In situ rumen degradation kinetics as affected by type of pasture and date of harvest

Botanical composition affects the nutritive value and nutrient degradation kinetics of pastures. However, there is little information about pastures composed of various species. The present study was conducted to evaluate in situ degradation kinetics of dry matter (DM), crude protein (CP) and neutral detergent fiber (NDF) of a naturalized non fertilized pasture (NNF), naturalized fertilized pasture (NF), sown pasture with Lolium perenne L. and Trifolium repens L. (RGWC), and sown pasture with Bromus valdivianus Phil., Dactylis glomerata L., Holcus lanatus L., L. perenne and T. repens (MIXED); at three dates of harvest (early-spring, mid-spring and autumn). Duplicate bags were incubated in the rumen of three cannulated cows for 2, 4, 8, 12, 24 or 48 h. Zero-hour bags were washed with water. Ruminal degradation was evaluated according to exponential models with or without lag phase. Pasture chemical composition showed on average low DM, high CP and digestible organic matter (DOMD). Degradation parameters of DM, CP and NDF exhibited strong pasture type*date of harvest interactions. In general, high potential degradability and fractional degradation rates of DM, CP and NDF were observed. Nutrient supply was generally higher for RGWC and NF, especially compared with NNF during autumn. Naturalized fertilized pastures supply similar amounts of nutrients to grazing cattle as sown pastures. However, it is necessary to consider other variables that determine the productive performance such as DM intake and rumen fermentation products.


Introduction
In humid-temperate regions, perennial ryegrass (Lolium perenne L.) pastures are the main source of nutrients for dairy cattle; however alternatives that might improve farm sustainability have been evaluated (Nie et al., 2004).Naturalized fertilized pastures may be suitable for cattle production systems as they have similar herbage mass yield and nutrient content as sown pastures; and lower production costs (Balocchi et al., 2010).Nutrient content does not assure certain animal performance levels, as productive responses depend on the quantity of nutrients available to the animal, which in turn is determined by fermentation processes (Hackmann et al., 2010).Therefore, alternative evaluations such as nutrient ruminal digestibility (Dulphy et al., 1999) are necessary to establish the feasibility of naturalized pastures as a forage source for grazing dairy systems.
Ruminal degradability of pastures is affected by the fertilization, season of the year, grazing management, botanical composition, stage of maturity, and method of conservation (Aufrere et al., 2003;Kasuya et al., 2008).In addition, forage species-stage of maturity interactions may also occur.Stages of maturity may vary among grass species at a given date of harvest and differences in digestibility occur due to the variable age of leaves or tillers, the timing of ear emergence and stem elongation (Bruinenberg et al., 2002).From this point of view, naturalized pastures are a unique combination of species, whose nutritive value should be determined.
The ruminal in situ incubation technique is considered to be a reference method to estimate degradation parameters, such as soluble, insoluble but degradable and undegradable fractions, potential and effective de-gradability, when adjusted to suitable nonlinear models (Ørskov and McDonald, 1979).These parameters are used by feeding evaluation models to estimate the nutritive value, nutrient supply and animal performance (Hackmann et al., 2010).For example, high rates of degradation have been implied to result in high voluntary intake and thus higher performance (Sun et al., 2012).
To our knowledge, there is little information about the degradation kinetics of permanent pastures with different botanical composition at a vegetative stage of maturity.So, the aim of this study was to assess the rumen degradation dynamics of dry matter, crude protein and neutral detergent fi ber of four permanent pastures with different botanical composition, at three different dates of harvest in southern Chile.

