J Sci Food Agric 1998, 76, 289È297 Eþects of Sodium and Potassium Fertilisers on the Composition of Herbage and its Acceptability to Dairy Cows Paul C Chiy,” Abdul-latif A Al-Tulihan, Mohammed H Hassan and Clive J C Phillips*¤ School of Agricultural and Forest Sciences, University of Wales, Bangor, Gwynedd, UK (Received 20 January 1997 ; revised version received 14 May 1997 ; accepted 30 June 1997) Abstract : Fertilisation of herbage with Na can increase acceptability to cows, but the inÑuence of fertiliser rate and fertilisation by K is unknown. In experiment 1, ten cows were grazed on pasture plots that had just been fertilised with 0È132 kgNa ha~1 (current Na) and had received 0È64 kg-Na ha~1 in the previous grazing season (residual Na). Herbage Na concentration increased in proportion to current Na from 2É7 to 4É9 g-Na kg~1 dry matter (DM) and also increased with increasing residual Na from 2É2 to 4É5 g-Na kg~1 DM. Herbage K concentrations were low (10 g kg~1 DM at 0 kg-Na ha~1) and were only slightly reduced by Na fertiliser. Herbage Mg and Ca concentrations and DM digestibility were maximum at 66È99 kg-current-Na ha~1. Cows grazed current-Na-fertilised plots to a lower height and spent more time grazing them. In experiment 2, pasture plots received no fertiliser, low and high isomolar and independent applications of Na and K or a combination of the two. The herbage was more mature than in experiment 1 and Na concentration of the herbage without Na fertilizer was high (5 g kg~1 DM). Na fertiliser, therefore, only slightly increased Na concentration, more in clover than in grass, and had little e†ect on K concentration. K fertiliser increased K concentration from 16 to 20 g kg~1 DM and reduced Na concentration to 3É5 g kg~1 DM. Sodium fertiliser, therefore, only increased the acceptability of herbage to cattle when herbage Na concentrations were initially low (less than 5 g kg~1 DM) and were increased substantially by the application of the fertiliser. ( 1998 SCI. J Sci Food Agric 76, 289È297 (1998) Key words : herbage ; fertiliser ; palatability ; sodium ; potassium ; ryegrass ; white clover INTRODUCTION the plant. The e†ects on magnesium (Mg) and calcium (Ca) concentrations and dry matter digestibility (DMD) vary between the di†erent plant fractions and depend on initial herbage Na concentration (Chiy and Phillips 1996, 1997). At herbage concentrations between 2É9 and 4É9 g-Na kg~1 DM cows graze for longer on Na-fertilised pastures (Chiy et al 1993a) and prefer them to unfertilised pasture, leading to increased milk production (Chiy and Phillips 1991). Apart from this, no information is available on the e†ect of Na or K concentrations in pasture on the grazing preference of dairy cows. Schneider et al Mammals have an instinctive appetite for sodium (Na) (Richter 1956), and the ability to taste it is inherent in ruminants (Denton 1966). Applying Na fertiliser to herbage increases herbage Na, but decreases its potassium (K) concentration (Chiy and Phillips 1993). Na and K, therefore, are competitors for absorption sites in * To whom correspondence should be addressed. ” Present address : Department of Clinical Veterinary Medicine, University of Cambridge, Cambridge, CB3 0ES, UK. 289 ( 1998 SCI. J Sci Food Agric 0022-5142/98/$17.50. Printed in Great Britain P C Chiy et al 290 (1986) found that the DM intake of cows o†ered Nasupplemented silage was greater at 5É5 than at 8É8 or 1É8 g-Na kg~1 DM. The extent to which herbage Na concentrations can be increased by Na fertiliser without adversely a†ecting herbage nutrient concentration has not been reported previously, but it is relevant when Na fertiliser is used to correct Na deÐciencies in grazing ruminants. Recent genetic improvements that have increased dairy cow milk yields, and hence the output of Na in milk, may predispose dairy cows to an increased risk of Na deÐciency (Hemingway 1995). Both