Pasture intake and milk production of dairy cows rotationally grazing on multi-species swards

1 Production, intake and behaviour of dairy cows rotationally grazing on multi1 species swards 2 A.I. Roca-Fernández1,2, J.L. Peyraud1,2, L. Delaby1,2 and R. Delagarde1,2 3 1INRA, UMR1348 Physiologie, Environnement et Génétique pour l’Animal et les 4 Systèmes d’Elevage, F-35590 Saint-Gilles, France; 2Agrocampus Ouest, UMR1348 5 Physiologie, Environnement et Génétique pour l’Animal et les Systèmes d’Elevage, 6 F-35000 Rennes, France. 7 Corresponding author: [email protected] 8 Short title: Grazing dairy cow intake on multi-species swards 9 10 Abstract 11 Increasing plant species diversity has been proposed as a means for enhancing 12 productivity and seasonal production stability of grazing lands facing drought stress. 13 Few studies have examined how botanical complexity of sown swards affects cow 14 performance. A two-year experiment was conducted to determine how sward 15 botanical complexity, from monocultures of grass to multi-species swards (MSS) 16 (grass-legume-forb), affect pasture chemical composition and nutritive value, milk 17 production and milk solids production, pasture dry matter (DM) intake and grazing 18 behaviour of cows. Five sward species: perennial ryegrass (L as Lolium), white clover 19 and red clover (both referred to as T as Trifolium because always sown together), 20 chicory (C as Cichorium) and tall fescue (F as Festuca) were assigned to four grazing 21 treatments by combining 1 (L), 3 (LT), 4 (LTC) or 5 (LTCF) species. Thereafter, the 22 LT swards were called mixed swards as a single combination of ryegrass and 23 clovers, while LTC and LTCF swards are called multi-species swards (MSS) as a 24 combination of at least four species from three botanical families. The experimental 25 Manuscript 2 area (8.7 ha) was divided into block replicates with a mineral nitrogen fertilisation of 26 75 kg N/ha/year in each treatment. The weather conditions were favorable for 27 pasture growth in both years, particularly in early summer. In total, 13 grazing 28 rotations were carried out by applying the same grazing calendar for all treatments 29 and the same pasture allowance of 19 kg DM/cow/day above 4 cm. Clover 30 represented 20% of DM in mixed and MSS swards; chicory represented 30% of DM 31 in MSS and tall fescue represented 10% of DM in LTCF swards. Greater milk 32 production (+1.1 kg/day) and milk solids production (+0.08 kg/day) were observed in 33 mixed swards than in ryegrass swards. Pasture quality and pasture DM intake were 34 unaffected by the inclusion of clover. Pasture DM (P<0.001), organic matter 35 (P<0.001) and neutral detergent fibre concentrations (P<0.001) were lower in MSS 36 than in mixed swards. Greater milk production (+0.8 kg/day), milk solids production 37 (+0.04 kg/day), pasture DM intake (+1.5 kg DM/day) and grazing time (+57 min/day) 38 were observed in MSS than in mixed swards. These positive effects of MSS swards 39 were observed in all seasons, but particularly during summer. In conclusion, 40 advantages of grazing MSS on cow performance were mainly due to the presence of 41 chicory, and to the cumulative effect of enhanced sward quality, increased pasture 42 DM intake that raised milk production and milk solids production. 43 Keywords: chicory, grass-legume mixtures, grazing, multi-species, milk production 44 45 Implications 46 Little is known about how increasing sward species complexity may affect milk 47 performance in grazing cows. The aim of this study was to determine how pastures 48 sown with 1, 3, 4 and 5 species - including grasses, legumes and chicory - affect 49 pasture chemical composition and nutritive value, milk production, pasture intake and 50 3 the grazing behaviour of cows. Pastures sown with greater sward species complexity 51 induced greater sward quality, pasture intake, milk production and milk solids 52 production than simpler pastures, in all seasons. Both legumes and chicory had 53 positive effects on sward quality and milk production. 54 55 Introduction 56 Grasslands cover approximately 40% of the agricultural area in Europe, and supply 57 most of the feed used by cattle and other ruminants (78 million livestock units) 58 (Huyghe et al., 2014). Traditionally, sown grasslands have been based on a two59 species (grass-legume) mixture (Høgh-Jensen et al., 2006). Legumes offer important 60 potential benefits when mixed with perennial ryegrass by (i) increasing pasture 61 production, (ii) substituting mineral N-fertiliser inputs by symbiotic N2 fixation, (iii) 62 mitigating and facilitating adaptation to climate change, as elevated atmospheric 63 CO2, warmer temperatures and drought-stress periods rise, (iv) increasing pasture 64 nutritive value and voluntary intake with a less marked decline of sward quality with 65 advancing maturity than grasses, leading to (v) higher cow performance (Lüscher et 66 al., 2014). However, grass-legume swards may be limited by the intensive yet 67 relatively short growth period for legumes, mainly clovers, compared to other sward 68 species, adapted to a larger range of weather conditions (Høgh-Jensen et al., 2006). 69 Increasing the botanical complexity of swards has been suggested as a means of 70 raising pasture productivity and seasonal production stability facing drought stress 71 (Sanderson et al., 2005). Nevertheless, much of the research to date in Europe has 72 been focused on single mixed swards. Enhanced milk production (10% to 25%) of 73 grazing dairy cows in grass-legume swards compared to perennial ryegrass swards 74 is generally related to the increase in pasture intake rather than in pasture nutritive 75 4 value (Harris et al., 1997; Ribeiro Filho et al., 2003; Dewhurst et al., 2009). There is a 76 rising interest to examine the effect of multi-species swards (MSS) on sward quality, 77 pasture intake (Sanderson et al., 2006; Deak et al., 2009; Sanderson, 2010) and milk 78 production in grazing dairy cows (Sanderson et al., 2005; Soder et al., 2006; 79 Chapman et al., 2008). Evidence from these studies suggests that the yield benefit 80 mainly results from including drought-tolerant species (i.e., forbs such as chicory). 