Grass herbage samples
All herbage samples were obtained in Valdivia, Chile (39º47' S, 73º13' W).The four evaluated pastures comprised a naturalized non fertilized pasture (NNF), naturalized fertilized pasture (NF), a sown pasture with L. perenne and Trifolium repens (RGWC), and a sown pasture with Bromus valdivianus Phil., D. glomerata, Holcus lanatus L., L. perenne and T. repens (MIXED), which was sown aiming to represent a naturalized pasture in good condition (according to Grime, 1989), and a similar proportion between grass species (17.5 %) and 30 % of T. repens.The layout of the pastures was organized in a randomized block design with three replicates (paddocks with an area of 400 m 2 ).Each treatment was grazed by dairy cows, once it had reached an average pre-grazing herbage mass of 2,400 -2,600 kg DM ha −1 , estimated by  Jenquip, Feilding, New Zealand).Cows left the paddocks with an average herbage mass of 1,400 -1,600 kg DM ha −1 .
Treatments were comprised of the four pastures during three dates of harvest (DH1: early-spring [5 Oct 2009], DH2: mid spring [5 Nov 2009] and DH3: autumn  [5 Apr 2010]).Fertilizers were applied to meet the requirements of a pasture with a potential production of 12 t DM ha −1 (i.e. 150 kg N ha −1 ; 50 kg P 2 O 5 ha −1 ; 120 kg K 2 O ha −1 ; 115 kg MgO ha −1 ; 1,000 kg CaCO 3 ha −1 ).Five applications of 30 kg N ha −1 were done in Aug, Oct and Nov 2009, and Mar and Apr 2010.The other nutrients were applied once a year in Apr 2009.All forage samples were harvested at the same time of the day (10h00) from each paddock, the day before a grazing event; cutting at fi ve cm above ground level with shearing scissors in ten different places in the paddock, at random.One pooled sample (ca.250 g) was obtained for each paddock (fi eld replicate) to determine botanical composition and nutrient content.The remaining material was freeze-dried and then ground to 5 mm using a Wiley mill.

In situ incubations
Three dry cows (635.5 ± 35.6 kg live weight) fi tted with a ruminal cannula (4" Pliable Rumen Cannula w/ Stopper and U Bolt, ANKOM technologies, NY, USA) were employed for the rumen degradability study.Cows grazed on a grass (main species: Agrostis capillaris, B. valdivianus, D. glomerata, and L. perenne) pasture (187 g kg −1 DM; 164 g CP kg −1 DM; 511 g NDF kg −1 DM) while incubations were performed.A new allocation was offered daily with a herbage allowance of 7 kg DM for each cow.An apparent intake of 5 kg DM day −1 was estimated as the difference between pre-and post-grazing pasture herbage masses and the area grazed.One day prior to the incubation, the rumen pH of cannulated cows was measured every two hours to assure optimal conditions for microbial growth; this resulted in an average pH level of 6.29 ± 0.16.
Samples of approximately 4 g DM were placed in dacron bags (10 × 20 cm; pore size of 40 -60 µm) and sealed.Dacron bags (n: 144) were incubated in the rumen for 2, 4, 8, 12, 24 and 48 h using a "gradual in/ all out" sequence, with duplicate bags at each time for each forage sample.All bags were placed within three polyester laundry bags (30 × 40 cm) to ensure similar location within the rumen and to facilitate location for removal (Vanzant et al., 1996).After removal from the rumen, dacron bags containing forage residues were washed under running cold water until no further color appeared, frozen at -20 ºC for 24 h to stop fermentative activity and rinsed in a domestic washing machine with cold tap water for 30 minutes.The 0 h bags (n = 6 for each treatment) were not placed in the rumen and the solubility was determined by soaking them in warm water (40 ºC) for 20 minutes.Residues were dried at 60 ºC for 48 h, individually weighed and duplicates were pooled by incubation time to analyze their CP and NDF content.From these data, nutrient disappearance was calculated.

