Introduction
Offering hay as the sole source of forage to dairy cows during the winter has become a more popular feeding strategy in the alpine regions of Europe, especially in organic farming. Besides financial advantages for dairy farmers when producing milk under such conditions, feeding hay as compared to silage enhances cows' forage dry matter intake (DMI) and milk energy output (Haselmann et al., Reference Haselmann, Wenter, Fuerst-Waltl, Zollitsch, Zebeli and Knaus2020). From a broader perspective, increasing the forage DMI of dairy cows also improves the net human food supply of the dairy industry by reducing or even eliminating the use of potentially human-edible ration components (i.e., grain, pulses) (Ertl, Reference Ertl2016).
Because the content of digestible nutrients is lower in forages than in concentrates, dairy cows fed heavily forage-based rations, as is more often the case with organic production, must achieve a higher forage intake level to better fulfill their nutrient and energy requirements. However, DMI is restricted, mainly by the physical fill of the rumen (Allen, Reference Allen2000) and the cows' chewing capacity (Beauchemin, Reference Beauchemin1991, Reference Beauchemin2018; Mertens, Reference Mertens1997).
Water-soluble carbohydrates (WSC) are valuable sources of readily available energy for rumen microbiota (Van Soest, Reference Van Soest1994), resulting in greater amounts of available nutrients from forages for dairy cows. One practical strategy to improve the WSC content of a given forage sward, resulting in improved performance and N utilization efficiency of dairy cows, is to cut forage later in the day, i.e., when plants have accumulated higher amounts of nonstructural carbohydrates (Brito et al., Reference Brito, Tremblay, Bertrand, Castonguay, Bélanger, Michaud, Lapierre, Benchaar, Petit, Ouellet and Berthiaume2008, Reference Brito, Tremblay, Lapierre, Bertrand, Castonguay, Bélanger, Michaud, Benchaar, Ouellet and Berthiaume2009). However, after plants are cut, WSC are respired by the plant, as well as by microbes present on the surface (Barr et al., Reference Barr, Smith and Brown1995). These researchers calculated that up to 40% of the total dry matter (DM) losses during field wilting of hay can be attributed to microbial activity. In addition, McGechan (Reference McGechan1989) observed that the respiratory rate increases linearly with rising temperatures, but quadratically with the incremental moisture content of plants.
In light of this, agricultural engineers developed the so-called ‘macerator’, which facilitates drying rates (amount of water per unit of time) that are one and a half to three times greater than observed for conventionally field-dried forages (Koegel et al., Reference Koegel, Straub, Shinners, Broderick and Mertens1992). In short, a set of intermeshing, profiled rolls crush the plants' stems, releasing moisture from the plant. This is highly effective for thick-stemmed plants such as alfalfa, but not for grasses, because the clearance between the rolls is not small enough to crush the thin stems. Therefore, impeller conditioners were developed that scratch the waxy cuticle of plants and break the stems at regular distances (2–3 cm). Although these processes speed up the drying rate, thus reducing respiration losses, they could potentially lead to higher DM losses, e.g., via shattering leaves (Greenlees et al., Reference Greenlees, Hanna, Marley, Bailey and Shinners2000).
However, it has not been studied if the mechanical stress of a mowing conditioner (macerator, impeller) also affects the feed intake behavior of cows (e.g., feed selection and chewing behavior). The shattering of leaves during the conditioning process might stop net photosynthesis in cut plants earlier as compared to intact plants. Studies have shown that excised leaves stop photosynthesis 30–40 min after being cut (Clark et al., Reference Clark, Prioul and Couderc1977), whereas intact plants (alfalfa) continue photosynthesis until they reach a DM content of about 35%. The mature leaf sheath is the major storage site for carbohydrate reserves in grasses (Thomas and James, Reference Thomas and James1999). We speculate that, in cut but not conditioned forage plants, the sheath might remain on the stems of grasses and probably continue to accumulate WSC for some time. This might affect the feed intake behavior of cows and could lead to a selective consumption of these plant parts. In addition, the mechanical stress caused by the conditioner also disrupts the physical integrity of stems, potentially reducing cows' mastication efforts while chewing (Susenbeth et al., Reference Susenbeth, Mayer, Koehler and Neumann1998) thus improving forage intake.