81 Skinner (2008) showed that including chicory in grass-legume swards improved 82 pasture production, especially in mid-summer when drought stress reduces 83 productivity of dominant cool-season sward species. However, chicory is 84 disadvantaged by its low persistency in swards beyond 3 or 4 years (Sanderson et 85 al., 2003). 86 Chicory is well known as a highly productive species, with high nutritive value (Li and 87 Kemp, 2005). It has been found to enhance sward quality by improving the seasonal 88 availability of high quality pasture (Marley et al., 2013). It shows a variable crude 89 protein concentration and contains higher water soluble carbohydrates (Hoskin et al., 90 1995), greater digestibility, lower fibre (Barry, 1998), and higher mineral and trace 91 element concentrations (Høgh-Jensen et al., 2006; Marley et al., 2013) than 92 perennial ryegrass, irrespective of the grazing season. Greater voluntary intake in 93 steers (Morel et al., 2014) and sheep (Niderkorn et al., 2014) has been found in 94 mixed swards including chicory compared with perennial ryegrass swards. Under 95 grazing management, the effect of including chicory in mixed swards on dairy cow 96 performance is unclear. Pasture DM intake and milk production were unaffected by 97 inclusion of chicory in the studies of Soder et al. (2006) and Muir et al. (2014), but 98 increases in milk production and pasture DM intake were found in the studies of Li 99 and Kemp (2005), and Chapman et al. (2008). In order to investigate the effect of 100 5 MSS on intake and milk production of grazing dairy cows, a two-year study was 101 conducted to test the hypothesis that increasing botanical complexity from 1 (grass) 102 to 5 sward species (grass-legume-forb) would affect pasture chemical composition 103 and nutritive value, milk production, milk solids production, pasture DM intake and 104 grazing behaviour of dairy cows. The duration of the study allowed for grazing in all 105 seasons, so as to investigate whether the effect of the sward’s botanical complexity 106 on cow performance would be different depending on the season. 107 108 Materials and methods 109 Location, treatments and experimental design 110 The experiment was conducted over two years: September 2011 – August 2012 111 (year 1) and September 2012 – August 2013 (year 2) at the INRA experimental dairy 112 farm of Méjusseaume (1.71°W, 48.11°N, Le Rheu, France). The soils are loamy with 113 a pH value of around 6.0, an organic matter content of 3%, and are particularly 114 sensitive to water stress during summer. The swards were sown in September 2010, 115 one year before starting the experiment. Four treatments were compared by seeding 116 pastures with a combination of 1 to 5 sward species as described in Table 1. The 117 sward species were: perennial ryegrass (L, Lolium perenne L.), white clover 118 (Trifolium repens L.) and red clover (Trifolium pratense L.), both referred to as T for 119 Trifolium since they are always sown together, chicory (C, Cichorium intybus L.), and 120 tall fescue (F, Festuca arundinacea Schreb.). The four treatments were applied by 121 combining 1 (L), 3 (LT), 4 (LTC) or 5 (LTCF) of these sward species, to increase 122 botanical complexity of pastures. Treatment L was considered as the control, and LT 123 as a commonly used grass-legume mixture in temperate pasture-based milk 124 production systems for its potential to supply consistent forage yields with low mineral 125 6 N-fertiliser due to symbiotic N2 fixation by legumes. In LTC, chicory was added as a 126 deep-root forb well adapted to dry summers and considering its good nutritive value. 127 Finally, in LTCF, tall fescue was added, as a more drought-resistant grass than 128 perennial ryegrass, increasing tolerance to dry soil conditions. The experiment was a 129 randomized complete block design with four replicates of each treatment. The total 130 area (8.7 ha) was divided into four blocks (replicates) and each block was subdivided 131 into four paddocks (treatments) with random distribution of treatments within each 132 block. During the two-year study, this area was dedicated solely to grazing and no 133 silage or hay was harvested. 134 135 Cows 136 Treatments within each block were simultaneously grazed by four homogeneous 137 groups of 7 to 10 autumn-winter-calving Prim’Holstein dairy cows, using a rotational 138 grazing system. The same cows could not be used during the 2 years of the study, 139 and several reference periods were needed to allocate cows in the 4 groups. A total 140 of 6 pre-experimental periods were considered throughout the trial, during which all 141 cows were managed similarly as a single herd (Table 2). Cows were balanced at 142 each pre-experimental period according to lactation stage (181 ± 85.4 days in milk), 143 milk production (27.4 ± 7.52 kg/day), milk fat concentration (37.1 ± 3.79 g/kg), milk 144 protein concentration (31.2 ± 2.39 g/kg), body weight (BW) (627 ± 15.2 kg) and body 145 condition score (BCS) (2.16 ± 0.19). On average, cows were 100, 164 and 280 days 146 in milk and yielded 35.4, 26.4 and 20.3 kg/day of milk respectively in the pre147 experimental spring, summer and autumn periods considered. 