments. This was indicated by the rate of decline (ρi) of the asymptotic curve depicting water disappearance at decreasing flow rates: ρi = 0.00378a, 0.00274ab, 0.00122b and 0.00275ab for D, LT, S and V, respectively. This suggests that liquid fed pigs work equally hard (LT and V) or less hard (S) to obtain water from an extra drinking nipple, compared to dry fed pigs (D). © 2009 Elsevier B.V. All rights reserved.
1. Introduction Water is an essential part of the nutritive and welfare requirements of pigs. Mroz et al. (1995) stated in a review that water intake mainly depends on body weight, feed intake and temperature. Heavier pigs need more water to maintain their body. The daily water intake of ad libitum fed finishing pigs increases on average from 2 l at 25 kg to 6 l at about 110 kg live weight (Nagai et al., 1994). Mroz et al. (1995) also found in their review that the water to feed ratio decreases with increasing age or weight and increases with the ambient temperature. When ambient temperature increases from 10 to 25 °C, the need for water for evaporative cooling, mainly via respiration,
⁎ Corresponding author. Animal Sciences Group of Wageningen UR, P.O. Box 65, 8200 AB Lelystad, The Netherlands. Tel.: +320 293 378. E-mail address: [email protected] (H.M. Vermeer). 1871-1413/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.livsci.2009.01.009
increases from 2.2 to 4.2 l/d for finishing pigs (Vandenheede and Nicks, 1991). Water can be offered to the pig as part of a liquid diet, or as plain, fresh water from a drinker. The welfare of pigs is compromised if water is unavailable (Kyriazakis and Savory, 1997). The EU minimum standards for the protection of pigs (N.N., 2001) state: “All pigs over two weeks of age must have permanent access to a sufficient quantity of fresh water.” Although intuitively logical, this requirement warrants further investigation following the development of new feeding systems as well as the current pressures on environmental aspects of pig husbandry. An increase in water consumption inevitably leads to an increase in urine production. It can be questioned if the need for fresh water does indeed exist, providing the physiological and behavioural needs for water uptake have been met by the ration. Part of the growingfinishing pigs is fed liquid feed, instead of dry feed with additional fresh water. Fresh water differs considerably from
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
liquid feed. Liquid feed has a high energy content and can incidentally have high concentrations of salt and acid. However, the water to feed ratio in these systems is much higher than in dry feeding systems: around 3.0: 1 for wet feed with 25% dm, compared to 2.0:1–2.5:1 for dry feed (Brooks et al., 1989; Van der Peet-Schwering and Plagge, 1995; Smolders and Hoofs, 2000). Even water to feed ratios of 1.5: 1 are reported without negative effects (Brumm et al., 2000).This suggests that the physiological requirement for water is met, but does not necessarily mean that no additional fresh water is needed to satisfy other (behavioural) requirements or covers individual variation in requirements (Brooks et al, 1989). Information about additional fresh water intake of liquid fed growing-finishing pigs is scarce. Smolders and Hoofs (2000) found an additional fresh water use of 0.86 l per pig per day, by pigs from 25 to 109 kg with the water to feed ratio decreasing from 3.1: 1 in the beginning and 2.9: 1 at the end of the growing-finishing period. Geary et al. (1996) compared different liquid feeds ranging from 15 to 25% dm for weaned piglets and measured the additional water use. At lower dry matter levels pigs ate more feed to keep their feed intake at the same level. Below 22.4% dm the additional fresh water use was stable at 0.22 l/pig/day. They conclude that even at low dry matter levels pigs keep drinking fresh water. However, it is unclear whether this water intake is associated with a nutritional or behavioural need, or whether it is redirected exploratory or ‘playing’ behaviour. Pigs in otherwise barren environments will tend to investigate and play with any objects enriching their environment. Stalled sows will develop behavioural routines or stereotypies directed at the nipple drinker, leading to excessive use of water (Rushen, 1984; Terlouw et al., 1991) which indicates that measuring water disappearance (as opposed to water intake) may be introducing important errors in the assessment of water requirements. Kyriazakis and Savory (1997) conclude that more meaningful methods to assess the motivation to obtain water are operant methods or aversion methods. Matthews and Ladewig (1994) used an operant technique to determine how hard pigs are willing to work for food in comparison to social contact. They produced so called demand curves by measuring the effort (i.e. 1 to 30 pushes on a button) required to get access to the “reward”. Similar to
113
Fig. 2. Layout of a pen, indicating the location of the four feeding systems (D = Dry feed; LT = Long trough; S = Sensor; V = Variomix).
demand curves in macro economics, these curves can be elastic or inelastic. Elasticity is defined by Lea (1978) as Elasticity = ð ð ΔyÞ = yÞ = ððΔxÞ = xÞ In which y = reward obtained and x = fixed ratio of effort to be made. In economic terms, the expenditure of a product is unaffected by price if elasticity is 1. This means that with increasing effort to be made, the amount of reward obtained will reduce linearly. Coefficients below or above 1 indicate less or more elasticity in demand, respectively (Lea, 1978; Jensen and Pedersen, 2008). Thus an elastic curve shows a rapid decline in level of access to the reward, when increasing effort is needed to obtain it. Inelastic curves are closer to a horizontal line: even when the effort required increases a lot, the animal will still try to get a similar level of access to the reward (and thus work increasingly hard). In Fig. 1, based on data from Matthews and Ladewig (1994), the amount of food received (solid line) and social contact received (dashed line) is projected against increasing effort (increasing ‘fixed ratio’) to obtain the rewards. The rate of decline of the demand curve for food is much closer to 0 (therefore ‘inelastic’) compared to the demand curve for social contact, which is more than 1 and can be called ‘elastic’. The present experiment aimed to test the hypothesis that the demand for additional fresh water will be less elastic in finishing pigs on wet feeding systems, compared to dry feeding. The assessment was made using a consumer — demand technique, but instead of pushing a button more or less often, the flow rate of the drinker was reduced, so the pigs had to work harder (drink longer) for the same amount of water. 2. Material and methods The effort finishing pigs will make to obtain additional fresh drinking water was assessed in a trial comparing three wet feeding systems and one dry feeding system. 2.1. Animals
Fig. 1. An increase in effort required to obtain a resource (via increased fixed ratio) will lead to a greater decrease in obtaining social contact (elastic demand) compared to food (inelastic demand). (From: Matthews and Ladewig, 1994).