pregnancy and lactation increase the physiological requirement for Na, and restricted DM intakes at the end of pregnancy limit voluntary Na intake at this time (ARC 1980). The objective of the present experiments was to examine the e†ects of Na fertiliser application with and without K fertiliser on herbage chemical composition and the grazing preference of pregnant dairy cows. MATERIALS AND METHODS Experimental design Experiment 1 Ten non-lactating, autumn-calving British Friesian cows with similar calving dates, lactation number, live weight and milk yield were selected at drying o† for grazing on the treatment area. This was a 3-ha Ðeld of predominantly perennial ryegrass (L olium perenne L cv S23). Herbage comprised 964 g perennial ryegrass per kilogram herbage freshweight. The area was divided into Ðve main plots of equal area (0É6 ha), identiÐed with corner posts (Fig 1). Water was available in all of these areas. The following Na fertiliser treatments were allocated at random to the experimental plots : 0 (nil treatment), 33 (low treatment), 66 (medium treatment), 99 (high treatment) and 132 (very high treatment) kg-Na ha~1 (Fig 1). Fertiliser Na was applied on 17 July 1991 as commercial granular NaCl (994 g-NaCl kg~1 fertiliser applied) with a 3-mm particle size. Application was by broadcasting with a Vicon land-geared fertiliser spinner, with a 5 m spread that ensured minimum inaccuracy at edges and corners of the plots. There were no bu†er zones between plots. All plots additionally received 85 kg-N ha~1 in March, 65 kg-N ha~1 in April and May and 85 kg-N, 43 kg-P O and 85 kg-K O ha~1 in 2 5 2 June. Prior to the experimental grazing period between 25 July and 25 August the entire area had been uniformly grazed by dairy cows. Each main plot was further subdivided by corner posts to indicate the three equal areas of residual Na treatments (1 ha treatment~1) applied in the previous year (Fig 1). These consisted of 0 (RN), 32 (RL) or 64 (RH) kg-Na ha~1, applied as NaNO in Ðve isonitrogenous applications at 3 monthly intervals in the year prior to the current trial, beginning from 9 April 1990 (Chiy and Phillips 1993). All plots had received a uniform management regime and fertiliser application of 300 kg-N, 30 kg-P O and 2 5 40 kg-K O ha~1 yr~1 in the previous 7 years before 2 1990. Experiment 2 Twelve cows were selected as in experiment 1 for grazing on a 3É5-ha pasture area of similar botanical composition. The area was divided into 14 equal plots of 0É25 ha each, identiÐed with corner posts, and allocated at random to seven fertiliser treatments : no fertiliser, Na low rate (40 kg-Na ha~1), Na high rate (80 kgNa ha~1), K low rate (69 kg-K ha~1), K high rate (138 kg-K ha~1), Na ] K low rate (20 kg-Na and 35 kg-K ha~1, respectively) or Na ] K high rate (40 kg-Na and 69 kg-K ha~1, respectively). This gave isomolar applications of the monovalent cations Na and K at the low and high application rates. Fertilisers used were NaCl, KCl and a combination of the two (Silvinit, Cleveland Potash Ltd, Whitby, UK) (Table 1). The fertilisers were applied on 30 August as in experiment 1 and the swards grazed from 3 September to 7 October. Prior to the start of the experiment the entire area had received 130 kg-N ha~1 in March before a Ðrst cut of silage in May. After this silage cut, 125 kg-N and 80 kg-K O were applied and a second silage cut 2 taken in August. The area was then grazed by dairy cows between the second silage cut and the start of the experiment, when the twelve cows were introduced. Sward measurements and chemical analysis Herbage height was measured using a sward stick on the day before the experiment began and subsequently TABLE 1 Fertiliser application rates in experiment 2 (kg ha~1) T reatment Fertiliser application rate L ow Fig 1. Layout of current and residual treatment applications in the Ðeld in experiment 1. Na K Na ] K High NaCl KCl