148 149 Grazing management and pastures 150 7 Due to limited grazing area compared to herd size (high global stocking rate), 151 permanent grazing within the experimental area was not possible. Consequently, 152 grazing was organised by rotations and cows grazed non experimental pastures as a 153 single herd between 2 experimental rotations. In total, 13 grazing rotations (7 in year 154 1 and 6 in year 2) were carried out during the two-year study, with a yearly average 155 of 2 rotations in autumn, 2.5 rotations in spring and 2 rotations in summer. Within a 156 grazing rotation, the 4 blocks were grazed successively, always in the same order. 157 Within each block, the 4 herds grazed simultaneously in 1 of the 4 paddocks, 1 herd 158 being dedicated to 1 treatment for the entire rotation and until the next reference 159 period (Table 2). During the 8 rotations without intake measurement (see Animal 160 measurements section below), a rotational grazing system was used (2 to 6 days of 161 residence time per paddock according to season and pasture availability). During the 162 5 rotations with intake measurement (see Animal measurements section below), a 163 strip-grazing system was employed using temporary electric fences. Fresh pasture 164 was allocated once daily after the a.m. milking and the pasture access time was 165 approximately from 9.00 a.m. to 3.30 p.m. and from 5.00 p.m. to 6.30 a.m. Cows 166 were thus at grazing 20 hours daily and received no supplement. The total residence 167 time per paddock and therefore area on offer each day was calculated 1 or 2 days 168 before grazing from the pre-grazing pasture mass estimated as described in the 169 Sward measurements section, and considering the two following management rules: 170 1) same grazing calendar (i.e. same dates) between treatments to avoid time lag, 171 and 2) similar pasture allowance (19 kg DM/cow/day > 4 cm) between treatments, to 172 define a medium to high grazing pressure (Pérez-Prieto and Delagarde, 2013) for 173 controlling post-grazing sward height. To combine both rules, additional, mobile, non174 experimental dairy cows were needed to adjust grazing pressure within each block 175 8 and between treatments based on differences in pre-grazing pasture mass. Pasture 176 refusals were mowed once per paddock and per year in late spring to a 5-6 cm 177 stubble height, and the clipped residues were left in place. The nitrogen fertilisation 178 level was similar between treatments (75 kg N/ha/year) by implementing 3 equal 179 applications of 25 kg N/ha/rotation of ammonium nitrate in spring and early summer 180 after grazing. Water and mineral blocks were always available to each herd during 181 grazing. The walking distance from paddocks to the milking parlour averaged 610 m. 182 183 Sward measurements 184 Pre-grazing pasture mass, pre-grazing and post-grazing sward heights, sward bulk 185 density and pasture allowance were determined for each treatment in each block and 186 rotation. Preand post-grazing sward heights were measured with an electronic rising 187 plate meter (30 × 30 cm and 4.5 kg/m2, AGRO-Systèmes, La Membrolle, France) on 188 the days before and after grazing by taking 60 and 50 measurements per treatment 189 at random, respectively, across 4 diagonals of each paddock, i.e. 100-120 190 measurements per ha. Pre-grazing sward height was adjusted by pasture daily 191 growth rate estimated by weekly measurement of sward height, and considering lag 192 time between measurement days and average grazing days within the paddock. 193 Adjusted pre-grazing pasture mass above 4 cm was calculated by multiplying the 194 adjusted pre-grazing sward height by the sward bulk density above 4 cm. To 195 determine sward bulk density, 4 strips of 8 m × 0.5 m per treatment were cut with a 196 motor scythe to a post-cutting sward height of 4 cm above ground level. The pasture 197 height on each strip was measured with a rising plate meter, before and after mowing 198 (15 measurements per strip), making it possible to estimate bulk density by dividing 199 pasture mass by cutting depth. The total quantity of pasture collected in each strip 200 9 was weighed and a fresh, representative 500-g subsample was oven-dried to 201 determine pasture DM concentration. Another fresh 500-g subsample was collected 202 at the same time, washed and oven-dried before being analysed for ash and crude 203 protein in blocks 1 to 4, for fibre in blocks 2 and 4, and for pepsin-cellulase 204 digestibility in blocks 2 and 4, but only during rotations with intake measurement. 205 Pasture botanical composition was determined for each treatment in 12 of the 13 206 grazing rotations carried out during the experiment (except in the grazing rotation 13 207 in late summer). A fresh pasture subsample of approximately 1,000 g was taken from 208 blocks 2 and 4 before grazing. Handfuls of pasture were randomly collected at each 209 of the 4 steps across 4 diagonals in each paddock to a cutting height of 3-4 cm 210 above ground level. A fresh representative 500-g pasture subsample was 211 immediately separated into 7 botanical items (Lolium perenne L., Trifolium repens L., 212 Trifolium pratense L., Cichorium intybus L., Festuca arundinacea Schreb, unsown 213 species and senescent material). Each constituent was oven-dried for 48 h at 80°C to 214 determine botanical composition on a DM basis. 215 216 Animal measurements 217 Cows were milked twice daily at 7.30 a.m. and 4.00 p.m. Milk production per cow was 218 recorded at each milking throughout the experiment. Milk fat and milk protein 219 concentrations were measured using individual milk samples, collected during 6 220 consecutive milkings per week, by near infrared spectrophotometry using a 221 Milkoscan instrument (Foss Electric, DK-3400, Hillerød, Denmark). Production of 4% 222 fat-corrected milk (4% FCM) was calculated according to INRA (2007). The body 223 weight (BW) of each cow was recorded automatically once daily after morning 224 milking. 