Digestion kinetics
Digestion curves for each data set were used to evaluate in situ degradation kinetics of pasture treatments in the rumen.Data were fi tted into exponential models with (McDonald, 1981) and without (Ørskov and McDonald, 1979) lag time.The model with highest correlation coeffi cient was selected for each nutrient.Thus, the disappearance of DM and CP were determined using the non-linear model described by Ørskov and McDonald (1979) to determine the constants and potential degradation (PD) according to the exponential model: where: A is the soluble fraction (g kg -1 ; fraction washed out at t = 0; this value resulted from the incubation of 0 h bags and fi xed into the model); B is the insoluble degradable fraction (g kg -1 ), k is the fractional degradation rate (h −1 ) and t is the time (h).
The effective degradability (ED; g kg -1 ) was calculated from the afore mentioned parameters assuming fractional passage rates (kp) of 2, 5 and 8 % h −1 : The passage rates are based on NRC (2001), which recommends those values for low, medium and highproducing dairy cows, respectively.
The disappearance of NDF was determined using the non-linear model with a lag phase described by Mc-Donald (1981): where: PD (g kg -1 ) is the degraded NDF at time t, A the intercept with the Y-axis; B (g kg -1 ) the insoluble but potentially degradable NDF fraction, k (h −1 ) the fractional degradation rate constant, t the incubation time (h) and L is the lag phase (h).In the parameterization of the model we assumed NDF to be insoluble in water, the A-fraction represented feed particle losses, and the fractional degradation rate was similar in the lost and remaining particles.The effective degradability in the rumen (ED; g kg -1 ) was calculated through the equation of McDonald (1981): where: A (g kg -1 ) is the soluble or rapidly degradable fraction; B (g kg -1 ), the degradation between A and the asymptote; k and kp (h -1 ), the fractional degradation and passage rate, respectively; and L, the lag time.where: Y ijk is an observation of the dependent variable; µ is the population mean for the variable;  i is the random effect of the fi eld replicate;  j is the effect of type of pasture;  k is the effect of date of harvest;  jk is the effect of the interaction between type of pasture and date of harvest; and  ijk is the random error associated with the observation ijk.
Values with p  0.05 were considered as signifi cant.Means were compared using the least squares mean linear hypothesis test.

Botanical and chemical composition
The naturalized non fertilized pasture was composed mainly of browntop (Agrostis capillaris L.), yorkshire fog (Holcus lanatus) and broad leaf species (BLS), with similar proportions among dates of harvest.The main species of the naturalized fertilized pasture were perennial ryegrass (L.perenne) and bromegrass (B. valdivianus).The improvement in soil nutrient content has shown to favor valuable grasses at the expense of forbs and less desirable grasses such as browntop (Butkuviene and Butkute, 2008).The perennial ryegrass and white clover mixture had over 900 g kg -1 of ryegrass; whereas orchard grass (D. glomerata) followed by bromegrass dominated the MIXED pasture (Table 1).
Pastures had on average low DM (174 g kg −1 ), high CP (216 g CP kg −1 DM) and DOMD (> 700 g kg −1 ; Table 2), which is characteristic of vegetative high quality pastures of temperate regions (Kolver and Muller, 1998).Pastures DM content was affected by the interaction between type of pasture and date of harvest (p < 0.01).Surprisingly, NNF and RGWC harvested during early spring showed higher DM content than when Model parameters were estimated with the nonlinear procedure of Graphpad Prism v5.0.Separate curves were estimated for each replicate.

Botanical and Chemical Analyses
Botanical composition of the pastures was determined by separating each species from fresh pasture samples by hand, then dried at 60 ºC for 48 h and weighed to estimate the proportion of each species in relation to the total dry mass.
Forages and bag residues were dried at 60 ºC for 48 h in a forced-air oven and ground in a Wiley mill through a 1mm sieve.Analytical dry matter content was determined by weighing before and after drying at 105 ºC for 12 h (AOAC, 1996).Analytical fractions for each pasture sample comprised crude protein (CP; AOAC, 1996; method 978.10), neutral detergent fi ber (NDF; Van Soest et al., 1991), acid detergent fi ber (ADF; AOAC, 1996), digestible organic matter in the dry matter (DOMD) by two-stage in vitro digestibility (Tilley and Terry, 1963), water-soluble carbohydrates (WSC; MAFF, 1985) and soluble protein contents (SP; Licitra et al., 1996).