Two batches of 288 pigs each were used between May and August 2006 (batch 1) and November 2006 to February 2007 (batch 2). All animals had a Great Yorkshire × Dutch Landrace mother and a Tempo terminal boar father (synthetic line, Topigs
114
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
Table 1 Composition of the dry and liquid diets for the start (week 1–6) and the end (week 7–17) of the finishing period. Dry feed
Energy (MJ NE/kg dm) Crude protein (% in dm) Ash (% in dm) Crude fiber (% in dm)
Liquid feed
Start
End
Start
End
11.0 19.9 6.8 5.6
10.6 17.3 8.8 5.2
10.7 18.5 5.7 5.2
10.8 17.5 5.3 5.3
Nederland). Gilts and barrows were kept separately in groups. During the experiment the animals were between 10 and 27 weeks old, weighing between 25 and 115 kg. Before the experiment they were weaned at an age of 4 weeks and kept in groups of 25 and liquid fed (batch 1) and in groups of 25 or 50 and dry fed (batch 2). The animals were randomly allocated to a pen in one of the two finishing rooms used in this experiment. 2.2. Housing In each batch two rooms with 12 pens for 12 pigs per pen were used. A room had a central inspection alley with 6 pens on each side. A pen was 2.5 m wide and 5.0 m deep (1.0 m2 per pig). The floor had a 0.6 m deep concrete slatted floor in the front, a convex concrete floor of 2.9 m in the middle and a 1.5 m deep metal (tribar) slatted floor in the back. Pen partitions were solid in the front and fenced in the back on both sides of the slatted floor. Every pen had a chain suspended from the pen partition as environmental enrichment in the front of the pen. Fig. 2 shows the pen layout. The rooms were mechanically ventilated aiming for a start temperature of 24 °C steadily declining towards a temperature of 21 °C at the end of the finishing period. Lights were on from 7.00 h to 17.00 h. 2.3. Feed and water All feeders were positioned along the side pen partitions, starting from one of the front corners of the pen. A detailed description follows under “Treatments”. An additional drinker bowl was mounted to the fence above the slatted part in the back of the pen. The drinker was a nipple/bowl combination (Drik-O-Mat Standard, Egebjerg International A/S, Nykøbing, Denmark) with an adjustable flow rate. The height of the drinker nipple was 30 cm above the floor and the rim of the bowl 28 cm above the floor. The present generation of
drinkers with a nipple-bowl combination restrict spillage of water best (Li et al., 2005). Because water consumption and water spillage could not be distinguished in this experiment we will use the term “water disappearance”. The drinking water and water in the liquid feed was ground water originating from a well, analysed periodically on microbial and other pollutions by the Dutch Animal Health Service and assessed as “suitable as drinking water for animals”. Table 1 shows the composition of the feeds used. Diet changed gradually from protein rich grower feed to energy rich finisher feed in week 5 and 6 of the growing finishing period. The energy content of the dry matter in the end feed was 10.8 MJ NE/kg. The composition of the liquid feed was for 50% based on liquid by-products and for 50% on dry components. The calculated dry matter content of the liquid feed was 23.8% in week 1–6 and 23.4% in week 7–17. This is a water to feed ratio of 4.2:1. Where feed intake in kg dry matter per pig is mentioned, dry matter is standardised as feed with 88% dry matter, comparable to dry pelleted feed. 2.4. Treatments and experimental design Four treatments were used each with 3 pens per room. Every room had 3 blocks of 4 pens. The two batches with two rooms each resulted in 12 pens per treatment in the experiment. The four treatments were randomised per block and were not changed between the batches. The treatments were the following feeding systems: Dry feed (D) — the pigs were fed using a dry wet feeder, with dry feed and a standard nipple drinker (600 ml/min) in the feeder, 1 eating place per pen of 12 pigs; feed was available ad libitum; the feeder was 0.36 m wide and 0.36 m long, with the front side of the trough 0.15 m high (Verbakel BV, St. Oedenrode, Netherlands); Long Trough (LT) — the pigs were semi ad libitum fed liquid feed in a long trough with 12 feeding places, resulting in simultaneous eating; the feed is distributed three times per day; the trough was 3.60 m wide, 0.30 m long and the front side was 0.15 m high (Grow Feeder BV, Uden, Netherlands); Sensor (S) — the pigs were fed semi ad libitum liquid feed in a short trough with 4 feeding places; the feeder is filled 5 to 10 times per day (frequency increasing during the trial), feeder is empty just before new refill; the trough was 1.20 m wide, 0.30 m long and the front side was 0.15 m high (Grow Feeder BV, Uden, Netherlands);
Table 2 Performance of the pigs per treatment.
Number of pens⁎ Number of pigs culled Starting weight (kg) End weight (kg) Age at slaughter (d) Carcass weight (kg) Daily gain (g/d) Lean meat (%) Feed intake (kg 88%dm) Feed conversion (kg feed/kg weight gain) ⁎12 pigs per pen.
Dry feed
Long trough
Sensor
Variomix
SEM
Significant differences (p)
12 4 22.5 115.3 186.2 90.0 757 56.6 2.10 2.77
12 4 22.2 114.8 187.2 89.5 748 56.7 2.07 2.77
12 6 21.9 113.7 187.6 88.5 740 56.0 2.14 2.91
12 10 22.4 114.5 190.3 89.2 728 56.7 2.14 2.94
11.0 0.38 0.021 0.055
ns (0.29) ns (0.57) # (0.08) # (0.06)
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
115
Fig. 3. Relationship between room temperature in batch 1 and 2 per week with additional water disappearance per feeding system.
Variomix (V)— the pigs were fed semi ad libitum liquid feed in a single feeder; the feeder is filled 5 to 10 times per day (frequency increasing during the trial), feeder is empty just before new refill; the feeder was 0.36 m wide and 0.36 m long, with the front side of the trough 0.15 m high (Verbakel BV, St.Oedenrode, Netherlands). In addition to the water available as described above (including the standard nipple drinker for the Dry feed treatment), the pigs could drink water from a separate drinking bowl. The flow rate of the nipple in this additional drinker could be adjusted to four different levels. Each week the flow rate was changed according to a randomised scheme. A batch of pigs was monitored for 12 weeks and every flow rate was used during one week per block of 4 weeks. The average flow rates were 134, 356, 733 and 1041 ml/min. The flow rates were randomly sampled by collecting water during 1 min with one sample per flow rate per pen per batch.
2.5. Water measuring system The water disappearance was measured at each of the 30 drinking points: 24 additional drinking bowls and 6 standard nipple drinkers in the feeders of the dry feed treatments. Each drinking point had a little 3 l container suspended on the wall 2 m above the slatted floor. This container had an own valve, which was controlled by a process computer (Fancom F92, Fancom BV, Panningen, Netherlands). Two sensors in the container measured the minimal and the maximal water level, switching the valve on and off. Only one of the 30 valves could be open at a time. Every filling pulse was registered by the central process computer and accounted for 3.0 l. The number of litre per hour per valve was recorded and transferred to a PC. The hourly figures were used to describe 24 h disappearance patterns. Per day the total water disappearance of a pen was calculated. The day of the change of flow rate and the day after (adaptation) were excluded, so five days per week
Fig. 4. Water disappearance pattern over 24 h per batch and per treatment (error bars = SEM).
116
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
Table 3 Use of water and feed per source per treatment, averaged over all flow rates. Treatment
Dry feeder (dry)
Long trough (liquid)
Sensor feeding (liquid)
Variomix (liquid)
SEM
p
Total water (l/pig/d) Additional bowl (l/pig/d)⁎ Standard nipple (l/pig/d) Water in feed (l/pig/d) Feed intake (kg dm/pig/d)
4.72a 3.39a 1.09 0.24 2.10
7.18b 0.74b – 6.44 2.07
7.36b 0.56bc – 6.80 2.14
7.27b 0.42c – 6.85 2.14
0.10 0.74
p b 0.001 p b 0.001
0.02
p = 0.076)
Different superscripts in a row mean a significant difference (p b 0.05). ⁎p b 0.001.
were used for the analysis. All the data of 16 days with technical problems in the measuring system were excluded from the data set with 120 days, resulting in 104 remaining days (87%). 2.6. Observations To calculate performance, weekly feed intake per pen was registered, as well as the start and the end weight of the animals. Veterinary treatments and mortality were also recorded. The observations on water disappearance started in the second week of a batch and continued to the 13th week. The flow rates were not changed after week 13. The pigs were transported to the slaughterhouse between week 15 and 18 at a live weight around 115 kg.
changing water disappearance. In the first batch the room temperature was 27.9 °C with 1.28 l per pig/d water disappearance from the additional test drinker and in batch 2 the room temperature was 24.7 °C with a water disappearance of 1.29 l per pig/d. 3.2. 24 h pattern water disappearance The water disappearance patterns from the additional drinking bowls over 24 h showed different patterns in the two batches. The first in the summer had typically two peaks (morning and afternoon), and the second batch in the autumn/ winter typically only had one high peak in the afternoon (Fig. 4). The total daily additional water disappearance in batch 1 and batch 2 did not differ.