Therefore, the aim of this study was to test the effects of grass hay harvested with the use of an impeller mowing conditioner as the sole source of forage, supplemented with 3.6 kg d−1 of concentrate (DM basis), on the feed intake behavior, nutrient digestibility, nutrient and energy supply, and performance of dairy cows. With the exception of the immediate conditioning of the freshly cut grass, hay harvest was performed under ceteris paribus conditions (e.g., cutting date, wilting times, tedding and pick-up techniques, residual barn-drying).
Animals, materials and methods
Preparation of the control and experimental hay
For this experiment, both the control hay and the experimental hay were harvested in June 2019 at the organic farm (European Commission, 2008) of the Secondary School for Agriculture, HBLA Ursprung (Elixhausen, Austria; altitude 570 m, annual precipitation 1250 mm, average temperature 8.5°C). Both types of hay used in this experiment were harvested in equal shares of the first regrowth of 4 ha perennial clover grass as well as 3.6 ha permanent grassland. The sward of the perennial clover grass was composed of roughly 40% grass (mainly Lolium perenne, Phleum pratense, Festuca pratensis), 55% legumes (Trifolium repens, Medicago lupulina, Medicago sativa) and 5% herbs (mainly Plantago lanceolata and Taraxacum officinale). The sward of the permanent grassland was composed of roughly 85% grass (mainly Dactylis glomerata, Trisetum flavescens, L. perenne, P. pratense, L. multiflorum), 10% legumes (T. repens and pratense) and 5% herbs (mainly P. lanceolata, T. officinale, Galium mollugo). At the time of cut, the inflorescences of most grasses had emerged.
On June 25, 2019, in the late morning, half of the acreage of both swards was cut by a front-and-rear disc mower combination with an integrated impeller conditioner (Alpha Motion, Pöttinger, Grieskirchen, Austria; Disco 3200c, Class, Harsewinkel, Germany); this was for the experimental hay. The remaining halves were cut using the same mowing technique (Easy Cut Butterly, Krone, Spelle, Germany) but without a conditioner; this was for the control hay.
The average climatic conditions during the harvest, June 25 to June 27, were as follows: ambient temperature 27.7°C, soil temperature 25.3°C and wind velocity 27 km h−1.
The steel tines of the impeller conditioner had a length of roughly 13 cm and a width of 2 cm. They were arranged pairwise in a Y-shaped manner, pivoting freely around a rotating shaft. Detailed information about the structure of the impeller is given by Greenlees et al. (Reference Greenlees, Hanna, Marley, Bailey and Shinners2000). The hood position, which regulates how closely the cut forages are held to the impeller, was adjusted to 2.5 cm. During mowing, the driving speed of the tractor was about 6–8 km h−1 and the rotational speed of the power-take-off shaft was roughly 450 rpm. Hood position and speed settings were intended to provide an intensive conditioning of the cut forages. Cut forages were kept in windrows, with an approximate width of 1 m (front mower) to 1.5–2 m (rear mower) for 1–2 h, and were then spread by tedding (HIT 8.9T, Pöttinger).
Additionally, cut forages for both types of hay were tedded twice on the following day (24 and 28 h after mowing) to ensure optimal wilting conditions on the field, because the density and yield of both swards were very high. Wilted forages were windrowed roughly 45 h after being cut. Windrows of wilted forages were harvested with a loading wagon (Primo 401, Pöttinger). Before being introduced into one of the two ventilation boxes, the respective forages harvested from both swards were thoroughly mixed using the hay crane. Both ventilation boxes had a basal area of 120 m2 and an iron grate hung about 50 cm above the floor of the box. Control and experimental forages were sampled five times each before barn drying. Forages were dried for about 2 d at ambient air temperature until they reached a DM content of 88%.
When barn drying was completed, ventilation boxes were emptied, and the two different types of hay were mixed for a second time before being stored in separate piles in the same barn for about 7 months until the onset of the experimental feeding period. Three samples were taken from each type of hay for chemical analyses in order to perform a ration calculation according to GfE (2001) prior to the onset of the feeding trial.
Animals, experimental setup and feeding regime
The feeding trial was conducted in January and February 2020 on the dairy farm at HBLA Ursprung (Elixhausen, Austria), and included seven primiparous and 12 multiparous Holstein cows. Cows were kept in a freestall housing system and the feed bunk was equipped with CALAN gates (American CALAN Inc., Northwood, NH, USA) to guarantee individual feeding. The experiment was approved by the Institutional Ethics and Welfare Committee of the University of Veterinary Medicine Vienna (Vetmeduni Vienna) in accordance with good scientific practice guidelines and national legislation (protocol no. ETK-192/12/2019).