NaCl KCl 102 È 51 È 130 65 204 È 102 È 260 130 E†ects of Na and K fertiliser on cattle pasture 291 TABLE 2 E†ects of current Na fertiliser treatment in experiment 1 on herbage chemical composition (g kg~1 DM, except where otherwise stated) T reatment Na K Mg Ca DM (g kg~1 fresh weight) DMD MADF CP Ash Nil L ow Medium 2É73 10É17 1É79 4É42 3É63 9É61 1É95 5É48 4É10 9É38 2É16 5É55 5É18 9É34 1É87 5É62 4É93 9É13 1É69 5É10 219 712 212 246 91É1 229 744 222 250 92É2 221 854 214 244 94É6 210 807 222 245 94É1 209 735 218 242 91É7 every 2 days during the experimental period. Twelve (experiment 1) and 16 (experiment 2) measurements were taken in a W-pattern across each plot. Ten hand-plucked herbage samples of approximately 200 g were collected at random in each plot at the time of fertiliser application and at three (experimental 1) and four (experiment 2) day intervals during the experiment. An additional sample was taken by the same method at the end of experiment 2 and separated into clover and grass leaf fractions for individual analysis to investigate the response of these two species to the fertiliser treatments. Individual samples were dried at High SED Prob 0É286 0É312 0É079 0É204 \0É001 0É02 \0É001 \0É001 V ery high 2É1 15É6 4É8 2É6 0É44 0É04 0É001 0É16 0É06 \0É001 90È100¡C for 16 h to determine DM concentration. They were then hammer-milled through a 1 mm screen and a sub-sample used for the determination of crude protein (CP, N ] 6É25) and modiÐed acid-detergent Ðbre (MADF) according to the methods of AOAC (1984), and DMD by a modiÐed technique (Omed et al 1989) of incubation in faecal liquor (El Shaer et al 1987). The remainder was ashed in a muffle furnace at 600¡C and a solution prepared by dry combustion, which was used for the analysis of Na, K, Mg and Ca by atomic absorption spectrophotometry (Model 951, Instrumentation Laboratory, Lexington, USA). TABLE 3 Residual e†ects of Na fertiliser applied in the previous year at a low (RL) or high (RH) rate, compared with no fertiliser (RN), on herbage chemical composition in experiment 1 (g kg~1 DM, except where otherwise stated) T reatment Na Na start Na end K Mg Ca Ca start Ca end DM (g kg~1 fresh weight) DMD MADF CP Ash RN RL 2É24 1É22 3É25 9É73 2É00 6É06 6É76 5É35 199 837 216 256 95É7 3É15 2É70 3É59 9É71 2É02 6É92 8É29 5É55 188 815 216 265 97É2 SED Prob 0É186 \0É001 0É264a \0É001a 0É517 0É158 0É327 NSb NS NS 0É462a \0É10a RH 4É54 4É08 3É19 9É81 1É86 6É20 7É90 4É49 195 784 217 272 98É2 a `Interaction between residual and start/end sampling dates. b NS \ P [ 0É10. H H 5É43 37É5 3É50 3É32 1É52 NS NS NS \0É005 NS P C Chiy et al 292 Animal measurements Animals had free access to the entire grazing area, and the location of each grazing cow in the treatment grid was recorded at 5-min intervals for 90 min each day. The 90-min period was randomly assigned within the main grazing period of the day, from 09.00 to 18.00 h. The sward biting rate was recorded for each cow in the morning, afternoon and evening for 5 days per week and the treatment area occupied by the cow on each occasion. The number of bites was counted over a 90 s period of continuous biting. Cows were allowed 30 s after the start of biting before counting commenced. The record was abandoned if the cow stopped grazing for more than 10 s. The distance walked by each cow during 60 s grazing was measured once every 2 days using a calibrated wheel and the treatment area occupied by the cow during the recording noted. The position of the hind right foot was used as the reference point. Statistical analysis The signiÐcance of di†erences between treatments was determined by analysis of variance of a generalised linear model using Genstat 5 (Lawes Agricultural Trust 1980). Factors included in the model were the treatment and sampling date, and cow in grazing behaviour measurements. Residual Na treatment was also included in experiment 1. If the probability