225 16 The chemical composition differed between the 7 botanical items considered (Table 375 5). Perennial ryegrass and tall fescue showed similar chemical composition. White 376 clover and red clover showed also similar chemical composition, except for fibre 377 concentrations, greater in red clover than in white clover. On average, legumes were 378 characterised by lower DM and NDF concentrations, and by greater CP and ADL 379 concentrations than grasses. Chicory had a very specific chemical composition, with 380 lower DM and OM concentrations than grasses and legumes, low NDF concentration, 381 close to that of legumes, similar CP concentration to that of grasses, and similar ADF 382 and ADL concentrations as those in red clover. 383 384 Milk production, milk composition and body weight 385 Milk production averaged 17.3 kg/day and was greater by 1.1 kg/day in mixed swards 386 than in ryegrass swards (P<0.01), and greater by 0.8 kg/day in MSS as opposed to 387 mixed swards (P<0.05) (Table 6). On average, milk fat concentration was lower in 388 MSS than in mixed swards (-1.0 g/kg, P<0.05), but this mainly occurred in autumn (- 389 2.7 g/kg) and not in spring (+0.5 g/kg) (interaction treatment × season, P<0.001, 390 Table 7). Milk protein concentration was unaffected by treatment. Fat-corrected milk 391 production, milk fat production, milk protein production and milk solids production 392 followed the same trends as milk production, and were greater in mixed swards than 393 in ryegrass swards (P<0.01), and greater in MSS than in mixed swards (P<0.05) 394 (Table 6). Milk production and milk solids production decreased from spring to 395 autumn due to the advancing lactation stage of the herd (Table 2). The treatment had 396 no effect on body weight (596 kg). 397 398 Faecal output, digestibility, intake, energy and protein balance, and uremia 399 17 Faecal OM output was on average unaffected by treatment (Table 6). In summer, 400 however, faecal output was greater in mixed swards than in ryegrass swards, and 401 greater in MSS than in mixed swards (interaction treatment × season, P<0.001, Table 402 7). Pasture OM digestibility averaged 792 g/kg and was greater in MSS than in mixed 403 swards, regardless of season (+10.0 g/kg, P<0.001). Pasture DM intake averaged 404 15.6 kg/day and was 1.6 kg greater in MSS than in mixed swards (P<0.01, Table 6). 405 This positive effect of MSS compared to mixed swards on pasture DM intake was at 406 its lowest in autumn (+0.7 kg DM/day) and its greatest in summer (+2.3 kg DM/day) 407 (interaction treatment × season: P<0.05, Table 7). Pasture OM intake and pasture 408 digestible OM intake followed the same trends as pasture DM intake, with greater 409 values in MSS than in mixed swards, and greater positive effect of MSS in summer 410 than in autumn and spring (Tables 6 and 7). Pasture intake (in DM, OM, or digestible 411 OM) was only greater in mixed swards than in ryegrass swards in summer (no 412 average effect, but interaction treatment × season: P<0.05, Table 7). Considering 413 only the 5 rotations with pasture DM intake measurement, milk production, fat414 corrected milk production and milk solids production were greater in MSS than in 415 mixed swards (P<0.05), and tended to be greater in LTCF than in LTC (P<0.10) 416 (Table 6), with no interaction based on the season (Table 7). On average, there were 417 no difference between treatments on UFL and PDIE balances, which averaged 107% 418 and 122% of requirements, respectively. Grazing MSS in summer enabled an 419 increase in energy balance compared to mixed swards, which did not occur in 420 autumn nor in spring (interaction treatment × season, P<0.001, Table 7). The 421 treatment had no effect on blood urea nitrogen concentration (238 mg/l). 422 423 Feeding behaviour 424 18 Grazing time averaged 532 min/day and was greater by 57 min/day in MSS than in 425 mixed swards (P<0.01, Table 6). This difference tended to be greater in summer as 426 opposed to spring (+74 vs. 41 min/day, respectively; interaction treatment × season: 427 P=0.07). The average duration of a grazing bout tended to be greater in MSS than in 428 mixed swards (+16 min/bout, P=0.06), and particularly in summer (interaction 429 treatment × season: P<0.05). No treatment effect was observed on the first grazing 430 bout duration (157 min), nor the number of grazing bouts (5.9 bouts/day), nor pasture 431 intake rate (31.7 g DM/min). 432 433 Discussion 434 The aim of this study was to determine how pastures sown with increasing botanical 435 complexity from 1 (grass) to 5 sward species (grass-legume-forb) affect pasture 436 chemical composition and nutritive value, milk production, milk solids production, 437 pasture DM intake and the grazing behaviour of dairy cows in different seasons. This 438 was successfully achieved, as botanical composition clearly differed between 439 treatments, even if clover and fescue proportions were lower, and chicory proportions 440 were greater than expected. Although some significant interactions between season 441 and treatment occurred for pasture botanical and chemical composition, the overall 442 effects of sward type on animal performance did not strongly differ between seasons. 443 This could be related to the weather conditions, globally favorable for pasture growth 444 and quality in both years, including early summer in the ryegrass swards. Finally, pre445 grazing pasture mass and pre-grazing pasture allowance were similar between 446 treatments, enabling us to compare sward types under similar grazing conditions and 447 management. 448 449 19 Effect of introducing clover in ryegrass swards 450 Including clovers in perennial ryegrass swards had no effect on pasture nutritive 451 value and pasture DM intake, yet presented a small positive effect on daily milk 452 production (+1.1 kg/day) and milk solids production. These results may be regarded 453 as consistent with the literature given the small difference in clover content between 454 ryegrass and mixed swards observed in our experiment (6 vs. 22%), along with the 455 high quality of the ryegrass pastures. 