2.7. Analysis There were three replicates (pens) of each feeding system randomly allocated in each housing unit. The water flow treatment was randomly allocated to week number within pen (Latin square). Response parameters (total water and additional water from bowl) were analyzed statistically using REML (Residual Maximum Likelyhood variance components analysis) with Genstat 8 (2002), with pen as experimental unit, according to the following model: Y P ijkl
− p 4flow ej+P e jk + P el+P e ijkl = α 0 + α 1i 4 1 − e i +P
where: Y = response parameter (additional water disappearance), α0 = intercept in experiment i, (α0 + α1i) = asymptotic value in feeding system i, ρi = rate parameter in feeding system i, Block (e_j) = effect of housing unit, Block (e_jk) = effect of pen within housing unit, Block (e_l) = effect of week number, Block (e_jl)=effect week number within housing unit, and Error=Error term (e_ijkl)The performance parameters (daily gain, lean meat, feed intake, feed conversion ratio) were analysed using Analysis of Variance with the model: Y=µ+treatment+batch+e.
3.3. Total water disappearance and water dispensed from the drinkers The total water disappearance is the sum of three sources: water in the feed, water from the additional test drinker and water from the standard nipple in the feeder of the dry feed treatment. Table 3 shows the totals for the treatments. 3.4. Water disappearance per flow rate The water disappearance decreases with decreasing flow rate in the additional drinker in all feeding systems. The level between systems is different as shown in Table 3. The predicted means were used to determine asymptotic curves as shown in Fig. 5. These curves reach their asymptote (maximum) at significantly different levels (p b 0.001) with Dry
3. Results The pigs started at an average age and bodyweight of 63.4 days and 22.2 kg respectively. They were slaughtered at an average age of 27 weeks (115 kg). Table 2 shows the performance per treatment. There were no significant differences (at pb 0.05). 3.1. Water disappearance pattern in the finishing period The 12-week measuring period started in week 2 and ended in week 13 of the growing finishing period. Fig. 3 shows a decrease in room temperature in time, but not resulting in a
Fig. 5. Predicted means and fitted curves of additional water disappearance per flow rate in four feeding systems.
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
117
Table 4 Parameters used in the formula to describe the asymptotic curve per feeding system. α1i ⁎ (1 − e− ρi ⁎ flowrate)
Dry feed
Long trough
Sensor
Variomix
Significance
α1i (asymptotic value)⁎ ρi (rate parameter)⁎
4.361a 0.00378a
1.028b 0.00274ab
1.187bc 0.00122b
0.550c 0.00275ab
p b 0.001 p = 0.087
Different superscripts in a row mean a significant difference (p b 0.05).
being the highest, followed by Long Trough and Sensor, and finally Variomix being the lowest (Table 4). The rate parameter, the value of ρi in the formula α1i ⁎ (1 − e− ρi ⁎ flowrate), determines the steepness of the curves. The slope of the curve for the Sensor treatment is less than for Dry and not different from Variomix and Long Trough. Log transformations of both x and y values of Fig. 5 result in Fig. 6 with linear regression lines. This showing that Sensor is more elastic than the two other liquid feed treatments. The slope of the Dry feed line did not differ from the slope of the liquid feed regression lines. The pigs in the Dry feed treatment had the choice to drink from the nipple in the feeder or to drink from the additional drinker, whereas the animals in the liquid feeding systems had a fixed water to feed ratio in the feed. The standard nipple in the feeder had fixed flow rate of 600 ml/min. In the Dry fed pigs the water disappearance from this standard nipple increased as the flow rate in the additional drinker decreased. At the highest flow rate pigs preferred the additional drinker, and at the lowest flow rate the water disappearance of both drinkers was almost equal. Table 5 shows the water disappearance in the Dry fed pigs per flow rate from the additional drinking bowl. The pigs kept the total water to feed ratio on a standard level between 2.0:1 and 2.5:1 at the different flow rates (Brooks et al., 1989; Van der Peet-Schwering and Plagge, 1995; Smolders and Hoofs, 2000). 4. Discussion In Dutch systems with ad libitum dry feed and free water availability the ratio between water and feed is between 2.0:1 and 3.0:1 (Van der Peet-Schwering and Plagge, 1995; Plagge and Van Leuteren, 1989; Plagge, 1991). In liquid feed the dry
matter content is around 25%, this means a water to feed ratio of 3:1. In this experiment the average water to feed ratio in the Dry feed treatment was 2.2:1 and in the liquid treatment 3.5:1. Even when the water disappearance from the additional drinker is excluded, the water to feed ratio in the liquid treatments was 3.2:1. This indicates a physiologically sufficient water supply for the average pig (Mroz et al., 1995). The water to feed ratio for the Dry fed pigs varied from 2.0:1 to 2.5:1 at the four flow rates, indicating that the additional drinker for the Dry fed pigs did not lead to a higher water disappearance compared to conventional systems with one drinker (despite the relative high room temperature of 26 °C). Overall it can be assumed that for Dry fed pigs there was no physiological need for additional water, because the water disappearance levels appeared to be in the normal range. This was supported by the absence of differences in performance and carcass data. The additional water disappearance was low for the liquid feed treatments with a water disappearance between 0.42 and 0.74 l/pig ⁎ day compared to 3.39 l/pig ⁎ day in the Dry fed pigs. This is a water disappearance of only 8% of the total daily water disappearance of the liquid fed pigs. The majority of the additional water disappearance occurred in the early morning and late afternoon. During the night the water disappearance was on a continuous low level and there were no indications of drinking behaviour compensating the meals during the daylight period. This was confirmed by the water disappearance by the Dry fed pigs during the early evening being higher than that of the liquid fed pigs. The pattern in the summer showed two drinking peaks and in the winter only one in all treatments, which is comparable to findings by Brumm (2006). Although only one drinker was provided, this did not appear to restrict access to additional water. From the data it can be calculated that the occupation of the drinker at the highest flow rate was around 12 min per day for the liquid fed pigs and 48 min for the Dry fed pigs. At the lowest flow rates this was 30 min and 2.5 h respectively, always leaving the additional drinker free for more than 21 h per day. The Dry fed pigs had a preference for the additional drinker at high flow rates. This preference moved gradually to the drinker in the feeder, as flow rate of the additional drinker decreases. However, access to the drinker in the feeder was influenced by feeding behaviour of other pigs. In wet/dry feeders the pigs need 45 to 85 min to eat, depending on form of the feed and availability of water in the trough (Ramaekers, Table 5 Water disappearance of the Dry fed pigs from the additional drinking bowl and from the nipple drinker in the feeder per flow rate.
Fig. 6. Relation between logarithmic (ln) transformations of both flow rate and water disappearance for the four treatments with linear regression lines; the figures near the lines are the lineair regression coefficients with different superscripts indicating p b 0.05.