Statistical Analyses
Prior to ANOVA, assumptions of normality and homogeneity of variance of the data were checked.Botanical and chemical composition, fi tted NLIN degradation parameters and ED were analyzed as a randomized complete block design with a 4 × 3 factorial arrangement, with four pastures and three dates of harvest as main factors and the random effect of fi eld replicate as a block; using the MIXED procedure of SAS (SAS Institute, 2006).The model used for the analysis was: Within fertilized treatments, NF had higher levels of CP than RGWC and MIXED for DH1 and DH2, and lower CP content than RGWC and MIXED for DH3.In general terms, WSC content was below 120 g kg −1 DM and the lowest levels were observed for MIXED (p < 0.001), which may be explained by the high proportion of orchard grass in this pasture, a species which is characterized by its low WSC content (Fulkerson et al., 2007;Turner et al., 2007).NDF content ranged between 343 and 518 g kg −1 DM, with highest levels observed for mid-spring pastures.This may be due to higher cell wall and lignin contents in forage plants as the growing season progressed (Aufrere et al., 2003;Kasuya et al., 2008).However, the interaction between type of pasture and date of harvest (p < 0.01) indicates that the NDF content decreased from late spring to autumn for NNF, NF and RGWC, whereas it remained similar for MIXED.
Acid detergent fi ber was only affected by the date of harvest (p < 0.001), with higher levels for late spring and autumn than for early spring, which is in accordance with the increase in cell wall content as plants become reproductive (Chaves et al., 2006b).