of treatment e†ects was 10% or greater, they were assumed not to be signiÐcant. RESULTS Experiment 1 E†ects of the current Na treatments on herbage chemical composition are given in Table 2. E†ects of the residual Na treatments are given in Table 3, and the signiÐcant interactions between current and residual treatments (herbage Na concentration, grass height and proportion of time spent grazing) are shown in Fig 2. Other residual ] current treatment interactions were not signiÐcant. Herbage chemical composition Current Na fertiliser application increased herbage Na concentration from the nil to high application rates, with no further e†ect at the very high rate (Table 2). Herbage Na concentration also increased in proportion to residual Na fertiliser application (Table 3). There appeared to be a maximum herbage Na concentration of approximately 5 g kg~1 DM. The signiÐcant interaction between current and residual Na application (SED 0É26, P \ 0É001) (Fig 2) demonstrated that, in the Fig 2. E†ect of current Na fertiliser application level on herbage Na content, herbage height and proportion of time cows spent grazing for residual nil (L), low (K) and high (|) treatments in experiment 1. E†ects of Na and K fertiliser on cattle pasture 293 TABLE 4 Current treatment e†ects on herbage height and grazing behaviour in experiment 1 T reatment Herbage height (cm) Grazing time (proportion of total time) Biting rate (no min~1) Walking rate (m min~1) Nil L ow Medium High V ery high 8É65 0É137 8É03 0É150 8É01 0É159 7É53 0É160 7É66 0É152 69É0 2É47 73É6 3É64 residual high treatment, Na concentration in herbage was greater at low and medium application rates than in residual nil and low. There was also a signiÐcant interaction between the time of the recording and the residual treatment e†ect (Table 3). Herbage at the start of the experiment showed a response in Na concentration to residual Na fertiliser application, but by the end of the experiment this response had disappeared. Herbage K concentration was signiÐcantly reduced by current Na fertiliser application at the medium, high and very high application rates (Table 2). Residual Na treatment did not a†ect herbage K concentration (Table 3). Herbage Mg concentration increased up to the medium application rate but thereafter decreased. It was not a†ected by residual Na treatment. Ca concentration was increased at all levels of Na fertiliser application, but especially at the medium and high rates. There was a signiÐcantly increased herbage Ca in residual low and high treatments at the start but not at the end of the experiment. Herbage DM content was increased at the low application rate, una†ected at the medium rate and thereafter decreased. Herbage DM digestibility increased linearly up to the medium rate and thereafter declined. Ash concentrations were increased in the medium and high treatments. There were no e†ects of residual Na treatments on herbage DM, DMD or ash concentrations. There were no TABLE 5 Residual treatment e†ects on herbage height and grazing behaviour in experiment 1 T reatment SED Prob RN RL RH 8É0 8É7 8É3 0É15 \0É001 0É161 0É149 0É143 0É0043 \0É001 78É0 3É80 76É8 3É09 67É6 3É11 SED Prob 0É219 0É005 \ 0É001 \ 0É001 1É46 0É313 0É03 0É03 current or residual Na treatment e†ects on herbage MADF concentration, but CP concentrations tended to be slightly increased in the low treatment. There was a linear increase in CP concentration with increasing residual Na fertiliser. Herbage height and cow behaviour Current and residual treatment means of herbage height and grazing behaviour are given in Tables 4 and 5, respectively. Cows grazed herbage to a lower height in the current treatments where Na was applied, particularly in high and very high treatments. There was a signiÐcant current ] residual Na treatment interaction in herbage height (SED \ 0É332, P \ 0É001). In residual nil treatment, and to a