At the beginning of the feeding trial, nine and ten cows were assigned to a control and an experimental group, respectively. The assignment to the groups was intended to ensure well-balanced groups based on cows' previous milk yield, body weight (BW), days in milk and parity. Transponders were attached to the cows' necks to allow them to open their assigned CALAN gate. Cows were given 7 d to adapt to the activated CALAN gates, followed by a 9-d covariate feeding period and the experimental feeding period of 35 d.
Two weeks before the onset of the feeding trial and until the end of the covariate feeding period, all cows were fed the same total mixed ration (TMR) which contained, on a DM basis, 30% grass silage, 20% artificially dried alfalfa hay, 26% corn silage and 24% concentrate including a mineral and vitamin premix. The chemical composition of the ration was 167, 352, 223 and 144 g kg−1 DM for crude protein (CP), neutral detergent fiber (aNDFom), acid detergent fiber (ADFom) and starch, respectively, meeting the requirements of a 750-kg cow producing roughly 30 kg of milk (GfE, 2001).
Data collected daily during the covariate feeding period were averaged for each cow and used as covariates in the statistical model. During the covariate feeding period, the mean (±SD) DMI, energy-corrected milk (ECM) yield and BW of the control and experimental group were 21.3 ± 3.7 and 21.6 ± 4.0 kg, 31.1 ± 5.9 and 32.7 ± 5.1 kg, 705 ± 74 and 675 ± 79 kg, respectively. Parities of the cows in the control and experimental groups were 3.9 ± 3.1 and 3.1 ± 2.3, and days in milk at the end of the covariate feeding period were 131 ± 72 and 124 ± 93, respectively. A t-test of these parameters revealed no significant (P ≥ 0.05) differences between groups.
After the covariate feeding period, the experimental feeding period (35 d) began, during which cows received either the control or the experimental hay depending on their group assignment. Each day, appropriate amounts of both types of hay were transferred from the barn, where they were dried and stored, to the feed bunk using the hay crane and a wheel loader. It was then manually offered to cows several times a day (06:00–23:00) in amounts calculated to ensure that 10–15% (as-fed) would remain as feed refusals. In addition, each cow was fed 3.6 kg d−1 of concentrate (DM basis; Table 1), which was offered in equal shares at the CALAN gates at 06:00, 11:30 and 16:00. In order to ensure that each portion was fully consumed, hay was removed before offering the concentrate and returned after feeding.
Table 1. Chemical composition of the control and experimental hay and the concentrate [g (kg DM)−1, unless stated otherwise; arithmetic mean ± standard deviation]. Forages have been cut either with or without the use of a mowing conditioner (experimental hay and control hay, respectively). Differences between the treatments were analyzed by a two-folded t-test
uCP, utilizable crude protein at the duodenum (GfE, 2001); RNB, ruminal N balance (DLG, 1997); aNDFom, neutral detergent fiber assayed with a heat-stable amylase and expressed exclusive of residual ash; ADFom, acid detergent fiber expressed exclusive of residual ash; Lignin (sa), lignin determined by solubilization of carbohydrates in ADFom with sulfuric acid; NFC [g (kg DM)−1] = 1000 − (Ash + CP + aNDFom + ether extracts); WSC, water-soluble carbohydrates; NEL, net energy lactation, determined by pepsin-cellulase solubility of OM, according to GfE (2008).
a Chemical composition of the control and the experimental forages before barn drying [n = 5; g (kg DM)−1, unless stated otherwise]: DM [g (kg fresh matter)−1] (833, 799), CP (126, 133), aNDFom (506, 503), NFC (256, 251), WSC (104, 89.6) and NEL [MJ (kg DM)−1] (5.90, 5.76), respectively.
b Concentrate, pelleted, composed of [g (kg DM)−1] wheat bran (250), sugar beet pulp (250), soybean cake (250), maize middlings (182), molasses (30), minerals [NaCl, 24; Ca(H2PO4)2, 12] and a vitamin premix (2; Spurvit-Prämix Raiffeisenverband Salzburg, Austria).
Data collection and calculations
Data were collected during the covariate feeding period and during the final 21 d of the experimental feeding period. The first 14 d of the experimental feeding period were considered as an adaptation period. Data on DMI were collected on seven consecutive days during the covariate feeding period (9 d) and during the final 21 days of the experimental feeding period (35 d). DM content of fresh feed and refusals was determined daily by drying samples at 105°C for 24 h.