Dry Matter degradation kinetics
The PD, 'A' and 'B' fractions, 'k' coeffi cient, ED2, ED5 and ED8 exhibited strong 'pasture'*'date of harvest' interactions (p  0.05; Table 3), which was also reported by Dohme et al. (2007).The PD for the RGWC-DH1 was the highest (958 g kg -1 DM) above all other pasture treatments, while NNF-DH3 had the lowest PD (730 g kg -1 DM).These values are similar to others (Dohme et al., 2007;Trujillo et al., 2010;Valderrama and Anrique, 2011).The higher PD of RGWC than NNF, may be related to the higher degradability of L. perenne than Agrostis spp.(Bruinenberg et al., 2004), the main species  In general terms PD tended to decrease from DH1 to DH3.It was expected that because of changes in the ratio of cell walls to cell contents, the digestibilities of pastures would be higher in the early vegetative stage (early spring for this study) due to a high cell soluble content (Groot et al., 1999).
The DM readily soluble fraction (A) of RGWC-DH1 was higher (464 g kg -1 DM) than that of the other treatments, with MIXED-DH3 having the lowest value (255 g kg -1 DM).RGWC showed the lowest NDF content during the three dates of harvest, resulting in a higher readily soluble fraction.Higher soluble fractions for perennial ryegrass than for orchard grass and bromegrass have been reported earlier (Dohme et al., 2007;Hoffman et al., 1993;Le Goffe et al., 1993).Our values are similar to those reported by Valderrama and Anrique (2011) for vegetative mixed and ryegrass pastures (346 and 442 g kg -1 DM, respectively), Trujillo et al. (2010) for perennial ryegrass (380 g kg -1 DM) and Sun et al. (2010) for vegetative perennial ryegrass harvested in early spring (435 -470 g kg -1 DM).The soluble fraction was higher for RGWC, NNF and NF during DH1, DH2 and DH3, respectively.MIXED showed a consistently lower soluble fraction for the three dates of harvest, which is consistent with the lower WSC content (Table 2).
The readily soluble fraction of DM was higher for early spring than for late spring and autumn.This agrees with results obtained by Dohme et al. (2007).However, Le Goffe et al. (1993) did not fi nd effects for spring, summer and autumn collected forages; suggesting that the effect of date of harvest on degradability parameters depends on the forage species.On the other hand, insoluble but degradable fraction 'B' of DM refl ected the inverse of the readily soluble fraction and was consistently higher for MIXED across the three dates of harvests, whereas NNF had the lowest values during DH2 and DH3.Although insoluble degradable fraction was higher for DH2, such difference was not considered biologically important as it represented less than 5 % of differences among the highest and lowest values.The values of RGWC and MIXED agree with those reported by Le Goffe et al. (1993) for perennial ryegrass (496 and 531 g kg -1 DM) and orchard grass (564 g kg -1 DM); and Trujillo et al. (2010) for perennial ryegrass (520 g kg -1 DM).
The naturalized fertilized pasture had the highest fractional degradation rates among pastures across the three dates of harvest.The highest and lowest fractional degradation rates of DM were observed for NF-DH1 (0.22 h −1 ) and RGWC-DH2 (0.11 h −1 ), respectively; and are in accordance with the fractional degradation rates of 0.123 and 0.182 h −1 for ryegrass and mixed (mainly composed by grasses) pastures (Valderrama and Anrique;2011); and of 0.148 -0.173 h −1 for vegetative ryegrass cultivars with a similar regrowth period (Sun et al. 2010).Faster fractional degradation rates of DM were found for early spring than late spring and autumn.This might be related to a lower leaf:stem ratio, as stems degrade slower than leaves (Chaves et al., 2006a).
Pastures with higher fractional degradation rates had higher effective degradability (ED) values, and were less affected by passage rate.For example, the lowest effect of passage rate was observed with NF-DH1, which had the highest fractional degradation rate (0.219 h −1 ) with a reduction of 7.9 units of percentage of total degradability considering passage rates of 0.02 and 0.05 h −1 .
The effective degradability estimated with ruminal passage rates of 0.02, 0.05 and 0.08 h −1 were consistently higher for RGWC-DH1 (910; 853; and 822 g kg -1 DM, respectively) and lower for NNF-DH3 (670; 605; and 574 g kg -1 DM, respectively).All pastures showed high ED values during DH1 and DH2; however, it was considered that differences of biological importance among pastures only occurred during autumn, where NF showed a consistently higher ED compared to the other pastures (875, 814 and 781 g kg -1 DM for rumen outfl ows of 0.02, 0.05 and 0.08 h −1 ).In general terms, increasing ruminal outfl ow reduced the ED, but did not alter the patterns observed in the different treatments.Naturalized non fertilized had a lower ED during DH3 despite the passage rate used for the estimation.Naturalized fertilized pasture had similar ED values in DH1 and DH3 with a reduction in DH2, which was higher considering faster passage rates (39; 63; and 73 g kg -1 DM for passage rates of 0.02, 0.05 and 0.08 h −1 ).Effective degradability was highest for early spring, followed by late spring and autumn.The highest ruminal effective degradability for RGWC was observed in DH1 and diminished to similar values in DH2 and DH3.Finally ED of MIXED decreased from DH1 to DH3.Similar values for effective degradability for vegetative ryegrass pastures have been reported by Trujillo et al. (2010) with 680 g kg -1 DM considering a passage rate of 0.06 h −1 %, and Valderrama and Anrique (2011) with 878, 802 and 749 g kg -1 DM for 0.02, 0.05 and 0.08 h −1 rumen outfl ows.