lesser extent residual low, herbage height declined with increasing current treatment Na application rate (Fig 2). However, in residual high herbage height increased for the medium, high and very high current treatments. There was a signiÐcant interaction between current and residual treatments in grazing time (Fig 2) (SED 0É174, P \ 0É001). Where there was no residual Na fertiliser, grazing time increased with current Na fertiliser. In the residual low treatment, grazing time increased up to the medium application rate and then declined ; in residual high, grazing time decreased with current Na application rate above the low level. Biting rate was increased in the low, medium and high current treatments (Table 4). It was not a†ected by residual treatment and there was no current ] residual treatment interaction. Walking rate was increased with Na application rate up to the medium rate and declined thereafter. Experiment 2 Herbage height (cm) Grazing time (proportion of total time) Biting rate (no min~1) 72É8 73É9 72É2 1É13 0É31 Herbage chemical composition Herbage Na concentration was high (5 g kg DM~1) when no Na fertiliser was applied (Table 6). Na fertiliser increased and K fertiliser decreased herbage Na concentration, with greater responses at the high application rate. The combined fertiliser had little e†ect on herbage P C Chiy et al 294 TABLE 6 Treatment e†ects on herbage chemical composition in experiment 2 (g kg~1 DM unless otherwise stated) Cation level None Na L ow Na Herbage 5É01 Grass leaf 5É31 Clover 4É99 K Herbage 16É1 Grass leaf 16É0 Clover 20É3 Mg Herbage 1É28 Grass leaf 2É05 Clover 2É20 Ca Herbage 4É83 Grass leaf 7É69 Clover 9É50 P Herbage 3É16 Grass leaf 3É09 Clover 3É77 Cl Herbage 38É4 Grass leaf 43É9 Clover 41É5 DM (g kg~1 freshweight) Herbage 212 Grass leaf 198 Clover 127 MADF Herbage 277 Grass leaf 234 Clover 182 CP Herbage 106 Grass leaf 146 Clover 263 Ash Herbage 83É9 Grass leaf 78É5 Clover 118 5É37 5É58 4É28 16É1 19É8 13É9 Na ] K K High 5É43 6É07 8É85 15É4 18É5 18É2 L ow 4É62 4É28 4É79 18É8 19É5 22É4 High 3É52 4É72 4É62 20É1 22É4 24É9 L ow 5É18 5É65 4É04 18É2 20É6 18É5 SED Prob 0É331 0É550 0É979 \0É001 0É10 0É02 0É67 1É02 2É79 \0É001 0É01 0É08 High 4É61 5É96 4É76 19É3 21É3 20É9 1É22 1É84 1É69 1É17 2É13 1É93 1É24 1É54 2É14 1É21 1É76 2É06 1É21 1É38 1É78 1É25 1É88 1É86 0É035 0É366 0É271 0É11 0É46 0É49 4É64 6É51 8É61 4É42 7É43 9É52 4É56 6É64 9É52 4É60 7É69 9É00 4É46 7É04 8É08 4É66 6É77 7É82 0É190 0É840 0É728 0É42 0É66 0É20 2É86 2É95 3É05 3É09 3É48 3É58 3É17 2É85 4É08 2É89 3É02 4É09 3É05 2É94 3É92 3É00 3É18 4É38 0É144 0É259 0É598 0É22 0É36 0É47 2É96 8É25 12É10 \0É001 0É23 0É49 39É4 52É5 28É1 44É6 50É7 44É4 46É6 59É0 50É9 49É7 62É4 53É9 47É9 59É2 39É6 54É3 67É2 49É4 212 205 169 205 191 155 215 200 142 211 200 146 206 197 162 201 191 160 5É3 6É1 19É8 0É09 0É33 0É47 260 228 167 266 227 184 271 224 193 271 236 193 260 221 177 268 228 174 4É6 4É6 15É3 0É01 0É10 0É60 101 140 228 106 148 242 102 143 269 102 141 270 106 147 247 112 157 256 2É9 7É1 24É6 0É06 0É37 0É61 86É1 80É5 136 92É9 81É9 140 91É7 87É5 135 94É0 91É4 133 93É2 84É7 133 99É6 90É5 134 2É76 2É54 28É1 \0É001 0É01 0É60 Na concentration at the low application rate and slightly decreased it at the high rate. Grass leaf and clover tended to respond in a similar manner. Herbage K concentration showed little change with Na fertiliser, and was increased by K fertiliser, particularly at the high rate. K fertiliser increased the K concentration of grass leaf and to a lesser extent clover, and Na fertiliser increased the K concentration of grass leaf. The combined fertiliser increased herbage and grass leaf K concentrations, but not clover. There was a trend for Mg, Ca and P concentrations to be reduced with all fertiliser treatments compared with no fertiliser, but this was not statistically signiÐcant. Herbage Cl concentration was increased by all fertiliser treatments, particularly