For chemical analyses, samples of the TMR (n = 2) offered during the covariate feeding period and the collected refusals (n = 2) were pooled over two consecutive days. Additional samples from both types of hay were drawn on d 16, d 18, d 20, d 22, d 24, d 25, d 26, d 28, d 30, d 32, d 33 and d 34 of the experimental feeding period. Each type of hay was sampled 15 times in total (Table 1). Each hay sample resulted from several deep grab samples at different places of the hay piles that were brought to the feed bunk daily. Hay grab samples were mixed thoroughly before lab analyses. Throughout the data collection period, individual grab samples of hay refusals were taken, relative to the amount of refusals, on nine regularly spaced days, and pooled for each day within each group (n = 9). In addition, four samples were taken of the concentrate.
Dry samples (i.e., hay and concentrate) were stored at room temperature in paper bags and wet feed samples (i.e., TMR and feed refusals) were vacuum packed and stored at −20°C until analysis in a commercial feed laboratory, where each sample was analyzed separately. The analytical methods applied were in accordance with VDLUFA (2012), method numbers: 8.1. (ash), 4.1.2. (CP), 5.1.1. (ether extracts), 6.5.1. (aNDFom; assayed with heat-stable amylase and expressed exclusive residual ash), 6.5.2. (ADFom; expressed exclusive residual ash), 6.5.3. [lignin (sa)], 7.3.1. (WSC) and 6.6.2. (pepsin-cellulase solubility of organic matter). Estimated energy content was evaluated based on the equation by GfE (2008). Mean values of the analytical results were used to calculate the cows' nutrient and energy intake.
Sorting of forage nutrients (i.e., CP, aNDFom, WSC) was expressed by the selection index according to DeVries et al. (Reference DeVries, Holtshausen, Oba and Beauchemin2011). The selection index was calculated as the percentage of the actual relative to the predicted forage nutrient intake for each cow. The actual forage nutrient intake is the difference between the amount of nutrients offered and refused. The predicted forage nutrient intake is the product of the forage DMI and the nutrient concentration in the offered forage. For this purpose, samples of individual forage refusals were pooled over three consecutive days for each cow, during the experimental feeding period, and analyzed for CP, NDF and WSC.
Chewing behavior was measured using 10 RumiWatch halters for four and seven consecutive days during the covariate and experimental feeding periods, respectively. During each period, halters were worn simultaneously by five or four cows in each group for the defined period of time (4 or 7 d, respectively) and afterwards switched to the remaining animals. Detailed information about the operation mode and a validation of the system is presented by Kröger et al. (Reference Kröger, Humer, Neubauer, Kraft, Ertl and Zebeli2016).
Milk yield was recorded automatically during milking (06:00 and 16:30) in a 2 × 3 herringbone milking parlor for nine consecutive days in the covariate feeding period and 21 consecutive days in the experimental feeding period. Individual milk samples were taken from two consecutive milkings, twice in the covariate feeding period and four times during the experimental feeding period (d 15, d 21, d 28, d 35). Samples were conserved with Bronysolv (ANA.LI.TIK, Vienna, Austria) and analyzed for fat, protein, lactose and urea by a MilkoScan (MilkoScan 7, Foss Electric, Hillerød, Denmark) according to the procedure DIN ISO 9622:2017-04. Nitrogen use efficiency of cows was calculated based on the factors 6.38 and 6.25 for milk protein yield and CP intake, respectively.
BW was measured after four consecutive milkings, once in the covariate and twice in the experimental feeding period (d 1 and d 35). Backfat thickness (BFT) was determined using a portable ultrasound generator (Linear VET transrectal probe, SV3513 VET; Esaote SpA, Genova, Italy), according to the procedure described by Schröder and Staufenbiel (Reference Schröder and Staufenbiel2006). Body condition score (BCS) was assessed according to the procedure described by Edmonson et al. (Reference Edmonson, Lean, Weaver, Farver and Webster1989). The changes in BW, BFT and BCS were calculated as the differences between values observed at the beginning and the end of the experimental feeding period (35 d).