Crude Protein degradation kinetics
All dependent variables were affected by the 'pasture'*'date of harvest' interaction (p < 0.05; Table 4).Hoffman et al. (1993) reported signifi cant forage species by stage of maturity interactions on all protein fractions except for degradation rate.The soluble fraction of crude protein was higher for RGWC during the three dates of harvest (459, 399 and 458 for DH1, DH2 and DH3, respectively), while the lowest values were estimated for MIXED during early spring, and NNF and MIXED during mid spring and autumn.The lower soluble fraction in MIXED may be explained by the proportion of orchard grass, a species that contains the Sci.Agric.v.70, n.6, p.405-414, November/December 2013 enzyme polyphenol oxidase, which reduces proteolysis by binding proteins with quinones forming complexes resistant to the digestive enzymes (Niderkorn and Baumont, 2009).Except for RGWC-DH1 and RGWC-DH2, predicted soluble fractions were lower than the 520 g kg -1 CP predicted for a perennial ryegrass pasture by Le Goffe et al. (1993), who observed that the CP soluble fraction was lower in autumn than in spring, whereas we observed higher 'A' fraction for autumn and early spring than for late spring.Although these differences were signifi cant, their relative small range (50 g kg -1 CP) suggest that seasonal variation in soluble CP fraction for NNF, NF and RGWC has little biological importance, which agrees with Steg et al. (1994) who found little effect of cutting date on CP soluble fraction.
The insoluble potentially degradable fraction of CP was higher for MIXED-DH1 (702 g kg -1 CP) and lower for RGWC-DH3 (440 g kg -1 CP).MIXED had a high degradable fraction for all dates of harvest, whereas RGWC 'B' fraction was always lower.Pastures harvested during autumn had a higher insoluble degradable fraction than those harvested in early and mid spring.
The maximum fractional disappearance rate was observed in RGWC-DH2 (0.294 h −1 ), and the lowest in RGWC-DH3 (0.146 h −1 ).Valderrama and Anrique (2011) predicted 'k' coeffi cients of 0.314 and 0.157 h −1 for mixed and ryegrass pastures, respectively, whereas Sun et al. (2010) reported CP fractional degradation rates ranging from 0.122 to 0.347 h −1 for perennial ryegrass, depending on the cultivar and regrowth period.In general terms, the fractional degradation rate tended to decrease in autumn.The interaction showed that the fractional degradation rate was higher for RGWC in DH1, RGWC and MIXED in DH2, whereas no differences among pastures were observed in DH3.
Fractional degradation rates of CP were higher than those of DM and NDF; this asynchrony has been described previously for permanent pastures (Chaves et al., 2006b;Sun et al., 2010).For example, Valderrama and Anrique (2011) estimated that differences between CP and DM degradation rates were about 73 and 27 % higher in mixed and ryegrass pastures, respectively.
In contrast, as occurred with the DM effective degradability, the patterns of crude protein ED (cpED) were more affected by the rumen passage rate used for the estimation.For example, the ED estimated with a passage rate of 0.02 h −1 , decreased from DH1 to DH3 for all pasture types except for NF with the highest and lowest values for RGWC-DH1 (942 g kg -1 CP) and NNF-DH3 (786 g kg -1 CP), respectively.On the other hand, considering a 0.05 h −1 rumen passage rate, ED decreased from DH1 to DH3 for NNF, NF and RGWC; while no differences were observed between DH1 and DH2 for MIXED.Furthermore, ED with passage rate of 0.08 h −1 decreased from DH1 to DH3 for NNF; was similar between DH2 and DH3 for NF; similar for DH1 and DH2, and higher than DH3; and higher in DH2 for MIXED (DH2>DH1>DH3).These fi ndings disagree with Flores et al. (2008) and Salaun et al. (1999) who suggest that estimates of cpED for autumn are comparable to those observed typically for early spring growth.
The lower cpED may be related to environmental factors such as rainfall, temperature and light, e.g. higher temperatures are negatively related to protein degradability (Salaun et al., 1999).In our experiment (Southern Hemisphere) the mean daily temperatures in the week preceding sampling were 9.8 ºC in Oct, 10.2 ºC in Nov and 14.2 ºC in Apr.Le Goffe et al. (1993) estimated an ED (kp = 0.06 h −1 ) that ranged between 796 and 858 g kg -1 CP, which are similar to our values for ED considering a kp of 0.05 h −1 ; they also observed a lower ED for autumn in relation to spring.
Values of effective rumen degradable protein (ERDP) give an estimate of N supply to microbes for grazing ruminants (Bowen et al., 2008).Higher and lower supplies of ERDP were obtained for NF-DH2 (222.8 g kg −1 DM) and NNF-DH2 (114.9 g kg −1 DM), respectively.On the other hand, the amount of protein escaping rumen degradation (UDP) may be available at an intestinal level as part of the metabolizable protein (Flachowsky and Lebzien, 2006); UDP was higher for RGWC-DH3 (50.5 g kg −1 DM) and lower for RGWC-DH1 (21.8 g kg −1 DM).Protein escaping from the rumen was higher in autumn than in early and mid spring for NNF, RGWC and MIXED, whereas for NF it was higher during mid spring and autumn than early spring.The higher UDP might be related to an increase in the N that is linked to the cell walls, which usually occurs as the growing season progresses and plants mature, resulting in a lower leaf:stem ratio (Chaves et al., 2006a).
According to NRC (2001), ERDP estimates satisfy the requirements of a 500 kg LW Holstein Friesian lactating dairy cow (Milk production: 20 l day −1 , Fat: 4.4 %, milk crude protein: 3.5 %, without body weight change) i.e. 99.1 g RDP per kg of ingested DM, assuming a DM intake of 15 kg day −1 .Furthermore, RDP supply exceeds requirements by 16 to 248 %.This surplus can be excreted as urinary N and may have negative impacts on the environment such as nitrate leaching, ammonia volatilization and nitrous oxide generation (Ledgard, 2001).