the K and combined fertilisers and at the high level. The same trend was evident for grass leaf but this was not statistically signiÐcant. Clover Cl concentration was variable and not signiÐcantly a†ected by treatment. The DM and CP concentrations were not a†ected by treatment. MADF concentration tended to be reduced by the fertiliser treatments, especially Na. Ash concentration was increased by fertiliser treatment compared with no fertiliser, particularly the K and combined fertilisers and at the high rate. E†ects of Na and K fertiliser on cattle pasture 295 Herbage height and cow behaviour Herbage height was increased in the Na and to a lesser extent K and combined fertiliser treatments, compared to no fertiliser (Table 7). The proportion of time spent grazing and the walking rate on pasture treatments were not a†ected by treatment, but the biting rate tended to be reduced for the fertiliser Na and K treatments and the combined Na ] K treatments. DISCUSSION Herbage composition The herbage Na concentrations of approximately 5 g kg~1 DM in the high and very high Na current treatments in experiment 1 have previously been achieved at lower rates of Na fertiliser application and have been accompanied by reductions in herbage K and increases in Ca, DMD and ash concentrations (Chiy and Phillips 1991, 1993). This herbage Na concentration appears to be the maximum that can be achieved under these conditions and with the application of Na fertilisers. As a result, the herbage Na : K ratio was only increased up to the high rate of Na application (0É27, 0É38, 0É44, 0É56 and 0É54 for treatments low to very high, respectively). In experiment 2 the application of Na fertiliser had relatively little e†ect on herbage Na concentration because it was already high, probably due to the maturity of the herbage. In the absence of Na fertiliser, Na contents of herbage at this site had been observed to increase with herbage maturity (Chiy and Phillips 1997). The proximity of the sea (250 m) and inadequate drainage may also have contributed to the high herbage Na concentrations. Higher herbage Na concentrations of 6È 8 g kg~1 DM have been reported following Na fertiliser application at 30È100 kg-Na ha~1 (Moseley 1980 ; Chiy et al 1993b), but only where herbage is cut before being fed to livestock, thus avoiding urine and faecal K return to the pasture. The Na : K ratio in experiment 2 increased in clover in response to fertiliser Na, but not in grass (Fig 3). In a separate experiment, Chiy and Phillips (1997) found that there was only an increase in older clover leaves Fig 3. E†ect of Na fertiliser application on Na : K ratio in herbage (K), grass (È) and clover ()) in experiment 2. ([50% unfolded) and young grass leaves (at least 95% of the surface area being green). In accordance with this, the herbage in experiment 2, being tall and relatively mature, would have mainly contained older grass and clover leaves, and therefore the response was only seen in clover. In experiment 1 the herbage was not separated into grass and clover but was predominantly shorter, younger grass than in experiment 2 ; hence, there was a response in herbage Na, K and Na : K ratio. The reduction in Na : K ratio with application of K fertiliser was evident for grass and clover in experiment 2 (Fig 4). In terms of uptake kinetics, the binding sites on the plasma membrane of the root cells have a high affinity for K and low affinity for Na, and it is therefore Fig 4. E†ect of K fertiliser application on Na : K ratio in herbage (K), grass (È) and clover ()) in experiment 2. TABLE 7 Herbage height and grazing behaviour in experiment 2 Cation level Herbage height (cm) Grazing time (proportion of total time) Biting rate (no min~1) Walking rate (m min~1) None Na Na ] K K SED Prob L ow High L ow High L ow High 12É6 0É118 14É0 0É132 13É5 0É117 13É5 0É145 13É0 0É163 13É5 0É125 13É1 0É152 0É25 0É019 \0É001 