456 Clovers are recognized as highly digestible and protein-rich forage species (Peyraud, 457 1993; Ribeiro-Filho et al., 2003; INRA, 2007). Similarly, voluntary intake of legumes, 458 and particularly clovers, is known to be 10% to 20% greater than that of grasses 459 (Ribeiro-Filho et al., 2003; INRA, 2007), due to their lower fibre concentration, lower 460 resistance to chewing and faster rate of particle breakdown (Dewhurst et al., 2009). 461 For these reasons, inclusion of white clover in a grass-based diet generally enhances 462 diet quality, daily pasture DM intake and milk production of dairy cows, either at 463 grazing (Harris et al., 1997; Phillips and James, 1998; Ribeiro-Filho et al., 2003) or 464 indoors (Harris et al., 1998). In our study, the increase in milk production of 1.1 465 kg/day in mixed compared to ryegrass swards is within the range of 1-3 kg/day, as 466 reported by Ribeiro-Filho et al. (2003), in several short-term experiments at the same 467 daily pasture allowance. The amplitude of the positive effect of clover may depend on 468 the ratio between grass and clover quality, and on the proportion of clover in the 469 swards (Harris et al., 1997; Harris et al., 1998). The greatest milk production 470 response to clover inclusion, almost 3 kg/day, was observed with low grass quality, 471 and clover content of more than 50% (Harris et al., 1997), where clover largely 472 increases diet quality, which is not the case in this experiment. 473 20 No differences concerning grazing behaviour were found between ryegrass and 474 mixed swards, corroborating previous results of Phillips and James (1998), Ribeiro475 Filho et al. (2003) and Ribeiro-Filho et al. (2012). The latter found that the daily 476 patterns of grazing and ruminating activities were similar between cows grazing on 477 grass-legume swards or pure grass swards, and that the average daily pasture intake 478 rate was only slightly affected by the inclusion of 40% of clover. Including clover in 479 ryegrass swards had no clear effect on milk composition, similarly to the results of 480 Harris et al. (1997) and Ribeiro-Filho et al. (2003). 481 482 Effect of multi-species swards compared to single ryegrass/clover mixtures 483 In our study, MSS were characterised by a high proportion of chicory (30%) that 484 replaced perennial ryegrass, with no changes in the clover proportion compared to 485 the single ryegrass/clover mixture. The high quality of MSS observed must thus be 486 related to the high proportion of chicory and to the specific chemical composition of 487 this species. When analysed separately, chicory was mainly characterised by low DM 488 and NDF concentrations, and high ash concentration, as previously found in many 489 studies (Barry, 1998; Sanderson, 2010; Muir et al., 2014). Due to the high mineral 490 concentration of chicory, as in other forbs, MSS containing chicory may be regarded 491 as an interesting option for enhancing macroand micro-minerals supply (Barry, 492 1998; Marley et al., 2013). Low NDF concentration, related to high OM digestibility, 493 may also be regarded as a nutritional advantage, leading to high energy 494 concentration and nutritive value of pastures (INRA, 2007; Muir et al., 2014). Low 495 NDF concentration is also cited as one of the main factors explaining high voluntary 496 intake, such as in clovers, through faster ruminal particle breakdown and passage 497 rates (INRA, 2007; Dewhurst et al., 2009). Conversely, the low DM concentration of 498 21 chicory may be regarded as a potential factor limiting its intake (Tinworth et al., 1999) 499 and thus the intake of MSS containing chicory, given that internal water is known to 500 limit pasture DM intake rate and daily DM intake in dairy cows fed on fresh grass 501 (Cabrera-Estrada et al., 2004). In our experiment, the positive effect of the presence 502 of chicory on pasture DM intake suggests that the negative effect of chicory’s low DM 503 concentration is largely compensated by the positive effect of the low NDF 504 concentration or any other component on daily intake. This is in line with several 505 previous studies, where intake and/or milk production have been found to increase 506 when chicory was included in mixed pastures and fed to dairy cows, either at grazing 507 (Chapman et al., 2008; Totty et al., 2013) or indoors (Barry, 1998; Minnee et al., 508 2012). The increase in milk production after inclusion of chicory in the diet generally 509 ranges from 1 to 2 kg/day, but an increase in milk production as high as 6 kg/day has 510 been observed for cows grazing clover-chicory mixtures in summer when compared 511 to low quality grass-based pastures (Chapman et al., 2008). 512 Advantages of MSS on a per cow basis were also clear for milk production and milk 513 solids production, and were due to the cumulative effect of greater pasture quality 514 and greater pasture intake. An additional reason may be that a mixture of several 515 forages has positive associative effects on daily intake (Harris et al., 2004), probably 516 through an increased motivation to eat while no digestive interactions have been 517 detected when mixing several forages (Niderkorn et al., 2014). This may be also 518 related to the high ingestibility of chicory due to its low NDF concentration and rapid 519 ruminal particle breakdown (Niderkorn et al., 2014), as it occurs with legumes 520 compared to grasses. Pasture digestibility was only slightly affected by pasture type, 521 suggesting that the low NDF concentration in chicory is partly compensated by its 522 high lignin concentration and potentially by its lower fibre digestibility. 