Flow rate from drinking bowl (ml/min)
134
356
733
1041
Water from drinking bowl (l/d) Water from feeder, 600 ml/min (l/d)
1.96 1.73
3.23 1.03
4.01 0.93
4.31 0.68
118
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
1996; Gonyou and Lou, 2000). With a feeding duration of 1 h per pig per day and 12 pigs per pen, this feeder-drinker combination is occupied for more than half of the day, resulting in social pressure on accessibility to the feeder (including the standard drinker) during the preferred drinking and feeding periods. This pressure on the standard drinker may have been lower if it had been spatially separated from the feed. It can be hypothesised that if the standard drinker had been placed away from the feeder and was thereby made more accessible, the effort made by the pigs to obtain water from the additional drinker would have been less. The present experiment aimed to test the hypothesis that the demand for additional fresh water will be less elastic in finishing pigs on wet feeding systems, compared to dry feeding. To do this, we assessed the rate parameter ρ of the fitted curve, which is a measure of the perseverance of the animals to obtain water (given that flow rates were progressively reduced). The statistical analysis showed that the exponential fit was better than the more classical log–log fitted lineair regression lines used by Matthews and Ladewig (1994). As Jensen and Pedersen (2008) stated, the change in demand (slope) varies with the price. However the outcome of both methods was equal. As presented in Table 4 and Fig. 6 the rate parameter was lower (steeper) for Sensor than for Dry and not different from the other two treatments. This suggests a more ‘elastic’ demand for Sensor compared to Dry fed pigs (Matthews and Ladewig, 1994). This could be caused by a combination of two factors: the number of feeding places and the number of meals per system. The more elastic demand in Sensor can be caused by the 4 feeding places with 5–10 feeding times per day. In Variomix there was only one feeding place and three feeding starts per day with probably more competition and a more persistent water disappearance at low flow rates. In Long Trough the number of meals per day was also three and the need for additional water between the meals could have been higher. None of the liquid feeding systems resulted in a less elastic demand compared to the Dry feeding system. The liquid fed pigs did not work harder for additional water and the maximum water disappearance from the additional drinker was low. This suggests that pigs fed via Long Trough and Variomix systems are equally persistent in obtaining additional water from an extra drinking nipple compared to Dry fed pigs with a standard drinking nipple, and that Sensor fed pigs are less persistent compared to Dry fed pigs. Acknowledgement The authors thank the Dutch Product Board for Livestock and Meat for funding this research project.
References Brooks, P.H., Carpenter, J.L., Barber, J., Gill, B.P., 1989. Production and welfare problems relating to the supply of water to growing-finishing pigs. Pig J. 23, 51–66. Brumm, M., 2006. Patterns of drinking water use in pork production facilities. Nebraska Swine Report, pp. 10–13. Brumm, M.C., Dahlquist, J.M., Heemstra, J.M., 2000. Impact of feeders and drinker devices on pig performance, water use, and manure volume. J. Swine Health Prod. 8 (2), 51–57. Geary, T.M., Brooks, P.H., Morgan, D.T., Campbell, A., Russell, P.J., 1996. Performance of weaner pigs fed ad libitum with liquid feed at different dry matter concentrations. J. Sci. Food Agric. 72, 17–24. GenStat 8 Committee, 2002. Genstat 8 Reference Manual; Release 3. Clarendon Press, Oxford, UK. Gonyou, H.W., Lou, Z., 2000. Effects of eating space and availability of water in feeders on productivity and eating behaviour of grower/finisher pigs. J. Anim. Sci. 78, 865–870. Jensen, M.B., Pedersen, L.J., 2008. Using motivation tests to assess ethological needs and preferences. Appl. Anim Behav. Sci. 113, 340–356. Kyriazakis, I., Savory, C.J., 1997. Hunger and thirst. In: Appleby, M.C., Hughes, B.O. (Eds.), Animal Welfare. CABI, Wallingford, pp. 49–62. Lea, S.E.G.,1978. The psychology and economics of demand. Psychol. Bull. 85 (3), 441–466. Li, Y.Z., Chénard, L., Lemay, S.P., Gonyou, H.W., 2005. Water intake and wastage at nipple drinkers by growing-finishing pigs. J. Anim. Sci. 83, 1413–1422. Matthews, L.R., Ladewig, J., 1994. Environmental requirements of pigs measured by behavioural demand functions. Anim. Behav. 47, 713–719. Mroz, Z., Jongbloed, A.W., Lenis, N.P., Vreman, K., 1995. Water in pig nutrition: physiology, allowances and environmental implications. Nutr. Res. Rev. 8, 137–164. N.N., 2001. Minimum standards for the protection of pigs. Commission directive 2001/93/EC. Off. J. Eur. Communities L 316, 36–38. Nagai, M., Hachimura, K., Takahashi, K., 1994. Water consumption in suckling pigs. J. Vet. Med. Sci. 56, 181–183. Plagge, J.G., 1991. De invloed van beperking van de drinktijd op het waterverbruik en technische resultaten bij vleesvarkens (The effect of limiting drinking time on water use and performance of finishing pigs). Research Report P1.64, Research Institute for Pig Husbandry, Rosmalen, The Netherlands. Plagge, J.G., Leuteren, J. van, 1989. Waterverbruik bij onbeperkt gevoerde mestvarkens (Water use of ad libitum fed finishing pigs). Research Report P1.32, Research Instute for Pig Husbandry, Rosmalen, The Netherlands. Ramaekers, P.J.L., 1996. Control of individual daily growth in group-housed pigs using feeding stations. Thesis, Wageningen University, Wageningen, 133 p. Rushen, J., 1984. Stereotyped behaviour, adjunctive drinking and the feeding period of tethered sows. Anim. Behav. 32, 1059–1067. Smolders, M.A.H.H., Hoofs, A.I.J., 2000. Onbeperkte drinkwaterverstrekking naast een brijvoerrantsoen met bijproducten bij vleesvarkens [Unrestricted drinking water supply for finishing pigs on liquid feed with bye products]. Research Report P4.45, Research Institute for Pig Husbandry, Rosmalen, The Netherlands, 8 p. Terlouw, E.M.C., Lawrence, A.B., Illius, A.W., 1991. Influences of feeding level and physical restriction on development of stereotypies in sows. Anim. Behav. 42, 981–991. Vandenheede, M., Nicks, B., 1991. L'approvisionnement en eau des porcs: un élément a ne pas négliger [Water requirements and drinking water systems for pigs.]. Ann. Med. Vet. 135, 123–128. Van der Peet-Schwering, C.M.C., Plagge, J.G., 1995. Influence of multi phase feeding on the performance and water intake of sows and barrows. Research Report P1.140, Research Institute for Pig Husbandry, Rosmalen, The Netherlands.