0É12 68É4 2É72 66É2 2É97 67É6 2É88 66É6 2É65 66É1 2É69 65É7 3É02 65É7 2É73 1É01 0É198 0É07 0É43 P C Chiy et al 296 to be expected that K fertiliser will induce a response in Na : K ratio of the herbage under a range of Na : K ratios in the soil solute. The low herbage K concentrations (c 10 g kg~1 DM) may have prevented any reduction in herbage Mg uptake with Na fertiliser in experiment 1, which occurs mainly at high K levels due to competition by monovalent cations for plant absorption sites (Mengel and Kirby 1978 ; Whitehead et al 1978). In experiment 2 with higher K concentrations there was evidence of a reduction in Mg and Ca concentrations with fertiliser treatment. Cow grazing behaviour A preference by dairy cows for pasture fertilised with 18 kg-Na ha~1 compared with unfertilised pasture has been reported (Chiy and Phillips 1991). An increase in grazing time has also been reported when cows are grazed on pastures containing 5 compared to 3 g-Na kg~1 DM, accompanied by an increase in cow biting rate and a reduction in herbage height (Chiy and Phillips 1991 ; Chiy et al 1993a). The interaction between current and residual treatment e†ects on herbage Na concentrations in experiment 1 demonstrated that the e†ects were only additive up to 5 g-Na kg~1 herbage. Residual Na accelerated the response of herbage Na concentration up to a maximum level with additional current Na (Fig 2), but the residual e†ects were eliminated by the end of the experiment. Despite the similarity across residual treatments of herbage Na concentrations at high current Na application, grazing time decreased and herbage height increased on swards with both high current and residual Na treatments. The cause of this adverse e†ect on grazing preference is uncertain, but it may relate to the decline in DMD at high Na application rates. Previous research (Chiy and Phillips 1993) has related the change in DM digestibility with Na fertiliser to a change in water-soluble carbohydrate concentration, which, together with Na, is likely to be a major gustatory signal to regulate feeding preferences. It is assumed that the reduction in herbage height with increased sodium content of herbage in experiment 1 reÑected more grazing at a faster intake rate, rather than reduced herbage growth. Perennial ryegrass has been observed to show positive responses to Na fertiliser, and is generally accepted to be natrophillic (Chiy and Phillips 1993). The increased biting rates in the low, medium and high treatments in experiment 1 may partly represent an attempt by the animal to maintain intake rate at the reduced herbage heights. However, herbage height was also reduced in the very high treatment where biting rate was reduced, suggesting that Na might be having speciÐc e†ects on biting rate at the highest application rate independent of grass height. Biting rate may be related to herbage DMD if herbage of lower digestibility (and higher structural carbohydrate concentration) takes longer to sever from the sward. It is perhaps surprising that, in experiment 1, cattle walked faster on pasture on which they spent longer grazing, suggesting that, Ðrstly, they made the conscious decision to graze the Na-fertilised pasture for longer. To do this they would have a identify and learn which areas of the Ðeld had high Na concentrations. Secondly, less searching was done as the cow moved her head from side to side whilst walking. Thirdly, at the lower grass height they may have been walking faster in search of taller herbage of the same palatability. Evidence for the latter is reported by Fraser and Broom (1990), who found that walking rate increased from 1É9 m min~1 on long herbage to 2É5 m min~1 on short herbage. The increase in herbage height and tendency for slower biting with fertiliser treatments in experiment 2 suggests that the palatability of the treated herbages was not increased. This corresponds to the