Apparent total tract digestibility of nutrients was determined using acid-insoluble ash as an internal indicator. Nine fecal samples were taken from each cow's rectum. This was performed in 8-h intervals, at 04:30, 12:30 and 20:30, starting 1 week prior to the end of the experiment. Samples were immediately frozen at −20°C. In the laboratory, thawed samples (roughly 90 d after sampling) were pooled for each cow, thoroughly mixed, dried with a freeze dryer (CoolSafe 100-9 pro, LaboGene, Lillerød, Denmark) and ground to < 500 μm (ZM 200, Retsch GmbH, Haan, Germany). The chemical analysis of samples was performed in accordance with VDLUFA (2012). Fecal pH was measured with an electrode (InLab Expert Pro, Mettler Toledo Inc., Columbus, OH, USA) in thawed individual fecal samples (n = 9 per cow) according to the procedure described in Li et al. (Reference Li, Khafipour, Krause, Kroeker, Rodriguez-Lecompte, Gozho and Plaizier2012).
Statistical analysis
The statistical analysis was performed in SAS (version 9.4, SAS Institute Inc., Cary, USA), using the MIXED procedure. Data were checked for normal distribution and the presence of outliers. The statistical model included the fixed effects treatment and day of the experiment, the corresponding value of the response variable from the covariate feeding period as covariate, and as a random effect, the cow nested within each treatment. For intake parameters (DM, nutrient and energy intake), the average BW and ECM yield observed during the experimental feeding period were also included in the statistical model. For chewing data, cows were nested within the treatment and the wearing period of halters. A first-order autoregressive covariance structure and the option ‘repeated measures’ in SAS was used for data that had been measured repeatedly on the same animal (e.g., DMI, milk yield, chewing behavior). As only one observation per cow was available for the parameters BW, BFT, BCS and nutrient digestibility, no random effect was fitted in the model.
A two-tailed t-test was used to test the differences in chemical composition between the treatments and whether selection indices differed significantly from 100 (i.e., for each treatment and parameter). The level of significance was set at P ≤ 0.05, and 0.05 < P ≤ 0.10 was interpreted as indicating a tendency.
Results
Nutritive composition of the control and the experimental hay
The preparation of hay with or without the use of an impeller mowing conditioner under strict ceteris paribus conditions resulted in significantly higher (P ≤ 0.01) amounts (per kg of DM) of ash (+4.8 g), CP (+9 g) and lignin (+5.7 g), but lower (P < 0.01) amounts of WSC (−12 g) and ether extracts (−1.6 g) in the experimental hay as compared to the control hay (Table 1). Accordingly, the energy content of the experimental hay was also lower as compared to the control hay [5.54 vs 5.67 MJ NEL (kg DM)−1; P = 0.01].
Chewing and sorting behavior, intake of dry matter, nutrients and energy
Chewing behavior did not differ between groups (P ≥ 0.28; Table 2). Intake of DM from forage and aNDFom intake also remained constant (average 18.4 and 10.5 kg d−1, respectively; Table 3). In contrast, compared to the control group, cows in the experimental group showed a higher (P < 0.01) lignin intake (+0.12 kg d−1; +16%) and a tendency (P = 0.10) toward an increased ADFom intake (+0.3 kg d−1; +5%). Furthermore, cows in the experimental group ingested a higher (P < 0.01) amount of CP (+0.21 kg d−1; +7%) and the ingested ration of these cows clearly showed (P < 0.01) a more positive ruminal N balance (RNB) as compared to the ingested ration of cows in the control group (+19 g d−1). In contrast, the intake of WSC was lower (P < 0.01) in the experimental group (−0.19 kg d−1; −8%), although NFC intake did not differ between groups (P = 0.43).
Table 2. Least squares means ± SE for the effects of barn-dried hay harvested with the use of a mowing conditioner on chewing behavior of dairy cows
a NDFom, NDF assayed with a heat-stable amylase and expressed exclusive of residual ash.