523 22 Some studies reported no positive effects of MSS or chicory on daily pasture DM 524 intake or milk production in grazing dairy cows, but these studies were generally 525 carried out at high concentrate supplementation levels, namely 9 kg/day in Soder et 526 al. (2006) and 6 kg/day in Muir et al. (2014). In our study, the lower milk fat 527 concentration found in MSS compared to mixed swards during autumn may be 528 related to the greater milk production and to the highest chicory proportion in the 529 swards, leading to the lowest diet fibre concentration and probably greatest ruminal 530 degradation rate and associated changes in VFA profile. 531 Greater daily pasture DM intake in MSS was reached through more time spent 532 grazing while no change occurred in pasture DM intake rate between MSS and mixed 533 swards. This suggests no fundamental changes in the short-term feeding behaviour 534 of cows when grazing chicory-based pastures, although large structural differences 535 exist between grasses, clovers and chicory. Similarly, no effect has been reported on 536 daily pasture intake rate of grazing dairy cows by including 30 to 40% of chicory in an 537 orchardgrass and white clover pasture (Soder et al., 2007) or in a perennial ryegrass 538 sward (Gregorini et al., 2013). 539 540 Effect of introducing tall fescue in multi-species swards 541 In our study, the fact that introducing tall fescue had no overall effect on pasture DM 542 intake nor on milk production may be related to the low tall fescue proportion in 543 swards (10%); as fescue partly replaced perennial ryegrass, while clover and chicory 544 proportions were unaffected. Grazing pure tall fescue generally had no effect, or else 545 decreased milk production in dairy cows by 1 or 2 kg/day when compared to grazing 546 pure perennial ryegrass (Lowe et al., 1999), due to greater fibre concentration and 547 lower digestibility, pasture intake being only slightly affected (INRA, 2007). There is 548 23 evidence that replacing 10% of DM diet from ryegrass to fescue would only have 549 small effects on cow nutrition, which is in line with the results of Chapman et al. 550 (2008) in mixed swards. Greater effects would be expected with older swards, or 551 greater development and proportion of tall fescue in the swards, as tall fescue is well 552 known for its relatively low rate of establishment compared to other grass species. 553 554 Conclusion 555 The comparison of perennial ryegrass monoculture, grass-legume mixed swards, and 556 multispecies swards containing grasses, legumes and chicory, only grazed by 557 lactating dairy cows during 2 years, has shown that increasing sward botanical 558 complexity from 1 to 5 species has a positive effect on a per cow performance basis 559 under similar grazing management. All sown species were of good quality, but 560 inclusion of both clovers and of chicory made it possible to enhance sward quality, 561 milk production and milk solids production on a per cow basis. The advantages of 562 multispecies grazing over mixed swards on milk production and milk solids 563 production are due to the cumulative effect of enhanced sward quality and increased 564 pasture DM intake, in all seasons, with relation to the very specific chemical 565 composition of chicory. How increasing sward botanical complexity may increase the 566 grazing system’s resilience to climate events such as drought, and may affect 567 pasture utilisation and milk production on a per hectare basis still remains to be 568 investigated. 569 570 Acknowledgments 571 The research leading to these results received funding from the European 572 Community's Seventh Framework Programme under the grant agreement no. FP7573 24 244983 (MultiSward). The financial support of the Fundación Juana de Vega in the 574 form of the first author’s post-doc fellowship is also gratefully acknowledged. The 575 authors would also like to acknowledge all the staff at the INRA dairy farm of 576 Méjusseaume (UMR1348 PEGASE, Le Rheu, France) and the technicians at 577 UMR1348 PEGASE (Saint-Gilles, France) for their chemical laboratory analyses. 578 579 References 580 Association Française de Normalisation 1997. Aliments des animaux - Dosage de 581 l’azote - Méthode par combustion (DUMAS) - NF V18 - 120 : Dosage des cendres 582 brutes - NF V18 - 101. AFNOR Editions, Saint-Denis La Plaine, France. 583 Aufrère J and Michalet-Doreau B 1988. Comparison of methods for predicting 584 digestibility of feeds. Animal Feed Science and Technology 20, 203-218. 585 Barry TN 1998. The feeding value of chicory (Cichorium intybus) for ruminant 586 livestock. Journal of Agricultural Science (Cambridge) 131, 251-257. 587 Cabrera-Estrada JI, Delagarde R, Faverdin P and Peyraud JL 2004. Dry matter 588 intake and eating rate of grass by dairy cows is restricted by internal, but not external 589 water. Animal Feed Science and Technology 114, 59-74. 590 Chapman DF, Tharmaraj J and Nie ZN 2008. Milk-production potential of different 591 sward types in a temperate southern Australian environment. Grass and Forage 592 Science 63, 221-233. 593 Deak A, Hall MH and Sanderson MA 2009. Grazing schedule effect on forage 594 production and nutritive value of diverse forage mixtures. Agronomy Journal 101, 595 408-414. 596 25 Delagarde R and Lamberton P 2015. Daily grazing time of dairy cows is recorded 597 accurately using the Lifecorder Plus device. Applied Animal Behaviour Science 165, 598 25-32. 599 Dewhurst RJ, Delaby L, Moloney A, Boland T and Lewis E 2009. Nutritive value of 600 forage legumes used for grazing and silage. Irish Journal of Agricultural and Food 601 Research 48, 167-187. 602 Gregorini P, Minnee EMK, Griffiths W and Lee JM 2013. Dairy cows increase 603 ingestive mastication and reduce ruminative chewing when grazing chicory and 604 plantain. Journal of Dairy Science 96, 7798-7805. 