Neutral Detergent Fiber degradation kinetics
Concentration and degradation characteristics of NDF are important factors determining the energetic value of forages for ruminants (Moharrery et al., 2009).All NDF degradation parameters were affected by the 'pasture'*'date of harvest' interaction (p < 0.05), which agrees with Dohme et al. (2007) and Hoffman et al. (1993).More than 600 g kg -1 NDF was present in the insoluble and potential degradable fraction, which is in accordance with Hoffman et al. (1993) for perennial grasses at the second node stage of maturity, and Sun et al. (2010) for vegetative perennial ryegrass cultivars.
Insoluble degradable NDF was similar between early and mid-spring and higher than autumn.The highest B fraction was predicted for RGWC in DH1 (920 g kg -1 NDF; Table 5); no differences among pastures were observed in DH2, whereas RGWC and NF were higher in DH3; which agrees with Cherney et al. (1993), who did not fi nd differences in NDF degradable fractions between early and late spring.Salaun et al. (1999) found no differences for perennial ryegrass samples collected in late spring (908 g kg -1 NDF) and autumn (935 g kg -1 NDF); although their B fractions were numerically higher, which might be explained by the higher N fertilization applied in their experiment (250 kg N ha −1 ), as N fertilization increases theoretical degradation in the rumen (Ferri et al., 2004).
The degradation rate of NDF was higher for NF-DH1 (0.20 h −1 ) and lower for NNF, NF and RGWC in DH3 with values of around 0.09 h −1 .In general terms, fractional degradation rate tended to be higher in DH1 and DH2 than DH3.The interaction arose from highest degradation rates for NNF, NF and RGWC during DH1 and DH2, whereas MIXED was highest during DH3.Our results partially agree with Uden (2006), who estimated an insoluble degradable fraction of 883 g kg -1 NDF for perennial ryegrass harvested in spring with a short lag time (0.8 h), however the predicted fractional degradation rate was slower (0.042 h −1 ).On the other hand, Sun et al. (2010) predicted similar fractional degradation rates of NDF in perennial ryegrass for early spring (0.099 -0.183 h −1 ) and Hoffman et al. (1993) for autumn samples (0.09 h −1 ).Degradation rates for DH3 were similar to those predicted by Chaves et al. (2006a) for perennial ryegrass leaves.However, values obtained for DH1 and DH2 were higher, probably because of the differences in the stage of maturity between our (vegetative stage) and their (reproductive stage) samples.
The high 'k' coeffi cients suggest that pasture NDF, during spring, may be composed mainly by hemicelluloses with a low degree of lignifi cation, resulting in a faster fermentation (Cajarville et al., 2006), as hemicelluloses degrade faster than the other constituents of NDF or ADF (Hackmann et al., 2010).The same authors reported fractional degradation rates of 0.173 h −1 and 0.234 h −1 for NDF and hemicelluloses of alfalfa hay, respectively.These values may be comparable to vegetative grass pastures, as Hoffman et al. (1993) indicated that perennial ryegrass exhibits low lignifi cation, fast degradation rate and a high NDF effective degradability.Martineau et al. (2006) reported lower NDF 'B' fractions, 'k' coeffi cients and ED for timothy silages and hay harvested in reproductive stages, thus indicating the effect of the stage of maturity on NDF degradation kinetics.