Table 3. Least squares means ± SE for the effects of barn-dried hay harvested with the use of a mowing conditioner on dairy cows' intake of DM, nutrients, energy, their sorting behaviora, and on energy and ruminal nitrogen balance
uCP, utilizable crude protein at the duodenum (GfE, 2001); RNB, ruminal N balance (DLG, 1997); aNDFom, neutral detergent fiber assayed with a heat-stable amylase and expressed exclusive of residual ash; ADFom, acid detergent fiber expressed exclusive of residual ash; Lignin (sa), lignin determined by solubilization of carbohydrates in ADFom with sulfuric acid; NFC [g (kg DM)−1] = 1000 − (Ash + CP + aNDFom + ether extracts); WSC, water-soluble carbohydrates; NEL, net energy lactation, determined by pepsin-cellulase solubility of OM, according to GfE (2008).
a According to DeVries et al. (Reference DeVries, Holtshausen, Oba and Beauchemin2011): Sorting % = 100 × (actual intake of Y/predicted intake of Y), where Y = forage nutrient (CP, aNDFom, WSC); values equal to 100% indicate no sorting, <100% sorting against (selective refusal) and >100% sorting for (preferential consumption); *P <0.05, **P <0.01 (significant differences from 100%; based on a t-test).
Feed sorting occurred in both feeding groups. Based on a t-test (H0: selection index = 100) and following DeVries et al. (Reference DeVries, Holtshausen, Oba and Beauchemin2011), cows in the control group sorted in favor of WSC (101.6; P < 0.01) and aNDFom (100.7; P < 0.01), while cows in the experimental group sorted for CP (100.5; P = 0.01) (Table 3). With respect to the treatment effect, we found that cows in the control group selectively preferred WSC (P < 0.01) as compared to cows in the experimental group. Although less pronounced (P = 0.07), a similar pattern was found for the sorting of aNDFom (Table 3).
Apparent total tract digestibility of nutrients, performance and efficiency parameters
Digestibility data (Table 4) did not show any significant differences between groups. Although numerically relatively small, we found a significant difference for fecal pH, being 0.12 points higher in cows of the experimental group. We did not observe any effects on milk yield or changes in body condition parameters (BW, BFT, BCS) when feeding the experimental instead of the control hay (Table 5). With respect to milk composition, the only difference we found was for milk urea content. The concentration was 5.4 mg (100 mL)−1 higher (+30.2%; P < 0.01) in the experimental as compared to the control group. Furthermore, nitrogen efficiency was 2.2% points lower (P = 0.01) in cows fed the experimental hay.
Table 4. Least squares means ± SE for the effects of barn-dried hay harvested with the use of a mowing conditioner on apparent total tract digestibility of nutrients as well as on DM content and pH value of feces in dairy cows
aNDFom, neutral detergent fiber assayed with a heat-stable amylase and expressed exclusive of residual ash; ADFom, acid detergent fiber expressed exclusive of residual ash; NFC [g (kg DM)−1] = 1000 − (Ash + CP + aNDFom + ether extracts).
Table 5. Least squares means ± SE for the effects of barn-dried hay harvested with the use of a mowing conditioner on dairy cows' performance, body condition, energy and protein balance, and efficiency parameters
ECM, energy-corrected milk yield (GfE, 2001); NEL and uCP balance = (intake/requirement) × 100 (GfE, 2001, 2008); uCP, utilizable crude protein (GfE, 2001); NEL, net energy lactation, determined by pepsin-cellulase solubility of OM, according to GfE (2008).
a Values represent means throughout the 35-d experimental feeding period; Body condition score was evaluated according to Edmonson et al. (Reference Edmonson, Lean, Weaver, Farver and Webster1989) on a 1–5 scale with 0.25 increments, 1 point: severe underconditioning, 5 points: severe overconditioning.
The day of the experiment had a significant but negligible effect on the response traits shown in Tables 2, 3 and 5 and is therefore not reported in detail.
Discussion
To the best of our knowledge, this is the first study dealing with the effects of hay harvested with the use of an impeller mowing conditioner on organic dairy cows' feeding behavior, feed intake and milk yield. Most of the previous studies on this topic have been on the use of a macerator to intensively condition freshly cut forages (Hong et al., Reference Hong, Broderick, Koegel, Shinners and Straub1988; Mertens and Koegel, Reference Mertens and Koegel1996; Broderick et al., Reference Broderick, Koegel, Mauries, Schneeberger and Kraus1999, Reference Broderick, Koegel, Walgenbach and Kraus2002).
The overall objective of this study was to test the effects of feeding grass forages treated with an impeller mowing conditioner to lactating dairy cows. Grass forages were subjected to the same cutting date, field curing technique and wilting times, field pickup, and barn drying method (ceteris paribus conditions), and the resulting hay made up 84% of the daily ration (DM basis). Cows of both groups responded with similar feeding behavior and forage intake, resulting in a similar milk yield.