Contents lists available at ScienceDirect
Livestock Science j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i v s c i
Motivation for additional water use of growing-finishing pigs Herman M. Vermeer ⁎, Nienke Kuijken, Hans A.M. Spoolder Animal Sciences Group of Wageningen UR, Lelystad, The Netherlands
a r t i c l e
i n f o
Article history: Received 5 November 2007 Received in revised form 23 December 2008 Accepted 14 January 2009 Keywords: Pigs Water Drinking Motivation Behaviour
a b s t r a c t Liquid fed growing-finishing pigs receive an amount of water in their ration that is more than their physiological requirement. For welfare reasons it can be argued that in addition to this diet, pigs may be motivated to obtain additional fresh water. The aim of the present experiment was to test the hypotheses that liquid fed pigs will work harder to obtain extra fresh water, compared to dry fed pigs which receive water in a conventional way. A consumer-demand technique was used, in which flow rate from an extra (test) drinker determined the ease with which pigs could obtain the water. The more persistent pigs were to obtain water from the test drinker (with declining flow rates), the harder they were assumed to work for it. Four treatments were divided over 48 pens of 12 finishing pigs in 2 batches (566 pigs). There was one Dry Feed treatment (D, with standard drinking nipple in a single space trough) and three liquid feeding systems: Long trough (LT); Sensor Feeding (S) and Variomix (V). Each pen had an additional drinker with a weekly randomly changing flow rate of 134, 356, 733 or 1041 ml/ min. From the extra drinker pigs used on average 3.39a (D), 0.76b (LT), 0.58bc (S) and 0.44c (V) litre per day (different superscripts indicate differences P b 0.05). The persistence to obtain water differed between the four treatments. This was indicated by the rate of decline (ρi) of the asymptotic curve depicting water disappearance at decreasing flow rates: ρi = 0.00378a, 0.00274ab, 0.00122b and 0.00275ab for D, LT, S and V, respectively. This suggests that liquid fed pigs work equally hard (LT and V) or less hard (S) to obtain water from an extra drinking nipple, compared to dry fed pigs (D). © 2009 Elsevier B.V. All rights reserved.
1. Introduction Water is an essential part of the nutritive and welfare requirements of pigs. Mroz et al. (1995) stated in a review that water intake mainly depends on body weight, feed intake and temperature. Heavier pigs need more water to maintain their body. The daily water intake of ad libitum fed finishing pigs increases on average from 2 l at 25 kg to 6 l at about 110 kg live weight (Nagai et al., 1994). Mroz et al. (1995) also found in their review that the water to feed ratio decreases with increasing age or weight and increases with the ambient temperature. When ambient temperature increases from 10 to 25 °C, the need for water for evaporative cooling, mainly via respiration,
⁎ Corresponding author. Animal Sciences Group of Wageningen UR, P.O. Box 65, 8200 AB Lelystad, The Netherlands. Tel.: +320 293 378. E-mail address: [email protected] (H.M. Vermeer). 1871-1413/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.livsci.2009.01.009
increases from 2.2 to 4.2 l/d for finishing pigs (Vandenheede and Nicks, 1991). Water can be offered to the pig as part of a liquid diet, or as plain, fresh water from a drinker. The welfare of pigs is compromised if water is unavailable (Kyriazakis and Savory, 1997). The EU minimum standards for the protection of pigs (N.N., 2001) state: “All pigs over two weeks of age must have permanent access to a sufficient quantity of fresh water.” Although intuitively logical, this requirement warrants further investigation following the development of new feeding systems as well as the current pressures on environmental aspects of pig husbandry. An increase in water consumption inevitably leads to an increase in urine production. It can be questioned if the need for fresh water does indeed exist, providing the physiological and behavioural needs for water uptake have been met by the ration. Part of the growingfinishing pigs is fed liquid feed, instead of dry feed with additional fresh water. Fresh water differs considerably from
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
liquid feed. Liquid feed has a high energy content and can incidentally have high concentrations of salt and acid. However, the water to feed ratio in these systems is much higher than in dry feeding systems: around 3.0: 1 for wet feed with 25% dm, compared to 2.0:1–2.5:1 for dry feed (Brooks et al., 1989; Van der Peet-Schwering and Plagge, 1995; Smolders and Hoofs, 2000). Even water to feed ratios of 1.5: 1 are reported without negative effects (Brumm et al., 2000).This suggests that the physiological requirement for water is met, but does not necessarily mean that no additional fresh water is needed to satisfy other (behavioural) requirements or covers individual variation in requirements (Brooks et al, 1989). Information about additional fresh water intake of liquid fed growing-finishing pigs is scarce. Smolders and Hoofs (2000) found an additional fresh water use of 0.86 l per pig per day, by pigs from 25 to 109 kg with the water to feed ratio decreasing from 3.1: 1 in the beginning and 2.9: 1 at the end of the growing-finishing period. Geary et al. (1996) compared different liquid feeds ranging from 15 to 25% dm for weaned piglets and measured the additional water use. At lower dry matter levels pigs ate more feed to keep their feed intake at the same level. Below 22.4% dm the additional fresh water use was stable at 0.22 l/pig/day. They conclude that even at low dry matter levels pigs keep drinking fresh water. However, it is unclear whether this water intake is associated with a nutritional or behavioural need, or whether it is redirected exploratory or ‘playing’ behaviour. Pigs in otherwise barren environments will tend to investigate and play with any objects enriching their environment. Stalled sows will develop behavioural routines or stereotypies directed at the nipple drinker, leading to excessive use of water (Rushen, 1984; Terlouw et al., 1991) which indicates that measuring water disappearance (as opposed to water intake) may be introducing important errors in the assessment of water requirements. Kyriazakis and Savory (1997) conclude that more meaningful methods to assess the motivation to obtain water are operant methods or aversion methods. Matthews and Ladewig (1994) used an operant technique to determine how hard pigs are willing to work for food in comparison to social contact. They produced so called demand curves by measuring the effort (i.e. 1 to 30 pushes on a button) required to get access to the “reward”. Similar to
113
Fig. 2. Layout of a pen, indicating the location of the four feeding systems (D = Dry feed; LT = Long trough; S = Sensor; V = Variomix).
demand curves in macro economics, these curves can be elastic or inelastic. Elasticity is defined by Lea (1978) as Elasticity = ð ð ΔyÞ = yÞ = ððΔxÞ = xÞ In which y = reward obtained and x = fixed ratio of effort to be made. In economic terms, the expenditure of a product is unaffected by price if elasticity is 1. This means that with increasing effort to be made, the amount of reward obtained will reduce linearly. Coefficients below or above 1 indicate less or more elasticity in demand, respectively (Lea, 1978; Jensen and Pedersen, 2008). Thus an elastic curve shows a rapid decline in level of access to the reward, when increasing effort is needed to obtain it. Inelastic curves are closer to a horizontal line: even when the effort required increases a lot, the animal will still try to get a similar level of access to the reward (and thus work increasingly hard). In Fig. 1, based on data from Matthews and Ladewig (1994), the amount of food received (solid line) and social contact received (dashed line) is projected against increasing effort (increasing ‘fixed ratio’) to obtain the rewards. The rate of decline of the demand curve for food is much closer to 0 (therefore ‘inelastic’) compared to the demand curve for social contact, which is more than 1 and can be called ‘elastic’. The present experiment aimed to test the hypothesis that the demand for additional fresh water will be less elastic in finishing pigs on wet feeding systems, compared to dry feeding. The assessment was made using a consumer — demand technique, but instead of pushing a button more or less often, the flow rate of the drinker was reduced, so the pigs had to work harder (drink longer) for the same amount of water. 2. Material and methods The effort finishing pigs will make to obtain additional fresh drinking water was assessed in a trial comparing three wet feeding systems and one dry feeding system. 2.1. Animals
Fig. 1. An increase in effort required to obtain a resource (via increased fixed ratio) will lead to a greater decrease in obtaining social contact (elastic demand) compared to food (inelastic demand). (From: Matthews and Ladewig, 1994).