absence of major e†ects on herbage Na concentration of Na fertilisers, but the reduction, from 5É0 to 3É5 g kg~1 DM, in herbage Na concentration with K fertiliser suggests that palatability was maintained by the increase in K concentration, from 16 to 20 g kg~1 DM. These experiments conÐrm that young grass swards with low Na contents will respond in herbage Na content more to fertiliser Na than taller, older swards. If herbage Na content is increased to approximately 5 g kg~1 DM by Na fertiliser, cattle will prefer that herbage to herbage with a more normal Na content and will bite it faster, leaving a sward grazed down to a lower height. ACKNOWLEDGEMENTS The authors are grateful to the University Farm technical sta† for assistance in the running of the experiment and to Cleveland Potash Ltd for the supply of fertilisers in Experiment 2. REFERENCES ARC 1980 T he Nutrient Requirements of Ruminant L ivestock. Agricultural Research Council, Commonwealth Agricultural Bureaux, Farnham Royal, UK. AOAC 1984 Official Methods of Analysis (14th edn). Association of Official Analytical Chemists, Washington, DC, USA. Chiy P C, Phillips C J C 1991 The e†ects of sodium chloride application to pasture, or its direct supplementation, on dairy cow production and grazing preference. Grass For Sci 46 325È331. Chiy P C, Phillips C J C 1993 Sodium fertilizer application to pasture. 1. Direct and residual e†ects on pasture production and composition. Grass For Sci 48 189È202. E†ects of Na and K fertiliser on cattle pasture Chiy PC, Phillips C J C 1996 E†ects of sodium fertilizer on the chemical composition of grass and clover leaves, stems and inÑorescences. J Sci Food Agric 72 501È510. Chiy P C, Phillips C J C 1997 E†ects of sodium fertilizer on the chemical composition of ryegrass and white clover leaves of di†erent physiological ages. J Sci Food Agric 73 337È348. Chiy P C, Phillips C J C, Bello M 1993a Sodium fertilizer application to pasture. 2. E†ects on dairy cow production and behaviour. Grass For Sci 48 203È212. Chiy P C, Phillips C J C, Omed H 1993b Sodium fertilizer application to pasture. 3. Rumen dynamics. Grass For Sci 48 249È259. Denton D A 1966 Some theoretical considerations in relation to innate appetite for salt. Conditional ReÑex 1 144È177. El Shaer H M, Omed H, Chamberlain A G, Axford R F E 1987 Use of faecal organisms for the in vitro determination of digestibility. J Agric Sci (Camb) 109 257È259. Fraser A F, Broom, D M 1990 Farm Animal Behaviour and W elfare (3rd Edn). Balliere Tindall, London, UK. Hemingway R G 1995 Requirements for sodium by livestock and dietary allowances. In : Sodium in Agriculture, eds Phillips C J C & Chiy P C. Chalcombe Publications, Canterbury, UK, pp 145È161. 297 Lawes Agricultural Trust 1980 Genstat V Mark 4 03. Rothamsted Experimental Station, Harpenden, UK. Mengel K, Kirby E A 1978 Principles of plant nutrition. Occasional paper, International Potash Institute, Berne, Switzerland. Moseley G 1980 E†ects of variation in herbage sodium levels and salt supplementation on the nutritive value of perennial ryegrass for sheep. Grass For Sci 35 105È113. Omed H M, Axford R F E, Chamberlain A G, Givens D I 1989 Development of a method for the in vitro estimation of digestibility of forage in ruminants. In : New T echniques in Cattle Production, pp 227, ed Phillips C J C. Butterworth, London, UK. Richter C P 1956 Salt appetite in mammals : its dependence on instinct and metabolism. In : Instinct in the Behaviour of Animals and Humans, ed Richter C P. Masson, Paris, France, pp 557È629. Schneider P L, Beede D K, Wilcox C J 1986 Responses of lactating cows to dietary Na source and quantity and potassium quantity during heat stress. J Dairy Sci, 69 99È110. Whitehead D C, Jones L H P, Barnes R J 1978 The inÑuence of fertilizer N plus K on N, S and other mineral elements in perennial rye grass at a range of sites. J Sci Food Agric 29 1È11.