Cows in both groups reached an appropriate feed intake level (Table 3) when comparing the current results with previous studies using similar amounts of concentrate in the ration (Ertl et al., Reference Ertl, Zebeli, Zollitsch and Knaus2016; Haselmann et al., Reference Haselmann, Wenter, Fuerst-Waltl, Zollitsch, Zebeli and Knaus2020). In contrast to our hypothesis, the use of the mowing conditioner for haymaking did not alter the cows' feed intake. Based on this observation and together with the fact that chewing behavior was not affected by the treatment (Table 2), we can speculate that similar mastication efforts were required for eating and ruminating (Susenbeth et al., Reference Susenbeth, Mayer, Koehler and Neumann1998). The similar chewing times observed in both groups lead to the assumption that the mechanical stress caused by the mowing conditioner was not able to significantly modify the physical effectiveness of the fiber (Mertens, Reference Mertens1997). It is likely that other factors limiting feed intake, such as the ruminal passage rate of digesta, were also not affected by the treatment, given the similar results in apparent total tract nutrient digestibility (Table 4).
Given the animals' similar forage DMI, most of the differences in the cows' nutrient intake (Table 3) resulted from the differences in the nutrient composition of the two types of hay (Table 1). In general, losses of nutrients on the field during wilting arise through post-cutting plant respiration, tedding leaf loss, leaching (i.e., due to rainfall) and the activity of microbes present on the plants' surface (Pizarro and James, Reference Pizarro and James1972; Barr et al., Reference Barr, Smith and Brown1995). The main driver for post-cutting plant respiration and the microbial activity is the water content of the plant material. According to McGechan (Reference McGechan1989), these respiratory losses increase quadratically with the moisture content of forages. Our data showed that the moisture content and its variation were numerically higher in the experimental [201 ± 22.7 g water (kg fresh matter)−1] as compared to the control forage before barn drying [167 ± 5.63 g water (kg fresh matter)−1]. This observation appeared to contradict our expectation of higher moisture values for the control group. Although conditioning of freshly cut forages should have caused a faster initial drying rate (Savoie, Reference Savoie2001), rewetting processes after dew formation overnight might have been more pronounced for this treatment if the cuticle layer was damaged. Rotz (Reference Rotz, Moore and Peterson1995) reported that conditioned forages absorb more water after being rewetted as compared to intact plants (i.e., control forages). Sundberg and Thylen (Reference Sundberg and Thylen1994) noted that absorption of surface water by the plants' tissue increases as they contain less water. Pizarro and James (Reference Pizarro and James1972) concluded that rewetting restores respiration back to the initial rate (i.e., during wilting), leading to additional respiratory losses.
Wilted forages were windrowed in the morning to reduce field losses, but this could have led to a higher moisture content of the experimental forage before barn drying, potentially leading to higher respiratory losses which could have contributed to the slightly decreased content of WSC in the experimental hay. The above observations suggest that the diverging changes in the nutrient profile between the forages cut with or without the use of an impeller mowing conditioner most likely occurred during field curing. Under the experimental conditions of the current study, the resulting differences in nutrient content between the two types of hay had only minor biological effects on the animals.
Our data also showed that cows fed the control hay selectively preferred aNDFom (P = 0.07) as well as WSC (P < 0.01) as compared to cows in the experimental group (Table 3). Although statistically significant, the selection index for WSC indicated that these cows ingested only 1.6% (DM basis) more of these substances than expected (i.e., predicted intake). However, the selection toward WSC in the control group might have contributed to the increased WSC intake in this group.
The higher availability of WSC for rumen microbiota in cows in the control group (Table 3) could have led to greater amounts of rumen-escapable microbial glycogen (Hall, Reference Hall2017). It was reported that, even in cows on all-forage diets, 10% of the microbial DM reaching the lower digestive tract consists of polysaccharides, e.g., glycogen (Owens et al., Reference Owens, Zinn and Kim1986; Branco et al., Reference Branco, Harmon, Bohnert, Larson and Bauer1999). The control hay most likely led to a greater availability of degradable nutrients for gut microbiota, which probably altered the fecal fermentation profile and resulted in a somewhat lower fecal pH value (−0.12 units; Table 4) in cows in the control group.