Two batches of 288 pigs each were used between May and August 2006 (batch 1) and November 2006 to February 2007 (batch 2). All animals had a Great Yorkshire × Dutch Landrace mother and a Tempo terminal boar father (synthetic line, Topigs
114
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
Table 1 Composition of the dry and liquid diets for the start (week 1–6) and the end (week 7–17) of the finishing period. Dry feed
Energy (MJ NE/kg dm) Crude protein (% in dm) Ash (% in dm) Crude fiber (% in dm)
Liquid feed
Start
End
Start
End
11.0 19.9 6.8 5.6
10.6 17.3 8.8 5.2
10.7 18.5 5.7 5.2
10.8 17.5 5.3 5.3
Nederland). Gilts and barrows were kept separately in groups. During the experiment the animals were between 10 and 27 weeks old, weighing between 25 and 115 kg. Before the experiment they were weaned at an age of 4 weeks and kept in groups of 25 and liquid fed (batch 1) and in groups of 25 or 50 and dry fed (batch 2). The animals were randomly allocated to a pen in one of the two finishing rooms used in this experiment. 2.2. Housing In each batch two rooms with 12 pens for 12 pigs per pen were used. A room had a central inspection alley with 6 pens on each side. A pen was 2.5 m wide and 5.0 m deep (1.0 m2 per pig). The floor had a 0.6 m deep concrete slatted floor in the front, a convex concrete floor of 2.9 m in the middle and a 1.5 m deep metal (tribar) slatted floor in the back. Pen partitions were solid in the front and fenced in the back on both sides of the slatted floor. Every pen had a chain suspended from the pen partition as environmental enrichment in the front of the pen. Fig. 2 shows the pen layout. The rooms were mechanically ventilated aiming for a start temperature of 24 °C steadily declining towards a temperature of 21 °C at the end of the finishing period. Lights were on from 7.00 h to 17.00 h. 2.3. Feed and water All feeders were positioned along the side pen partitions, starting from one of the front corners of the pen. A detailed description follows under “Treatments”. An additional drinker bowl was mounted to the fence above the slatted part in the back of the pen. The drinker was a nipple/bowl combination (Drik-O-Mat Standard, Egebjerg International A/S, Nykøbing, Denmark) with an adjustable flow rate. The height of the drinker nipple was 30 cm above the floor and the rim of the bowl 28 cm above the floor. The present generation of
drinkers with a nipple-bowl combination restrict spillage of water best (Li et al., 2005). Because water consumption and water spillage could not be distinguished in this experiment we will use the term “water disappearance”. The drinking water and water in the liquid feed was ground water originating from a well, analysed periodically on microbial and other pollutions by the Dutch Animal Health Service and assessed as “suitable as drinking water for animals”. Table 1 shows the composition of the feeds used. Diet changed gradually from protein rich grower feed to energy rich finisher feed in week 5 and 6 of the growing finishing period. The energy content of the dry matter in the end feed was 10.8 MJ NE/kg. The composition of the liquid feed was for 50% based on liquid by-products and for 50% on dry components. The calculated dry matter content of the liquid feed was 23.8% in week 1–6 and 23.4% in week 7–17. This is a water to feed ratio of 4.2:1. Where feed intake in kg dry matter per pig is mentioned, dry matter is standardised as feed with 88% dry matter, comparable to dry pelleted feed. 2.4. Treatments and experimental design Four treatments were used each with 3 pens per room. Every room had 3 blocks of 4 pens. The two batches with two rooms each resulted in 12 pens per treatment in the experiment. The four treatments were randomised per block and were not changed between the batches. The treatments were the following feeding systems: Dry feed (D) — the pigs were fed using a dry wet feeder, with dry feed and a standard nipple drinker (600 ml/min) in the feeder, 1 eating place per pen of 12 pigs; feed was available ad libitum; the feeder was 0.36 m wide and 0.36 m long, with the front side of the trough 0.15 m high (Verbakel BV, St. Oedenrode, Netherlands); Long Trough (LT) — the pigs were semi ad libitum fed liquid feed in a long trough with 12 feeding places, resulting in simultaneous eating; the feed is distributed three times per day; the trough was 3.60 m wide, 0.30 m long and the front side was 0.15 m high (Grow Feeder BV, Uden, Netherlands); Sensor (S) — the pigs were fed semi ad libitum liquid feed in a short trough with 4 feeding places; the feeder is filled 5 to 10 times per day (frequency increasing during the trial), feeder is empty just before new refill; the trough was 1.20 m wide, 0.30 m long and the front side was 0.15 m high (Grow Feeder BV, Uden, Netherlands);
Table 2 Performance of the pigs per treatment.
Number of pens⁎ Number of pigs culled Starting weight (kg) End weight (kg) Age at slaughter (d) Carcass weight (kg) Daily gain (g/d) Lean meat (%) Feed intake (kg 88%dm) Feed conversion (kg feed/kg weight gain) ⁎12 pigs per pen.
Dry feed
Long trough
Sensor
Variomix
SEM
Significant differences (p)
12 4 22.5 115.3 186.2 90.0 757 56.6 2.10 2.77
12 4 22.2 114.8 187.2 89.5 748 56.7 2.07 2.77
12 6 21.9 113.7 187.6 88.5 740 56.0 2.14 2.91
12 10 22.4 114.5 190.3 89.2 728 56.7 2.14 2.94
11.0 0.38 0.021 0.055
ns (0.29) ns (0.57) # (0.08) # (0.06)
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
115
Fig. 3. Relationship between room temperature in batch 1 and 2 per week with additional water disappearance per feeding system.
Variomix (V)— the pigs were fed semi ad libitum liquid feed in a single feeder; the feeder is filled 5 to 10 times per day (frequency increasing during the trial), feeder is empty just before new refill; the feeder was 0.36 m wide and 0.36 m long, with the front side of the trough 0.15 m high (Verbakel BV, St.Oedenrode, Netherlands). In addition to the water available as described above (including the standard nipple drinker for the Dry feed treatment), the pigs could drink water from a separate drinking bowl. The flow rate of the nipple in this additional drinker could be adjusted to four different levels. Each week the flow rate was changed according to a randomised scheme. A batch of pigs was monitored for 12 weeks and every flow rate was used during one week per block of 4 weeks. The average flow rates were 134, 356, 733 and 1041 ml/min. The flow rates were randomly sampled by collecting water during 1 min with one sample per flow rate per pen per batch.
2.5. Water measuring system The water disappearance was measured at each of the 30 drinking points: 24 additional drinking bowls and 6 standard nipple drinkers in the feeders of the dry feed treatments. Each drinking point had a little 3 l container suspended on the wall 2 m above the slatted floor. This container had an own valve, which was controlled by a process computer (Fancom F92, Fancom BV, Panningen, Netherlands). Two sensors in the container measured the minimal and the maximal water level, switching the valve on and off. Only one of the 30 valves could be open at a time. Every filling pulse was registered by the central process computer and accounted for 3.0 l. The number of litre per hour per valve was recorded and transferred to a PC. The hourly figures were used to describe 24 h disappearance patterns. Per day the total water disappearance of a pen was calculated. The day of the change of flow rate and the day after (adaptation) were excluded, so five days per week
Fig. 4. Water disappearance pattern over 24 h per batch and per treatment (error bars = SEM).