605 Harris SL, Clark DA, Auldist MJ, Waugh CD and Laboyrie PG 1997. Optimum white 606 clover content for dairy pastures. Proceedings of the New Zealand Grassland 607 Association 59, 29-33. 608 Harris SL, Auldist MJ, Clark DA and Jansen EBL 1998. Effect of white clover content 609 in the diet on herbage intake, milk production and milk composition of New Zealand 610 dairy cows housed indoors. Journal of Dairy Research 65, 389-400. 611 Høgh-Jensen H, Nielsen B and Milan-Thamsborg S 2006. Productivity and quality, 612 competition and facilitation of chicory in ryegrass/legume-based pastures under 613 various nitrogen supply levels. European Journal of Agronomy 24, 247-256. 614 Hoskin S, Stafford K and Barry T 1995. Digestion, rumen fermentation and chewing 615 behavior of red deer fed fresh chicory and perennial ryegrass. Journal of Agricultural 616 Science 124, 289-295. 617 Huyghe C, De Vliegher A, van Gils B and Peeters A 2014. Grassland and herbivore 618 production in Europe and effect of common policies. Ed Quae, 287 pp. 619 32 716 717 718 Figure 1. Monthly mean temperature and cumulated rainfall of the experimental 719 years (year 1, September 2011 – August 2012 and year 2, September 2012 – August 720 2013) and of the last 30-year average 721 See specific file 722 723 724 33 Table 3. Mean pasture botanical composition (12 rotations) and pasture 725 characteristics (13 rotations) of multi-species swards rotationally grazed by dairy 726 cows (2 years) 727 Sward treatments1 Contrasts3 L LT LTC LTCF SDt2 T M F Botanical composition of pasture offered (proportion of DM) Lolium perenne L. 0.70 0.59 0.44 0.34 Trifolium repens L. 0.06 0.13 0.10 0.10 Trifolium pratense L. 0 0.09 0.10 0.07 Cichorium intybus L. 0 0 0.29 0.30 Festuca arundinacea Schreb. 0 0 0 0.10 Unsown species 0.18 0.13 0.05 0.04 Senescent material 0.09 0.08 0.06 0.06 Pasture mass (kg DM/ha, > 4 cm) 2218 2436 2388 2354 971.1 0.259 0.700 0.853 Pre-grazing sward height (cm) 12.6 13.7 15.8 15.2 4.23 0.204 0.013 0.509 Sward bulk density (kg DM/ha/cm, > 4 cm) 267 270 219 229 37.5 0.646 0.001 0.195 Pasture allowance (kg DM/day, > 4 cm) 18.7 19.1 19.3 19.4 2.13 0.384 0.418 0.881 Post-grazing sward height (cm) 4.8 5.0 4.7 4.7 0.78 0.125 0.030 0.798 Chemical composition and nutritive value of pasture offered (> 4 cm) DM (g/kg) 185 173 138 141 20.7 0.004 0.001 0.395 OM (g/kg DM) 897 895 866 869 9.4 0.230 0.001 0.228 CP (g/kg DM) 187 190 189 200 23.3 0.504 0.247 0.019 NDF (g/kg DM) 535 530 469 484 29.6 0.563 0.001 0.081 ADF (g/kg DM) 257 263 263 268 15.7 0.161 0.523 0.244 ADL (g/kg DM) 32 38 51 49 9.1 0.012 0.001 0.435 PCd (g/kg DM) 752 734 776 758 24.9 0.159 0.005 0.168 UFL (/kg DM) 0.88 0.86 0.88 0.86 0.024 0.135 0.580 0.153 PDIE (g/kg DM) 96 94 95 94 2.7 0.168 0.696 0.337 DM = Dry Matter; OM = Organic Matter; CP = Crude Protein; NDF = Neutral Detergent Fibre; 728 ADF = Acid Detergent Fibre; ADL = Acid Detergent Lignin; PCd = Pepsin-cellulase 729 Digestibility; UFL = Unité Fourragère Lait (Feed unit for milk production; 1UFL = 7.115 MJ 730 NE); PDIE = Protein truly digestible in the intestine, with energy-limiting microbial synthesis in 731 the rumen. 732 1 See Table 1. 733 2 Standard deviation of the model for the effect of the treatment. 734 3 Orthogonal contrasts: T (L vs. LT), M (LT vs. LTC/LTCF) and F (LTC vs. LTCF). 735 34 Table 4. Seasonal pasture botanical composition (12 rotations) and pasture characteristics (13 rotations) of multi-species swards rotationally 736 grazed by dairy cows (2 years) 737 1 See Tables 1 and 3. 738 2 Standard deviation of the model for the effect of the season. 739 3 Above 4 cm 740 741 Grazing season Autumn Spring Summer p-value Sward treatments1 L LT LTC LTCF L LT LTC LTCF L LT LTC LTCF SDs2 s t ˣ s Botanical composition of pasture offered (proportion of DM) Lolium perenne L. 0.62 0.49 0.24 0.18 0.77 0.71 0.59 0.47 0.65 0.50 0.38 0.33 Trifolium repens L. 0 0.19 0.10 0.09 0.05 0.09 0.08 0.07 0.13 0.17 0.12 0.15 Trifolium pratense L. 0 0.14 0.12 0.10 0 0.06 0.09 0.06 0 0.10 0.09 0.06 Cichorium intybus L. 0 0 0.46 0.46 0 0 0.19 0.22 0 0 0.29 0.26 Festuca arundinacea Schreb. 0 0 0 0.08 0 0 0 0.07 0 0 0 0.11 Unsown species 0.27 0.14 0.06 0.03 0.15 0.11 0.04 0.04 0.16 0.17 0.07 0.04 Senescent material 0.16 0.14 0.11 0.09 0.06 0.05 0.03 0.05 0.06 0.06 0.05 0.05 Pasture mass (kg DM/ha)3 1138 1357 1581 1474 3452 3722 3482 3549 2064 2228 2101 2038 4791.7 0.001 0.946 Pre-grazing sward height (cm) 8.3 8.9 11.3 10.7 17.3 18.8 20.2 20.0 12.3 13.3 15.8 14.9 82.22 0.001 0.998 Sward bulk density (kg DM/ha/cm)3 268 296 240 253 275 268 234 243 257 246 184 191 220.9 0.001 0.054 Pasture allowance (kg DM/day)3 16.5 17.6 18.2 17.8 20.7 20.9 20.3 21.1 18.9 18.8 19.5 19.3 6.88 0.001 0.539 Post-grazing sward height (cm) 3.9 3.9 3.7 3.6 5.4 5.8 5.2 5.3 4.9 5.3 5.1 5.3 8.93 0.001 0.464 Chemical composition and nutritive value of pasture offered 3 DM (g/kg) 182 176 131 133 176 167 140 139 197 175 142 152 133.9 0.003 0.206 OM (g/kg DM) 889 888 849 850 905 902 879 882 897 895 871 874 54.1 0.001 0.001 CP (g/kg DM) 210 206 211 230 164 169 178 183 185 194 178 187 84.6 0.001 0.077 NDF (g/kg DM) 539 527 439 445 525 528 472 498 541 536 497 508 72.5 0.001 0.011 ADF (g/kg DM) 249 256 259 259 255 262 252 263 267 272 278 282 53.8 0.001 0.695 ADL (g/kg DM) 31 38 55 54 32 37 47 44 32 40 50 48 28.9 0.111 0.611 PCd (g/kg DM) 788 763 861 840 756 748 771 741 711 690 697 694 57.1 0.001 0.066 UFL (/kg DM) 0.92 0.89 0.95 0.93 0.88 0.87 0.87 0.84 0.84 0.82 0.80 0.80 0.165 0.001 0.213 PDIE (g/kg DM) 101 98 104 103 95 94 94 90 92 91 88 89 17.4 0.001 0.075 35 742 743 Table 5. Chemical composition (g/kg DM: 12 rotations) of the botanical items in multi-species swards rotationally grazed by dairy cows 744 Lolium Trifolium Trifolium Cichorium Festuca Unsow n Senescent p-value Contrasts2 perenne repens pratense intybus arundinacea species material SD1 sp sp x s C vs. L C vs. G G vs. L L. L. L. L. Schreb. DM (g/kg) 196 156 177 102 202 148 271 26.6 0.001 0.820 0.001 0.001 0.002 OM 892 891 896 833 889 875 865 11.5 0.001 0.021 0.001 0.001 0.428 CP 196 253 239 193 204 210 114 18.5 0.001 0.004 0.001 0.386 0.001 NDF 520 345 389 350 559 375 619 30.8 0.001 0.501 0.190 0.001 0.001 ADF 229 208 244 244 249 232 342 22.9 0.001 0.451 0.064 0.644 0.089 ADL 26 56 91 94 21 77 58 10.1 0.001 0.187 0.001 0.001 0.001 1 Standard deviation of the model for the effect of the year and the season. 