Milk performance was not affected by treatment. Of the milk constituents analyzed, only the milk urea content (Table 5) was significantly affected [+5.4 mg (100 mL)−1] in cows fed the experimental hay. This was caused by a greater feed protein intake (+0.210 kg d−1; Table 3) (Amanlou et al., Reference Amanlou, Amirabadi Farahani and Eslamian Farsuni2017), but also by a lower intake of readily fermentable energy (−0.190 kg d−1 WSC; Table 3) (Oltner et al., Reference Oltner, Emanuelson and Wiktorsson1985; Hof et al., Reference Hof, Vervoorn, Lenaers and Tamminga1997). In addition, the feed protein of the conditioned hay might have been also somewhat more susceptible to microbial fermentation in the rumen, resulting in a higher amount of rumen-degradable feed protein. Due to the potential effects of rewetting, degradation of protein via plant-derived proteases and/or other microbial activity during wilting could have also been more pronounced in the conditioned forage as compared to the control forage (McGechan, Reference McGechan1989; Barr et al., Reference Barr, Smith and Brown1995). In combination with a lower supply of readily fermentable carbohydrates, this could have resulted in less capture of dietary nitrogen as microbial nitrogen, compromising the efficiency of nitrogen utilization. Higher proportions of soluble carbohydrates in the forages or the supplemented concentrate (e.g., from sugar beet pulp) could have led to a higher milk production level in the experimental group.
General considerations
In Austria, about 20% of the hay milk-producing dairy farms use impeller mowing conditioners (M. Kittl, personal communication). The major argument for using this technology is that conditioned forages reach the desired DM content for barn drying (60–65%) 2–3 h sooner as compared to untreated forages (Pöllinger, Reference Pöllinger2000). This makes hay harvesting less dependent on weather conditions. However, it should also be considered that mowing machines with integrated conditioners are approximately 40% heavier and require about 20% more power per unit of cut grassland as compared to standard rotary mowers (Pöllinger, Reference Pöllinger2000).
Regardless of the observations made in this study, the impact on biodiversity, in particular on field fauna, has to be considered prior to issuing a general recommendation for the widespread use of mowing conditioners. Humbert et al. (Reference Humbert, Ghazoul, Sauter and Walter2010) estimated that impeller conditioners increase the damage of field invertebrates by a factor of two to three as compared to standard rotary mowers. In the same study, mortality in caterpillars increased from 38 to 69% when using the mowing conditioner. Delaying cutting time, meaning cutting later in the day (Frick and Fluri, Reference Frick and Fluri2001), and/or leaving uncut grass strips on the meadow could potentially mitigate this problem (Humbert et al., Reference Humbert, Ghazoul, Sauter and Walter2010).
Conclusions
This is the first study investigating the effects of barn-dried hay harvested with an impeller mowing conditioner on the feeding behavior and performance of cows fed a mainly hay-based diet. Under ceteris paribus conditions, including the same cutting and wilting times, number of teddings, and harvesting techniques, the use of the conditioner did not improve forage DMI or performance of organic dairy cows fed only approximately 16% concentrate (DM basis) and producing around 28 kg milk d−1. The similar chewing times observed in both groups suggest that the physical effectiveness of the fiber was not affected under the current feeding conditions. Experimental hay contained less WSC, reducing the supply of readily fermentable carbohydrates for rumen microbiota. In combination with an increased availability of feed protein, this might have compromised the efficiency of N use in the experimental hay group. A somewhat higher proportion of soluble carbohydrates in the concentrate (e.g., sugar beet pulp) for cows in the experimental group might have counteracted this effect and probably could have led to a higher level of milk performance. Future research should consider investigating the effects of conditioner use with shortened field wilting times for forages before barn drying as well as the effects of feeding organic cows with less concentrate or without any concentrate at all.
Acknowledgements
We are thankful to the Secondary School for Agriculture, HBLA Ursprung (Elixhausen, Austria) for housing the experiment, and to the farm staff, especially Franz Griessner and Florian Gollhofer. We would also like to thank Anita Dockner and Sabine Leiner (Institute of Animal Nutrition and Functional Plant Compounds at the University of Veterinary Medicine Vienna) for performing chemical analyses and Kathleen Knaus for editing assistance.
Financial support
This study is part of a larger research project on forage efficiency (research project no. 101210, BMLFUW-LE.1.3.2/0073-PR/8/2017) and funded by the Austrian Federal Ministry for Sustainability and Tourism, the Provincial Government of Salzburg (Austria), Raiffeisenverband Salzburg and the ‘Ja! Natürlich!’ brand of the REWE group (Wiener Neudorf, Austria).