116
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
Table 3 Use of water and feed per source per treatment, averaged over all flow rates. Treatment
Dry feeder (dry)
Long trough (liquid)
Sensor feeding (liquid)
Variomix (liquid)
SEM
p
Total water (l/pig/d) Additional bowl (l/pig/d)⁎ Standard nipple (l/pig/d) Water in feed (l/pig/d) Feed intake (kg dm/pig/d)
4.72a 3.39a 1.09 0.24 2.10
7.18b 0.74b – 6.44 2.07
7.36b 0.56bc – 6.80 2.14
7.27b 0.42c – 6.85 2.14
0.10 0.74
p b 0.001 p b 0.001
0.02
p = 0.076)
Different superscripts in a row mean a significant difference (p b 0.05). ⁎p b 0.001.
were used for the analysis. All the data of 16 days with technical problems in the measuring system were excluded from the data set with 120 days, resulting in 104 remaining days (87%). 2.6. Observations To calculate performance, weekly feed intake per pen was registered, as well as the start and the end weight of the animals. Veterinary treatments and mortality were also recorded. The observations on water disappearance started in the second week of a batch and continued to the 13th week. The flow rates were not changed after week 13. The pigs were transported to the slaughterhouse between week 15 and 18 at a live weight around 115 kg.
changing water disappearance. In the first batch the room temperature was 27.9 °C with 1.28 l per pig/d water disappearance from the additional test drinker and in batch 2 the room temperature was 24.7 °C with a water disappearance of 1.29 l per pig/d. 3.2. 24 h pattern water disappearance The water disappearance patterns from the additional drinking bowls over 24 h showed different patterns in the two batches. The first in the summer had typically two peaks (morning and afternoon), and the second batch in the autumn/ winter typically only had one high peak in the afternoon (Fig. 4). The total daily additional water disappearance in batch 1 and batch 2 did not differ.
2.7. Analysis There were three replicates (pens) of each feeding system randomly allocated in each housing unit. The water flow treatment was randomly allocated to week number within pen (Latin square). Response parameters (total water and additional water from bowl) were analyzed statistically using REML (Residual Maximum Likelyhood variance components analysis) with Genstat 8 (2002), with pen as experimental unit, according to the following model: Y P ijkl
− p 4flow ej+P e jk + P el+P e ijkl = α 0 + α 1i 4 1 − e i +P
where: Y = response parameter (additional water disappearance), α0 = intercept in experiment i, (α0 + α1i) = asymptotic value in feeding system i, ρi = rate parameter in feeding system i, Block (e_j) = effect of housing unit, Block (e_jk) = effect of pen within housing unit, Block (e_l) = effect of week number, Block (e_jl)=effect week number within housing unit, and Error=Error term (e_ijkl)The performance parameters (daily gain, lean meat, feed intake, feed conversion ratio) were analysed using Analysis of Variance with the model: Y=µ+treatment+batch+e.
3.3. Total water disappearance and water dispensed from the drinkers The total water disappearance is the sum of three sources: water in the feed, water from the additional test drinker and water from the standard nipple in the feeder of the dry feed treatment. Table 3 shows the totals for the treatments. 3.4. Water disappearance per flow rate The water disappearance decreases with decreasing flow rate in the additional drinker in all feeding systems. The level between systems is different as shown in Table 3. The predicted means were used to determine asymptotic curves as shown in Fig. 5. These curves reach their asymptote (maximum) at significantly different levels (p b 0.001) with Dry
3. Results The pigs started at an average age and bodyweight of 63.4 days and 22.2 kg respectively. They were slaughtered at an average age of 27 weeks (115 kg). Table 2 shows the performance per treatment. There were no significant differences (at pb 0.05). 3.1. Water disappearance pattern in the finishing period The 12-week measuring period started in week 2 and ended in week 13 of the growing finishing period. Fig. 3 shows a decrease in room temperature in time, but not resulting in a
Fig. 5. Predicted means and fitted curves of additional water disappearance per flow rate in four feeding systems.
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
117
Table 4 Parameters used in the formula to describe the asymptotic curve per feeding system. α1i ⁎ (1 − e− ρi ⁎ flowrate)
Dry feed
Long trough
Sensor
Variomix
Significance
α1i (asymptotic value)⁎ ρi (rate parameter)⁎
4.361a 0.00378a
1.028b 0.00274ab
1.187bc 0.00122b
0.550c 0.00275ab
p b 0.001 p = 0.087
Different superscripts in a row mean a significant difference (p b 0.05).
being the highest, followed by Long Trough and Sensor, and finally Variomix being the lowest (Table 4). The rate parameter, the value of ρi in the formula α1i ⁎ (1 − e− ρi ⁎ flowrate), determines the steepness of the curves. The slope of the curve for the Sensor treatment is less than for Dry and not different from Variomix and Long Trough. Log transformations of both x and y values of Fig. 5 result in Fig. 6 with linear regression lines. This showing that Sensor is more elastic than the two other liquid feed treatments. The slope of the Dry feed line did not differ from the slope of the liquid feed regression lines. The pigs in the Dry feed treatment had the choice to drink from the nipple in the feeder or to drink from the additional drinker, whereas the animals in the liquid feeding systems had a fixed water to feed ratio in the feed. The standard nipple in the feeder had fixed flow rate of 600 ml/min. In the Dry fed pigs the water disappearance from this standard nipple increased as the flow rate in the additional drinker decreased. At the highest flow rate pigs preferred the additional drinker, and at the lowest flow rate the water disappearance of both drinkers was almost equal. Table 5 shows the water disappearance in the Dry fed pigs per flow rate from the additional drinking bowl. The pigs kept the total water to feed ratio on a standard level between 2.0:1 and 2.5:1 at the different flow rates (Brooks et al., 1989; Van der Peet-Schwering and Plagge, 1995; Smolders and Hoofs, 2000). 4. Discussion In Dutch systems with ad libitum dry feed and free water availability the ratio between water and feed is between 2.0:1 and 3.0:1 (Van der Peet-Schwering and Plagge, 1995; Plagge and Van Leuteren, 1989; Plagge, 1991). In liquid feed the dry
matter content is around 25%, this means a water to feed ratio of 3:1. In this experiment the average water to feed ratio in the Dry feed treatment was 2.2:1 and in the liquid treatment 3.5:1. Even when the water disappearance from the additional drinker is excluded, the water to feed ratio in the liquid treatments was 3.2:1. This indicates a physiologically sufficient water supply for the average pig (Mroz et al., 1995). The water to feed ratio for the Dry fed pigs varied from 2.0:1 to 2.5:1 at the four flow rates, indicating that the additional drinker for the Dry fed pigs did not lead to a higher water disappearance compared to conventional systems with one drinker (despite the relative high room temperature of 26 °C). Overall it can be assumed that for Dry fed pigs there was no physiological need for additional water, because the water disappearance levels appeared to be in the normal range. This was supported by the absence of differences in performance and carcass data. The additional water disappearance was low for the liquid feed treatments with a water disappearance between 0.42 and 0.74 l/pig ⁎ day compared to 3.39 l/pig ⁎ day in the Dry fed pigs. This is a water disappearance of only 8% of the total daily water disappearance of the liquid fed pigs. The majority of the additional water disappearance occurred in the early morning and late afternoon. During the night the water disappearance was on a continuous low level and there were no indications of drinking behaviour compensating the meals during the daylight period. This was confirmed by the water disappearance by the Dry fed pigs during the early evening being higher than that of the liquid fed pigs. The pattern in the summer showed two drinking peaks and in the winter only one in all treatments, which is comparable to findings by Brumm (2006). Although only one drinker was provided, this did not appear to restrict access to additional water. From the data it can be calculated that the occupation of the drinker at the highest flow rate was around 12 min per day for the liquid fed pigs and 48 min for the Dry fed pigs. At the lowest flow rates this was 30 min and 2.5 h respectively, always leaving the additional drinker free for more than 21 h per day. The Dry fed pigs had a preference for the additional drinker at high flow rates. This preference moved gradually to the drinker in the feeder, as flow rate of the additional drinker decreases. However, access to the drinker in the feeder was influenced by feeding behaviour of other pigs. In wet/dry feeders the pigs need 45 to 85 min to eat, depending on form of the feed and availability of water in the trough (Ramaekers, Table 5 Water disappearance of the Dry fed pigs from the additional drinking bowl and from the nipple drinker in the feeder per flow rate.