Lag times were fi tted for NF-DH3 and MIXED-DH3 with 0.9 and 2.8 h, respectively, indicating a delay in the degradation of fraction B. This suggests a higher lignin content of these treatments, as lignin acts as a barrier to ruminal microbes and thus, hinders microbial attachment and colonization of the degradable fi ber.This is relevant because ruminal microbes can only ferment cell walls if they come into direct contact with them (Martineau et al., 2006).Effective disappearance of NDF was relatively extensive for all pastures.Effective NDF degradability with passage rates of 0.02, 0.05 and 0.08 h −1 , was consistently higher for RGWC-DH1 (851, 704 and 613 g kg -1 NDF, respectively) and lower for NNF-DH3 (490, 384 and 315 g kg -1 NDF for ED2, ED5 and ED8, respectively).Slightly lower NDF effective degradability (540 g kg -1 NDF, estimated assuming a ruminal passage rate of 0.06 h −1 ) for perennial ryegrass pasture have been reported previously (Trujillo et al., 2010).
Regardless of the rumen passage rate considered to estimate ED (0.02, 0.05 and 0.08 h −1 ), ED was similar between DH1 and DH2 and higher than DH3 for NNF and NF, while it decreased from DH1 to DH3 (DH1>DH2>DH3) for RGWC and MIXED.Similar neutral detergent fi ber ED values were reported by Dohme et al. (2007) for a passage rate of 0.02 h -1 (662 and 759 g kg -1 NDF for orchard grass and ryegrass) and 0.08 h −1 (482 and 585 g kg -1 NDF for orchard grass and ryegrass).Using a slower passage rate (0.02 instead of 0.05) increased the degradation of NDF by 140-300 g kg -1 NDF.Such increase has been already reported by Chaves et al. (2006a) and occurs as a consequence of a longer residence in the rumen.

Conclusions
High potential degradabilities and fractional degradation rates of DM, CP and NDF were observed, which is typical of forages harvested at immature stages of growth.Nutrient supply was generally higher for RGWC and NF, especially in comparison with NNF during autumn.Pastures harvested in early spring tended to have higher degradabilities and faster fractional degradation rates, followed by mid spring, and fi nally autumn.
Interactions between type of pasture and date of harvest occurred for all DM, CP and NDF degradation parameters, suggesting that it is advisable to have specific estimations of degradation patterns for different pastures at each season of utilization, hence a more accurate data for ration formulation.

Table 1 -
Mean botanical composition of pastures.

Table 2 -
Mean nutrient content of pastures.

Table 3 -
In situ dry matter degradation parameters of pastures across dates of harvest.RGWC and NNF, respectively.The 'pasture'*'date of harvest' interaction indicates that ruminal in situ DM degradation parameters of pastures were affected in different ways by the date of harvest.For example, the potential degradability was higher for RGWC during DH1; NNF and NF during DH2; and NF during DH3.