745 2 Orthogonal contrasts: chicory vs. legumes (C vs. L), chicory vs. grasses (C vs. G) and grasses vs. legumes (G vs. L). 746 747 748 36 Table 6. Milk production, milk composition and body weight (13 rotations), pasture intake, nutrient balance and uremia (5 rotations), and grazing 749 behaviour (2 rotations) of dairy cows rotationally grazing on multi-species swards 750 Sward treatments1 Contrasts L LT LTC LTCF RSDt2 T M F Milk production (kg/day) 16.1 17.2 17.8 18.2 1.94 0.005 0.022 0.357 4% FCM production (kg/day) 15.9 17.0 17.5 17.9 1.81 0.003 0.026 0.377 Milk fat concentration (g/kg) 41.0 41.0 40.3 39.8 2.75 0.989 0.049 0.430 Milk protein concentration (g/kg) 32.8 32.6 32.4 32.0 1.57 0.481 0.152 0.237 Milk solids production (kg/cow) 1.14 1.22 1.25 1.28 0.129 0.003 0.039 0.439 Milk fat production (g/day) 633 673 696 708 73.6 0.006 0.034 0.474 Milk protein production (g/day) 511 542 559 568 59.8 0.006 0.049 0.474 Body weight (kg) 594 599 593 598 14.5 0.083 0.210 0.221 Faecal output OM (kg/day) 2.73 2.84 2.92 2.94 0.308 0.298 0.370 0.854 Pasture OM digestibility (g/kg) 788 786 798 794 0.3 0.518 0.001 0.123 Pasture OM intake (kg/day) 13.0 13.4 14.5 14.3 1.65 0.424 0.054 0.774 Pasture DM intake (kg/day) 14.4 15.0 16.6 16.5 1.95 0.388 0.008 0.835 Pasture digestible OM intake (kg) 10.2 10.6 11.6 11.4 1.37 0.463 0.034 0.695 Milk production (kg/day) 15.3 15.8 17.2 18.5 1.95 0.481 0.012 0.064 4% FCM production (kg/day) 15.4 16.0 17.0 18.1 1.76 0.384 0.007 0.080 Milk solids production (kg/cow) 1.12 1.15 1.21 1.29 0.128 0.445 0.012 0.084 UFL balance (% of requirements) 1.05 1.06 1.09 1.06 0.104 0.720 0.580 0.352 PDIE balance (% of requirements) 1.23 1.19 1.25 1.20 0.117 0.295 0.322 0.138 Blood urea nitrogen concentration (mg/l) 245 224 233 251 20.3 0.198 0.209 0.298 Grazing time (min/day) 503 503 570 550 30.7 0.978 0.002 0.306 First grazing bout duration (min) 146 161 164 155 17.5 0.271 0.931 0.508 Number of grazing bouts (/day) 5.7 6.1 5.5 6.2 0.80 0.436 0.595 0.124 Mean grazing bout duration (min) 97 90 114 98 16.3 0.432 0.056 0.113 Pasture intake rate (g DM/min) 31.9 32.1 31.1 31.5 2.38 0.903 0.619 0.856 1 See Tables 1 and 3. 2 Residual standard deviation of the model for the effect of the treatment. 751 37 Table 7. Seasonal variation of milk production, milk composition and body weight (13 rotations), pasture intake, nutrient balance and uremia (5 752 rotations), and grazing behaviour (2 rotations) of cows rotationally grazing on multi-species swards 753 1 See Tables 1, 3 and 6. 2 Residual standard deviation of the model for the effect of the season. 754 Grazing season Autumn Spring Summer p-value Sward treatments1 L LT LTC LTCF L LT LTC LTCF L LT LTC LTCF RSDs2 s t ˣ s Milk production (kg/day) 10.5 11.9 12.0 13.2 22.5 23.0 24.0 23.7 15.3 16.5 17.3 17.6 0.27 0.001 0.405 4% FCM production (kg/day) 11.8 13.3 13.1 13.9 21.3 21.8 22.9 22.7 14.7 15.9 16.5 17.1 0.25 0.001 0.431 Milk fat concentration (g/kg) 49.1 48.1 46.9 43.9 36.6 36.5 36.7 37.3 37.4 38.5 37.2 38.1 0.41 0.001 0.001 Milk protein concentration (g/kg) 38.0 37.0 37.3 35.9 29.4 29.7 29.7 29.6 31.0 31.0 30.1 30.4 0.21 0.001 0.010 Milk solids production (kg/cow) 0.90 1.00 1.00 1.04 1.48 1.52 1.61 1.58 1.04 1.13 1.16 1.20 0.018 0.001 0.464 Milk fat production (g/day) 506 564 557 574 821 835 889 881 572 621 642 667 10.5 0.001 0.335 Milk protein production (g/day) 396 438 439 470 663 684 717 699 473 505 520 534 8.1 0.001 0.492 Body weight (kg) 598 607 599 603 589 599 593 599 594 592 588 591 2.5 0.001 0.116 Faecal output OM (kg/day) 2.71 2.77 2.50 2.70 2.80 2.71 2.90 2.80 2.69 3.05 3.35 3.31 0.064 0.001 0.001 Pasture OM digestibility (g/kg) 779 773 790 788 797 800 809 805 789 786 796 789 2.0 0.001 0.307 Pasture OM intake (kg/day) 12.2 12.3 11.9 12.8 13.9 13.7 15.1 14.4 12.8 14.3 16.4 15.7 0.34 0.001 0.010 Pasture DM intake (kg/day) 13.4 13.6 13.7 14.9 15.5 15.2 17.4 16.5 14.4 16.1 18.8 18.0 0.39 0.001 0.019 Pasture digestible OM intake (kg) 9.5 9.5 9.4 10.1 11.1 10.9 12.2 11.6 10.2 11.3 13.1 12.4 0.28 0.001 0.028 Milk production (kg/day) 10.1 9.7 11.3 14.4 19.8 20.5 21.9 22.2 16.0 17.2 18.4 18.8 0.41 0.001 0.578 4% FCM production (kg/day) 12.1 12.3 12.7 12.5 19.0 19.3 20.8 20.9 15.3 16.3 17.5 17.9 0.38 0.001 0.659 Milk solids production (kg/cow) 0.94 0.96 0.97 1.18 1.33 1.34 1.45 1.45 1.08 1.14 1.21 1.24 0.027 0.001 0.603 UFL balance (% requirements) 1.10 1.06 1.02 1.00 1.01 0.98 1.04 0.98 1.03 1.13 1.21 1.19 0.018 0.001 0.001 PDIE balance (% requirements) 1.32 1.20 1.21 1.40 1.17 1.09 1.16 1.09 1.20 1.29 1.37 1.36 0.064 0.001 0.001 Blood urea nitrogen (mg/l) - - - - 225 203 206 236 266 244 261 265 9.2 0.005 0.882 Grazing time (min/day) - - - - 513 512 557 548 492 494 583 552 8.3 0.738 0.064 First grazing bout duration (min) - - - - 142 154 148 143 150 167 180 167 5.5 0.001 0.149 Number of grazing bouts (/day) - - - - 5.8 6.3 6.0 6.8 5.7 5.9 5.1 5.6 0.19 0.001 0.212 Mean grazing bout duration (min) - - - - 102 89 104 90 93 91 124 107 4.1 0.041 0.023 Pasture intake rate (g DM/min) - - - - 30.4 30.1 29.9 30.3 33.4 34.1 32.3 32.6 0.81 0.001 0.681 Figure 1 Roca-Fernandez et al, Animal, 2015 0 20 40 60 80 100 120 140 0 5 10 15 20 25 Sept Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Rainfall (mm/month) Mean Temperature (°C) Month Rainfall year 1 Rainfall year 2 Rainfall last 30 years Temperature year 1 Temperature year 2 Temperature last 30 years Figure