Fig. 6. Relation between logarithmic (ln) transformations of both flow rate and water disappearance for the four treatments with linear regression lines; the figures near the lines are the lineair regression coefficients with different superscripts indicating p b 0.05.
Flow rate from drinking bowl (ml/min)
134
356
733
1041
Water from drinking bowl (l/d) Water from feeder, 600 ml/min (l/d)
1.96 1.73
3.23 1.03
4.01 0.93
4.31 0.68
118
H.M. Vermeer et al. / Livestock Science 124 (2009) 112–118
1996; Gonyou and Lou, 2000). With a feeding duration of 1 h per pig per day and 12 pigs per pen, this feeder-drinker combination is occupied for more than half of the day, resulting in social pressure on accessibility to the feeder (including the standard drinker) during the preferred drinking and feeding periods. This pressure on the standard drinker may have been lower if it had been spatially separated from the feed. It can be hypothesised that if the standard drinker had been placed away from the feeder and was thereby made more accessible, the effort made by the pigs to obtain water from the additional drinker would have been less. The present experiment aimed to test the hypothesis that the demand for additional fresh water will be less elastic in finishing pigs on wet feeding systems, compared to dry feeding. To do this, we assessed the rate parameter ρ of the fitted curve, which is a measure of the perseverance of the animals to obtain water (given that flow rates were progressively reduced). The statistical analysis showed that the exponential fit was better than the more classical log–log fitted lineair regression lines used by Matthews and Ladewig (1994). As Jensen and Pedersen (2008) stated, the change in demand (slope) varies with the price. However the outcome of both methods was equal. As presented in Table 4 and Fig. 6 the rate parameter was lower (steeper) for Sensor than for Dry and not different from the other two treatments. This suggests a more ‘elastic’ demand for Sensor compared to Dry fed pigs (Matthews and Ladewig, 1994). This could be caused by a combination of two factors: the number of feeding places and the number of meals per system. The more elastic demand in Sensor can be caused by the 4 feeding places with 5–10 feeding times per day. In Variomix there was only one feeding place and three feeding starts per day with probably more competition and a more persistent water disappearance at low flow rates. In Long Trough the number of meals per day was also three and the need for additional water between the meals could have been higher. None of the liquid feeding systems resulted in a less elastic demand compared to the Dry feeding system. The liquid fed pigs did not work harder for additional water and the maximum water disappearance from the additional drinker was low. This suggests that pigs fed via Long Trough and Variomix systems are equally persistent in obtaining additional water from an extra drinking nipple compared to Dry fed pigs with a standard drinking nipple, and that Sensor fed pigs are less persistent compared to Dry fed pigs. Acknowledgement The authors thank the Dutch Product Board for Livestock and Meat for funding this research project.
References Brooks, P.H., Carpenter, J.L., Barber, J., Gill, B.P., 1989. Production and welfare problems relating to the supply of water to growing-finishing pigs. Pig J. 23, 51–66. Brumm, M., 2006. Patterns of drinking water use in pork production facilities. Nebraska Swine Report, pp. 10–13. Brumm, M.C., Dahlquist, J.M., Heemstra, J.M., 2000. Impact of feeders and drinker devices on pig performance, water use, and manure volume. J. Swine Health Prod. 8 (2), 51–57. Geary, T.M., Brooks, P.H., Morgan, D.T., Campbell, A., Russell, P.J., 1996. Performance of weaner pigs fed ad libitum with liquid feed at different dry matter concentrations. J. Sci. Food Agric. 72, 17–24. GenStat 8 Committee, 2002. Genstat 8 Reference Manual; Release 3. Clarendon Press, Oxford, UK. Gonyou, H.W., Lou, Z., 2000. Effects of eating space and availability of water in feeders on productivity and eating behaviour of grower/finisher pigs. J. Anim. Sci. 78, 865–870. Jensen, M.B., Pedersen, L.J., 2008. Using motivation tests to assess ethological needs and preferences. Appl. Anim Behav. Sci. 113, 340–356. Kyriazakis, I., Savory, C.J., 1997. Hunger and thirst. In: Appleby, M.C., Hughes, B.O. (Eds.), Animal Welfare. CABI, Wallingford, pp. 49–62. Lea, S.E.G.,1978. The psychology and economics of demand. Psychol. Bull. 85 (3), 441–466. Li, Y.Z., Chénard, L., Lemay, S.P., Gonyou, H.W., 2005. Water intake and wastage at nipple drinkers by growing-finishing pigs. J. Anim. Sci. 83, 1413–1422. Matthews, L.R., Ladewig, J., 1994. Environmental requirements of pigs measured by behavioural demand functions. Anim. Behav. 47, 713–719. Mroz, Z., Jongbloed, A.W., Lenis, N.P., Vreman, K., 1995. Water in pig nutrition: physiology, allowances and environmental implications. Nutr. Res. Rev. 8, 137–164. N.N., 2001. Minimum standards for the protection of pigs. Commission directive 2001/93/EC. Off. J. Eur. Communities L 316, 36–38. Nagai, M., Hachimura, K., Takahashi, K., 1994. Water consumption in suckling pigs. J. Vet. Med. Sci. 56, 181–183. Plagge, J.G., 1991. De invloed van beperking van de drinktijd op het waterverbruik en technische resultaten bij vleesvarkens (The effect of limiting drinking time on water use and performance of finishing pigs). Research Report P1.64, Research Institute for Pig Husbandry, Rosmalen, The Netherlands. Plagge, J.G., Leuteren, J. van, 1989. Waterverbruik bij onbeperkt gevoerde mestvarkens (Water use of ad libitum fed finishing pigs). Research Report P1.32, Research Instute for Pig Husbandry, Rosmalen, The Netherlands. Ramaekers, P.J.L., 1996. Control of individual daily growth in group-housed pigs using feeding stations. Thesis, Wageningen University, Wageningen, 133 p. Rushen, J., 1984. Stereotyped behaviour, adjunctive drinking and the feeding period of tethered sows. Anim. Behav. 32, 1059–1067. Smolders, M.A.H.H., Hoofs, A.I.J., 2000. Onbeperkte drinkwaterverstrekking naast een brijvoerrantsoen met bijproducten bij vleesvarkens [Unrestricted drinking water supply for finishing pigs on liquid feed with bye products]. Research Report P4.45, Research Institute for Pig Husbandry, Rosmalen, The Netherlands, 8 p. Terlouw, E.M.C., Lawrence, A.B., Illius, A.W., 1991. Influences of feeding level and physical restriction on development of stereotypies in sows. Anim. Behav. 42, 981–991. Vandenheede, M., Nicks, B., 1991. L'approvisionnement en eau des porcs: un élément a ne pas négliger [Water requirements and drinking water systems for pigs.]. Ann. Med. Vet. 135, 123–128. Van der Peet-Schwering, C.M.C., Plagge, J.G., 1995. Influence of multi phase feeding on the performance and water intake of sows and barrows. Research Report P1.140, Research Institute for Pig Husbandry, Rosmalen, The Netherlands.